Doug Bandow

President Trump has repeatedly threatened to attack North Korea. Prior to the announcement in March of a prospective summit between President Trump and Kim Jong-un, policymakers were becoming increasingly convinced that war was a serious possibility. However, there is no military solution to the Korean crisis at an acceptable cost and risk.

The Kim regime has tested both nuclear weapons and long-range missiles. North Korean officials have long contended that America’s “hostile policy” required such a capability and insisted Pyongyang would never negotiate away its deterrent.

Now South Korean officials say that Kim is ready to denuclearize. Even if true, the security guarantees he reportedly expects may be more than the administration is prepared to grant. A hastily pulled together summit based on differing expectations could produce a clash and push the president back to his militarily aggressive strategy. If diplomacy is seen to fail, war could become more likely.

War, however, offers no guarantee of effective denuclearization. The United States likely does not know the location of all of the North’s nuclear facilities; even if it did, many targets may be buried too deep to reach. Moreover, military action risks a full-scale regional conflagration.

Kim could see even a limited military action as a prelude to attempted regime change. North Korea has the conventional capability to wreak destruction on Seoul. Use of weapons of mass destruction could, in worst-case scenarios, result in millions of casualties.

Kim appears to be seeking nuclear weapons to deter America. The greatest danger of war is not that Pyongyang would deliberately initiate nuclear war, but that U.S. policy convinces Pyongyang an American attack is imminent, thus tempting the North to strike preemptively. U.S. conventional superiority might pressure the North Koreans to “use it or lose it.”

The Trump administration should formalize diplomatic channels to Pyongyang to defuse tensions and explore nonmilitary alternatives. In pressing for tougher action against North Korea, Washington should address China’s interests. If and when tensions ease, the Trump administration should begin a process of military disengagement, turning defense responsibilities over to South Korea and its neighbors and withdrawing military forces, which entangle the United States and provide North Korea with American targets. Should a diplomatic resolution remain deadlocked, Washington should consider accepting South Korea’s development of its own deterrent.

Introduction

North Korea has challenged the United States and destabilized Northeast Asia since its creation in 1948. Emerging from the Soviet Union’s post-World War II occupation zone, the Democratic People’s Republic of Korea (DPRK) attempted to forcibly reunite the peninsula in 1950, triggering the Korean War. What began as a civil war drew in the United States and sundry allies, as well as China, and even, though less fully, the Soviet Union.

Although the conflict was nominally “limited”—the Truman administration rejected proposals to use nuclear weapons—the consequences were horrendous. Battle deaths alone exceeded 1.2 million; even more civilians likely died.¹ Much of the peninsula was ravaged by ground combat and repeated bombing. As American troops prepared to extirpate the North Korean regime, the People’s Republic of China (PRC) intervened. The conflict became an entirely new war, ending with the combatants back near the original boundary.

On October 1, 1953, the United States signed the Mutual Defense Treaty with the Republic of Korea (ROK), backed by a garrison that remains to this day. At that time, the ROK could not survive without American military support. However, the balance of power gradually shifted. In the 1960s, the South’s economy began to grow at an impressive pace. Today, South Korea fields a capable military.² It is better trained than the North’s and equipped with superior and more modern weapons. Moreover, the ROK enjoys roughly 45 times the GDP of the North.³ Add to that twice the population, extensive diplomatic and economic relations with the outside world, and a vast technological edge, and the ROK has the means to defeat North Korea militarily.

The collapse of the Soviet Union left the United States as the globe’s unrivaled military leader. China’s relationship with both America and North Korea also changed. While the PRC and Washington remain at odds on a number of important issues, the former no longer sees America’s presence on the Korean Peninsula as posing an existential threat, in contrast to 1950.

Today, Beijing is not interested in going to war for the North or against the United States. The PRC has warned the DPRK that China would not back its nominal ally if the latter started a conflict. It is not clear how Beijing would react if the United States initiated war, as it has deliberately kept its policy ambiguous.⁴ Most analysts doubt that the PRC would intervene militarily on behalf of Pyongyang even if Washington started the conflict, though Beijing might act to contain the consequences, perhaps using force to secure a rump buffer state in the north of the peninsula.

Since the end of the Cold War, the two Koreas have been locked in a cold war of their own, which has occasionally flared hot. A quarter century ago, the DPRK’s nuclear weapons program became a source of growing concern. Over the years, Washington attempted both engagement and isolation, buttressed by threats of war, to dissuade Pyongyang from moving forward. However, the nuclear development program that was largely a symbolic exercise under founder Kim Il-sung became a practical reality under his grandson, Kim Jong-un. The North is soon likely to possess the ability to target the American homeland with nuclear weapons.

That is a worrying thought, but Pyongyang’s intention is almost certainly defensive. Nuclear weapons serve several objectives. For instance, they yield international status and bolster the military’s prestige. Most importantly, though, nuclear weapons are the ultimate deterrent. The DPRK lacks reliable friends while facing a coalition led by the world’s military superpower, which in recent years has initiated regime change against several vulnerable governments. Most strikingly, Washington made a deal with Libya’s Muammar el-Qaddafi to eliminate the latter’s missile and nuclear programs before ousting him.

North Korea has repeatedly insisted that it will never negotiate away its nuclear arsenal. Few U.S. analysts believe otherwise. However, in bilateral discussions with South Korea in March, Pyongyang reportedly claimed it would consider denuclearization only if the United States “eliminated” the “military threat to the North.”⁵ There is little agreement as to what that might entail from the North Korean perspective. For its part, the Trump administration insisted it would not negotiate unless Pyongyang first surrendered on the essential issue and agreed to abandon its nuclear weapons. Only sticks were on the table. The announcement in March of a possible summit between the two leaders left unsaid which set of presumptions, if either, would prevail.

The administration is still pushing for tougher sanctions. Additional and more intensive economic penalties would damage the DPRK’s economy and the well-being of its citizens, but are unlikely to change Pyongyang’s course. After all, a half-million or more North Koreans died of starvation in the late 1990s, with no discernible impact on official policy.⁶ Nor is China likely to impose the trade embargo sought by Washington absent meaningful U.S. concessions to Beijing.

Until the dramatic announcement of a planned Trump-Kim meeting, the administration appeared to believe it had a choice either to accept a nuclear DPRK or launch a preventive war. President Trump then suggested he would choose the second rather than allow Pyongyang to gain the ability to hit the United States. “Denuclearization is [the North’s] only acceptable future,” the president declared in September 2017.⁷

The ever-hawkish Sen. Lindsey Graham (R-SC), who reportedly talks regularly with President Trump, predicted in December 2017 that the chances of war were 30 percent—and would rise to 70 percent if the North were to conduct another nuclear test.⁸ Even more worrying, in January 2018 the administration dropped its appointment of Victor Cha, who had served as director of Asian affairs in George W. Bush’s National Security Council, as ambassador to the ROK because he opposed military action.⁹

In December and January, sensationalist media reports fomented public concern, even near hysteria, in Hawaii, exacerbated by the false warning that an intercontinental ballistic missile (ICBM) was headed toward the islands.¹⁰ A majority of Americans say they back military strikes against North Korea, at least if other strategies fail to halt its nuclear program.¹¹ A recent Quinnipiac University poll found that nearly half of Republicans support preventive war.¹²

The risk of war would be high if only one country were represented by a blustering, impetuous, and unpredictable leader. The risk is far greater when both nations suffer under such leadership.¹³ One shudders at the thought of adding today’s mano a mano personal insults to the tenser moments of the Cold War, such as the Berlin airlift or the Cuban missile crisis. Contributing to the tension, in March Trump fired Lt. Gen. H. R. McMaster as national security adviser and picked former United Nations ambassador John Bolton, an extreme hawk who argued for preventive war against North Korea in February 2018, as his replacement.¹⁴ Assessing the likelihood of war is impossible with any precision, but until the recent South Korea-North Korea discussions that led to the summit proposal, Washington policymakers appeared to view an American attack as a possible and even reasonable option.¹⁵ Bolton is on record as saying President Trump should use the summit as an opportunity to deliver an ultimatum to Kim and refuse to negotiate a compromise arrangement, making military action in the aftermath of failed talks that much more attractive.¹⁶

Although hope is widespread that a successful summit could lead to a denuclearization agreement, there is a significant chance the meeting will never occur, or that if it does it will end in failure. The administration would then be likely to declare diplomacy a dead option and return to its highly confrontational strategy, including the threat of preventive strikes.

Unfortunately, almost any U.S. military action would likely trigger full-scale war. The resulting conventional combat alone would transcend any of America’s recent conflicts, including the Iraq War. The impact could expand far outside Korea’s borders, reaching China, Russia, Japan, and beyond. In initiating military conflict, Washington would be triggering the very war it has spent more than six decades trying to prevent.

War should be a last resort, reserved for preempting an imminent threat of attack. Otherwise, Washington should adopt a mix of negotiation, deterrence, containment, and disengagement to address the North’s challenge. The lack of a single, simple “silver bullet” is unfortunate, but it reflects the complexity of the North Korea problem. Ultimately, the most important objective should be to preserve peace, however tenuous, on the Korean Peninsula and prevent any conflict that might occur from spreading to the North’s neighbors and the United States.

North Korea’s Desire for Deterrence

North Korea is essentially friendless. Pyongyang’s enemies, most importantly, the United States and the ROK, are growing ever more powerful. Washington also became far more militarily aggressive in the aftermath of the Cold War. The United States has regularly engaged in regime-change wars, for example in Afghanistan, Iraq, and Libya, and in a range of other significant military actions, including in Bosnia, Serbia, and Somalia, and against the Islamic State.

The DPRK has long had a substantial conventional deterrent capability, with thousands of artillery pieces and abundant missiles targeting Seoul, which sits roughly 35 miles south of the North-South border, or the demilitarized zone (DMZ). Nevertheless, the overall conventional military balance continues to shift against the North. Under these conditions, further destabilization is possible. U.S. officials could come to believe they are capable of suppressing North Korean fire against Seoul, or Washington could decide that advancing America’s security outweighed whatever damage might be done to the South. In either case, deterrence would fail, leaving the United States free to attack, whether to take out the North’s missiles and nuclear weapons, decapitate North Korea’s leadership, or undertake even broader operations.

In August, then-secretary of state Rex Tillerson tried to allay Pyongyang’s fears by stating that the administration was not seeking regime change.¹⁷ However, his comments were hardly reassuring. For one thing, President Trump publicly undercut Secretary Tillerson numerous times on matters of diplomacy, tweeting in October that Tillerson’s attempts to negotiate with Pyongyang were a waste of time.¹⁸

Even if Pyongyang believed the secretary or other U.S. officials spoke for President Trump despite his dismissive treatment of them, it could not count on this or future administrations living up to such promises. The president repudiated the nuclear nonproliferation agreement with Iran negotiated by the Obama administration, and American intervention in Libya offered a particularly dramatic example of Washington’s cynicism: after pocketing the Qaddafi regime’s missile and nuclear concessions, U.S. and European officials initiated regime change when Tripoli faced a rebel uprising. Qaddafi suffered a particularly gruesome and painful death. At the time, Pyongyang noted that it would never make the same mistake.¹⁹

In short, any nation in as vulnerable a position as the DPRK would likely desire a more effective deterrent against America. The North could never obtain enough conventional weapons to match its superpower enemy, but nuclear weapons are the great equalizer. Even the hawkish Max Boot of the Council on Foreign Relations observed, “There is nothing to indicate that Kim Jong-un is suicidal or even expansionist; he is simply building nuclear weapons to ensure the survival of the regime.”²⁰ Boot dismissed fears that Pyongyang cannot be deterred: despite possessing nuclear weapons, Kim has not used them.²¹ Neither his father nor grandfather reignited the Korean War in another attempt at coercive reunification. That the regime is evil does not mean it is irrational or suicidal.

Nuclear weapons offer several benefits for North Korea. The DPRK gains status: otherwise, why would nations pay the slightest attention to the isolated, desperately poor, and virulently hostile regime? Nuclear weapons also may be a tool for extortion. Rivals may offer aid in an attempt to calm a nuclear power that appears bent on brinkmanship. Even more important, the missile and nuclear programs reward North Korea’s armed services, helping to cement the military leadership’s loyalty to the Kim dynasty.

Most fundamentally, though, nuclear weapons are the best military deterrent. They compensate for conventional weakness and therefore can prevent even a global superpower like the United States from taking advantage of its overwhelming strength. North Korea is not the first nation to see nukes this way. China insulated itself from Soviet and U.S. attack by developing nuclear weapons. Israel’s unacknowledged nuclear arsenal eliminated the possibility of another concerted conventional attack by its neighbors. Pakistan’s arsenal deters a far stronger India from striking. Nuclear weapons are the only military realm in which Russia competes equally with America.

The Kim regime is quite explicit in its reasoning, both publicly and privately. When I visited Pyongyang in June, foreign ministry officials said nuclear weapons were necessary because of Washington’s “hostile policy,” and especially its “military threats” and “nuclear threats.”²² Without question, America’s attitude is hostile, and the United States does make military and nuclear threats. Washington justifies this approach as a necessary response to the DPRK’s unremitting antagonism toward South Korea, going back to the Korean War.²³ However, particularly in the post-Cold War context, the United States has often militarized its opposition to weak regimes for reasons having nothing to do with defending itself or its allies, resulting in a perverse incentive for proliferation.

Indeed, Pyongyang demonstrates how even paranoids have enemies. In 1994, President Bill Clinton seriously considered military strikes against the North.²⁴ President George W. Bush lumped the DPRK in with Iran and Iraq as a member of the infamous “axis of evil.” He also said he “loathed” Kim Jong-il.²⁵ U.S. aircraft carriers—an “armada” in Trump-speak—regularly sail around the peninsula and U.S. bombers routinely overfly it. The American and South Korean militaries conduct annual military exercises that Pyongyang finds deeply provocative. President Trump has matched Kim’s rhetoric in threatening to unleash “fire and fury” and “destroy North Korea.”²⁶

In October, Ahn Dong-chun, deputy chairman of North Korea’s rubber-stamp parliament, explained that “Our country is being threatened, the very existence of the DPRK is at stake.” He added, “Our nuclear programs are nuclear deterrence programs aimed at protecting our independence; we have no choice but to develop our nuclear programs.”²⁷ One can argue that, under the circumstances, the North’s supreme leader would be a fool not to develop a reliable nuclear deterrent.

Indeed, knowledgeable Washington officials understand the North’s position. Scott W. Bray, national intelligence manager for East Asia in the Office of the Director of National Intelligence, stated in June:

We believe North Korea’s strategic objective is the development of a credible nuclear deterrent. Kim Jong-un is committed to development of a long-range nuclear-armed missile capable of posing a direct threat to the continental United States to complement his existing ability to threaten the region. Kim views nuclear weapons as a key component of regime survival and a deterrent against outside threats. Kim probably judges that once he can strike the U.S. mainland, he can deter attacks on his regime and perhaps coerce Washington into policy decisions that benefit Pyongyang and upset regional alliances—possibly even to attempt to press for the removal of U.S. forces from the peninsula.²⁸

Understanding that deterrence is the main motivation behind North Korea’s nuclear program is necessary in order to devise appropriate strategies and correct for Washington’s tendency to inflate the threat posed by North Korea.

Assessing the North Korean Threat

The world would certainly be a better place if North Korea did not possess nuclear weapons. However, the president mistakenly conflates evil with undeterrable. “We need only look at the depraved character of the North Korean regime to understand the nature of the nuclear threat it could pose to America and our allies,” Trump declared in his State of the Union address in January. Yet both Joseph Stalin’s Soviet Union and Mao Zedong’s People’s Republic of China were evil and deterrable. Dictators, including Kim Jong-un, usually prioritize self-preservation, which strengthens deterrence.²⁹

Proliferation inevitably creates uncertainty and instability. Acquisition of nuclear weapons by a regime that is simultaneously isolated, opaque, authoritarian, brutal, and hostile magnifies the danger. Nevertheless, Supreme Leader Kim is unlikely to start a war with America. Despite the North Korean regime’s malevolence, there is no evidence that Kim or anyone under him is suicidal.³⁰ Kim has devoted extraordinary effort to consolidating his power and enhancing his nation’s status. He has demonstrated no interest in departing this earth in a massive radioactive funeral pyre arranged by Washington.

Might Kim believe that his nation could defeat America in a war? Intelligence services report no evidence that he is this disconnected from reality. As noted earlier, deterrence evidently worked against his grandfather and father, neither of whom acted on their many threats against South Korea and the United States.³¹ Indeed, his commitment to creating missiles and nuclear weapons suggests cold calculation and a realization that his country lags well behind America in conventional military strength.³² He almost certainly will not attack the United States unless he sees war as inevitable and believes his country would have the best chance of survival by preempting an American assault.³³

The North’s essentially defensive posture does not mean there is no danger of conflict. The possibilities of mistake and misjudgment are real.³⁴ However, if President Trump acts responsibly, an admittedly questionable expectation, the likelihood of combat remains low. Kim knows starting a war would end his dynasty.

North Korea’s nuclear weapons are a concern for the United States primarily because of Washington’s military entanglement in Northeast Asia. Indeed, the most likely consequence of a North Korean nuke is to undermine America’s security guarantee of South Korea. If Pyongyang is capable of striking the U.S. homeland, America’s commitment to defend South Korea puts U.S. territory at risk. Involvement in even a conventional conflict could go nuclear: if U.S. forces threatened the Kim regime’s survival, Pyongyang would have reason to threaten nuclear war.

But alliances are, or at least should be, a means to an end. Namely, they should enhance U.S. security interests. America’s military guarantee for the South was initially based on Seoul’s weakness; the changing security environment has made this arrangement obsolete. Absent America’s involvement, a Korean conflict would be a horrific war but would remain a regional struggle far from U.S. territory. Moreover, the ROK, with a vastly larger economy, bigger population, and a far more extensive network of allies and trading partners, could defend itself. Dropping Washington’s promise to go to war on behalf of Seoul actually would make America more secure, reducing its chance of being drawn into other nations’ conflicts.

The Military Option: Feasibility and Risks of Escalation

Nevertheless, Washington has good reason to want to disarm the North. Defending the nonproliferation regime has been an important U.S. foreign policy priority for decades. Denuclearizing North Korea would prevent further proliferation by North Korea and limit the potential harm should war break out in Northeast Asia. Taking military action, however, will not serve this interest constructively.

Since the Korean War, Washington has viewed military action as an option, mainly because there are only imperfect solutions to the North Korea problem. Various pundits, analysts, and politicians proposed military action well before Pyongyang possessed much of a missile or nuclear capability. However, decisionmakers have repeatedly concluded that war is a bad option.

In 1969, President Richard Nixon apparently considered ordering a tactical nuclear strike against North Korea in retaliation for the downing of an American plane but quickly dropped the plan.³⁵ A quarter century ago, when a North Korean nuclear weapon was embryonic, the Clinton administration developed plans for war, but officials appreciated the risk that limited strikes would escalate to full-scale war.³⁶ South Korean president Kim Young-sam believed only his opposition prevented a second military cataclysm, though Clinton administration officials would later deny conflict was imminent.

In 1994, then-secretary of defense William Perry and assistant secretary of defense Ashton Carter (who later served as President Barack Obama’s secretary of defense) prepared plans to destroy the North’s Yongbyon facility. They acknowledged, however, that “a strike on Yongbyon, while surgical in and of itself, would hardly be surgical in its overall effect. The likely result of such a strike would be a spasmodic lashing out by North Korea’s antiquated, but large and fanatical, military.”³⁷ Robert Gallucci, assistant secretary of state for political-military affairs at the time, said, “I believe it would have resulted almost certainly in war.”³⁸ Retired general Gary Luck, who served as U.S. commander in South Korea during the Clinton administration, believed that “if we pull an Osirak,” referring to Israel’s 1981 bombing of Iraq’s Osirak nuclear reactor, “they will be coming south.”³⁹

The George W. Bush and Obama administrations also eschewed military action.⁴⁰ Former national security adviser and secretary of state Condoleezza Rice called the idea “lunacy.”⁴¹ Susan E. Rice, who served President Obama both as UN ambassador and national security adviser, warned that President Trump risked “tipping the Korean Peninsula into war.” Echoing Condoleezza Rice, she argued that a “decision to start a pre-emptive war on the Korean Peninsula, in the absence of an imminent threat, would be lunacy.”⁴²

Only a full-scale invasion could permanently end the North Korean threat, but few openly advocate such a drastic course of action. Neither commentators nor policymakers are willing to openly argue for inaugurating general war for that purpose. As was evident in the George W. Bush administration’s selling of the Iraq War, Washington knows that wars for regime change need to be viewed as cheap—a “cakewalk,” as one former government official put it—to receive popular support.⁴³ Advocates of U.S. military action have been reluctant to describe a prospective conflict on the Korean Peninsula that way.

Instead, proposals for limited military action to achieve lesser ends are more typical. One possibility is an air or naval blockade.⁴⁴ As an act of war, it would be highly provocative, but it would be less dangerous than a direct attack. However, it would have only limited effect, doing nothing to eliminate and little to slow the North’s development of nuclear weapons.

Most policy recommendations involve airstrikes targeting North Korean nuclear or missile assets or Kim and the DPRK leadership. The specifics differ and have grown more complex over time. Absent an objective to preempt an immediate existential threat, however, the arguments for a U.S. assault are weak.

The first challenge of a limited military strike is accurately identifying and destroying targets. Much has changed since the Clinton administration considered war in 1994. There are more nuclear facilities dispersed throughout the country in a greater number of locations, and many are buried deep underground, protected by bunkers or mountains.⁴⁵ The RAND Corporation’s Bruce Bennett explained: “There is no such thing as a surgical strike in North Korea” since “we don’t really know for sure where all their weapons are.”⁴⁶ Even if Washington knew their location, it could not guarantee their destruction. Max Boot comments that anyone imagining a successful “surgical” strike “has been watching too many Jason Bourne movies.”⁴⁷

Moreover, Pyongyang’s missile and nuclear capabilities have steadily advanced, including employment of mobile launch systems.⁴⁸ The survival of only a few launch vehicles would leave the North with a substantial retaliatory capacity, at least against the ROK and Japan.⁴⁹ Newer missiles rely on solid fuel, which renders them less vulnerable. Moreover, the DPRK may have located some weapons and facilities close to its border with China, meaning American military action there could carry high risks of an inadvertent clash with Beijing.⁵⁰ Even successful military strikes might only slow Pyongyang’s efforts while increasing the incentive to continue its nuclear program. Furthermore, the North would retain its conventional capability for targeting Seoul and other South Korean targets. Only an overwhelming, immediate threat could justify action under such circumstances.

The second difficulty is avoiding a general war, however limited the initial military objective. In theory, the United States could restrict its attacks to missile development and launch facilities, nuclear development activities, military command and control facilities, and political leaders.⁵¹Washington could then announce that “this is a limited defensive strike on a military target—an operation that poses no threat to civilians—and that America does not intend to bring about regime change,” as University of Texas professor Jeremi Suri proposed.⁵² The United States could accompany bombing runs with the promise that it planned no additional military action but would act with overwhelming force in response to North Korean retaliation. According to this argument, a reasonable, rational DPRK leadership would accept the resulting losses, exercise restraint in the face of offensive military attacks, and survive.

However, as Victor Cha asked, “If we believe that Kim is undeterrable without such a strike, how can we also believe that a strike will deter him from responding in kind?” The notion that Kim will not escalate undermines the very premise on which the argument for preventive U.S. military action is based—that Kim is irrational and undeterrable.⁵³ Moreover, acquiescence might not be the “rational” course for Kim in this scenario. Even if he somehow could be confident that the United States would not seek regime change, his inaction could ignite internal opposition to his rule and threaten his own survival.⁵⁴ In fact, it would be difficult for any leader of any government to accept supinely the destruction of the regime’s most important military assets acquired at such great effort and expense. The loss of face and increased military vulnerability would create a powerful incentive to retaliate. An attempt to decapitate the regime would be even more threatening and harder to ignore.⁵⁵

Restraint by Pyongyang under these circumstances seems highly unlikely. After all, based on America’s previous behavior, the Kim regime would be foolish to accept such reassurances from Washington. Given President Trump’s dismissal of negotiations and apparent willingness to toss aside the Iran nuclear deal, faith in the promises of the Trump administration would seem imprudent at best. Anyway, what would matter is not Washington’s intentions, but the Kim regime’s perceptions.

The North would likely see initial U.S. attacks as an attempt to disable the regime’s best weapons and top leadership before launching a more general military effort. Having watched the effectiveness of American arms in Iraq and elsewhere, the North Koreans almost certainly realize that yielding the initiative to Washington would guarantee defeat.

If the Kim government believed an American attack was imminent, the regime would have every reason to use all available weapons and do as much harm as possible. Any effective U.S. military action would require massive preparations of its conventional in-theater forces beforehand. The required buildup and movement would be impossible to conceal, thus serving as a visible signal for Pyongyang to strike first.⁵⁶

Even successful decapitation might not prevent retaliation. In the event of war, North Korean units are supposed to begin firing at Seoul without orders from above.⁵⁷ Thae Young-ho, a high-level diplomat who defected in 2016, contends that North Korean commanders “are trained to press the button without any further instructions from the general command” in the event of attack. “Artillery and short-range missiles will fire against South Korea,” Young-ho said.⁵⁸

With artillery and Scud missiles, the DPRK forces could bombard Seoul and its environs.⁵⁹ The capital is “a massive agglomeration of everything that is important in South Korea,” said Robert Kelly of Pusan National University.⁶⁰ Nearly 26 million South Koreans, roughly half the ROK’s population, live in the Seoul metropolitan area and are therefore within range of North Korean artillery and Scud missiles.⁶¹ Conventional bombardment of Seoul alone could result in tens of thousands of casualties.⁶²

Analysts warn that even interception of a North Korean test missile could lead to an unpredictable response from Pyongyang.⁶³ At the very least, the North would be inclined to retaliate in kind, perhaps by launching a limited strike on, for example, the Yongsan Garrison, the headquarters for the U.S. Army and United States Forces Korea, located in Seoul. (Scheduled for relocation in 2019, the facility is home to about 26,000 Americans, including active-duty soldiers, civilians, and family members.⁶⁴) Other targets, civilian or military, would also be at risk.

It is unclear how effective the DPRK’s nuclear and ballistic missile capabilities are. For instance, estimates of the number of artillery pieces, rate of fire, and reliability of munitions are in dispute. Many of the North’s artillery can only hit the northern third of Seoul—and the ROK has substantial counterbattery capabilities.⁶⁵ Still, former secretary William Perry warned that the North’s artillery force is better protected today than in 1994, when he drafted plans for U.S. military strikes. Even then, he argued, North Korean forces “could kill tens of thousands before we could stop them.” He added, “The price was very high then, but the price today would be very much higher.”⁶⁶

Some analysts foresee several hundred thousand shells per hour hitting the city and predict that only a few hours would be necessary for its destruction.⁶⁷ But even an attack of lesser intensity would cause enormous damage.⁶⁸ North Korea also could target South Korea’s two dozen nuclear power facilities, which could generate substantial nuclear fallout.

Efforts to eliminate the North’s artillery and missiles, however effective, would take time, leaving a window during which North Korea could target the country’s political, business, and economic centers. Pyongyang’s large, mostly antiquated tank fleet would be vulnerable to air attack but still might be capable of reaching Seoul. Finally, the North’s possession of an array of chemical and biological weapons could augment any conventional strike.⁶⁹ Such weapons could be used to attack Seoul and other South Korean cities. To assume casualties could be minimized would be to play a reckless game of international chicken with hundreds of thousands or even millions of lives.

Possible Chinese involvement would complicate any conflict. Beijing could intervene directly in North Korea’s defense against America, at least if the latter attacked the former.⁷⁰ However, the PRC has no love for the Kim regime and is unlikely to directly confront the United States militarily because it would risk a major war with a superior power. Nevertheless, Beijing still might attempt to preempt American conquest by intervening before Washington had conquered the North, perhaps preserving a rump state and preventing an outcome that would leave U.S. forces on the Yalu River.

Moreover, there is good reason to doubt that Washington could keep to its own limited aims once war breaks out. It is inconceivable that Washington would stop with the restoration of the current border, having once before left a surviving DPRK as unfinished business. Policymakers’ desire to forestall any future conflict would be strong.

Conquering North Korea, with its extensive mountainous terrain, also would be bloody and difficult. Without outside support, any insurgency almost certainly would be doomed, but irregular resistance could greatly extend the conflict. Over time, strong powers have won a diminishing share of asymmetric wars.⁷¹ There are numerous recent examples of failure, or at least only partial victory, and at very high cost: Vietnam, Iraq, Afghanistan, Sri Lanka, Chechnya, Burma, Colombia, and Turkey (vs. Kurdish separatists). Each case obviously differs dramatically in its specifics, but all illustrate the folly of assuming a quick and easy victory against North Korean resistance.

The United States could launch a far more aggressive and wide-ranging attack, even using nuclear weapons in its own first strike,⁷² or as retaliation for even “one [artillery] round going into Seoul,” in the words of retired general Tom McInerney.⁷³ But nuclear weapons are no panacea. A seemingly unprovoked, or at least unjustified, nuclear strike would have extraordinary geopolitical implications. Nuclear fallout would likely affect South Korea, China, Japan, and Russia. A war of choice, killing millions of North Koreans, could trigger global political blowback against the United States. Barry Posen, political scientist at the Massachusetts Institute of Technology, warned that “the United States would make itself an international pariah for decades, if not centuries,” adding that the nuclear option “should not even be on the table.”⁷⁴

While the United States and its allies might be able to blunt a North Korean conventional assault, the North could use nuclear weapons and other unconventional weapons against U.S. military forces in Japan as well as South Korea.⁷⁵ Even worse, Pyongyang could target both Seoul and Tokyo to maximize civilian casualties.

Even U.S. nuclear strikes might not eliminate all of the North’s retaliatory capabilities. In that case, the Kim regime would have no reason to hold back any remaining military forces. Mark Bowden of the Atlantic warned of retaliation by Pyongyang: “With only a few of its worst weapons, North Korea could, probably within hours, kill millions. This means an American first strike would likely trigger one of the worst mass killings in human history.”⁷⁶

The Catastrophic Costs of War

The costs of a war on the Korean Peninsula, even if it remained a purely conventional contest, are widely acknowledged. One would have to look back to the original Korean War for anything comparable. Gen. Joseph Dunford, chairman of the Joint Chiefs of Staff, said such a conflict would be “horrific,” with “a loss of life unlike any we have experienced in our lifetimes.” Defense secretary James Mattis similarly warned that combat would be “catastrophic” and represent “probably the worst kind of fighting in most people’s lifetimes.”⁷⁷

In advocating for the use of force against North Korea, Sen. Lindsey Graham (R-SC) admitted that “Japan, South Korea, China would all be in the crosshairs of a war if we started one with North Korea,” but also said, “If there’s going to be a war to stop [Kim], it will be over there. If thousands die, they’re going to die over there. They’re not going to die here. And [President Trump] told me that to my face.”⁷⁸ Columnist Ralph Peters was equally dismissive of the lives of others: “Better a million dead North Koreans than a thousand dead Americans.”⁷⁹ He left unaddressed what he thinks about the lives of South Koreans, Japanese, and Chinese. The U.S. government’s first responsibility is to its own citizens, but it should not ignore the consequences for foreign populations. In any case, war would result in many dead Americans, as well as foreigners.

Casualty estimates reflect numerous unpredictable factors, since great uncertainties surround any military operation. Although South Korea, backed by the United States, enjoys an enormous military advantage, the North has a quantitative manpower and weapons edge and has invested in asymmetric capabilities.⁸⁰ Pyongyang would ultimately lose a conflict, but the war could evolve in unpredictable ways involving special operations forces, forward-deployed conventional arms, and more destructive unconventional, including nuclear, weapons.⁸¹

Moreover, casualties likely would not be limited to the Korean Peninsula. The DPRK could hit Guam and Japan with missiles and perhaps nuclear weapons. Conflict also could spill over the Chinese and Russian borders. If the belligerents used nuclear weapons, fallout could contaminate surrounding nations. While the worst case is not inevitable, the vagaries of war are unpredictable and most conflicts turn out worse than predicted.

In this case, the forecasts are awful enough. According to a Congressional Research Service report: “Conservative estimates anticipate that in the first hours of a renewed military conflict, North Korean conventional artillery situated along the Demilitarized Zone (DMZ) could cause tens of thousands of casualties in South Korea,” where nearly a quarter-million Americans, soldiers and citizens, are living or visiting every day.⁸²“A protracted conflict,” the report continues, “particularly one in which North Korea uses its nuclear, biological, or chemical weapons—could cause enormous casualties on a greater scale, and might expand to include Japan and U.S. territories in the region.”⁸³

When the Clinton administration considered military action in 1994, it predicted as many as one million casualties in the larger conflict.⁸⁴ Ashton Carter, who as assistant secretary of defense helped prepare the Clinton plan, said, “the loss of life in that war—God forbid that kind of war ever starts on the Korean Peninsula—loss of life is horrific.”⁸⁵ In Foreign Policy magazine, a military war gamer who prepared conflict scenarios for the Pentagon explained that the U.S. armed services “expect a massive humanitarian crisis, enormous loss of life, and economic disaster.”⁸⁶

A single 100-kiloton nuclear weapon dropped on a major city could kill hundreds of thousands of people.⁸⁷ Many estimates for a Second Korean War go into the millions.⁸⁸ Adm. Michael J. Dumont, vice chairman of the Joint Chiefs of Staff, observed: “Invading North Korea could result in a catastrophic loss of lives for U.S. troops and U.S. civilians in South Korea. It could kill millions of South Koreans and put troops and civilians in Guam and Japan at risk.”⁸⁹ Victor Cha has also warned of “millions of casualties.”⁹⁰

One worst-case analysis from the Bulletin of the Atomic Scientists projected potentially 2.5 million deaths and 6.9 million injuries of varying severity.⁹¹ A nuclear attack on Seoul and Tokyo could result in as many as 2.1 million deaths and 7.7 million injuries. If the bombs were thermonuclear, the death toll would be even higher.⁹² One war game that considered nukes hitting the American homeland estimated 8 million deaths.⁹³

The economic costs of conflict also would be colossal, reaching into the hundreds of billions or trillions of dollars and would extend well beyond the Korean Peninsula. Washington would face extraordinary pressure to underwrite occupation and finance reconstruction across the entire battle zone. The postconflict expense of reconstructing North and South Korea, along with other possible targets—Japan and Guam, for instance—would be enormous.⁹⁴ Occupying and rebuilding a devastated North Korea and combatting any insurgency would be far more difficult than in Iraq and Afghanistan, where U.S. troop levels peaked at 170,000 and 100,000, respectively. Those two operations together cost almost $5 trillion.⁹⁵

The bottom line is that any war is likely to escalate to cataclysmic proportions. In September, President Trump warned that if the United States attacks the DPRK,” it will be a very sad day for North Korea.”⁹⁶ That is true, but not only for North Korea.

Threatening War as a Substitute for Going to War

Are President Trump’s warnings justified as an attempt to intimidate the Kim government?⁹⁷ Threatening war is better than triggering war, but the first could lead to the second. Backing a proposal to negotiate with the threat of force makes negotiation more likely, at least in theory, but if that is the present administration’s strategy, it risks misfiring badly. First, the president has often undercut the efforts of his own officials to promote negotiation.⁹⁸ Second, his casual talk of war has heightened tensions and increased the prospect of conflict. Warned Condoleezza Rice: “I can remember in crisis after crisis, your rhetoric gets hotter and hotter, escalates more and more and more. Really, it just puts an environment around the problem [that makes it] hard to get it solved.”⁹⁹

The president’s foreign policy views and willingness to use military force remain ambiguous if not mysterious more than a year into his presidency. That makes it more difficult for other governments, including Pyongyang, to discern his intentions, understand his rhetoric, and predict his actions. None of Trump’s predecessors, including President Clinton, who actually considered a preventive strike, threatened war as explicitly as Trump has. Pyongyang may therefore view an American attack as being a much greater possibility today. In that case, the North is likely to be much more sensitive to any military moves that might be a prelude to such a strike, such as bolstering existing U.S. forces and evacuating civilians.¹⁰⁰

Belief that Washington is ready to strike would influence how the DPRK responds to individual U.S. military moves, short of an actual attack, that it finds threatening. Two years ago, Pyongyang asserted that “the right to nuclear preemptive attack is by no means the U.S. monopoly.”¹⁰¹ North Korean foreign minister Ri Yong-ho said in September: “Since the United States declared war on our country, we will have every right to make countermeasures, including the right to shoot down United States strategic bombers even when they are not inside the airspace border of our country.”¹⁰² Because North Korean units are reportedly authorized to respond to attack without orders from Pyongyang, even a mistaken DPRK response to a presumed threat could trigger American or South Korean retaliation (or both) and quickly spiral into full-scale war.

More broadly, given North Korea’s military vulnerability and Washington’s obvious incentive to seek and destroy Pyongyang’s weapons as part of any attack, the DPRK could—and arguably should, if it is acting in its perceived interest—adopt a “use it or lose it” approach. The regime might seek to preempt what it believes to be an imminent U.S. attack, starting a war that neither nation, let alone South Korea and neighboring states, desired.¹⁰³ While the North is likely to lose in any case, waiting for the United States to build up its conventional forces and launch a concentrated bombing campaign would guarantee a speedier defeat. The first Gulf War serves as an illustrative example to an adversary like North Korea of the importance of preempting expected U.S. military action.¹⁰⁴ Not only have rapid deployment capabilities advanced considerably since 1991, but also the United States is postured in Northeast Asia in such a way as to make North Korea more vulnerable than Iraq was. “The base infrastructure in the region and the accessibility of North Korea from the sea should make it possible to generate around 4,000 sorties a day compared to the 800 a day that were so effective in Iraq,” write former CIA director James Woolsey and retired U.S. Air Force general Thomas McInerney. “We must be prepared to win a war, not execute a strike,” they contend.¹⁰⁵

The president’s bluster has also undercut Washington’s relationship with the ROK. South Koreans say they are used to the North’s over-the-top rhetoric but expect U.S. policy to be measured and responsible. The Moon government has been devoting substantial effort to convincing its people that the Trump administration will not act without its approval, which may or may not be true.¹⁰⁶

However, President Trump has not confirmed that he would limit his options to satisfy America’s allies, and other officials insist that U.S. interests are paramount, irrespective of the costs to others, including South Koreans. Should Seoul perceive that the United States is moving toward a preventive strike, allied dissension could disrupt American military efforts, especially if Washington intends to use ROK bases. In fact, Pyongyang’s recent gambit at the 2018 Winter Olympics was an effort to drive a wedge between the allies. The move was mainly directed at Seoul, not Washington, though the DPRK was reportedly prepared to meet with Vice President Mike Pence. That meeting did not occur and, if reports are accurate, would have been of little use since the vice president merely planned to reaffirm Washington’s maximalist demand for denuclearization.

Alternative Policy Options

The Trump administration and the Kim regime are on a collision course. North Korea believes a deliverable nuclear weapon is the only sure guarantee for regime preservation. However, as former national security adviser H. R. McMaster said, “Kim Jong-un should recognize that if he thinks the development of this nuclear capability is keeping him safer, it’s actually the opposite.”¹⁰⁷ Observers increasingly fear the administration is serious about its threats of war.

There remain alternatives. None are perfect, but all are preferable to starting a conflict.

The administration’s priority should be de-escalating the current crisis. The North’s past offer, backed by China, to freeze missile and nuclear testing in exchange for suspending joint U.S.-ROK military exercises would break the cycle of increasing North Korean tests and American threats.¹⁰⁸Reducing tensions, as well as the frightened urgency that has dominated public debate of the issue, is essential. However useful the exercises might be, they are not as important as moving the potential belligerents back from the brink.

The United States should maintain sanctions on the Kim regime to reduce its resources and push it toward the negotiating table. However, sanctions alone are unlikely to be sufficient. The historical experience with economic warfare is mixed, but generally negative.¹⁰⁹ In the case of North Korea, sanctions are even less likely to work. The regime, buttressed by Kim’s ruthlessness, appears stable. Moreover, as noted earlier, it sees nuclear weapons as a tool of regime survival, the most important objective of the DPRK government, which means Pyongyang is willing to pay a high price to retain them.

Washington also should establish diplomatic contacts with Pyongyang. The initial process should be modest, though even that could pose a challenge for an administration that has lagged in staffing the State Department.¹¹⁰ Communication with an adversary should not be seen as a reward. It would have been foolish not to talk to the Soviet Union during the Cold War. Indeed, the more threatening the situation, the greater the need for contact. Dialogue would help reduce the North’s sense of isolation and vulnerability, which has encouraged Pyongyang to prioritize its nuclear weapons development. In preparation for a Trump-Kim summit in May, the administration should craft an offer to the North involving a peace treaty, diplomatic recognition, security guarantees, and economic relief.

Along with an opening to the DPRK, the United States should hold serious discussions with China. To win stronger Chinese action against the North, Washington must address the PRC’s concerns over having either a failed state next door or a reunified Korea, allied with America, and with U.S. troops on its border.¹¹¹ Possible confidence-building measures include offering aid for refugees, accepting possible Chinese military intervention in the aftermath of a North Korean collapse, and guaranteeing that U.S. forces would leave a reunited peninsula. Washington should also seek Beijing’s support for a U.S. peace offer to the North, along with a promise to impose harsher economic sanctions if Pyongyang rejects it.

Finally, Washington should reconsider both its conventional security guarantee and nuclear umbrella over South Korea. The need for the “mutual” defense treaty disappeared long ago, since South Korea can now carry the burden for its own defense and America’s commitment to South Korea is more of a liability than a security asset.¹¹² The growing nuclear threat makes the alliance increasingly dangerous. Is this or any other administration really prepared to risk American cities to protect Seoul?¹¹³ There are good reasons to support nonproliferation, but acquiescing to South Korea’s development of its own nuclear deterrent may be in U.S. national security interests.¹¹⁴ Seoul started down that path three decades ago, abandoning its program only under strong U.S. pressure. Today, two-thirds of South Koreans support obtaining nuclear weapons.¹¹⁵ Moreover, the mere possibility of a South Korean (and Japanese) nuclear arsenal would likely spur China to do more.¹¹⁶

Conclusion

The United States has spent nearly 65 years attempting to prevent a rerun of the Korean War. It would be extremely foolish to deliberately reignite the same conflict today, especially when nonmilitary options have yet to be exhausted, a fact dramatically illustrated by the prospective Trump-Kim summit. Ironically, when I visited the ROK last year, several policymakers indicated that they were more concerned about the Trump administration starting a war than they were about the Kim regime starting one. They have spent decades living with North Korea’s military threat, while relying on the United States to be the responsible party. This familiar arrangement no longer appears to prevail.

At least before the president agreed to meet the North Korean leader, McMaster had made conflict sound inevitable: President Trump “is not going to permit this rogue regime, Kim Jong-un, to threaten the United States with a nuclear weapon. So he is going to do anything necessary to prevent that from happening.” Exactly what that would be went unstated, though McMaster said, “our military leaders are refining, improving plans every day.”¹¹⁷ That is, plans for war.

Preventive strikes are highly unlikely to end the North Korean threat at an acceptable cost. As the Atlantic’s Mark Bowden put it, “any effort to crush North Korea flirts not just with heavy losses, but with one of the greatest catastrophes in history.”¹¹⁸ America’s losses likely would be in the tens of thousands. A nuclear strike on both Seoul and Tokyo could yield total casualties in the millions.

To trigger such a cataclysm in order to prevent an unlikely attack by a nation that can be contained and deterred would be, frankly, mad. The United States has dealt with similar challenges already. For decades, Washington confronted the Soviet Union, a far more dangerous power than North Korea. It did the same with the People’s Republic of China, under dictator Mao Zedong (apparently even more unhinged than Kim), who cited his nation’s large population as reason not to be concerned about the possibility of nuclear war. Washington considered preventive strikes in both instances but decided against that course.¹¹⁹ Looking back, we should thank policymakers for taking the long path.

Former secretary of state Tillerson vowed that “diplomatic efforts will continue until the first bomb drops.”¹²⁰ However, it is essential that the first bomb not drop. The Trump administration appears to be following the George W. Bush administration’s misbegotten strategy in Iraq, only with less thought and preparation.¹²¹ There are risks to containing and deterring North Korea, but they pale beside the costs of plunging the peninsula into the abyss of war. The United States has helped keep the peace in Northeast Asia for decades. Whether or not the hoped-for summit occurs, new thinking is essential to maintain that peace.

Notes

1. “Korean War Fast Facts,” CNN, June 10, 2017, https://www.cnn.com/2013/06/28/world/asia/korean-war-fast-facts/index.html.
2. The Military Balance: The Annual Assessment of Global Military Capabilities and Defence Economics, 2017 (London: International Institute for Strategic Studies, 2017), pp. 303-10.
3. Benjamin Katzeff Silberstein, “North Korea’s Economy Grew by Almost 4 percent in 2017, BOK Says,” North Korean Economy Watch, July 21, 2017, http://www.nkeconwatch.com/category/organizaitons/bank-of-korea/.
4. See, for example, Kristin Huang, “China’s Nuclear Get-Out Clause over Defense of North Korea,” South China Morning Post, April 13, 2017; http://www.scmp.com/news/china/diplomacy-defence/article/2087320/china-not-obliged-defend-n-korea-if-its-attacked-say; “Reckless Game over the Korean Peninsula Runs Risk of Real War,” editorial, Global Times, August 10, 2017, http://www.globaltimes.cn/content/1060791.shtml; and Ankit Panda, “China and North Korea Have a Mutual Defense Treaty, But When Would It Apply?,” DiplomatOnline, August 14, 2017, https://thediplomat.com/2017/08/china-and-north-korea-have-a-mutual-defense-treaty-but-when-would-it-apply/.
5. Choe Sang-Hun and Mark Landler, “North Korea Signals Willingness to ‘Denuclearize,’ South Says,” New York Times, March 6, 2018.
6. Daniel Goodkind et al., “A Reassessment of Mortality in North Korea, 1993-2008,” paper presented to Population Association of America, Population Bureau, U.S. Census Bureau, March 28, 2011, http://paa2011.princeton.edu/papers/111030.
7. Jacob Pramuk, “Read Trump’s Combative Full Remarks to the UN on North Korea,” CNBC, September 19, 2017, https://www.cnbc.com/2017/09/19/read-trumps-combative-full-remarks-to-the-un-on-north-korea.html.
8. Uri Friedman, “Lindsey Graham: There’s a 30 Percent Chance Trump Attacks North Korea,” AtlanticOnline, December 14, 2017, https://www.theatlantic.com/international/archive/2017/12/lindsey-graham-war-north-korea-trump/548381/.
9. David Nakamura and Anne Gearan, “Disagreement on North Korea Policy Derails White House Choice for Ambassador to South Korea,” Washington Post, January 30, 2018.
10. Alex Wellerstein, “The Hawaii Alert Was an Accident. The Dread It Inspired Wasn’t.” Washington Post, January 16, 2018.
11. Lydia Saad, “More Back U.S. Military Action vs. North Korea Than in 2003,” Gallup.com, September 15, 2017.
12. Aaron Blake, “Almost Half of Republicans Want War with North Korea, a New Poll Says. Is It the Trump Effect?,” Washington Post, October 15, 2017.
13. Although President Trump is no murderous dictator, he seems to share some personality traits with Supreme Leader Kim. Scott D. Sagan of Stanford University argued that “both are prone to lash out impulsively at perceived enemies, a tendency that can lead to reckless rhetoric and behavior.” Scott D. Sagan, “The Korean Missile Crisis: Why Deterrence Is Still the Best Option,” Foreign AffairsOnline, November 10, 2017, https://www.foreignaffairs.com/articles/north-korea/2017-09-10/korean-missile-crisis.
14. John Bolton, “The Legal Case for Striking North Korea First,” Wall Street Journal, February 28, 2018, https://www.wsj.com/articles/the-legal-case-for-striking-north-korea-first-1519862374.
15. See, for example, Uri Friedman, “On North Korea, ‘We’re Fools if We Don’t Start Taking the President at His Word,’” AtlanticOnline, October 27, 2017, https://www.theatlantic.com/international/archive/2017/10/chris-murphy-north-korea-trump/544203/; Evan Osnos, “Is the Political Class Drifting toward War with North Korea?,” New Yorker, November 8, 2017, https://www.newyorker.com/news/daily-comment/is-the-political-class-drifting-toward-war-with-north-korea; and David E. Sanger, “Talk of ‘Preventive War’ Rises in White House over North Korea,” New York Times, August 20, 2017, https://www.nytimes.com/2017/08/20/world/asia/north-korea-war-trump.html?_r=0. Sen. Lindsey Graham (R-SC) claimed that President Trump said he would strike the North rather than allow the DPRK to possess the ability to target the American homeland. Quoted in Zack Beauchamp, “Lindsey Graham: Trump ‘Told Me’ He Will Bomb North Korea if It Keeps Testing Missiles,” Vox, August 1, 2017, https://www.vox.com/world/2017/8/1/16075198/trump-lindsey-graham-north-korea-war.
16. John Bolton, interview, FoxNews.com, March 9, 2018, http://video.foxnews.com/v/5748784936001/?#sp=show-clips.
17. “North Korea: US Not Seeking Regime Change, Says Rex Tillerson,” BBC, August 2, 2017, http://www.bbc.com/news/world-us-canada-40797613.
18. David Nakamura, “Trump Says Tillerson Is ‘Wasting His Time’ Trying to Pursue Negotiations with North Korea,” Washington Post, October 1, 2017.
19. Mark McDonald, “North Korea Suggests Libya Should Have Kept Nuclear Program,” New York Times, March 24, 2011.
20. Max Boot, “Can the U.S. Wait North Korea Out?,” Commentary, July 5, 2017, https://www.commentarymagazine.com/foreign-policy/asia/can-u-s-wait-north-korea/.
21. Max Boot, “The ‘Fire and Fury’ Can Wait,” Commentary, August 9, 2017, https://www.commentarymagazine.com/foreign-policy/asia/north-korea/north-korea-deterrence-containment-fire-fury/.
22. Conversations in Pyongyang, June 12-15, 2017. See, for example, Doug Bandow, “Pyongyang Speaks: An Interview inside North Korea,” Raddington Report, September 26, 2017, https://raddingtonreport.com/pyongyang-speaks-an-interview-inside-north-korea/.
23. Of course, deterrence does not have to be provided by America. South Korea also has the ability to do so. See, for example, Doug Bandow, Tripwire: Korea and U.S. Foreign Policy in a Changed World (Washington: Cato Institute, 1996); and Ted Galen Carpenter and Doug Bandow, The Korean Conundrum: America’s Troubled Relations with North and South Korea (New York: Palgrave Macmillan, 2004).
24. Although Clinton administration officials denied U.S. military action was imminent, South Korea’s President Kim Young-sam contended that an attack was thwarted only because he rejected the idea in a phone call with President Clinton. Choe Sang-hun, “Korean Crisis Is Different This Time,” New York Times, August 3, 2009; and Choe Sang-hun, “Kim Young-sam, South Korean President Who Faced Down Military, Dies at 87,” New York Times, November 21, 2015.
25. Reuters Staff, “Previous U.S. Comments about Kim Jong-Il,” Reuters, December 6, 2007, https://www.reuters.com/article/us-korea-north-nuclear-bush/factbox-previous-u-s-comments-about-kim-jong-il-idUSN0619890220071206.
26. Peter Baker and Choe Sang-hun, “Trump Threatens ‘Fire and Fury’ against North Korea if It Endangers U.S.,” New York Times, August 8, 2017.
27. Quoted in “North Korea Lawmaker: We Need Nukes Because of US Threat,” Associated Press, October 16, 2017, http://abcnews.go.com/amp/International/wireStory/north-korea-lawmaker-nukes-us-threat-50495737.
28. Scott W. Bray, “North Korea’s Nuclear Weapons and Missile Capability,” presented to ICAS Special Summer Symposium, June 26, 2017, Institute for Corean-American Studies [sic], http://www.icasinc.org/2017/2017p/2017pswb.pdf.
29. The United States faced a similar problem in Iraq and chose badly, argued Columbia University’s Richard Betts. He believes deterrence would have been better than preventive war. Richard K. Betts, “The Lost Logic of Deterrence: What the Strategy That Won the Cold War Can—and Can’t—Do Now,” Foreign AffairsOnline, March 1, 2013, https://www.foreignaffairs.com/articles/united-states/2013-02-11/lost-logic-deterrence.
30. Former national security adviser H. R. McMaster curiously contended that deterrence didn’t work with “a regime that imprisons and murders anyone who seems to oppose the regime,” apparently ignoring the experiences of China, the Soviet Union, and others. John Delury, “Take Preventive War with North Korea Off the Table,” Foreign AffairsOnline, August 2017, https://www.foreignaffairs.com/articles/north-korea/2017-08-22/take-preventive-war-north-korea-table.
31. See, for example, David C. Kang, “International Relations Theory and the Second Korean War,” International Studies Quarterly 47, no. 3 (September 2003): 301-24.
32. The North has amassed a significant missile arsenal, though its capabilities remain unclear. Eleanor Albert, “North Korea’s Military Capabilities, in Review,” Defense One, December 2, 2017, http://www.defenseone.com/threats/2017/12/north-koreas-military-capabilities-review/144232/?oref=d-river.
33. North Korean casualties in any conflict would be high, and it is not certain that DPRK forces could reach Seoul, though if they did the city would face ruin even if subsequently liberated. Any North Korean attack would be an act of desperation, but victory for the United States and South Korea would be dearly bought in the best of cases. Franz-Stefan Gady, “What Would the Second Korean War Look Like?,” Diplomat Online, April 19, 2017, https://thediplomat.com/2017/04/what-would-the-second-korean-war-look-like/.
34. Exactly how great is a matter of dispute. At least one analyst believes that it would be better to accept the certainty of millions of dead in a preventive war today than to “take the deterrence gamble.” Kevin R. James, “North Korea: Why War Is the Only Option Now,” National InterestOnline, November 29, 2017, http://nationalinterest.org/blog/the-buzz/north-korea-why-war-the-only-option-now-23414. Yet more than seven decades into the nuclear age, with multiple nuclear powers, there has yet to be an accidental or misguided launch. That weighs against starting a war guaranteeing massive death and destruction to forestall the possibility of mistake or misjudgment in the future.
35. Blake Stilwell, “That Time a Drunk Richard Nixon Tried to Nuke North Korea,” Business Insider, Oct. 28, 2017.
36. Jamie McIntyre, “Washington Was on Brink of War with North Korea 5 Years Ago: Pentagon Had Predicted up to 1 Million Deaths,” CNN, October 4, 1999, http://edition.cnn.com/US/9910/04/korea.brink/.
37. Ashton B. Carter and William J. Perry, “Back to the Brink,” Washington Post, October 20, 2002.
38. Quoted in McIntyre, “Washington Was on Brink of War with North Korea.”
39. Quoted in Doug Bandow, “Thank Kim Young-sam for Preserving the Peace,” Japan Times, December 1, 2015, https://www.japantimes.co.jp/opinion/2015/12/01/commentary/world-commentary/thank-kim-young-sam-for-preserving-the-peace/.
40. Marjorie Cohn, “A Preemptive Strike on North Korea Would Be Catastrophic and Illegal,” Huffington Post, August 14, 2017.
41. Quoted in Evan Osnos, “The Risk of Nuclear War with North Korea,” New Yorker, September 18, 2017, https://www.newyorker.com/magazine/2017/09/18/the-risk-of-nuclear-war-with-north-korea.
42. Susan E. Rice, “It’s Not Too Late on North Korea,” New York Times, August 10, 2017.
43. Ken Adelman, “Cakewalk in Iraq,” Washington Post, February 13, 2002.
44. See, for example, Greg Keeley, “A Naval Blockade Is Just What We Need to Contain North Korea,” The Hill, November 18, 2017, http://thehill.com/opinion/national-security/361055-a-naval-blockade-is-just-what-we-need-to-contain-north-korea; James Stavridis, “A Naval Blockade Can Rein in North Korea,” Gulf News, September 15, 2017, http://gulfnews.com/opinion/thinkers/a-naval-blockade-can-rein-in-north-korea-1.2090645; and Corey Charlton, “Showdown with Kim: Donald Trump Should Blockade North Korea and Shoot Down Any Missile Test by ‘Madman’ Kim Jong Un, ex Navy Seal Says,” Sun, April 15, 2017, https://www.thesun.co.uk/news/3337533/donald-trump-should-blockade-north-korea-and-shoot-down-any-missile-test-by-madman-kim-jong-un-ex-navy-seal-says/.
45. See, for example, Victor Cha, “Victor Cha: Giving North Korea a ‘Bloody Nose’ Carries a Huge Risk to Americans,” Washington Post, January 30, 2018.
46. Quoted in Sanger, “Talk of ‘Preventive War.’”
47. Boot, “Can the U.S. Wait North Korea Out?” Fellow neoconservative and former deputy defense secretary Paul Wolfowitz also said that military options “are extremely dangerous.” Paul Wolfowitz, interview, MSN, August 7, 2017, https://www.msn.com/en-us/news/crime/no-good-options-in-north-korea-fmr-deputy-defense-secretary-paul-wolfowitz/vp-AApBZ4c.
48. Mark Bowden, “How to Deal with North Korea,” Atlantic, July/August 2017, https://www.theatlantic.com/magazine/archive/2017/07/the-worst-problem-on-earth/528717/.
49. See, for example, Barry R. Posen, “The Price of War with North Korea,” New York Times, December 6, 2017.
50. One recent missile launch came from less than 20 miles from the PRC. Kim Young-hie, “Pre-empting Peace,” JoongAng Ilbo, August 14, 2017, http://mengnews.joins.com/view.aspx?aId=3037105.
51. Kathleen J. McInnis, “The North Korean Nuclear Challenge: Military Options and Issues for Congress,” Congressional Research Service, CRS Report 7-5700, R44994, November 6, 2017, pp. 1, 25-31. Before he became defense secretary, Ashton Carter joined former defense secretary William Perry in advocating the destruction of North Korean missiles before they could be test launched. Ashton B. Carter and William J. Perry, “If Necessary, Strike and Destroy,” Washington Post, June 22, 2006.
52. Jeremi Suri, “Bomb North Korea, Before It’s Too Late,” New York Times, April 12, 2013.
53. Cha, “Victor Cha: Giving North Korea a ‘Bloody Nose’ Carries a Huge Risk to Americans.”
54. His grandfather survived the debacle of the Korean War, but Kim Jong-un’s brutal rule likely has created opponents ready to strike if the opportunity arises.
55. South Korea plans to establish what it describes as a “decapitation unit.” Choe Sang-hun, “South Korea Plans ‘Decapitation Unit’ to Try to Scare North’s Leaders,” New York Times, September 12, 2017. Then-CIA director Mike Pompeo (whom Trump nominated in March 2018 to be secretary of state) also has suggested an interest in regime change, contrary to the statements of former secretary Tillerson. Eli Watkins, “CIA Chief Signals Desire for Regime Change in North Korea,” CNN, July 21, 2017, http://www.cnn.com/2017/07/20/politics/cia-mike-pompeo-north-korea/index.html. The actual impact might be to make Pyongyang more trigger-happy.
56. Former defense secretary William Perry believed the administration “could dramatically reduce the casualties if we would do some judicious reinforcements to begin with.” “Examining the Lessons of the 1994 U.S.-North Korea Deal,” Frontline, undated, http://www.pbs.org/wgbh/pages/frontline/shows/kim/themes/lessons.html. In 2006, Perry and Ashton Carter proposed reinforcing U.S. air and naval forces around the peninsula when preparing to strike North Korean test missiles. Carter and Perry, “If Necessary, Strike and Destroy.”
57. Andrei Lankov, “North Korea’s Sixth Nuke Test: Who Will Win the War of Words?,” NK News, September 6, 2017, https://www.nknews.org/2017/09/north-koreas-sixth-nuke-test-who-will-win-the-war-of-words/.
58. Quoted in Daniel L. Davis, “Time Running Out in Korea,” Real Clear Defense, November 16, 2017, https://www.realcleardefense.com/articles/2017/11/16/time_running_out_in_korea__112649.html.
59. Bowden, “How to Deal with North Korea.”
60. Quoted in Motoko Rich, “In North Korea, ‘Surgical Strike’ Could Spin into ‘Worst Kind of Fighting,’” New York Times, July 5, 2017, https://www.nytimes.com/2017/07/05/world/asia/north-korea-south-us-nuclear-war.html.
61. See, for example, McInnis, “The North Korean Nuclear Challenge,” 2.
62. Osnos, “The Risk of Nuclear War.”
63. See, for example, Mark Fitzpatrick and Michael Elleman, “Deterring North Korea the Wrong Way,” Politics and Strategy: The Survival Editors’ Blog, August 11, 2017, https://www.iiss.org/en/politics%20and%20strategy/blogsections/2017-6dda/august-b877/deterring-north-korea-the-wrong-way-4bd9.
64. “USAG Yongsan, Republic of Korea,” Military Installations (website), http://apps.militaryonesource.mil/mos/f?p=132:
    content:0::no::p4_inst_id,p4_inst_type:2935,installation.
65. Reid Kirby, “Sea of Sarin: North Korea’s Chemical Deterrent,” Bulletin of the Atomic Scientists, June 21, 2017, https://thebulletin.org/sea-sarin-north-korea%E2%80%99s-chemical-deterrent10856.
66. Quoted in Michael D. Shear and Michael R. Gordon, “How U.S. Military Actions Could Play Out in North Korea,” New York Times, August 11, 2017, https://www.nytimes.com/2017/08/11/world/asia/north-korea-trump-military.html?_r=0. The Pentagon figured on more than a half-million military casualties alone during the first three months of war. “US Believed It Would ‘Undoubtedly Win’ War with North Korea in 1994—but with Huge Casualties,” Guardian, December 8, 2017.
67. David Majumdar, “5 Ways North Korea’s Military Could Start World War III,” National InterestOnline, October 4, 2017, http://nationalinterest.org/blog/the-buzz/5-ways-north-koreas-military-could-start-world-war-iii-22593.
68. See, for example, Rich, “In North Korea, ‘Surgical Strike’ Could Spin into ‘Worst Kind of Fighting.’”
69. Bowden, “How to Deal With North Korea”; and Theo Emery, “The Other Threat from North Korea,” New York Times, October 28, 2017.
70. See, for example, Hannah Parry, “Chinese State-Owned Paper Says China Will Intervene and Stop America if It Attacks North Korea First—and Will Only Stay Neutral if Kim Attacks the States First,” Daily Mail, August 11, 2017.
71. See, for example, Ivan Arreguin-Toft, “How the Weak Win Wars: A Theory of Asymmetric Conflict,” International Security 26, no. 1 (Summer 2001): 97.
72. Bowden, “How to Deal with North Korea.”
73. Quoted in Nick Giampia, “US Could Wipe Out North Korea in 15 Minutes, Says Gen. Tom McInerney,” Fox Business, August 7, 2017, https://www.foxbusiness.com/politics/us-could-wipe-out-north-korea-in-15-minutes-says-gen-tom-mcinerney.
74. Posen, “The Price of War with North Korea.”
75. Jeffrey Lewis, “North Korea Is Practicing for Nuclear War,” Foreign PolicyOnline, March 9, 2017, http://foreignpolicy.com/2017/03/09/north-korea-is-practicing-for-nuclear-war/.
76. Bowden, “How to Deal with North Korea.”
77. Quoted in Uri Friedman, “Lindsey Graham Reveals the Dark Calculus of Striking North Korea,” AtlanticOnline, August 1, 2017, https://www.theatlantic.com/international/archive/2017/08/lindsey-graham-north-korea/535578/.
78. Quoted in Friedman, “Lindsey Graham.”
79. Ralph Peters, “The Moral Answer to North Korea Threats: Take Them Out!” New York Post, September 4, 2017. Fox News analyst Thomas McInerney similarly argued that the dead will “be mostly North Koreans.” “Fox Analyst Dismisses That Millions Would Die in Nuclear Strike Because ‘They’ll Be Mostly North Koreans,’” Media Matters, August 10, 2017.
80. McInnis, The North Korean Nuclear Challenge,” 15-16; and David Majumdar, “North Korea’s Land Forces Pack a Big Punch (4,300 Tanks, for Starters),” National Interest Online, April 18, 2017, http://nationalinterest.org/blog/the-buzz/north-koreas-land-forces-pack-big-punch-4300-tanks-starters-20257.
81. McInnis, “The North Korean Nuclear Challenge,” 18, 38-47.
82. McInnis, “The North Korean Nuclear Challenge,” summary. For American population numbers, see Cha, “Victor Cha: Giving North Korea a ‘Bloody Nose’ Carries a Huge Risk to Americans.”
83. McInnis, “The North Korean Nuclear Challenge,” i-ii, 1-20.
84. McIntyre, “Washington Was on Brink of War with North Korea 5 Years Ago.”
85. “Examining the Lessons of the 1994 U.S.-North Korea Deal,” Frontline, undated, http://www.pbs.org/wgbh/pages/frontline/shows/kim/themes/lessons.html.
86. Chetan Peddada, “A Sneak Peek at America’s War Plans for North Korea,” Foreign Policy Online, September 7, 2017, http://foreignpolicy.com/2017/09/07/a-sneak-peak-at-americas-war-plans-for-north-korea/.
87. Sagan, “The Korean Missile Crisis.”
88. McInnis, “The North Korean Nuclear Challenge,” 3.
89. Quoted in Osnos, “Is the Political Class Drifting toward War?”
90. Quoted in Uri Friedman, “What Are America’s Options on North Korea?,” AtlanticOnline, April 7, 2017, https://www.theatlantic.com/international/archive/2017/04/trump-options-nuclear-north-korea/522075/.
91. Kirby, “Sea of Sarin.”
92. Michael J. Zagurek, Jr., “A Hypothetical Nuclear Attack on Seoul and Tokyo: The Human Cost of War on the Korean Peninsula,” 38 North, October 4, 2017, http://www.38north.org/2017/10/mzagurek100417/.
93. Harry J. Kazianis, “The Case for Containing North Korea,” National Interest, November-December 2017, http://nationalinterest.org/feature/the-case-containing-north-korea-22727.
94. Any North Korean collapse, even without a war, would be extraordinarily expensive. Even under “optimistic assumptions,” some 260,000 to 400,000 personnel would be needed to occupy the North. Bruce W. Bennett and Jennifer Lind, “The Collapse of North Korea: Military Missions and Requirements,” International Security 36, no. 2 (Fall 2011): 86. But a RAND Corporation report warned that collapse could bring civil war, “humanitarian disaster,” “huge numbers of refugees,” “horrendous” consequences, and more. Bruce W. Bennett, Preparing for the Possibility of a North Korean Collapse (Santa Monica, CA: RAND Corporation, 2013), pp. xvii-xviii.
95. See, for example, Neta C. Crawford, “US Budgetary Options through 2016: $4.79 Trillion and Counting, Summary of Costs of US Wars in Iraq, Syria, Afghanistan and Pakistan and Homeland Security,” Watson Center, Brown University, September 2016, http://watson.brown.edu/costsofwar/files/cow/imce/papers/2016/Costs%20of%20War%20through%202016%20FINAL%20final%20v2.pdf; and “Report: Nearly $5 Trillion Spent on Iraq and Afghanistan Wars So Far,” Stars and Stripes, September 13, 2016, https://www.military.com/daily-news/2016/09/13/report-nearly-5-trillion-spent-on-iraq-afghanistan-wars-so-far.html. Bruce Bennett warned that North Korean security personnel likely would be hostile to unification, especially since some of them would have been involved in criminal activity. Bennett, Preparing for the Possibility of a North Korean Collapse, xxi-xxiii. That hostility would be much greater toward foreign occupation.
96. Quoted in Anne Gearan and Emily Rauhala, “Trump Renews Threat of Force against North Korea over Nuclear Weapons,” Washington Post, September 7, 2017, https://www.washingtonpost.com/world/china-will-back-fresh-un-sanctions-on-north-korea-over-nuclear-tests/2017/09/07/afc6ac52-93a9-11e7-b9bc-b2f7903bab0d_story.html?utm_term=.61af5931986b.
97. Some observers speculate that the president believes he is reprising President Dwight Eisenhower’s supposed threat to use nuclear weapons in the Korean War, though in fact the armistice was reached without communication of such a threat. William I. Hitchcock, “Trump Threatened to Nuke North Korea. Did Ike Do the Same?,” Washington Post, August 11, 2017.
98. Peter Beinart, “What Bob Corker Really Fears,” AtlanticOnline, October 10, 2017, https://www.theatlantic.com/international/archive/2017/10/corker-north-korea-trump/542514/.
99. Quoted in Olivia Beavers, “Condi Urges Trump: ‘Watch Your Rhetoric’ on North Korea,” The Hill, October 10, 2017, http://thehill.com/homenews/administration/354789-condi-urges-trump-watch-your-rhetoric-on-north-korea.
100. See, for example, McIntyre, “Washington Was on Brink of War with North Korea 5 Years Ago.”
101. Quoted by Sagan, “The Korean Missile Crisis.”
102. Quoted in Max Boot, “Trump Needs to Watch Ken Burns’ ‘Vietnam War’ ASAP,” Foreign Policy Online, September 26, 2017, http://foreignpolicy.com/2017/09/26/trump-needs-to-watch-ken-burns-vietnam-war-asap/.
103. One possible scenario is presented by Jeffrey Lewis, “This Is How Nuclear War with North Korea Would Unfold,” Washington Post, December 8, 2017.
104. McIntyre, “Washington Was on Brink of War with North Korea 5 Years Ago.”
105. Quoted in Paul Reynolds, “North Korea: Looming Crisis?,” BBC, August 18, 2003, http://news.bbc.co.uk/2/hi/asia-pacific/3160823.stm.
106. See, for example, Mythili Sampathkumar, “South Korea in Pleas to Avoid North Korea War after Trump’s ‘Locked and Loaded’ Comment,” Independent, August 14, 2017, http://www.independent.co.uk/news/world/americas/us-politics/south-korea-north-donald-trump-non-military-nuclear-war-locked-loaded-us-president-a7892781.html.
107. Quoted in Kyle Feldscher, “H. R. McMaster: Threatening a Military Option in North Korea Keeps Kim Jong-un in Check,” Washington Examiner, October 16, 2017, http://www.washingtonexaminer.com/hr-mcmaster-threatening-a-military-option-in-north-korea-keeps-kim-jong-un-in-check/article/2637561.
108. When I visited the North in June, officials told me that the offer, having been rejected by Washington, was now off the table. However, China and Russia could help move it back to consideration. Indeed, Moscow claims that the DPRK recently “showed interest” in the proposal. Seungmock Oh, “North Korea ‘Showed Interest’ in Freeze-for-Freeze Deal: Russian Deputy FM,” NK News, December 6, 2017, https://www.nknews.org/2017/12/north-korea-has-showed-interest-in-freeze-for-freeze-deal-russian-deputy-fm/.
109. See, for example, Gary Clyde Hufbauer et al., Economic Sanctions Reconsidered: History and Current Policy, 2nd ed. (Washington: Institute of International Economics, 1990); Navin A. Bapat et al., “Determinants of Sanctions Effectiveness: Sensitivity Analysis Using New Data,” Empirical and Theoretical Research in International Relations 39, no. 1 (2013): 79-98; T. Clifton Morgan and Valerie L. Schwebach, “Fools Suffer Gladly: The Use of Economic Sanctions in International Crises,” International Studies Quarterly 41, no. 1 (March 1997): 27-50; T. Clifton Morgan and Valerie L. Schwebach, “Economic Sanctions as an Instrument of Foreign Policy: The Role of Domestic Politics,” Empirical and Theoretical Research in International Relations 21, no. 3 (1995): 247-63; Jaleh Dashti-Gibson et al., “On the Determinants of the Success of Economic Sanctions: An Empirical Analysis,” American Journal of Political Science 41, no. 2 (April 1997): 608-18; Robert A. Pape, “Why Economic Sanctions Still Do Not Work,” International Security 23, no. 1 (Summer 1998): 66-77; and Robert A. Pape, “Why Economic Sanctions Do Not Work,” International Security 22, no. 2 (Fall 1997): 90 -136.
110. For instance, after more than a year in office, President Trump had yet to nominate an ambassador to Washington’s ally, South Korea.
111. Doug Bandow, “Will China Solve the North Korea Problem? The United States Should Develop a Diplomatic Strategy to Convince China to Help,” Cato Institute Policy Analysis no. 806, December 6, 2016, https://www.cato.org/publications/policy-analysis/will-china-solve-north-korea-problem.
112. Bandow, Tripwire; and Carpenter and Bandow, The Korean Conundrum.
113. Even during the Cold War, analysts recognized that “extended deterrence” was problematic. The issue afflicts Europe and the Middle East, as well as East Asia. See, for example, Steven Pifer et al., “U.S. Nuclear and Extended Deterrence: Considerations and Challenges,” Brookings Institution, Arms Control Series Paper No. 3, May 2010, pp. 4-7, 18-43.
114. Doug Bandow, “Let Them Make Nukes: The Case for ‘Friendly’ Proliferation,” Foreign Affairs Online, July 26, 2016 https://www.foreignaffairs.com/articles/japan/2016-07-26/let-them-make-nukes.
115. Michelle Ye Hee Lee, “More Than Ever, South Koreans Want Their Own Nuclear Weapons,” Washington Post, September 13, 2017.
116. Nuclear proliferation can elicit balancing behaviors by neighbors. See Matthew Kroenig, “Approaching Critical Mass: Asia’s Multipolar Nuclear Future,” National Bureau of Asian Research Report, June 2016, http://nbr.org/publications/element.aspx?id=897.
117. Quoted in “Military Options for N. Korea ‘under Constant Refinement’: McMaster,” Yonyap News Agency, October 15, 2017, http://www.koreaherald.com/view.php?ud=20171016000122.
118. Bowden, “How to Deal with North Korea.”
119. Just four years after the wartime alliance against Nazi Germany, White House advisers urged President Harry S. Truman to authorize attacks on Soviet nuclear facilities. Other countries have considered taking similar actions: USSR. vs. China, China vs. India, and India vs. Pakistan. Only Israel did so, against Iraq and Syria, but neither of them was capable of retaliating effectively. Graham Allison, “Can North Korea Drag the U.S. and China into War?,” AtlanticOnline, September 11, 2017, https://www.theatlanticcom/international/archive/2017/09/north-korea-us-china/539364.
120. Sarah N. Lynch, “U.S. Diplomacy with North Korea to Continue until ‘First Bomb Drops’: Tillerson,” Reuters, October 15, 2017.
121. See, for example, Kori Schake, “The North Korea Debate Sounds Eerily Familiar,” AtlanticOnline, December 8, 2017, https://www.theatlantic.com/international/archive/2017/12/north-korea-iraq-war-george-w-bush-trump/547796/.

James Bacchus

In all too many minds, the relevance of the World Trade Organization (WTO) is much in doubt. The failure of the 11th Ministerial Conference of the WTO last December in Buenos Aires to complete any new multilateral or other agreements — or even to agree on the traditional declaration concluding the conference — has left many wondering if this is the beginning of the end for the WTO.

Expectations were low in Buenos Aires, and the ministers fulfilled those low expectations. Complicating matters, and further stirring doubts, are the continued assaults of the Trump administration on the WTO, both verbally and in the conduct of U.S. trade policy. Traditional U.S. leadership at the WTO is missing, as members struggle to find a way forward toward further trade liberalization and international economic integration.

Yet, despite the seemingly bleak assessments, some see not the beginning of the end for the WTO, but the end of the beginning. There were clear signs on several fronts in Buenos Aires that WTO members are ready to turn toward “plurilateral” solutions on trade that could, in time, become fully multilateral solutions. Multilateral trade agreements must always be the ultimate goal for the WTO. But there is more than one way to get to multilateralism. Starting with agreements among some, but not yet all, WTO members and then gradually transforming them into fully global agreements appears to be the most promising path to multilateralism in the 21st century.

Digital trade, services trade, fisheries subsidies, environmental goods, investment facilitation, and other issues are all ripe for negotiation and agreement. By taking a plurilateral approach toward multilateralism, the members of the WTO can ensure that this is the end of the beginning — and not the beginning of the end — for the World Trade Organization.

Introduction

Was the 11th Ministerial Conference of the World Trade Organization (WTO) last December the beginning of the end or the end of the beginning for the WTO? Trade negotiators came away from the Buenos Aires, Argentina, conference wondering about the future of the rules-based world trading system. The headlines following the conference were not encouraging. One publication proclaimed, “The WTO May Have Reached Its Breaking Point.”¹ Another read, “It’s the End of the WTO as We Know It — and Trump Feels Fine.”² Perhaps the most optimistic of all the headline assessments was that of The Economist: “The WTO Remains Stuck in Its Rut.”³ In the public’s mind, and in all too many political and diplomatic circles, the centrality of the World Trade Organization to global trade is much in doubt.

Expectations going into the conference were low, and — by and large — those low expectations were fulfilled. The conference failed to produce solutions that the world very much needs to long-festering problems. There was no multilateral outcome at all. Members could not even agree on the customary conference-concluding declaration. Many considered it a major achievement that the U.S. Trade Representative, Robert Lighthizer, deigned to attend the conference — although he left early after lecturing the other delegates largely on all that he and President Donald Trump think is wrong with the WTO. The one hoped-for multilateral outcome — a long-sought agreement limiting fisheries subsidies that support overfishing and illegal fishing — failed once again. And the proposed plurilateral agreement freeing trade in environmental goods that had come so close to closure the year before was barely mentioned in Buenos Aires.

In a world threatened by the seeming retreat of the United States and other countries from the institutions of international cooperation toward the protectionism and mercantilism of “economic nationalism,” uncertainty about the future of the WTO is cause for grave concern. This concern has only been heightened by Trump’s imposition of new tariffs on imported steel and aluminum and his announced intention of imposing tariffs on other products. The purported basis of some of these tariffs is national security, but underlying all these measures is the accusation that China (and others) are cheating on trade rules. By acting unilaterally on this issue, however, the United States is circumventing the WTO rules that require all members to take all their trade-related disputes falling within the scope of the treaty to WTO dispute settlement for a multilateral judgment before taking any trade action.⁴

According to Trump, the WTO is a “catastrophe” and has been a “disaster” for the United States.⁵ At the very least, his tumultuous trade policy signals a turn away from the WTO as the fulcrum of world trade. Globally, a failure by the WTO to continue to provide the framework for rules-based trade would lead to a dangerous accumulation of economic disruption and confrontation as more and more countries emulate the United States and fall back into self-defeating acts of trade restriction and trade discrimination of all kinds. The ongoing stalemate over updating the global trade rules, evidenced anew in Buenos Aires, only reinforces the inclination of many WTO members to look elsewhere for the solutions they need and does nothing to counter the growing forces against trade and against globalization that are fueling commercial confrontation.

At the same time, however, despite the failure of the ministerial conference to produce any multilateral outcome, there were some encouraging signs in Buenos Aires of systemic evolution on several fronts. Subsets of like-minded members pledged to proceed with plurilateral negotiations on a variety of pressing new issues, including digital trade, investment facilitation, disciplines for fossil-fuel subsidies, trade opportunities for micro-, small-, and medium-sized enterprises, and more.⁶ Even amid current U.S. hostility toward the WTO, a number of these pledges offer real promise. Indeed, in some — such as the initiative on digital trade — the United States seems likely to participate.

Although most of the global trade liberalization accomplished under the multilateral system since the creation of the General Agreement on Tariffs and Trade (GATT) in 1947 has been the product of consensus-based multilateral negotiating rounds, the WTO agreements permit — indeed, they encourage — alternative, plurilateral approaches to liberalization in which some, but not all, WTO members agree to move ahead first with new agreements to liberalize trade. If this is to be the end of the beginning — and not the beginning of the end — of the WTO, the momentum for pursuing such plurilateral deals within the WTO framework must be encouraged and harnessed to establish and modernize world trade rules and to continue to sustain the centrality of the World Trade Organization.

The Logic of Multilateralism

In considering the future of the WTO, it is necessary to understand the importance of multilateralism and nondiscrimination. Central to the success of the international trading system, which gradually evolved from the GATT in 1947 into the WTO in 1995, has been the overarching principle that international trade negotiators describe routinely — and sometimes almost reverently — as “multilateralism.” The logic of multilateralism applies to all kinds of international cooperation, but in trade the enduring goal of multilateralism is to act globally to reduce trade barriers globally.

This has long been accomplished within the WTO-based trading system through the working of one fundamental rule, the rule that requires most-favored-nation (MFN) treatment, an idea that dates back nearly a thousand years to innovations by the Baltic traders of the Hanseatic League. The WTO most-favored-nation rule is widely misstated and is even more widely misunderstood. It is thought by many to mean that in lowering a tariff or another barrier to international trade, one country will give the products of another country more favorable treatment than it gives to those of all other countries. In fact, this basic rule of trade means precisely the opposite: products of every other country receive the same trade treatment as products of the most favored of all countries. The rule requiring MFN treatment forbids discrimination between and among the like traded products of other WTO members.

This fundamental trade principle is enshrined as a general obligation of all WTO members in the WTO agreements on goods and on services in the WTO treaty. Thus, whenever a concession on trade in a good or service is made by one WTO member to another, that same concession must be made to every other WTO member. In this way, the mutual reciprocity of trade concessions is “multilateralized,” meaning that all WTO members (and the entire world trading system) benefit from each and every trade concession. Through this mechanism, the reduction of trade barriers over the course of seven decades has been achieved globally. The gains from trade have thus been maximized again and again for all the participants in the system through the successive rounds of multilateral trade negotiations conducted first by the GATT and now by the WTO. Through the working of the MFN rule, a multilateral agreement has vastly more potential to lower trade barriers and, thus, to increase trade and prosperity than will any single or series of bilateral or regional agreements.

The potential global economic payoff from continued reliance on the multilateral approach to trade liberalization could be considerable. World Bank models have suggested that a global free trade agreement “could add $5 trillion to the world’s GDP by 2020, $3 trillion of which would go to developing countries. And by the close of this century, such a deal could increase GDP by more than $100 trillion, with most of the gains accruing outside developed nations.”⁷ Moreover, beyond these numbers, an even greater economic payoff could result from the role of freer trade as a catalyst for necessary economic change. Global free trade would be the equivalent of a global tax cut and could help jump-start much of the modernization that is needed globally to meet the new challenges of the 21st century.

We are, of course, a long way from concluding a global free trade agreement. As ambitious as it was, the liberalization implicit in the Doha Round of multilateral negotiations did not begin to approach global free trade. But we are much closer than we were decades ago to global free trade because of our past adherence to the multilateral approach, which has long worked well for the trading system. Now, unfortunately, it does not. Now there are vastly more participants in the trading system and thus many more negotiators at the trading table. Now the economic and related interests of the participants vary more than ever before. And now, therefore, WTO members find it increasingly difficult to negotiate multilaterally under the procedures they have long followed.

Since 1947, multilateral trade negotiations have aimed to produce an outcome characterized as a “single undertaking” — a situation where nothing is agreed until every country taking part in the negotiations agrees on everything. Moreover, multilateral trade negotiations have long sought to achieve a consensus, meaning that nothing can be concluded if any one negotiating party formally objects.⁸ Now, though, it has become exceedingly difficult for the members to reach a consensus in a single undertaking. The Doha Round, launched in 2001, achieved virtually none of its objectives after 14 years of frustration and impasse.

Despite these real obstacles, the logic of multilateralism remains. When attainable, multilateral approaches to market access and to rules for trade are still the best way to boost world trade and the prosperity that follows. On some significant issues, such as the global market distortions caused by agricultural subsidies, only multilateral solutions are available. Plurilateral solutions will not work. Politically, it would be impossible for the United States to agree to cut its agricultural subsidies if the European Union were not subject to similar terms. And how could the United States and the European Union agree to agricultural subsidies cuts in a plurilateral deal if China were not subject to similar terms? And so on.

Furthermore, changing a rule of general application in the trading system in a less than fully multilateral way would pose fundamental problems. Take, for example, the “national treatment” rule, which forbids discrimination in favor of local producers over foreign producers of like imported products. What would happen to the flow of world trade if national treatment were changed in a plurilateral agreement to mean one thing for some countries and something entirely different for others? As a result, there would be something considerably less than the desired security and predictability for the overall WTO-based world trading system.⁹

On these and on some other issues relating to market access and trade rules, the only solutions are multilateral ones. But, on other issues, including many of the new and emerging 21st-century trade issues, there is another approach, one that can produce partial trade solutions now that are potentially preludes to multilateral solutions later.

Plurilateralism, the Next-Best Option

Contrary to widespread perception, the multilateral agreement that established the WTO does not require that all WTO trade negotiations be conducted multilaterally. WTO members have chosen to continue to pursue new trade obligations multilaterally, but they are not required to do so. An option provided by the WTO treaty is to pursue new trade obligations plurilaterally through negotiations among a self-selected subset of WTO members seeking the perceived economic advantages of agreements within the WTO rules framework that are “WTO-plus,” which add to existing obligations and afford additional benefits to those members that choose to accept the obligations by becoming parties to the agreements.¹⁰

Under the WTO treaty, the WTO-plus benefits of these plurilateral trade agreements can be provided in one of two ways. The agreements can be MFN, which means their benefits can be provided inclusively to all WTO members, including those that have not accepted the additional obligations of the multilateral agreement, as is the case with the WTO Information Technology Agreement (ITA). Or the agreements can be non-MFN, which means their benefits can be provided exclusively only to those countries that negotiate and agree to comply with the additional obligations in the new agreements, as is the case with the WTO Government Procurement Agreement (GPA).

The MFN approach to a plurilateral agreement among an ad hoc and like-minded coalition of willing WTO members is more appealing if the agreement is a tariff-reducing market access agreement and if a critical mass of participants in the particular market sector in question is seen as having been assembled to agree to the plurilateral agreement, thus minimizing the potential effect of free riders who benefit from the agreement without adhering to its terms. This was the case with the ITA. The non-MFN approach is more appealing where the additional obligations are not conventional market access obligations and where the parties to the agreement do not constitute such a critical mass. This was the case with the GPA.

The shared expectation of many of us in the United States and elsewhere who helped establish the WTO was that such plurilateral approaches by like-minded WTO members desirous of deeper levels of liberalization and economic integration would be commonplace. We envisaged the WTO as a forum and as a framework for ongoing innovation in providing market access and devising world trade rules to accommodate and facilitate ongoing innovations in an ever-evolving world economy. We foresaw the new international institution called the World Trade Organization as addressing emerging trade issues through agreements relating to specific sectors of global commerce and to specific trade issues that — at least at the outset — would be less than fully multilateral.¹¹

It is an understatement to say that not every suggested solution to a trade problem commands immediate universal acceptance. The fact is, very few such suggestions do. In the WTO, it almost always takes time to build a critical mass of countries to move a new idea ahead. It often takes experience to discern just how that idea should move ahead. For some countries, there is, understandably, a natural reluctance to moving forward with new ideas without having the time and experience from which to have some notion of what will happen later.

One of the advantages of a plurilateral approach taken by some, but not all, WTO members in embracing a new idea in trade is that it provides a proving ground of trial and error within the rules-based framework of the WTO trading system. Ideas that fail can be abandoned. Ideas that work can be improved and scaled up over time to become part of fully multilateral agreements that bind all WTO members. In some ways, this approach resembles the traditional view of states in the United States as laboratories of democracy — as places where new ideas can be tried first locally and, if successful, adopted by other states and possibly at the national level.

The gradual evolution of the GATT into the WTO demonstrates the wisdom of plurilateralism. Several multilateral trade agreements in the WTO treaty — those dealing with antidumping, safeguards, subsidies, and technical regulations — began as plurilateral GATT codes that were accepted by some, but not all, GATT contracting parties. These codes only became fully multilateral with the establishment of the WTO. In the minds of those who anticipated a series of such plurilateral approaches by the WTO, the same incremental legal path would be pursued in numerous other areas of current and future global trade concern. At first this incrementalism happened: the inclusion of the GPA in the WTO treaty and the conclusion soon afterward of the ITA and the protocols on basic telecommunications services and on financial services. To many, this seemed to be the way forward for incrementally achieving the shared goal of more multilateralism.

But then came September 11, 2001. The 9^th WTO Ministerial Conference in Doha, Qatar, which convened shortly after the attacks, saw the launch of the Doha Development Round. Almost out of habit, the conference was a single undertaking bound by the consensus rule. The negotiations continued off and on for 14 years until they ended — not with a bang, but with a whimper — at the convening of the 10^th WTO Ministerial Conference in Nairobi, Kenya, in 2015. The only notable negotiating success resulting even tangentially from the Doha Round occurred when WTO members pulled the issue of trade facilitation out of the Doha negotiations and concluded the WTO Trade Facilitation Agreement separately and multilaterally in Bali, Indonesia, in 2013.

Failure, once again, to reach a multilateral outcome in Buenos Aires in December 2017 suggests that a course correction is urgently needed for the WTO.

The Costs of Rejecting Plurilateralism

Some WTO members have long been reluctant to support new plurilateral agreements, making it harder to conclude them as WTO agreements. For sound economic reasons, these reluctant members prefer the general — and generally bigger — payoff from multilateral deals. They are also hesitant to assume new obligations on top of those they already have. Sometimes they are unwilling even to permit other members to agree to new WTO obligations that they themselves do not wish to undertake for fear that, as has happened before, those new obligations will be negotiated without them and then eventually become fully multilateral. Not least, many developing countries are of the view that before agreeing to negotiate new agreements, they should get the benefits they feel they were promised, but have not yet received, from previous agreements.

In some respects, this reluctance is understandable. But a refusal, for whatever reason, to allow new plurilateral agreements to be concluded within the framework of the WTO only guarantees that the plurilateral undertakings of like-minded countries desirous of more ambitious trade liberalization and integration will occur outside the WTO. Developed countries, in particular, are eager to move ahead on many trade fronts. The reluctance of other WTO members to do so was a catalyst for the negotiations outside the WTO of such “megadeals” as the Trans-Pacific Partnership (TPP) among countries along the Pacific Rim and the proposed Transatlantic Trade and Investment Partnership (TTIP) between the United States and the European Union. All the participants in those negotiations are WTO members. Those negotiations could have — and should have — occurred within the WTO, but because the countries involved were unable to move ahead within the WTO, they sought to do so outside the WTO.

Negotiating these megadeals within the legal ambit of the WTO would have reduced the concern among nonparticipating WTO members that they would be locked out of these new trade arrangements. It would also have eliminated the geopolitical undertones that the new arrangements were intended for political reasons to exclude other WTO members. Plurilateral agreements within the WTO are open to all WTO members that choose to accept their obligations. With respect to the TPP, for example, the issue of whether it was intended to surround or isolate China would simply not have arisen; if the TPP were a non-MFN plurilateral agreement within the WTO, China (or any other WTO member) could join simply by agreeing to comply with the terms of the agreement.

Negotiating these megadeals and other regional arrangements within the WTO also would have helped minimize the growing concern that the proliferation of international trade arrangements outside the WTO threatens to undermine the basic nondiscrimination obligation of most-favored treatment that is at the core of the WTO. ¹² Inclusive plurilaterals that are, by definition, MFN would, of course, perpetuate the MFN obligation. Exclusive plurilaterals that are non-MFN would initially be discriminatory, but they need not establish trade discrimination permanently. Rather, they could become ever closer to MFN and eventually fully MFN as more WTO members agree to their terms.

Moreover, had these negotiations been conducted within the WTO and aimed at establishing new plurilateral WTO agreements like the GPA and the ITA, disputes arising under their provisions would be subject to the WTO dispute settlement system. Thus, the parties to those new agreements would have had the benefit of both seasoned trade jurists and a body of trade jurisprudence that will otherwise have to be recreated over time under new and untried dispute settlement systems. As it is, the new mega-arrangements outside the WTO are only reinventing the existing wheel of dispute settlement in international trade.

It may be asked, why would, say, the United States and Europe want to negotiate the TTIP within the WTO? Would they not want an exclusive bilateral arrangement? In answer, why would they? If other WTO members were willing and able to incur the obligations of a TTIP, would it not be to the benefit of both the United States and the European Union to have a broader terrain for their vision of freer mutual trade and further economic integration? And would not the flow of world trade and world investment be enhanced overall if the ambit of such obligations were not merely trans-Atlantic, but transglobal?

An imposing obstacle to non-MFN plurilateral agreements is the need to receive the approval of the WTO membership to include them as plurilateral trade agreements in “Annex 4” of the WTO agreement.¹³ Upon the request of the WTO members that are parties to a plurilateral agreement, the WTO Ministerial Conference “may decide exclusively by consensus to add that agreement to Annex 4.”¹⁴ Securing such a consensus will not be easy under any circumstances; it is certainly not easy in the current circumstances.

Generally, developed countries with advanced economies have been eager to negotiate new WTO obligations to meet new economic needs, even if the agreements reached are exclusive because they are non-MFN. In contrast, developing countries generally resist non-MFN agreements. In addition to their apprehension of new obligations and their focus on getting the benefits of current obligations, they have had little interest in innovations in rule making that do not address their core concerns of agricultural and manufacturing access to the developed world. A number of developing countries, “such as Brazil, China, India, and South Africa have openly expressed their rejection of a plurilateral alternative to the Doha impasse, preferring instead a multilateral approach.”¹⁵

At this point, the developing countries that are resisting allowing new non-MFN plurilateral agreements as part of the WTO legal framework are simply outsmarting themselves. Their resistance has only encouraged developed countries ambitious for freer trade and more economic integration to turn from the WTO to the alternative of bilateral, regional, and megadeals outside the WTO. The adoption of such new obligations outside the WTO will influence global commerce in the same ways that already make developing countries apprehensive, but they will be voiceless to assert their interests, which would not be so if the new obligations were part of the WTO.

Conceivably, megadeals could be negotiated outside the WTO and later become WTO deals. For instance, the members of the TPP could request that it be added to “Annex 4” and thus become a plurilateral trade agreement within the legal framework of the WTO. As with any other non-MFN deal, this would require approval by consensus of the WTO Ministerial Conference. This WTO-centered approach would certainly be preferable to the current approach. As it is, other countries can join the TPP only with the permission of the current parties to the TPP. If the TPP were already within the WTO legal framework, other WTO members could join the TPP just by agreeing to comply with, and be bound by, TPP obligations. This said, building a consensus to bring the TPP within the WTO would be an arduous political task. Better to pursue such mega-ambitions within the WTO in the first place.

The drift away from the WTO in search of such megadeals is decidedly not in the interest of the vast majority of WTO members, including the developing countries that increasingly have their rightful say in the councils of the WTO and that benefit enormously from the centrality of the WTO trading system. A world of competing trading blocs bound by megadeals would not be a world that benefits those countries that are still on the margins of the world economy. Few of the poorer countries in the world will ever be invited to participate in a megadeal, but if such a deal is concluded within the legal structure of the WTO, they will have the right to benefit from that deal if they are willing to accept its obligations.

A continued drift outside the WTO would bolster the conclusion that now is the beginning of the end for the WTO. A return to the WTO would advance the view that now is the end of the beginning.

Will the United States Support or Subvert a Shift to WTO Plurilateralism?

Alas, there is not much to suggest a turn back by the United States to the multilateralism manifested in the WTO. Under Trump, the United States is — so far — still showing up for WTO meetings and engaging in WTO dispute settlement (although mostly defensively). The United States is still making occasional WTO proposals, such as its recent and laudable proposal for more compliance with, and more transparency in, required subsidies notifications. At the same time, the United States is often uncharacteristically silent in WTO committee sessions. The U.S. delegates often cannot speak because they have no clear instructions. The traditional leadership of the United States is missing — and is missed — in the WTO.

Trump’s protectionist and unilateralist trade ambassador, Robert Lighthizer, is hardly a tribune for the WTO. One reason why members failed to agree on a unified statement for the customary concluding declaration at the Buenos Aires ministerial conference, and ended up with no final declaration at all, was reportedly because Lighthizer, on behalf of the United States, insisted on excluding language from the declaration describing the WTO as the center of the multilateral trading system. Like the vast majority of other WTO members, the United States has always insisted in the past on including this statement of mutual allegiance to multilateralism. No more. Not only do Trump and Lighthizer not see the WTO as central to world trade, it is not clear that they see the need for the WTO at all. They see international trade as a win-lose proposition and a zero-sum struggle of all against all. They do not see international trade as a win-win proposition for all who participate in trade, which is the motivating philosophical underpinning of the WTO trading system.

Among the Trump administration’s top trade priorities is to defend U.S. sovereignty over the making of U.S. trade policy, which is often portrayed as a necessary response to what it characterizes as WTO overreaching into the sovereign domain of domestic discretion in policymaking. Unlike past U.S. presidents and administrations of both parties, Trump and those who serve him do not seem to understand the concept of sharing sovereignty as an effective means of international cooperation toward the end of solving common global problems. As for international cooperation on trade, Trump and Lighthizer alike have, on many occasions, expressed antipathy toward the WTO and at times have hinted that withdrawing from the organization might be in the best interests of the United States.

Meanwhile, Trump has repeatedly expressed a preference for bilateral — over global and regional — trade deals, even though he has yet to secure a single negotiating partner for bilateral negotiations. The only one of 35 ongoing negotiations to which the United States is even nominally a party is the TTIP, which is in limbo because of Trump’s concern that it is a regional agreement and because of his lack of attention to advancing it. It is not hard to understand why few governments want to negotiate with the Trump administration, given that the president pulled out of the TPP, continues to threaten to pull the plug on the North American Free Trade Agreement (NAFTA), has imposed and threatened illegal unilateral trade restrictions and appears to have no coherent or consistent positions on trade policy (or, for that matter, on much else).

In the months following the Buenos Aires Conference, the president seemed to open the door to returning to the TPP, but then changed his mind again. His secretary of commerce, Wilbur Ross, is reported to have said that the TTIP negotiations are still alive. Meanwhile, juxtaposed to the administration’s expressions of displeasure with the WTO is its professed interest in continuing to use the WTO dispute settlement system and its intermittent efforts to ensure that U.S. trade laws are being applied in a WTO-consistent manner. Incongruities abound.

Given the president’s fickleness in overall policymaking and his predilection for saying one thing in the morning and another thing in the afternoon, who knows the extent to which he and his administration are committed over the long term to what seems a notable shift in U.S. trade strategy from more open to more constricted trade? Trump and his closest trade advisers rarely seem to think of the long term; they only seem to think of the short term. They also cause confusion with an endless stream of inconsistent statements. Having announced, for instance, in a defiant televised address that he was withdrawing the United States from the Paris climate agreement, Trump has also said that he is open to returning to that agreement. But, assuming this is so, at what price to ongoing global combat against climate change? So, too, with trade. If the Trump administration did return to the negotiating table on these megadeals, what would be the U.S. negotiating approach? If the my-way-or-the-highway tactics of the United States in the renegotiations of the NAFTA and the Korea-U.S. trade agreement are any indication, little might result from a turn by Trump back to the TPP and the TTIP.

All this said, Lighthizer, who is wrong on so much else about trade and the WTO, was right in declaring after his fly-by to Buenos Aires, “Many Members recognized that the WTO must pursue a fresh start in key areas so that like-minded WTO Members and their constituents are not held back by the few Members that are not ready to act.”¹⁶ The Office of the U.S. Trade Representative tweeted that “the new direction of the WTO is set: improving trade through sectoral agreements by like-minded countries.”¹⁷ (The Office of the U.S. Trade Representative confirmed, too, that, in the brave and unbridled new world of social media, its tweets are official statements by the United States of America.) Lighthizer also ventured boldly, “MC11 will be remembered as the moment when the impasse at the WTO was broken.”¹⁸

Reports of the demise of the WTO are premature. Day-to-day, the WTO works smoothly. Because of the enabling global framework of WTO rules, almost all of world trade flows easily and without dispute every day. Because of the existence of the WTO dispute settlement system, almost all WTO members choose to comply with almost all WTO rules almost all the time. Most important, because of this system for resolving disputes, WTO members can resolve their inevitable trade disputes with each other peacefully and according to rules on which all members have previously agreed. This remains true despite the shameful recent attacks by the Trump administration on the WTO dispute settlement system and especially on WTO judges.

Nevertheless, Lighthizer was right in telling members in Buenos Aires that the WTO is “becoming a litigation-centered organization.”¹⁹ The WTO has proven to be proficient at upholding existing rules — an achievement not to be underestimated. But the WTO has not yet proven that it can be equally proficient in agreeing on new rules or on changes in existing rules. Unless members soon learn how to negotiate successfully on trade for the 21st century, the weight of the burden of a backlog of dispute settlement decisions will eventually intensify the dysfunction at the WTO, while its members continue to drift elsewhere to resolve the ever more complex new issues of the global economy.

Toward the End of the Beginning

If it was not clear to WTO members before the ministerial conference in Buenos Aires, it surely should be clear now that if they are going to make progress on trade liberalization any time soon, they will have to do so plurilaterally. With many countries turning inward and with many more increasingly weary of endless global trade negotiations that never seem to produce results, plurilateralism may offer the most promising path to multilateralism in the WTO. Indeed, for now, it may be the only path.

Absent progress within the WTO system, the alternatives are more bilateral, regional, and mega-agreements on trade made outside the sheltering legal framework of the WTO among the more ambitious members. Already, hundreds of trade agreements have been concluded outside the WTO — the vast majority of them since the start of the deadlock in the Doha Development Round. What began as an aberration has become a preoccupation. At present, 35 new bilateral and regional trade pacts are under consideration around the world.²⁰

Looking past the apocalyptic headlines and the disappointments in Buenos Aires, there is scope for the more optimistic view that it is not the beginning of the end, but the end of the beginning for the WTO. The absence of a multilateral outcome in Buenos Aires may have triggered a psychological and tactical shift toward pursuit of plurilateral trade solutions. After years of ambivalence for fear of undermining the ongoing multilateral negotiations, like-minded members may now move forward as willing allies to modernize the rules framework of the WTO. In the absence — for now — of U.S. leadership, the European Union, Japan, China, Canada, Australia, Chile, New Zealand, and others can help fill the leadership void. On different issues, different leaders or combinations of leaders may emerge.

The list of topics that should be addressed by new trade rules is long and getting longer every day. Based on all that was said and done in Buenos Aires, several of the topics on this long list seem to be good prospects for immediate consideration, including digital trade and services.

Digital trade seems to be at the top of the list. When the WTO was established a quarter of a century ago, there was no such thing as digital trade. Thus, there were no WTO rules specifically concerning digital trade. All these years later, there are still no specific WTO rules on digital trade: WTO rules are analog, not digital. And while, overall, the growth in world trade has slowed in recent years, trade in digital goods and services has multiplied 45-fold in the past decade.²¹ In ample demonstration of the death of distance, the algorithms of digital trade are everywhere a potent accelerant for globalization and have changed the conduct of international commerce profoundly. Business-to-business digital commerce is estimated to account for 90 percent of all global electronic commerce.²² A WTO without rules on digital trade is not a “world trade” organization.

In Buenos Aires, a coalition of 71 WTO members (counting the EU member states individually) announced that they will begin exploratory work toward future negotiations on trade-related aspects of electronic commerce. They stressed that they are open to the inclusion of additional WTO members. The first session of their talks is to be held in 2018. Although these members have no negotiating mandate from the WTO, they have said they will conduct their talks within the WTO.²³ Significantly, the United States joined this coalition, with Ambassador Lighthizer saying, “Initiatives like this among like-minded countries offer a positive way forward for the WTO in the future.”²⁴

An issue that should also be near the top of the plurilateral list is services trade. For 15 years, Doha Round negotiators were unable to make any headway on services trade by expanding the scope of the General Agreement on Trade in Services, which is part of the WTO agreement. Ultimately, a group of like-minded countries began negotiating a separate accord on services on the sidelines of the WTO — but not actually in the WTO — that they called the “Trade in Services Agreement.” These negotiations halted after Trump’s election and have not resumed. The willingness of the United States to negotiate on digital trade and the fact that services account for 75 percent of the U.S. economy suggest that Trump and Lighthizer, who so far have focused mainly on manufacturing trade, may be willing to take another look at this critical issue for American workers and businesses and take part in new plurilateral negotiations aimed, ultimately, at a multilateral solution.

Commitments to liberalize some services, such as occupational licensing and legal services, may be difficult to achieve multilaterally, but many others areas are amenable to plurilateral solutions. Of concern to many WTO members in Buenos Aires was finding a way to advance negotiations on domestic regulation of services. If the United States, China, or any other WTO member is unwilling to join new services negotiations, then other like-minded WTO members interested in liberalizing trade in services should simply proceed, where they can, without them.

The WTO ministers agreed in Buenos Aires to continue talking about disciplining fisheries subsidies with hopes of adopting a multilateral agreement by the next ministerial conference in 2019. (In trade negotiations, agreeing to continue to talk is seen as a success; trade negotiators do not understand why the rest of the world is not impressed by such an obvious accomplishment.) The looming deadline on this topic is 2020, when the members of the United Nations — including all 164 members of the WTO — have agreed to discipline subsidies for overcapacity and overfishing and to eliminate subsidies of illegal, unreported, and unregulated fishing.²⁵ Encouragingly, the United States has agreed to continue to participate in negotiations (after rejecting, with its withdrawal from the TPP, the restrictions there on fisheries subsidies, which are much like those being sought in the WTO).²⁶ If these talks reach a multilateral roadblock, then a plurilateral alternative should be pursued within the WTO that could ultimately grow into a fully multilateral solution.

Not to be forgotten are the frustrating negotiations on freeing trade in environmental goods, which are defined by the WTO as “products that can help achieve environmental and climate protection goals, such as generating clean and renewable energy, improving energy and resource efficiency, controlling air pollution, managing waste, treating waste water, monitoring the quality of the environment, and combatting noise pollution.”²⁷ Annual global trade in environmental goods is currently estimated to be nearly $1 trillion and is growing rapidly along with a rising global demand.²⁸ However, tariffs on these products persist and, currently, some WTO members assess duties as high as 35 percent on environmental goods. Liberalizing trade in these products would do much to speed the spread of clean and more-efficient technologies throughout the world, including to the developing countries where they are much needed to promote clean energy.

Negotiations on an Environmental Goods Agreement have been in progress for several years. Having started in 2012 with a list of 54 environmental goods subject to tariffs, some 46 WTO members, accounting for most of the world’s trade in these goods, continue to try to conclude an inclusive plurilateral WTO agreement to eliminate those tariffs and extend the duty-free benefits to all other WTO members on an MFN basis. These negotiations have stalled as the negotiating countries quarrel over precisely which goods are “environmental goods” that should fall within the scope of the agreement. Is a bicycle an “environmental good”? If so, are all kinds of bicycles “environmental goods”? Should we distinguish between a child’s first bicycle with training wheels and a high-performance French racing bicycle? And so on. The negotiating countries have been endlessly creative in constructing arguments for defining virtually everything as “environmental goods.” These negotiations should be resumed and, once an agreement is reached to eliminate barriers to trade in environmental goods, talks should be started on making the new plurilateral agreement an Environmental Goods and Services Agreement by eliminating the barriers to trade in environmental services as well.

Yet another issue ripe for a plurilateral agreement is that of excess capacity for production — in steel especially, but also in some other basic traded products. Global oversupply in these products is depressing world prices, distorting world markets, and consequently intensifying pressures for imposing new unilateral trade restrictions worldwide. Unilateral trade restrictions — such as those imposed by Trump on steel and aluminum — not only violate WTO rules, but also could cause a spiral of global protectionism as countries impose reactive retaliatory measures. Far better to deal with the very real problem of excess capacity by negotiating rules than by a descent into unknown depths of global protectionism.

There are WTO rules to deal with situations of short supply.²⁹ There are, however, no WTO rules to deal with situations of oversupply. When the original 23 contracting parties of the GATT wrote the rules in 1947 in the hungry aftermath of World War II, oversupply was not a trade problem. Now it is, and now WTO rules are needed to help avoid the initiation of what could become a mutually destructive exchange of national trade restrictions and international trade disputes that would grip and, perhaps, paralyze the WTO. Proceeding from the work already being done in the steel sector under the auspices of the G20 group of leading economies, WTO members should negotiate a plurilateral sectoral agreement on steel that could ultimately become a multilateral WTO agreement.

Included in such a sectoral agreement could be guidelines on best practices reminiscent of those in the reference paper to the protocol on basic telecommunications services under the WTO services agreement.³⁰ Along with the agreement itself, those best practices could begin plurilaterally and then, with time, become multilateral. This negotiating approach of setting out best practices could be emulated in other new areas of trade concern, including two areas that drew much attention in Buenos Aires: gender equity and micro-, small-, and medium-sized businesses. The goal of these and other best practices efforts should not necessarily be to create new rules in these areas of concern; rather, it should be to encourage seeing trade policymaking through the lens of these concerns.

Another topic of discussion in Buenos Aires was investment facilitation. A large group of WTO members, comprising both developed and developing countries, endorsed a joint statement there agreeing to start “structured discussions with the aim of developing a multilateral framework on investment facilitation.”³¹ Examples of what an agreement on investment facilitation would contain include strengthened “electronic governance,” such as a “single electronic window” that would publish investment documents and help streamline applications and admissions procedures for incoming investments; a national focal point for mediating and facilitating investor concerns with public authorities; voluntary standards of corporate social responsibility; and guarantees of transparency.³²

Ideally, this new WTO framework on investment facilitation would accompany, and perhaps be an expansion of, the multilateral Trade Facilitation Agreement, which was concluded in Bali in 2013 and is now being phased into full implementation. It, too, could be phased in over time, and it could contain differing obligations for WTO members at different stages of development. Moreover, it could be accompanied by technical assistance. Should WTO members not be able to proceed multilaterally on this topic, then it should be the subject of a WTO plurilateral agreement that could evolve into a fully multilateral pact.

If optimism is to be justified, then these few initiatives must be only the start. The topics that now seem closest to successful plurilateral negotiations are far from the only trade topics that can and should be advanced through this approach. There are many others. Some have been included in innovations in some of the bilateral, regional, and mega-agreements, including regulatory coherence, technical regulations, sanitary and phytosanitary measures, intellectual property protections, disciplines for state-owned enterprises, and trade remedies. Others would address more broadly the role and rules of the WTO in a new world economy facing new opportunities and confronting great challenges extending far beyond traditional trade concerns.

Conclusion

Plurilateral solutions are not the only solutions for the World Trade Organization. Multilateral agreements containing multilateral rules must always be the ultimate goal for the WTO. The logic of multilateralism endures. Indeed, the need to multilateralize international cooperation grows with each passing day. But there is more than one approach to achieving multilateralism. If, after more than two decades of both historic accomplishment and accumulating frustration, now is to be the end of the beginning for the WTO, and not the beginning of the end, then members can no longer afford the illusion that progress consists of merely agreeing to talk or scheduling a meeting or putting a topic on a discussion agenda in Geneva. Progress means getting things done. And that must start now.

Members must begin to negotiate in new ways that will lead to new trade agreements — and soon. If WTO members wait, if they hesitate, if they simply talk without really negotiating, and if they fail to act immediately on their shared realization that new challenges necessitate a new way of doing things, then the next ministerial conference of the WTO in 2019 may be the last one that many involved in trade policymaking will bother to attend.

Notes

1. Alex Lawson, “The WTO May Have Reached Its Breaking Point,” Law 360>, December 14, 2017.
2. Megan Cassella, “It’s the End of the WTO as We Know It-and Trump Feels Fine,” >Politico, December 14, 2017.
3. “The WTO Remains Stuck in Its Rut,” The Economist, December 14, 2017.
4. Article 23.1, WTO Dispute Settlement Understanding.
5. David J. Lynch and Damian Paletta, “Trump Tilts Hard-Line Ahead of Trade Decisions,” Washington Post, February 27, 2018; and Jacob M. Schlesinger, “Trump to Impose Steep Aluminum and Steel Tariffs,” Wall Street Journal, March 2, 2018.
6. “WTO Ministerial: In Landmark Move, Country Coalitions Set to Advance on New Issues,” Bridges Special Update, December 13, 2017, p. 2.
7. Bjorn Lomborg, “Promises to Keep: Crafting Better Development Goals,” Foreign Affairs 93, no. 6 (November/December 2014): 133.
8. “The Marrakesh Agreement Establishing the World Trade Organization,” Marrakesh, Morocco, April 15, 1994 (the “Marrakesh Agreement”), 1n.
9. Article 3.2, WTO Dispute Settlement Understanding.
10. Article X.3, Marrakesh Agreement.
11. This statement is based on my personal recollections from the time. As one of the six original cosponsors of the implementing legislation for the Uruguay Round trade agreements in 1994, when I was a member of the U.S. House of Representatives, I held this expectation, and my firm recollection is that many others-especially Americans-held it as well.
12. See Jagdish Bhagwati, Termites in the Trading System: How Preferential Agreements Undermine Free Trade (Oxford: Oxford University Press, 2008).
13. Article X, Marrakesh Agreement.
14. Article X.9, Marrakesh Agreement.
15. Peter Draper and Memory Dube, “Plurilaterals and the Multilateral System,” International Centre for Trade and Sustainable Development and World Economic Forum, December 2013, p. 3.
16. Office of the U.S. Trade Representative, “USTR Robert Lighthizer Statement on the Conclusion of the WTO Ministerial Conference,” press release, December 2017, https://ustr.gov/about-us/policy-offices/press-office/press-releases/2017/december/ustr-robert-lighthizer-statement.
17. USTR (@USTradeRep), “Congratulations to DG @WTODGAZEVEDO and Minister @SusanaMalcorra on a successful #MC11. The new direction of the WTO is set: improving trade through sectoral agreements by like-minded countries,” Twitter, December 13, 2017, 2:09 p.m., https://twitter.com/USTradeRep/status/941067574267269122.
18. Office of the U.S. Trade Representative, “USTR Robert Lighthizer Statement on the Conclusion of the WTO Ministerial Conference,” press release, December 2017, https://ustr.gov/about-us/policy-offices/press-office/press-releases/2017/december/ustr-robert-lighthizer-statement.
19. Office of the U.S. Trade Representative, “Opening Plenary Statement of USTR Robert Lighthizer at the WTO Ministerial Conference,” press release, December 2017, https://ustr.gov/about-us/policy-offices/press-office/press-releases/2017/december/opening-plenary-statement-ustr.
20. Ana Swanson and Jim Tankersley, “As US Trumpets ‘America First,’ Rest of the World Is Moving On,” >New York Times, January 25, 2018.
21. Rana Forooher, “Trump’s Trade Policies Won’t Help My Town,” >Financial Times, March 6, 2017.
22. United Nations Conference on Trade and Development, “Information Economy Report,” 2015.
23. Luis Gil Abinader and Gus Rossi, “The WTO E-Commerce Agenda after the Buenos Aires Ministerial,” >Public Knowledge (blog), December 22, 2017, https://www.publicknowledge.org/news-blog/blogs/the-wto-e-commerce-agenda-after-the-buenos-aires-ministerial.
24. “USTR Robert Lighthizer on the Joint Statement on Electronic Commerce,” U.S. Trade Representative, press release, December 12, 2017.
25. See Goal 14 and Target 14.6 in United Nations, “Transforming Our World: The 2030 Agenda for Sustainable Development,” October 21, 2015, http://www.un.org/ga/search/view_doc.asp?symbol=A/RES/70/1&Lang=E.
26. “USTR Says It Will Work with EU, Others to Advance Disciplines on Fisheries Subsidies,” >World Trade Online, December 13, 2017.
27. World Trade Organization, “Environmental Goods Agreement (EGA),” World Trade Organization, “Environmental Goods Agreement (EGA),” https://www.wto.org/english/tratop_e/envir_e/ega_e.htm.
28. Office of U.S. Trade Representative, “Environmental Goods Agreement,” https://ustr.gov/trade-agreements/other-initiatives/environmental-goods-agreement.
29. Article XI:2(a), General Agreement on Tariffs and Trade.
30. Reference Paper, Protocol on Basic Telecommunications Services, General Agreement on Trade in Services.
31. “WTO Ministerial: In Landmark Move, Country Coalitions Set Plans to Advance on New Issues,” p. 2.
32. “Brazil Circulates Proposal for WTO Investment Facilitation Deal,” >Bridges 22, no. 4 (February 8, 2018), https://www.ictsd.org/bridges-news/bridges/news/brazil-circulates-proposal-for-wto-investment-facilitation-deal.

Corey A. DeAngelis

Is public schooling a public good, a merit good, or a demerit good? Public schooling fails both conditions specified in the standard economic definition of a public good. In order to place public schooling into one of the remaining two categories, I first assess all of the theoretical positive and negative externalities resulting from public schooling as opposed to publicly financed universal school vouchers. Then, in an original contribution to the literature, I quantify the magnitude and sign of the net externality of government schooling in the United States using the preponderance of the most rigorous scientific evidence.

While the counts of theoretical positive and negative externalities are about equal, the empirical evidence leads me to estimate that public schooling in the United States has a net negative externality of at least $1.3 trillion—over the lifetime of the current cohort of children in government schools—relative to publicly funded universal school vouchers. I conclude with three policy recommendations: (1) the U.S. government should not operate schools at the local, state, or federal level on the basis of schooling’s being a public good; (2) U.S. citizens should not fund government schooling indirectly through the tax system on the basis of schooling being a merit good; and (3) the United States should instead fund education directly—rather than schooling—through a universal Education Savings Account (ESA) program.

Introduction

“No schooling was allowed to interfere with my education.”
— Grant Allen, Rosalba: The Story of Her Development, p. 101

Horace Mann, often called the father of American public schooling, and others argued that government-run common schools were necessary to bring together children from diverse backgrounds and to inculcate a uniform set of American values that would contribute to a stable and cohesive democratic society.¹ In common schools, children from all backgrounds could learn how to interact with one another and become proper citizens.² Mann traveled to Prussia to examine its system of common schools in 1843.³ He helped pass the first modern compulsory schooling attendance law in the United States in his home state of Massachusetts in 1852.⁴ Within seven decades, every state had followed suit; Mississippi was the last state to pass a compulsory schooling attendance law in 1918.⁵

Taxpayer-funded and government-run schools exist in all 50 states. This likely is attributable to many people with good intentions, like Mann, thinking that common schools could improve society overall.⁶ In general, a better-educated populace should result in positive social effects, all else being equal.

However, there are opportunity costs to maximizing education. For example, someone who pursues 10 college degrees may achieve a well-rounded and advanced education without contributing much to other individuals in society.⁷ And, of course, schooling and education are not one and the same. The formal definition of “education” is “the act or process of imparting or acquiring general knowledge, developing the powers of reasoning and judgment, and generally of preparing oneself or others intellectually for mature life.”⁸ Since schooling is but one channel available for an individual to acquire an education, it is important for the current study to examine the externalities of the actual policy in place in the United States — traditional public schooling — rather than some ideal policy that could hypothetically increase education for all children.

Schooling as a Public Good

The economic argument for government using coercion to fund — and even operate — a specific good or service is strongest for a good or service deemed to be a “public good.” The formal definition of a public good is attributed to Nobel laureate economist Paul Samuelson. In a classic 1954 article he explained that such a good satisfies two necessary conditions: (1) it is nonexcludable, and (2) it is nonrivalrous in consumption.⁹

The nonexcludability provision means that the producer cannot prevent nonpayers from using the good without bearing costs that exceed the benefit of payment. This provision is important because nonexcludability leads to a potential free-rider problem: individual consumers can enjoy the benefits of a product without directly paying for it. Consequently, the market may underprovide the good in question, or even fail to provide it at all. A feasible policy solution is to provide and produce the good publicly. In other words, the free-rider problem could be eliminated if all members of society were forced to pay for the service indirectly through taxes.

The nonrivalry provision simply means that one individual’s consumption of the good does not diminish the abilities of others to consume it. A radio station can be thought of as a true public good. Because it would be extremely difficult to prevent anyone with a radio from listening, the good is considered nonexcludable. And because one person’s consumption of the service does not affect whether the rest of society can listen, the radio is considered nonrivalrous. One policy implication could be to have taxes fund local radio stations. However, the market avoids the potential free-rider problem with radio stations by using advertisements as a funding source.

If schooling were indeed a public good, there would perhaps be a stronger economic argument for government funding and operation of schools. However, schooling easily fails both parts of the economic definition. If one student occupies a seat in a classroom, another child is prevented from sitting in the same seat. In addition, if students are added to a given classroom, the teacher is less able to tailor the educational approach to each child, which could reduce the average amount of personalized education received by each student. Because of this, schooling fails the nonrivalrous part of the definition. Second and perhaps most important, because it is not difficult to exclude a person from a school — or any other type of institution with walls — schooling fails the nonexcludability condition. If someone does not pay me to educate the student, I can simply deny the student services. Fortunately, schools will never suffer from a true free-rider problem because they are not true public goods. That is precisely why private schools and tutoring services operate effectively today without government operating or funding them.

Schooling as a Merit Good

When people, including prominent education scholars, say that schooling is a public good, I believe they mean that schooling is “good for the public.”¹⁰ Or, as an economist would say, schooling is a “merit good” because it has net positive externalities.¹¹ An economic externality occurs whenever a voluntary transaction between two parties affects an involuntary third party in a positive or negative way. The original argument regarding economic externalities has to do with pollution — a negative externality.¹² When I buy a car from a factory, the car manufacturer benefits from the transaction because it gets my money, and I benefit from the transaction because I get a car. Because the transaction is voluntary (and not coerced), it would only occur if both parties perceived that expected benefits would exceed expected costs.¹³ However, the rest of society could be involuntarily harmed by the transaction because they must breathe air that is less clean. Consequently, the market may produce a number of automobiles that is higher than the socially optimal level, where total social costs exceed total social benefits. As Arthur Pigou pointed out, one way to internalize the negative externality of pollution is to reduce consumption of automobiles toward the socially optimal level by taxing each unit of production — what is now called a Pigouvian tax.¹⁴

In the case of education, the externality is expected to be positive, which would make education a merit good. If I purchase an education through a school or otherwise, I benefit from the transaction because I will be able to command a higher salary in the future, and I will feel good about being an educated citizen. The education provider benefits from the transaction financially. And the rest of society is better off because of the benefits I provide to society as a result of my education, but for which I don’t earn a market income. For instance, the educational blogs, lectures, and journal articles I post for free on the internet help society (I hope). Also, as an educated citizen, I am less likely to break the law and more likely to cast an informed vote on Election Day.¹⁵ According to economists, leaving education to purely private transactions would result in education falling below the socially optimal level.¹⁶ A feasible policy solution to move education levels up is a negative Pigouvian tax, also known as a Pigouvian subsidy. As Nobel laureate economist Milton Friedman concluded, government may have a role in funding schooling because of the theoretical positive externalities — or “neighborhood effects” — of education in general.¹⁷

While education itself seems to have net positive externalities, the case is less clear for the system of traditional public schooling we have in the United States today. After all, if the traditional public schooling system is reducing overall levels of education, or producing education very inefficiently, it would be considered a demerit good — a good that has net negative externalities. In this analysis, I examine all the theoretical externalities around the traditional public schooling system in the United States today.

In addition, I make the first attempt to calculate the net externality of traditional public schools relative to a realistic counterfactual: a private school of choice that could accept the public school’s per pupil funding amount as full payment for tuition and fees. Of course, the comparisons made here are not between traditional public schools and no schooling at all. If families were not forced to allocate 100 percent of their publicly raised educational resources to their assigned public schools, they could take those same funds to schools of their choosing. The three externalities that I examine are (1) an educated populace, (2) taxpayer costs, and (3) social cohesion.

Existing Literature

The most rigorous and relevant literature that we have comparing traditional public schools to private alternatives are analyses of private school choice programs. The best-known type of private school choice program, championed by Milton Friedman, uses vouchers that allow families to take their publicly raised education funds to the school of their choice.¹⁸ When parental demand for educational vouchers exceeds the supply of voucher funding, random lotteries are typically used to determine which families are able to exercise private school choice. The lottery setting allows social scientists to experimentally evaluate the effects of access to private school choice programs — and the effects of private schooling in general — on students. Since random chance determines who gets access to the program, the only difference between treatment and control groups is that one group received access to a private school choice program. Because several experimental evaluations exist on the effects of private school choice programs on student achievement, I exclude less rigorous studies that are not able to establish causal relationships from this review. For example, the empirical methodologies used in the 2013 book by Lubienski and Lubienski did not allow the authors to make causal claims because they simply examined the association between school type and math test scores after controlling for some observable characteristics such as race and gender.¹⁹

Educated Populace

Society benefits from a better-educated populace because individuals are more likely to interact with people who could teach them something new. In addition, better-educated citizens may produce high-quality goods and services that benefit the rest of society. For example, when a hard-working individual completes medical school, he or she benefits the rest of society by providing valuable services. The relevant positive externality can be thought of as the extent to which productive abilities are increased by the policy alternative (i.e., private school choice vs. residentially assigned public schooling).

A meta-analytic and systematic review of 19 experimental voucher studies around the world finds that, on average, private schools increase math scores by 15 percent of a standard deviation and reading scores by 27 percent of a standard deviation.²⁰ Out of the 17 voucher experiments in the United States, 11 find statistically significant positive test-score effects for some or all students, four find no statistically significant effects, while two find negative effects.²¹ The meta-analysis from 16 of the U.S. experimental studies finds that, on average, private schooling does not have a statistically significant effect on reading scores, but it increases math scores by around 7 percent of a standard deviation.²²

The scientific evidence on longer-term educational outcomes such as high school graduation rates is less abundant. Foreman’s summary of three rigorous studies linking private school choice programs to high school graduation finds positive effects.²³ The only U.S. experiment on the subject finds that attending a private school through the D.C. Opportunity Scholarship Program increased the likelihood of high school graduation by 21 percentage points.²⁴ The one quasi-experimental study on the subject finds that attending a private school using the Milwaukee Parental Choice Program increases the likelihood of high school graduation by 3 percentage points.²⁵ The final study included in the review finds that Milwaukee private schools graduate voucher students at a rate 12 percentage points higher than Milwaukee public schools; however, this study is merely observational.²⁶

Taxpayer Costs

In theory, all taxed funds are a negative externality if taxed individuals do not consent to the transaction. If citizens refuse to pay taxes, they must gain citizenship elsewhere or go to jail, both of which come with extraordinarily high transaction costs. Nonetheless, this analysis takes a conservative approach by comparing the taxpayer costs associated with traditional public schools to the policy-relevant counterfactual: the taxpayer costs incurred from a private school choice program.

We can examine the taxpayer effects of private school choice programs by looking at how current school choice laws affect statewide educational funding formulas. As shown in Forster’s review of the evidence, 25 out of 28 studies find that private school choice programs save taxpayer money, while 3 studies find no statistically significant fiscal effects.²⁷ Spalding finds that 10 voucher programs in the United States generated a cumulative savings of at least $1.7 billion between 1990 and 2011.²⁸ Since the 2016 Forster review, all other fiscal impact studies of private school choice programs that I know of have found taxpayer savings.²⁹

This savings happens for two main reasons: (1) school voucher laws usually mandate that the voucher amount must be a fraction of the total per pupil expenditure in traditional public schools; and (2) private school tuition fees are often below the state-mandated maximum voucher funding amount.

As shown by EdChoice, the average state-funding amount allocated toward voucher students is around 59 percent of the per pupil funding in traditional public schools.³⁰

Social Cohesion

A given educational setting can result in positive externalities if it results in a more cohesive society. An improved education could strengthen the character skills necessary to follow the law and tolerate the views of others. Furthermore, an educational setting can improve social cohesion through increasing racial diversity and integration. If someone is less likely to break the law because of character education, that person will be less likely to steal from others in society, and if someone is more tolerant of others, that person will be more likely to interact with society peacefully. Finally, if children grow up around diverse populations of students, they may be more likely to get along with people from different backgrounds as adults.

As shown in a review of 11 experimental and quasi-experimental studies, DeAngelis finds that private school choice programs in the United States increase these types of civic outcomes.³¹ None of the studies reviewed find negative effects. The only study linking private school choice to adult criminal behavior finds that the Milwaukee Parental Choice Program leads to a 7 percentage point reduction in felonies and a 6 to 9 percentage point reduction in misdemeanors for male students.³²

DeAngelis also finds that effects of private school choice are null to positive for tolerance of others, positive on charitable giving, positive on volunteering, and null to positive on political participation.³³ Wolf’s review of 21 quantitative studies similarly finds that private school choice increases civic outcomes overall.³⁴ Forster’s review of the empirical evidence also finds that private school choice in the United States has null to positive effects on civic values and practices.³⁵ Nine out of the 10 quantitative studies linking private school choice to racial integration find statistically significant positive effects, while one study finds no effects.³⁶ Notably, Egalite, Mills, and Wolf find that, by using the Louisiana Scholarship Program, 82 percent of student transfers increased racial integration for their former public schools and 45 percent of student transfers improved racial integration in their new private school.³⁷

According to the existing evidence, government schooling appears to have negative effects on society through a less-educated populace, higher taxpayer burden, less tolerance, more crime, and racial segregation. A vote count of the evidence can be found in Table 1 below:

Table 1: Government-schooling externalities and their signs

Note:“Null” indicates that the preponderance of the evidence suggests that government schools do not have statistically different effects on society than private schools of choice. “Negative” indicates that the preponderance of the scientific evidence suggests that government schools produce socially less-desirable outcomes than do private schools of choice.

Data and Analysis

Using data from the Digest of Educational Statistics, I can quantify public schooling externalities associated with a less-educated populace, a larger taxpayer burden, and less social cohesion relative to publicly financed universal school vouchers. Specifically, the data allow me to quantify the externalities associated with changes in test scores, high school graduation rates, taxpayer funding, and criminal activity.

Some 50.477 million children are expected to be enrolled in public elementary and secondary schools in the United States in the 2017-2018 school year.³⁸ This population is relevant for my calculation of nationwide externalities of public schooling.

Educated Populace

For the societal effects of government schooling’s ability to educate the populace, I examine two outcomes: test scores and high school graduation. Overall, Shakeel, Anderson, and Wolf find that private school choice programs increase reading scores by 4 percent of a standard deviation and math scores by 7 percent of a standard deviation.³⁹ Consequently, I estimate one model based on reading scores and the other based on math scores. However, the effect on reading scores is not statistically significant, so the externality associated with an educated populace is zero in the first model.

For math scores, I follow previous research linking standardized effect sizes with estimates found by Eric Hanushek.⁴⁰ Hanushek estimates that a one-standard-deviation increase in student cognitive ability leads to a 13 percent increase in lifetime earnings. Additionally, only 70 percent of learning gains are retained from year to year.⁴¹ By multiplying those two estimates together, I can find the learning gains relative to the average U.S. worker.⁴² I use Bureau of Labor Statistics data to find average earnings for U.S. employees ($49,630) and assume that current students will work between the ages of 25 and 70, or 46 years.⁴³ When I calculate the net present value of lifetime earnings, I assume a 1 percent yearly growth in average salaries and a 3 percent annual discount rate. Based on these assumptions, the net present value of lifetime earnings for the average U.S. worker coming from the public school system is $1,234,957. Using Hanushek’s estimates, the average lifetime earnings for U.S. students with access to 13 years of private school choice is $1,341,225.

Thus, the reduction in lifetime earnings for each student experiencing 13 years
of government schooling is $106,268 ($1,341,225 - $1,234,957). Multiplying this result by the number of students in government schools reveals an overall negative effect on lifetime earnings of $5.364 trillion ($106,268 × 50.477 million). Of course, one can argue that the lower amount of earnings is accrued to the individual rather than the rest of society. However, the decrease in earnings reflects a $5.364 trillion (in 2017 dollars) reduction in production within society overall. Since the lower level of production results from a less-educated populace and harms the rest of society as a whole, it is a negative externality of government schooling.

Alternatively, I can calculate this particular externality through the effects of private school choice programs on graduation rates. While the experimental study in Washington, D.C., finds that private schooling increases the likelihood of graduation by 21 percentage points, I use the much less substantial 3 percentage point increase in graduation rates found in the Milwaukee voucher analysis in order to provide a conservative estimate.⁴⁴ I also use evidence from Levin, finding that each high school graduate produces around $277,000 (in 2017 dollars) in social benefits derived from additional tax revenues and reductions in health, crime, and welfare costs.⁴⁵ Combining findings from Cowen (et al.) and Levin, I find that government schooling results in about 1,514,310 fewer high school graduates (50.477 million U.S. students multiplied by a 3 percentage point reduction in likelihood of graduation). This reduction leads to negative social effects of around $419.464 billion (1,514,310 fewer graduates multiplied by $277,000 in social costs each).⁴⁶

Taxpayer Costs

There are two ways to calculate the effects on taxpayers of government schooling relative to private schools of choice. First, I use data from EdChoice showing that the average state-funding amount allocated to voucher students is around 59 percent of the per pupil funding in traditional public schools.⁴⁷ According to the National Center for Education Statistics for 2013-2014, public education spending was around $625.016 billion in 2014 dollars. That is equivalent to around $656.019 billion in 2017 dollars. Multiplying this amount by the 59 percent found by EdChoice suggests that these students would cost $387.051 billion to educate, or around $268.968 billion less than in public schools. In other words, over the course of 13 years of k-12 schooling, the 50.477 million children in U.S. public schools would cost taxpayers an additional $3.497 trillion.

Second, I could compare the average tuition and fees charged in all private schools to the average per pupil expenditure in all public schools. According to the Digest of Education Statistics Table 205.50, average private school tuition was around $10,740 per student in 2011-2012, or around $11,633 in 2017 dollars. According to the Digest of Education Statistics Table 236.60, average public school per pupil expenditure was $11,991 in 2011-2012, or around $12,988 in 2017 dollars. In other words, it costs around $1,355 more ($12,988 - $11,633) to educate a child in a government school each year, on average. Over 13 years, this costs society an additional $17,615 per child. This costs taxpayers an additional $889.152 billion for 50.477 million children. This estimate is only about one-fourth the size of the taxpayer cost estimate in the previous paragraph because this estimate uses the average tuition level of all private schools rather than the tuition level of current private schools of choice in the United States.

Social Cohesion

This section is limited to the effects of government schooling on the future criminal activity of students because it is infeasible to quantify the effects of tolerance, political participation, and racial segregation on society overall. The only quasi-experimental study linking private school choice to crime finds that private schools reduce the likelihood that male students will commit felonies by 4 percentage points in Milwaukee.⁴⁸ Assuming these benefits only accrue to about half of the 50.477 million U.S. students (the males), we should expect around 1.01 million fewer felons. McCollister, French, and Fang find that the social cost of a felony is around $23,242 in 2017 dollars.⁴⁹ Thus, a 1.01 million increase in the number of felons, produced by government schools, leads to around a $23.474 billion increase in social costs. In order to provide conservative estimates, this analysis ignores the positive effects of the Milwaukee voucher program on reducing misdemeanors.

Table 2 : Conservative estimates of government-schooling externalities

Table 3 : Alternative estimates of government-schooling externalities

Overall Results

The most conservative estimates of the externalities of government schooling in the United States can be found in Table 2, while alternative estimates can be found in Table 3 below. The results in Table 2 are more conservative because (1) they assume zero benefits accrue from the positive effects found for math achievement; and (2) they assume that the per pupil funding amount would be equal to the average private school tuition level rather than the average amount currently spent on a private school of choice. As shown in Table 2, this study reveals a net negative externality of government schooling of at least $1.331 trillion. This is a significant effect, as it is over 7 percent of the nation’s entire gross domestic product (GDP) recorded in 2016.⁵⁰ Notably, this likely is a lower bound of the actual effect, as I have no monetized values for the social harms from less tolerance, political participation, and racial integration.

Table 3 indicates a net negative externality of around $9.303 trillion. By comparison, this would be equivalent to about half of the U.S. GDP in 2016. However, these estimates should be treated with caution because they combine the calculations of externalities from two academic outcomes, test scores and graduation rates. Nonetheless, even if the effect derived from changes in graduation rates is excluded from this model, the overall negative externality is $8.884 trillion, still about half of the 2016 U.S. GDP.

Conclusion and Policy Implications

Since schooling fails both the nonrivalry and nonexcludability conditions, there is no strong argument for government operation of schooling on the basis of the service being a public good.⁵¹ While public schooling is certainly not a public good, it may be “good for the public” if it increases overall education levels without any unintended consequences. Even Milton Friedman claims that, because schooling may be an economic merit good, a valid argument may be made for government funding of schools.⁵²

However, because public schooling may not maximize one’s education, it may have significant negative externalities relative to a universal voucher program for schools of choice. Indeed, the results of this study suggest that publicly funded government-run schools, relative to private schools of choice, have substantial negative effects on U.S. society overall associated with a less-educated populace, less social cohesion, and increased taxpayer burdens. In 2017, the most conservative model finds a net negative externality of government schooling of around $1.331 trillion, over 7 percent of the U.S. GDP recorded in 2016, while the alternative specification finds that public schooling results in a net negative externality of about half of U.S. GDP in 2016. Note that these are lifetime estimates of the effects of government schools on 50.477 million children relative to whether they would have attended private schools of choice for 13 years in the United States.

Since government schooling in the United States results in a net negative externality relative to private schools of choice, we should not subsidize government schooling based on the economic argument that it is a merit good. According to the evidence, we should eliminate the negative externalities of government schooling by allowing families to reallocate their educational resources to the private schools that best serve their children. Specifically, states should pass legislation to enact universally accessible Education Savings Accounts (ESAs) to allow families to customize their children’s educational experiences. An ESA would allow society to educate children — rather than simply school them — by allowing parents to allocate education dollars toward various educational services such as schooling, tutoring, online instruction, textbooks, and even college costs. In addition, a universal program may provide the demand necessary for market entry. Market entry and competitive pressures could improve the diversity and quality of educational options available to children while reducing average educational costs.

Of course, not all school choice programs are created equal. Recent studies find that highly regulated private school choice programs can reduce school quality.⁵³ In addition, regulation of private school choice programs could result in more negative externalities by incentivizing existing private schools to operate like current government schools.⁵⁴ In order to reduce the externalities associated with government schooling, we should allow private schools to continue their specialized approaches by reducing the quantity and intensity of regulations linked to private school choice program funding.

Notes

1. Horace Mann, Lectures on Education (Boston: L.N Ide, 1855); Benjamin Rush, “Thoughts upon the Mode of Education Proper in a Republic,” in Essays on Education in the Early Republic, ed. Frederick Rudolph (Cambridge, MA: Harvard University Press, 1965); and John Dewey, Democracy and Education (New York: Macmillan Company, 1916).
2. Mann, Lectures on Education.
3. Karl E. Jeismann et al., German Influences on Education in the United States to 1917 (Cambridge: Cambridge University Press, 2006), pp. 21-41.
4. Forest C. Ensign, Compulsory School Attendance and Child Labor: A Study of the Historical Development of Regulations Compelling Attendance and Limiting the Labor of Children in a Selected Group of States (Iowa City, IA: Athens Press, 1921); and Michael B. Katz, The Irony of Early School Reform: Educational Innovation in Mid-Nineteenth Century Massachusetts (New York: Teachers College Press, 1968).
5. William M. Landes and Lewis C. Solmon, “Compulsory Schooling Legislation: An Economic Analysis of Law and Social Change in the Nineteenth Century,” Journal of Economic History 32, no. 1 (1972): 54-91.
6. Amy Gutmann, Democratic Education (Princeton, NJ: Princeton University Press, 1999).
7. Bryan Caplan, The Case against Education: Why the Education System Is a Waste of Time and Money (Princeton, NJ: Princeton University Press, 2018).
8. Dictionary.com, “Education,” http://www.dictionary.com/browse/education.
9. Paul A. Samuelson, “The Pure Theory of Public Expenditure,” Review of Economics and Statistics 36, no. 4 (1954): 387-89.
10. Henry M. Levin, “Education as a Public and Private Good,” Journal of Policy Analysis and Management 6, no. 4 (1987): 628-41; and Chris Lubienski, “Whither the Common Good? A Critique of Home Schooling,” Peabody Journal of Education 75, no. 1-2 (2000): 207-32.
11. Richard A. Musgrave, “A Multiple Theory of Budget Determination,” FranzArchiv/Public Finance Analysis (1956/1957): 333-43.
12. R. H. Coase, “The Problem of Social Cost,” Journal of Law and Economics 56, no. 4 (2013): 837-77; and Arthur C. Pigou, The Economics of Welfare (London: Macmillan and Company, 1920).
13. Richard A. Epstein, Free Markets under Siege: Cartels, Politics, and Social Welfare (Stanford, CA: Hoover Institution, 2008).
14. Pigou, The Economics of Welfare.
15. Lance Lochner and Enrico Moretti, “The Effect of Education on Crime: Evidence from Prison Inmates, Arrests, and Self-Reports,” American Economic Review 94, no. 1 (2004): 155-89; and Andre Blais et al., “Where Does Turnout Decline Come From?,” European Journal of Political Research 43, no. 2 (2004): 221-36.
16. Musgrave, “A Multiple Theory of Budget Determination,” pp. 333-43.
17. Milton Friedman, “The Role of Government in Education,” in Economics and the Public Interest, ed. Robert A. Solo (New Brunswick, NJ: Rutgers University Press, 1955), pp. 123-44.
18. Friedman, “The Role of Government in Education,” pp. 123-44.
19. Christopher Lubienski and Sarah Lubienski, The Public School Advantage: Why Public Schools Outperform Private Schools (Chicago: University of Chicago Press, 2013).
20. M. Danish Shakeel et al., “The Participant Effects of Private School Vouchers across the Globe: A Meta-Analytic and Systematic Review,” University of Arkansas Department of Education Reform Working Paper No. 2016-07, May 2016, https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2777633.
21. Kaitlin P. Anderson and Patrick J. Wolf, Evaluating School Vouchers: Evidence from a Within Study Comparison, University of Arkansas Department of Education Reform Working Paper No. 2017-10, April 2017; John Barnard et al., “Principal Stratification Approach to Broken Randomized Experiments: A Case Study of Vouchers in New York City,” Journal of American Statistical Association 98, no. 462 (2003): 299-323; Joshua M. Cowen, “School Choice as a Latent Variable: Estimating the ‘Complier Average Causal Effect’ of Vouchers in Charlotte,” Policy Studies Journal 36, no. 2 (2008): 301-15; Jay P. Greene, “Vouchers in Charlotte,” Education Next 1, no. 2 (2001); Jay P. Greene et al., “Effectiveness of School Choice: The Milwaukee Experiment,” Education and Urban Society 31, no. 2 (1999): 190-213; William G. Howell et al., “School Vouchers and Academic Performance: Results from Three Randomized Field Trials,” Journal of Policy Analysis and Management 21, no. 2 (2002): 191-217; Hui Jin et al., “A Modified General Location Model for Noncompliance with Missing Data: Revisiting the New York City School Choice Scholarship Program Using Principal Stratification,” Journal of Educational and Behavioral Statistics 35, no. 2 (2010): 154-73; Cecilia E. Rouse, “Private School Vouchers and Student Achievement: An Evaluation of the Milwaukee Parental Choice Program,” Quarterly Journal of Economics 113, no. 2 (1998): 553-602.; Patrick J. Wolf et al., “School Vouchers and Student Outcomes: Experimental Evidence from Washington, DC,” Journal of Policy Analysis and Management 32, no. 2 (2013): 246-70; Eric Bettinger and Robert Slonim, “Using Experimental Economics to Measure the Effects of a Natural Educational Experiment on Altruism,” Journal of Public Economics 90, no. 8-9 (2006): 1625-48; Marianne P. Bitler et al., “Distributional Effects of a School Voucher Program: Evidence from New York City,” National Bureau of Economic Research Working Paper No. 19271, July 2013; Alan B. Krueger and Pei Zhu, “Another Look at the New York City School Voucher Experiment,” American Behavioral Scientist 47, no. 5 (2004): 658-98; Jonathan N. Mills and Patrick J. Wolf, “The Effects of the Louisiana Scholarship Program on Student Achievement after Three Years,” Louisiana Scholarship Program Evaluation Report #7, June 2017; Atila Abdulkadiroglu et al., “Free to Choose: Can School Choice Reduce Student Achievement?,” American Economic Journal 10, no. 1 (2018): 175-206; and Mark Dynarski et al., “Evaluation of the DC Opportunity Scholarship Program: Impacts after One Year,” National Center for Education Evaluation and Regional Assistance report, June 2017.
22. Shakeel et al., “The Participant Effects of Private School Vouchers across the Globe: A Meta-Analytic and Systematic Review.”
23. Leesa M. Foreman, “Educational Attainment Effects of Public and Private School Choice,” Journal of School Choice 11, no. 4 (2017): 642-54.
24. Wolf et al., “School Vouchers and Student Outcomes: Experimental Evidence from Washington, DC.”
25. Joshua M. Cowen et al., “School Vouchers and Student Attainment: Evidence from a State-Mandated Study of Milwaukee’s Parental Choice Program,” Policy Studies Journal 41, no. 1 (2013): 147-68.
26. John R. Warren, “Graduation Rates for Choice and Public School Students in Milwaukee, 2003-2009,” School Choice Wisconsin, January 2011.
27. Greg Forster, “A Win-Win Solution: The Empirical Evidence on School Choice,” EdChoice report, 2016, https://www.edchoice.org/research/win-win-solution/.
28. Jeff Spalding, “The School Voucher Audit: Do Publicly Funded Private School Choice Programs Save Money?,” EdChoice report, September 2014, https://www.edchoice.org/research/the-school-voucher-audit/.
29. Martin F. Lueken, “The Tax-Credit Scholarship Audit: Do Publicly Funded Private School Choice Programs Save Money?,” EdChoice report, 2016, https://www.edchoice.org/research/tax-credit-scholarship-audit/; Corey A. DeAngelis and Julie R. Trivitt, “The Fiscal Effect of Eliminating the Louisiana Program on State Education Expenditures,” University of Arkansas Department of Education Reform Working Paper no. 2016-06, August 2016; Julie R. Trivitt and Corey A. DeAngelis, “State and District Fiscal Effects of a Universal Education Savings Account Program in Arkansas,” University of Arkansas Department of Education Reform Working Paper no. 2017-04, March 2017; and Julie R. Trivitt and Corey A. DeAngelis, “State Fiscal Impact of the Succeed Scholarship Program 2016-2017,” University of Arkansas Department of Education Reform Policy Brief, March 2017, http://www.uaedreform.org/state-fiscal-impact-of-the-succeed-scholarship-program-2016-2017/.
30. “School Choice in America,” School Choice in America Dashboard, https://www.edchoice.org/school-choice/school-choice-in-america/#.
31. Corey A. DeAngelis, “Do Self-Interested Schooling Selections Improve Society? A Review of the Evidence,” Journal of School Choice 11, no. 4 (2017): 546-58.
32. Corey A. DeAngelis and Patrick J. Wolf, “The School Choice Voucher: A ‘Get Out of Jail’ Card?,” University of Arkansas Department of Education Reform Working Paper no. 2016-03, March 2016, https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2743541.
33. DeAngelis, “Do Self-Interested Schooling Selections Improve Society? A Review of the Evidence.”
34. Patrick J. Wolf, “Civics Exam,” Education Next 7, no. 3 (2007).
35. Forster, “A Win-Win Solution: The Empirical Evidence on School Choice.”
36. Forster, “A Win-Win Solution.”
37. Anna J. Egalite et al., “The Impact of Targeted School Vouchers on Racial Stratification in Louisiana Schools,” Education and Urban Society 49, no. 3 (2016): 271-96.
38. “Enrollment in Elementary, Secondary, and Degree-Granting Postsecondary Institutions, by Level and Control of Institution: Selected Years, 1869-70 through Fall 2025,” Digest of Education Statistics, https://nces.ed.gov/programs/digest/d15/tables/dt15_105.30.asp?current=yes.
39. Shakeel et al., “The Participant Effects of Private School Vouchers across the Globe: A Meta-Analytic and Systematic Review.”
40. Patrick J. Wolf et al., “The Productivity of Public Charter Schools,” School Choice Demonstration Project, University of Arkansas, July 2014; Corey A. DeAngelis and Ben DeGrow, “Doing More with Less: The Charter School Advantage in Michigan,” Mackinac Center for Public Policy report, 2018, https://www.mackinac.org/s2018-01; Corey A. DeAngelis et al., “Bigger Bang, Fewer Bucks? The Productivity of Public Charter Schools in Eight US Cities,” University of Arkansas Department of Education Reform report, 2018, http://www.uaedreform.org/bigger-bang-fewer-bucks-the-productivity-of-public-charter-schools-in-eight-u-s-cities/; and Eric A. Hanushek, “The Economic Value of Higher Teacher Quality,” Economics of Education Review 30, no. 3 (2011): 466-79.
41. Hanushek, “The Economic Value of Higher Teacher Quality.”
42. Since over 90 percent of U.S. children attend public schools, the overall average income in the nation should largely reflect their average income levels.
43. “Occupational Employment Statistics: May 2016 National Occupational Employment and Wage Estimates,” Bureau of Labor Statistics, https://www.bls.gov/oes/current/oes_nat.htm#00%E2%80%930000.
44. Wolf et al., “The Productivity of Public Charter Schools”; and Cowen et al., “School Vouchers and Student Attainment,” pp. 147-68.
45. Henry M. Levin, “The Economic Payoff to Investing in Educational Justice,” Educational Researcher 38 no. 1 (2009): 5-20.
46. Cowen et al., “School Vouchers and Student Attainment,” pp. 147-68; and Levin, “The Economic Payoff to Investing in Educational Justice,” pp. 5-20.
47. “School Choice in America,” School Choice in America Dashboard.
48. DeAngelis and Wolf, “The School Choice Voucher: A ‘Get Out of Jail’ Card?”
49. Kathryn E. McCollister et al., “The Cost of Crime to Society: New Crime-Specific Estimates for Policy and Program Evaluation,” Drug and Alcohol Dependence 108, no. 1 (2010): 98-109. I exclude the two most costly types of crime — rape and murder — from this calculation in order to provide a more conservative estimate.
50. “United States GDP 1960-2017,” Trading Economics, https://tradingeconomics.com/united-states/gdp.
51. Samuelson, “The Pure Theory of Public Expenditure.”
52. Friedman, “The Role of Government in Education,” pp. 123-44.
53. Yujie Sude et al., “Supplying Choice: An Analysis of School Participation Decisions in Voucher Programs in DC, Indiana, and Louisiana,” Journal of School Choice 12, no. 1 (2018): 8-33.
54. Corey A. DeAngelis and Lindsey Burke, “Does Regulation Induce Homogenization? An Analysis of Three Voucher Programs in the United States,” University of Arkansas Department of Education Reform (EDRE) Working Paper No. 2017-14, September 2017, https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3038201.

Patrick Porter

Recent political tumult and the election of Donald Trump to the U.S. presidency have driven anxious commentators to lament the collapse of a post-1945 “liberal world order.” Nostalgic for the institution building and multilateral moment of the early postwar era, they counsel Washington to restore a battered tradition, uphold economic and security commitments, and promote liberal values. On closer inspection, while it is true that the postwar world was more prosperous and peaceful than what came before, the claim that a unitary “liberal order” prevailed and defined international relations is both ahistorical and harmful. It is ahistorical because it is blind to the process of “ordering” the world and erases the memory of violence, coercion, and compromise that also marked postwar diplomatic history. It loses sight of the realities and limits of the exercise of power abroad, the multiplicity of orders that arose, and the conflicted and contradictory nature of liberalism itself. While liberalism and liberal projects existed, such “order” as existed rested on the imperial prerogatives of a superpower that attempted to impose order by stepping outside rules and accommodating illiberal forces. “Liberal order” also conflates intentions and outcomes: some of the most doctrinaire liberal projects produced illiberal results. This nostalgia is harmful because framing the world before Trump in absolute moral terms as a “liberal order” makes it harder to consider measures that are needed to adapt to change: the retrenchment of security commitments, the redistribution of burdens among allies, prudent war-avoidance, and the limitation of foreign policy ambitions. It also impedes the United States from performing an increasingly important task: to reappraise its grand strategy in order to bring its power and commitments into balance.

Introduction

According to a view popular in Washington, D.C., and other capitals around the world, the United States used its power and idealism for more than 70 years to create a security and economic order that transformed the world. This world order was liberal because the United States was liberal. “Liberal” in this context means the pursuit of security both through the spread of liberty, in the form of free markets and democratic constitutions, and the rule of law, in the form of rule-based international institutions. Today, defenders of that order fear that President Trump and a set of regressive forces are laying waste to it. They claim the consequences are grave: we are witnessing the “end of the West as we know it,”¹ the abandonment of “global leadership” by its “long-time champion,”² and a “coming Dark Age.”³Foreign Affairs, the house organ of the foreign policy establishment, recently asked 32 experts whether the “liberal order is in peril.” Most agreed it is, with 26 respondents registering a confidence level of 7 out of 10.⁴ Alarmed by the political tumult of our time, nostalgists recall the post-1945 moment of institution building and benign internationalism and call for its reclamation.

They are, however, in the grip of a fiction. Liberalism and liberal projects abounded in the past 70 years. But the dream of a unitary, integrated global system organized around liberalism is ahistorical. In truth, the pre-Trump world was a more brutal and messy place than the nostalgia allows. To be sure, there was liberalism, and it did help define postwar international relations. Broadly speaking, the post-1945 period was, on many measures, more prosperous, less violent, and more collaborative than what came before. One defect of “liberal order” nostalgia is that it exaggerates these qualities and simply leaves out too many contrary historical realities. Other critics have already noted the gap between nostalgia and history and that the postwar world was never “whole.” At times the liberal order was neither very liberal nor very orderly. There may be “islands of liberal order, but they are floating in a sea of something quite different.”⁵

Not only do nostalgists get the history wrong, they fail to confront what “world ordering” actually entails. The main critique in this paper is that the fetish for “liberal order” has obscured what is involved in the process of “ordering” — or attempting to order — the globe. The United States, as the leading actor in the orthodox narrative, emerges as a power that created order through a benign internationalist vision, consensus building, and institution creating. But the successes and failures of that order also flowed from coercion, compromise, and rougher power politics. As the ordering superpower, the United States did not bind itself with the rules of the system. It upended, stretched, or broke liberal rules to shape a putatively liberal order. Appeals to the myth of a liberal Camelot flow from a deeper myth, of power politics without coercion and empire without imperialism.

This fuller narrative is also a story of tragic limits. The world was not so easily subjugated. Efforts to spread liberalism often contained the seeds of illiberalism. Multiple orders collided and met the limits of their reach and power. Efforts to create a liberal order ended up accommodating illiberalism. Liberalism itself proved to be a conflicted thing. At times, projects to advance it had unexpected results. As it happens, the pursuit of “liberal order” is not just an antidote to the current difficulties suffered by the international system but a source of them.

Ideas about “order” matter and have weighty policy implications. Just as material power enables or forecloses certain choices, ideas condition and constrain a country’s grand strategic decisions. Those who lament the fall of the “liberal order” are saying, in effect, that some ideas are illegitimate and should be off the table. They worry that “populism” and “isolationism” endanger traditional ideas that were once dominant, leading America to abandon its manifold commitments overseas, in turn driving the world into disorder. When they call for the reclamation of the old order, they also call for the perpetuation of American primacy. By contrast, this paper argues that the exaggerated notion of the “liberal order” and its imminent collapse is a myth of the foreign policy establishment and leads America to overstretch.

This analysis is divided into three parts. First, I examine the lamentations for a lost world, unpacking what such lamentations claim about how the world “was” before its dissolution allegedly began. Liberal order nostalgia performs two functions: by denying the violent coercion, resistance, and unintended consequences of “world ordering,” it sanitizes history into a morality tale and delegitimizes arguments for revision and retrenchment. The lamentations also give an alibi to American primacy, attributing its demise to forces external to it. By reducing the issue to one of inadequate political will, and by blaming either elites or the public at large for failing to keep the faith, “liberal order” lamentations dodge the painful question of how such an excellent order could produce unsustainable burdens, alienate its own citizenry, and provoke resistance.

In the second section, I demonstrate that “liberal order” rhetoric is ahistorical and therefore largely mythical. The claim that a single, internally consistent, and consensual order predominated for more than 70 years, with liberal projects producing liberal results, fares poorly when compared with the major patterns of international relations from 1945, in the spheres of both security and commerce. Conversely, the claim that American statecraft is now being turned upside down is hyperbolic, and blind to the quiet victories that orthodox U.S. grand strategy is winning under the Trump presidency.

Lastly, I argue that “world order” nostalgia is harmful. There is a prudent case for retrenchment, and a diplomacy of deterrence, power sharing, and accommodation, through which the United States could pursue security in a multipolar world. For an overstretched superpower to address the imbalance of power and commitments, it will have to look beyond ritual incantations.

The Claim: The Liberal Order Is under Assault

The prospect of major change in the international system is attracting a new wave of literature about “world order.” Recent crises and political revolts have prompted security experts on both sides of the Atlantic to announce the coming of end times. The rise of pernicious “isms” — economic protectionism, authoritarian nationalism, political tribalism, superpower unilateralism — has triggered these fears, along with the gauntlets being thrown down by revisionist powers threatening U.S. hegemony in the Persian Gulf, Eastern Europe, and Asia. In the United States, the focal point of this eschatology is the presidency of Donald Trump. After the election of an erratic, coarse demagogue to the nation’s highest office in November 2016, security experts lamented the passing of a postwar structure that civilized international life, presided over by a benign American hegemon.

What is being threatened? The objects of anxiety are a “liberal world order,” which allegedly held sway for 70 years, and even the end of “the West” itself. The life of this order is normally periodized from the end of World War II in 1945 to the recent past. As the storyline goes, the United States as benevolent hegemon designed and underwrote a “global, rules-based” economic and security order that transformed the world.⁶

After its chief competitor, the Soviet Union, collapsed in 1989-1991, it extended this strategy globally. Proponents of liberal order draw on the logic of hegemonic stability theory.⁷ According to that theory, one dominant state exercises such a preponderance of power that it lessens the insecurities that lead to arms races and spirals of alarm, enabling other states to ease their security competitions with neighbors and rivals, relax their arms programs, and focus on economic growth. More ambitiously, it not only reshapes institutions and markets but remakes the preferences of other states. To its admirers, this order, for all its imperfections, achieved unprecedented general peace and prosperity. It was based on a harmony of interests between the United States and the rest of the world. It made the world a single system or “whole,” as Council on Foreign Relations president Richard Haass claims.⁸ Revision of the order, and retreats by the hegemon, will lead to increased disorder.

On the campaign trail Donald Trump explicitly threatened the status quo. He denounced allies as delinquent and threatened to shred alliances, tolerate nuclear proliferation, re-erect tariff walls, and abandon international agreements. To security traditionalists who oppose Trump, his revisionist challenge accelerates the collapse of a “liberal order” under a transnational assault by authoritarian forces. In a state of shock, they seek orientation in an ahistorical myth about the world before this dark age. As Princeton’s Aaron Friedberg tweeted, “After WWII US built a system of democratic states, tied together by trade, institutions and common values — a liberal order. Now it needs to defend that order against the illiberal powers it tried to incorporate after the Cold War.”⁹ Historian Jeremy Suri charges that Trump is plunging the world into a great regression by “launching a direct attack on the liberal international order that really made America great.” The elements of this order include “a system of multilateral trade and alliances that we built to serve our interests and attract others to our way of life.” Suri explains:

Through the European Recovery Program (the Marshall Plan), the General Agreement on Tariffs and Trade (now the World Trade Organization), the International Monetary Fund (IMF), and the World Bank, among other institutions, the United States led a postwar capitalist system that raised global standards of living, defeated Soviet communism, and converted China to a market economy. Through the North Atlantic Treaty Organization (NATO) in Europe and a web of alliances in Asia and the Middle East, the United States contained aggressive states, nurtured stable allies, and promoted democratic reforms when possible.¹⁰

This sunny “highlights package” offers a strangely bloodless retelling of history. It is a euphemistic rendering of the Cold War and the actual practice of anti-Soviet containment by the superpower and its proxies. The Bay of Pigs, napalm, East Timor, the shah of Iran, and the Contras fade into the background. That this pristine retelling should come from distinguished historians of American diplomacy like Friedberg and Suri suggests how seductive the vision of an earlier and better order has become. Trump, too, is complicit in this mythmaking. Like his opponents, he frames his own election in stark terms. Trump speaks of a dark prehistory of “globalism,” of open borders, predatory capitalism, futile wars, and general American victimhood, and a return to wholesome nationalism, industrial regeneration, civilizational rebirth, and, of course, making America “great.”¹¹

What was the liberal order, as its defenders define it? If an “order” is a coming together of power with social purpose, a “world order” is an international design of institutions, norms, and patterned relationships that defines the global balance of power.¹² Some commentators argue that for a viable world order to emerge in a time of turbulence, the United States may have to compromise. Amitav Acharya, Michael Mazarr, and Henry Kissinger seek to revive the concept of world order, but unlike those of other “world order” visionaries, their proposed designs are pluralistic and require the United States to temper its universalism for the sake of stability and negotiated coexistence in a polycentric world.¹³

By contrast, the liberal order is a missionary project that looks to extirpate rival orders and demands the perpetuation of American dominance. As an ideal type, the “liberal order” entails a copious number of norms and institutions, suggesting that good things go together. In accounts of the postwar liberal order, many or all of the following features appear, though with varying emphasis: the rule of law and the supremacy of “rules,” humanist globalism and humanitarian development, free trade, multilateral cooperation, the security provision of the United States (principally through its permanent alliances), and a commitment to liberal progress through the advocacy of democratic and market reforms. Its institutions span the United Nations, NATO, the North American Free Trade Agreement (NAFTA), the Trans-Pacific Partnership (TPP), the General Agreement on Trade and Tariffs (followed by the World Trade Organization), the IMF, and the World Bank. A commitment to protocols and the bridging of divides figure centrally. Hence phrases like “open, rule-based international economy” abound.¹⁴ The “order,” proponents argue, embodied also a pattern of behavior, or, as Jake Sullivan at the Carnegie Endowment for International Peace explains, a “system of norms, institutions, and partnerships” whereby, under the hegemon’s stewardship, collective cooperation trumped competition for relative advantage, significant shares of sovereignty were ceded for the benefits of collective action, and a global consensus spread.¹⁵ An “illiberal order” would presumably be the opposite of these things: politically and economically divided and closed, authoritarian, uncooperative, coercive, and disrespectful of rules and norms.

The liberal order is necessarily hierarchical. To speak of liberal order is to speak also of American primacy, with the former depending on the exercise of the latter. Nostalgists do not deny that the American superpower upheld it partly through overwhelming military strength. However, they emphasize the nonbloody uses of force, for example, deterring and dissuading adversaries, reassuring and uniting allies, and preventing conflict. And they stress the consensual, attractive quality of American hegemony. Postwar continental Europe therefore stands out as a favorite area of emphasis, as an “empire by invitation.”¹⁶

Most anxious observers agree that a significant “fall” is occurring.¹⁷ To explain it, they weigh heavily an alleged loss of political will within the West. Leading theorists of liberal order, such as Princeton’s G. John Ikenberry, have long warned that “the hallmarks of liberal internationalism — openness and rule-based relations enshrined in institutions such as the United Nations and norms such as multilateralism — could give way to a more contested and fragmented system of blocs, spheres of influence, mercantilist networks, and regional rivalries.”¹⁸ Once optimistic that the order would withstand geopolitical challenges and prove resilient, Ikenberry now fears a different kind of insurgent threat, flowing not from hostile subversive states but from within. Working- and middle-class populations, he suggests, may lose faith in the order as democracy degenerates.¹⁹

Similar complaints have arisen across the Atlantic. Warnings against U.S. disengagement are a staple of rhetoric from security thinkers in allied countries.²⁰ For Robin Niblett, director of the internationalist Chatham House, Trump replicates and feeds on the destructive forces that powered “Brexit,” forcing liberalism into retreat.²¹ For the University of Exeter’s Doug Stokes, as for Ikenberry and Niblett, domestic discontent may unravel the worldwide arrangements that best served America’s “globalized” interests. For the old order to reproduce itself, it must make a new settlement with the American working class.²²

Most of these diagnoses have a common premise. All offer an upbeat, potted history of the world created in and after 1945. Many then blame the crumbling of that world on agents or forces that are separate from it.²³ If the order is perishing, they argue, it is being assassinated rather than dying from its internal failures. They have little to say about the significant reverses that occurred while the order reigned. These included some of America’s most disastrous wars, geopolitical chaos in the Persian Gulf from the Iran-Iraq War to the present sectarian breakdown, resurgent jihadi Islamism, the greatest act of urban terrorism committed by a nonstate actor in history, the eurozone crisis, the economic regression of Russia under “shock therapy,” mounting and unsustainable debt, the global financial crisis, the entrenchment and immobility of wealth, and the growing underclass of working poor. Rather than attributing to the old order the failures that occurred on its watch, nostalgists blame mismanagement, or popular fatigue, or “populism” and demagogues that whipped up mass discontent. They give credit to the order and U.S. primacy for benign developments, discounting other possible contributors such as the deterrent effects of nuclear weapons or the memory of World War II. That the order may have been complicit in its own undoing is hardly considered.

Nostalgia for a lost order is not just the complaint of self-styled liberal internationalists of any particular faction. It has a wider provenance among those who believe today’s choice lies between continuing American primacy and chaos. The strength of the consensus is reflected in a Brookings Institution paper, coauthored by former high-ranking officials in the administrations of George W. Bush and Barack Obama, urging the White House to revert to a traditional posture, upholding an order favoring openness, human rights, and peace, and claiming that this is the only viable grand strategy for the United States.²⁴ The hawkish intellectual Robert Kagan argues that the order established after World War II was a “liberal enlightenment project” that is now “challenged by forces from within and without,” not only because of popular fatigue with the burdens of international leadership but because Americans have forgotten the reasons their country adopted the role as the world’s guarantor and stabilizer.²⁵ Primacists such as Robert Lieber, Thomas Wright, and Eliot Cohen issue similar warnings.²⁶

The prominence of neoconservatives among this chorus is ironic. Critics once accused neoconservatives of violating the principles of liberal order with their bellicose unilateralism, by agitating for preventive war in Iraq in March 2003 without an explicit UN mandate, and by justifying torture. But this reflects the paradoxical problem at the heart of liberal “world ordering.” On the one hand, under most popular visions of liberal order, the hegemon creates a world based on deference to institutions and rules. But actual international life includes hostile, noncooperative forces that refuse to defer. Thus the liberal order includes conflicting rather than complementary rules and principles. It contains “veto players” like Russia or China with different conceptions of order. The protection and enforcement of such an order, and the enforcer’s own preponderance, rests upon selection and the exercise of a hegemon’s privilege.

Liberal world orders typically involve several impulses, namely, internationalism, integration, and imperialism.²⁷ That last, imperialism, is the most contentious. Historically, world orders, with their trading protocols and monetary regimes, and control of sea lanes, commercial routes, and access to raw materials, are designed and imposed by the strong. The opening of Asian markets, a celebrated feature of liberal order, was also a pre-1945 byproduct of violent and imperial coercion, imposed on China by Great Britain through the Opium Wars and on Japan by American Commodore Matthew Perry with the threat of naval bombardment in 1853-1854.

Proponents of liberal order occasionally admit that what is sometimes framed straightforwardly as a rule-bound order is in fact a system of imperial power (and vigilante privilege) exercised by a hegemon. Robert Cooper, the former diplomat and adviser to Prime Minister Tony Blair, argued that if the world had a civilized core that deserved lawful conduct, there was also a barbarous periphery that warranted “rougher methods of an earlier era.”²⁸“Among themselves,” he wrote, “the postmodern states operate on the basis of laws and open co-operative security” but “in the jungle, one must use the laws of the jungle.”²⁹ One-time advocate of American empire Michael Ignatieff admitted that being an imperial power “means enforcing such order as there is in the world and doing so in the American interest. It means laying down the rules America wants (on everything from markets to weapons of mass destruction) while exempting itself from other rules (the Kyoto Protocol on climate change and the International Criminal Court) that go against its interest.”³⁰ At the height of the war on terror, jurists advising the George W. Bush administration used a similar logic to justify the suspension of the rule of law and Geneva conventions.³¹ One can defend these inconsistencies as necessary, or not. That they continually recur suggests that liberal “world ordering” is an inescapably compromised process.

The United States is an imperial power, even if it is distinct from former empires. It may lack the land hunger of empires past and look different from European or Asian imperialisms. It was averse to formal annexation, refused to claim the mantle of empire, and made a succession of retreats, from Vietnam, Lebanon, and Iraq. But it still throws its weight around in imperial ways: through coercion, subversion, or patronage, it penetrates the sovereign autonomy of other states to constrain their choices. The political economy underlying American interventions was at times coercive, as in the structural-adjustment programs visited on developing countries from Eastern Europe to Latin America. Leading primacists who speak up for liberal world ordering have earlier acknowledged that the American project overseas must be necessarily imperial, albeit in this case of a distinctively informal “American” kind, involving the forceful suppression of revolt, hard-nosed enforcement of parameters around other states’ policies, and the exercise of unequal bargaining influence.³² But for the most part, today’s lamentations for a dying liberal order do not acknowledge the necessarily imperial component. This is the imperialism that does not know itself.

Invocations of the liberal world order have become the ritualized language of the foreign policy establishment. In the academy, there is a well-established scholarly study of “world orders.”³³ But too often, especially among think-tank specialists and the foreign policy commentariat, the liberal order’s admirers assert what they ought to prove. In that respect, President Trump’s former deputy assistant for strategic communications at the National Security Council, Michael Anton, was right to argue that the foreign policy establishment has become a “priesthood,” fonder of recapitulation than argumentation.³⁴ It repeats its axioms but neglects to ground its vision in a careful reading of history or the present. A good example is an article by Edward Luce in June 2017, mourning the internal defeat of Western liberalism and a new world “disorder.” To demonstrate the imminence of disorder under Trump, Luce did not examine what Trump was actually doing beyond the decision to withdraw from the Paris climate change agreement, which is not an unprecedented pullback. The many Trump administration moves that have affronted Moscow go unmentioned: for example, the reinforcement of NATO through increased funding of the European Reassurance Initiative; and the bombing of the Assad regime, Putin’s ally, in April 2017. Instead, Luce cited the statements of two allies who were wondering where Trump’s behavior would lead, gossip about personality clashes in Trump’s court, and corroborating statements from other figures in the foreign policy establishment who take the “liberal order” as an article of faith, betraying a confirmation bias. Without evidence, he then accused President Obama of “global retrenchment.”³⁵ Typical of the genre, Luce contrasted these recent failures with the postwar internationalism and institution building of President Harry Truman.

To be reprimanded for violating established norms of American “global leadership,” against the exalted standards of Truman, is an occupational burden that comes with the presidency. Before Trump, critics had accused presidents Clinton, George W. Bush, and Obama of retreating from a long-standing commitment to a liberal world order. The charge that Washington is abandoning a noble Trumanite diplomatic past is less an observation than a political predisposition, substantively shallow yet part of the framework within which debate is conducted. International history after 1945 is more fraught.

Mythologizing the Postwar Order

Accusations that U.S. presidents are flouting a long-standing postwar liberal order rely upon a mythologized account of history. When presidents wage war unilaterally, topple governments, coerce allies, threaten abandonment, disregard the demands of international institutions, practice economic protectionism, or cultivate illiberal allies and clients, they may or may not be practicing prudent statecraft. Either way, their behavior is unexceptional. They are not departing from tradition.

Let us revisit history first in the area of trade and political economy. President Trump disturbs defenders of the liberal order partly through his economic nationalism, his promise to rebuild American manufacturing industries by erecting protective tariff walls, and his withdrawal from the Trans-Pacific Partnership (TPP). The extent to which U.S. membership in the TPP represented a long-standing commitment to liberal order is doubtful, given that the United States signed the agreement in February 2016 and withdrew in January 2017, and given that the agreement’s intent was to exclude and contain China as much as to integrate other Asian economies. Nonetheless, nostalgists assert that Washington’s commitment to free trade was a central pillar of liberal order. Identifying the marketing precepts of the Washington consensus with liberal order, they present Washington’s various “Open Door” policies as both a cause and symptom of the world’s liberalization.

Without question, the United States in the postwar period dismantled the economic architecture of the British Empire. It strove to reshape the international economic environment on its terms, to establish and exploit the dollar as the reserve currency, to promote open markets, and to make a world safe for the penetration of American capital. How far did that process, that long-standing commitment to the “Open Door,” represent a liberal drive for free trade?

In reality, there was not one but several postwar economic orders. As Michael Lind cautions:

Globalization on a large scale, characterized by the emergence of transnational corporations and supply chains, got underway only in the 1990s and 2000s. Mass immigration to the United States and Europe is also largely a post-Cold War phenomenon. The euro and the Eurozone date back only to 1999. Labor mobility within Europe is also a relatively recent policy. Controversial “megaregional” trade pacts like NAFTA, the TPP, and TTIP [Trans-Atlantic Trade and Investment Partnership], which go beyond old-fashioned tariff reductions to rewrite much domestic legislation, go back only to the 1990s.³⁶

Niall Ferguson, too, observes that “truly free trade, truly free capital flows and large-scale migration across borders did not begin until the 1990s.”³⁷ Earlier decades saw capital controls, fixed exchange rates, and periodic returns to tariff barriers. One major pillar of the postwar order was U.S. ally, democratic Japan. Under U.S. military protection, that same country instituted, in the words of Claremont McKenna College’s Leon Hollerman, “the most restrictive foreign trade and foreign-exchange control system ever devised by a major free nation.”³⁸ The long-running competition with the Soviet Union moved the United States to deliberately encourage the economic growth of its Asian allies, but under the shield of a neomercantilist state. In other words, in identifying the U.S.-led order with market and trade liberalization, nostalgists historicize what are in fact a quite recent set of post-1989 international arrangements. From this perspective, a liberal order did arrive, only it came later, in the age of post-1989 unipolarity.

A Less-Than-Completely Liberal Trading Order

There is a more basic defect in recollections of the liberal order. During the postwar era, the United States persistently flouted liberal economic principles and imposed restrictive measures when it suited. Indeed, major powers have not historically risen through free trade and passive governments. Ascending powers have typically risen partly through the deliberate, visible, and intervening hand of an activist state.³⁹ So too with America. All U.S. presidents have had to manage the tension between the commitment to the “Open Door” and the demand for industrial protection at home. The subsidy, the tariff, the quota, and the bilateral voluntary-restriction agreement have remained part of America’s repertoire. As a recent study of global data by Gowling WLG reveals, the United States is a “long-term and prolific proponent of protectionist policies,” and the world order it presides over is notably protectionist.⁴⁰ Since the 2008 financial crisis, the United States has imposed tariffs worth $39 billion, while the world’s top 60 economies have adopted more than 7,000 protectionist trade measures worth more than $400 billion. The United States and the European Union both accounted for the highest number of protectionist measures, each exercising more than 1,000, with India a distant third at 400.⁴¹ America’s trade protectionism has the highest impact on other countries. Foreign farmers would be baffled by the claim that the old order embodied free trade, when the United States persistently granted agricultural subsidies and other mechanisms limiting foreign governments’ access to U.S. consumers.⁴²

The reversion to protectionism has precedents from before the global financial crisis. Some of the most strident advocates of open markets and the dismantling of trade barriers have in practice done the opposite. One was President Ronald Reagan. Reagan had championed the cause of free trade as a foundation of progress and peace. Yet as president, he increased the proportion of imports subject to restrictions by 100 percent from 1980, as well as tightened quotas, introduced “voluntary restraint agreements” and new duties, raised tariffs, and strengthened the Export-Import Bank in order to protect the recovery of U.S. industries, especially automotive, computer-chip, and steel.⁴³ Reagan justified these steps on the grounds that he was forcing economic competitors to trade freely. Regardless, his policies were a long way from Adam Smith.

President Clinton also championed free trade, in words and deeds. He drove through NAFTA, a free trade zone uniting North America’s three largest economies, and pushed for China’s admission to the WTO under “most favored nation” status. Yet under Clinton, rice subsidies that continued during his administration enabled U.S. growers to dump their product onto the markets of vulnerable rural countries such as Haiti, Ghana, and Indonesia at depressed prices. Clinton has since apologized to Haiti for the devastation that these arrangements inflicted on the country.⁴⁴

President George W. Bush emulated Reagan rhetorically, invoking the principles of free trade and unfettered markets. Yet in 2002, he increased steel tariffs by 30 percent, only to back down 20 months later under threat of punitive countertariffs by the European Union, a protectionist bloc in its relations with many countries beyond its borders. Confronted with the prospect of economic meltdown in the crisis of 2008, Bush intervened in the market with strongly protectionist measures, including bailouts of major firms, claiming, “I have abandoned free-market principles to save the free-market system.”⁴⁵ The reintroduction of protectionist measures today, then, is not such a sudden or radical departure as is sometimes claimed, though Trump’s open enthusiasm for a “trade war” does mark a difference. There is a defensible logic to the position that in order to practice free trade a country needs a viable economy to practice it with. Reagan and Bush’s contortions on the issue reflect the inherent difficulty of liberal projects, whose architects often feel impelled to compromise with illiberal pressures. A world where even the most avowed exponents of free trade continually resorted to protectionism, though arguably more free and liberal than what had gone before or than what might have prevailed otherwise, was still not the “flat” free-market capitalist world we are being invited to be nostalgic for.

Nostalgists claim that one dividend of American hegemony has been the economic liberalization of the globe, or large parts of it. They could point, for example, to the transformation of China into a wealthy capitalist economy that has lifted its population out of “a dollar a day” poverty. But this liberalizing process is more conflicted than potted histories suggest. Beijing achieved rapid industrial revolution and the movement of its workforce from the field into the factory through authoritarian and illiberal measures: involuntary-resettlement urbanization schemes, population control through forced abortion and compulsory sterilization, severe working conditions, repression of civil society, including trade unions, labor, and human-rights activists, and internet surveillance, among other measures. The People’s Republic has consistently ranked low on the Freedom Index.⁴⁶

A more direct application of U.S.-backed liberalism happened in Russia. After communist rule collapsed in 1991, at the urging and advice of the United States’ government and economists, Moscow embarked on a program of “shock therapy” to restructure Russia around the principle of market exchange, adopting accelerated privatization of state industries, deregulation, fiscal discipline, and the shedding of price controls. This experiment was a major effort in the project to enlarge the global liberal order at a rapid clip. It had the support of the leading institutions of global capitalism, the IMF, World Bank, and U.S. Treasury Department. Harvard academic Jeffrey Sachs, one of Russian liberalization’s architects from 1991 to 1993, set out the program’s logic in TheEconomist, a journal that champions the cause of the liberal world order. “To clean up the shambles left by communist mismanagement, Eastern Europe must take a swift, dramatic leap to private ownership and a market system. West Europeans must help it do so.”⁴⁷“Swift, dramatic leap,” a vast program grounded in classical liberal economics, took on the tempo and zeal of the revolutionary communism it aimed to replace. These rapid reforms replaced an oppressive and failed communist system. They did so at Washington’s continual insistence that Russia reform itself on “our conditions.” But the results on many measures were disastrous: capital flight and deep recession; slumping industrial production; malnutrition; the rise of criminality — a criminalized economy, in fact — intertwined with a corrupt oligarchy enjoying a concentration of wealth; and the decline of health care and an increased rate of premature deaths.⁴⁸ As Nobel laureate and former World Bank chief economist Joseph Stiglitz observed, by eschewing the more gradualist path of Poland or China, the consequences of the program were profoundly illiberal.⁴⁹“Liberal order” visionaries are quick to give their ideas credit for the prosperity of nations from Western Europe to the Pacific Rim, finding causation in correlation. They deny such a direct link between their ideas and the problems of post-Soviet Russia.⁵⁰ Yet it is hard to accept that measures like sudden privatization and the rise of monopolies in a corrupt country were not related to asset stripping and capital flight or that “eliminating the housing and utilities subsidies that sustained tens of millions of impoverished families” did not play a major part in the social ruin that followed.⁵¹ Western technocrats, diplomats, and politicians were deeply implicated in the new order’s design.

The Hard Edge of the Liberal World Order

Lamentations for the end of the liberal order are also heard in the realm of “hard” security. The U.S. hegemon, nostalgists warn, is losing (or has lost) the political will to underwrite the international system through a commitment to permanent alliances and to intervene to bring order out of chaos. Part of the current intellectual confusion flows from the conflation of liberalism, which is supposedly peaceable, consensual, and benign, with the process of “world ordering.” It is here that defenders of the old order present their most misleadingly anodyne account of history. A review of the actual experience of the past 70 years suggests that the process of “world ordering” must at times be coercive. For all the attractions of American hegemony abroad, there has also been resistance and imposition.

To understand how the superpower met that resistance and imposed itself, we must go beyond the romanticized postwar moment of Trumanite internationalism in the late 1940s. Consider both ends of the chronology as it is usually presented, from 1945 until the recent past. Admirers trace the restructuring of international life in that first year to the visionary institution building that President Truman oversaw amid World War II, such as the United Nations Conference on International Organization in San Francisco and its main creation, the UN Charter. In this rendering, the founders conceived the liberal order through a collaborative process of institution building. The narrative is strikingly nonviolent.

In fact, to create the conditions for that visionary world making, the liberal order was conceived in blood. Only months later, the same U.S. president launched two atomic strikes on Imperial Japan, immolating and irradiating two of its cities after blockade, firebombing, and starvation had not broken its will. He did so to put down an adversary that had been brutally pursuing a rival vision for an Asian order of its own. In order to create an order, Washington swept aside a competitor by introducing a genocidal weapon into the world. There are powerful arguments that this was the “least bad” choice available.⁵² Tellingly, though, in panegyrics for a dying liberalism, the words “Hiroshima” and “Nagasaki” hardly appear.

If there were liberal principles that underpinned the UN as it was founded in 1945, they were at first self-determination and sovereignty rather than democracy and human rights. The world order was hardly born “liberal” in the sense implied today: recall that two of the permanent five members of the UN Security Council were totalitarian communist states, and two of the democracies were managing colonial empires that they would not relinquish for decades. Then and now, modern liberalism is antithetical to the grave exertion of state power still practiced in 58 countries, the death penalty. To be sure, the birth of the post-1945 world order did advance some liberal ideas broadly. The general norm against imperial aggression was one. This, however, was not strong enough to prevent or dislodge China’s seizure of Tibet, the bids of Turkey and Greece to grab Cyprus, Israel’s occupation of Gaza and the West Bank, India’s occupation of Kashmir and annexation of Goa, Indonesia’s occupation of East Timor, or indeed the Soviet Union’s occupation of Eastern Europe.

At the other end of the chronology, the present moment, consider that the U.S. hegemon has been waging a “war on terror” against Islamist jihadi groups since the 9/11 attacks of 2001. In pursuing the liberal cause of democratization as an antidote to terror, Washington entered the age of “enhanced interrogation” and preventive war. Now, with new weapons (drones) at hand, Washington conducts a sustained campaign of extrajudicial assassinations, often without the consent of host countries and without seeking formal permission or mandates. It has conducted renditions of suspected terrorists without trial. Reluctant to deal with live captives in indefinite detention, a more liberal president from 2009 increasingly avoided the dilemma by killing them. Meanwhile, whatever benefits it has wrought, American unipolarity was not peaceful or liberalizing for the unipolar power. The first two decades of the unipolar Pax Americana after 1989, which made up less than 10 percent of America’s history, generated 25 percent of the nation’s total time at war. That period is more bellicose by an order of magnitude than the preceding eras of bipolarity and multipolarity, in terms of frequency if not intensity.⁵³ Whether in Iraq and Libya, or now with U.S. assistance to Saudi Arabia’s indiscriminate bombardment of Yemen, this proclivity to continuous war making has not created a “liberal” condition of peaceful order. At home, there is a continuous state of alarm and vigilance, whereby “normality” is permanently suspended by an unending state of exception. This, combined with an encouraged state of paternalism where citizens are encouraged to be passive consumers of events, has helped weaken the checks and balances of the republican Constitution.⁵⁴ Detention without trial, secret, warrantless surveillance, unauthorized wars, torture, covert “black sites” — these are not the obvious features of a robust liberal constitutional order. If large parts of the world have not accepted liberalism in major areas of civic life, neither has the United States.

Instead of a full reckoning with diplomatic history, nostalgists frame history around the positive creation of new architectures and schemes. Thus the Marshall Plan (1948-1961) figures centrally in America’s postwar historic mission, based on, as Benn Steil puts it, “the moral primacy of democratic government and free economic exchange.”⁵⁵ This absolute, almost platonic account of the past has little room for other, less-celebrated events from the same era, such as the British- and U.S.-backed overthrow of Iranian primeministerMohammadMossadegh in 1953, deposeddespitehis commitment to national independence and secular democracy. In this picture, the violence and compromises of hegemony, moral and strategic, almost vanish.

Nostalgia for the liberal order also overlooks the reality that it was enforced through coercion. In the same era, a defining episode in the postwar assertion of American hegemony was the Suez crisis of 1956. In that hinge event of the Anglo-American relationship, the U.S. Sixth Fleet stalked and harassed British ships in the Mediterranean, fouling their radar and sonar, menacing them with aircraft and lighting them up at night with searchlights.⁵⁶ With the British pound and oil supplies under pressure, President Dwight Eisenhower threatened Britain with the simple formula of “no ceasefire: no loans.” Patronage could be rapidly withdrawn, regardless of recent history, blood ties, or shared visions of Western-enforced order. The United States enforced its interpretation of that order by targeting its ally’s vitals.⁵⁷

Between those two moments in time, the United States practiced geopolitics ruthlessly. It partly did so in the course of its long security competition with the Soviet Union. Strikingly, the Cold War as it was actually conducted and lived — where two superpowers did not allow rules, sovereignty, multilateralism, and institutions to constrain them when the stakes were high — does not occupy a prominent place in the mytho-history. Hardly anywhere in nostalgic reminiscences do there appear the numerous coups that were sponsored or supported by Washington. These interventions linked to the United States since 1945 may or may not have been defensible. They certainly violated one of the claimed core principles of “liberal,” “rules-based” order, that of self-determination.

The United States not only overthrew governments (sometimes democratically elected ones) — or attempted to — in Albania, Ghana, Guatemala, Greece, Cuba, Chile, Iran, El Salvador, Nicaragua, South Vietnam, Argentina, and Grenada. It also supported violently illiberal forces, from Islamist mujahideen in Afghanistan-Pakistan and President Hosni Mubarak’s oppressive state in Egypt to the Indonesian Suharto regime and its death squads. A mainstay of U.S. hegemony in the Persian Gulf is its partnership with Saudi Arabia, an absolutist state that beheads apostates and survives by making concessions to Wahhabi theocrats. It is currently waging a brutal campaign against rebels in Yemen that, according to Amnesty International, includes attacks that are “indiscriminate, disproportionate or directed against civilians and civilian objects, including funeral gatherings, schools, markets, residential areas and civilian boats.”⁵⁸ NATO allies on the European continent for decades included authoritarian Portugal and Greece. West Germany, the poster child of the liberal order, did not have elections during its first four years, and its proud social democracy retained officials who had been security elites in the Third Reich.⁵⁹ Former Nazi mandarins stuffed the highest levels of government, including the Foreign Office and the Interior and Justice Ministries. Several former Nazi generals would later become senior commanders in the Bundeswehr. And in the 1948 Italian elections, the CIA helped ensure the electoral defeat of communists by funding anti-communist parties, forging documents to discredit the Communist Party, and warning Italians that if they publicly supported the party they would be barred from entering the United States. For the sake of liberalism in the long term, the United States exercised its privileges. If the deliberate subversion of a democratic election abroad with “fake news,” bribes, and coercion represents the antithesis of liberal world order, as Trump’s critics now suggest, then Washington attacked that order in the period of its creation. Coups, partisan electoral interventions, the cooptation of illiberal actors, and the flouting of international law made American hegemony unexceptional.

In dismantling the power of old European colonial empires, the United States erected a form of domination that had an imperial quality of its own. Consider one of its more ambitious ventures in liberal ordering: the invasion and remaking of Iraq. The occupiers of Iraq regarded themselves as liberators. After invasion, though, the United States also projected power over Iraq’s interior governance in imperial fashion and with a liberal program, with all the tensions this implies. Director of the Coalition Provisional Authority Paul Bremer applied a program of rapid liberalization not only through the well-known de-Ba’athification and disbanding of the Iraqi Army, but through the order for “the full privatization of public enterprises, full ownership rights by foreign firms of Iraqi businesses, full repatriation of foreign profits … the opening of Iraq’s banks to foreign control, national treatment for foreign companies and … the elimination of nearly all trade barriers.”⁶⁰ The United States continued to impose itself on Iraqi politics when it wanted, demanding and receiving the resignation of elected prime minister Ibrahim al-Jaafari in May 2007. Intended to implant market democracy, these measures infringed the country’s sovereign democratic will. In other words, the liberators were freeing the Iraqis to conform with the occupier’s preferences.

It remains hard to have an empire without imperialism. Yet many visions of liberal order erase the historical process of imperialism, decentering, as Jeanne Morefield argues, “imperial violence while simultaneously positing the necessity of imperial action.”⁶¹ If liberalism at a basic level is an enlightenment project committed to liberty, equality, and limitations on state power, and if “world ordering” requires imperialist power projection, it is hard to fuse them without friction. Some may conclude from this historical record that, in the history of American hegemonic “world ordering,” liberalism was missing in action. On each occasion, critics have accused the United States of betraying its own liberal traditions in the pursuit of power. But it is hard to believe that a republic whose leaders so often and so intensely enunciate liberal principles is really driven by secret, amoral cynicism. A more troubling possibility should be considered. Liberalism is a powerful engine of American statecraft, but that statecraft often violates liberal principles. As a dogma of foreign policy, liberalism is jealous, intolerant, and messianic. Applied unchecked, it leads to its own illiberal opposite.⁶² The practitioners of rough geopolitics were not necessarily hypocrites. They often believed they were serving the ultimate cause of forging a liberal peace under American oversight but that to do so they had to accommodate illiberal allies and pitilessly destroy liberalism’s enemies. In this way, a superpower attempting to create a liberal order permits itself to employ unsentimental methods.

Thus in February 2017, David Petraeus could recall sincerely that “to protect freedom here at home, we adopted a foreign policy that sought to protect and, where possible, promote freedom abroad, along with human rights and rule of law,” invoking American values such as “political pluralism” and “a free and open society.”⁶³ Yet as commander in Iraq, Petraeus sought to reverse that country’s implosion and salvage victory by compromising these standards. To that fight, he brought pragmatic, byzantine divide-and-exploit methods, paying for the defection of former Iraqi insurgents and working with Shia paramilitary units not known for their commitment to the Hague conventions. As director of the CIA, Petraeus advocated and implemented a campaign of “signature” drone strikes, whereby the assailant knowingly targets a group gathering — at a funeral for an al Qaeda member, for instance — because of their suspicious behavior and association, rather than through verified identification of the presence of individual persons. Such strikes, therefore, can also threaten noncombatants and the innocent.⁶⁴ To bolster the struggling rebellion in Syria, Petraeus later in 2015 advocated luring away and recruiting “opportunistic” members of the jihadist Jabhat al-Nusra, then formally affiliated with al Qaeda.⁶⁵ This is not the place to arbitrate the wisdom and legitimacy of such measures. Dealing with conflicts in such places is a choice of agonies, and no doubt Petraeus and his peers regard themselves as guarding Americans while they sleep and trafficking with lesser evils to keep greater ones at bay. But note that a senior advocate of liberal order can also advocate measures that risk “crowd killing” and that involve enlistment of members of jihadi terrorist organizations and collaboration with sectarian governments. Champions of liberalism must somehow navigate their ideals through the illiberal demands of warfare.

Nostalgists for the liberal order also betray a shallow conception of their central idea, liberalism. They conflate liberalism with other desirable phenomena, like capitalism and democracy. They neglect the possibility of illiberal democracy, and illiberal capitalism. Majority democratic rule does not equate with, or necessarily produce, a liberal protection of individual rights such as the presumption of innocence or trial by jury, a liberal tolerance for opposition and dissent, or a constitutional order that separates powers and constrains government through an independent judiciary or a free press. Capitalism can also be illiberal, as the Chinese Communist Party demonstrates. One of America’s long-term allies, Singapore, evolved as a supervised market democracy that curtailed the right to dissent. South Korea, an ally and protectorate within America’s Asian system, evolved first as a dictatorship under authoritarian founding fathers who were also modernizers, Syngman Rhee and Park Chung Hee. These authoritarians nurtured the chaebol business groups, Hyundai, Daewoo, and Samsung. Free markets took root first as highly protected markets under unfree political conditions. Such contradictions are absent from liberal-order panegyrics.

As it is recalled, the “liberal order” embodies the permanent commitment of the United States to alliances and institutions without coercion. A broader historical perspective suggests, however, that Trump’s coercive treatment of allies is less of a break with the past than is often thought. In reality, the United States has often coerced allies with threats of abandonment and punishment.⁶⁶ In 1954, Secretary of State John Foster Dulles threatened Europe with an “agonizing reappraisal” of alliances. In 1973 and 1974, President Richard Nixon and his national security adviser Henry Kissinger suspended intelligence and nuclear cooperation with Britain to punish noncooperation over a U.S.-initiated declaration of principles and the privacy of bilateral and UK-European Economic Community discussions. The United States has used the threat of abandonment to persuade allies and clients to cancel their nuclear programs, including West Germany, Japan, and Taiwan, while threatening adversaries with sanctions or preventive war. The demand that European allies shoulder more of the burden of military expenditure has been a staple of U.S. diplomacy, from President Eisenhower to former secretary of defense Robert Gates. Despite Britain spending blood and treasure in Afghanistan and Iraq to support the war on terror and cement its standing in Washington, President Obama made a blunt threat that departing from the European Union would place the UK at the “back of the queue” when seeking a bilateral free trade agreement. Assured commitment to institutions and allies through only positive solidarity is a false memory. This underlines the pattern whereby Washington underwrites a liberal world order not by adhering to its principles but by stepping outside them, practicing punishment, threats, and bribes that it would not accept if directed at itself.

In “liberal order” litanies, another persistent claim is that the order was “rules-based.” It was not. Rules exist, and flouting them can have costs. But at critical moments for strong states such as the permanent five members of the UN Security Council, rules proved to be slippery; they were invoked, stretched, arbitrarily altered, or ignored, as interest permitted. The unreality of nostalgic legalism was illustrated in the summer of 2016 by two adversaries who both at different times have appealed to “rules” as the arbiter of international order. China defied the unanimous ruling of the Permanent Court of Arbitration, which found against its territorial claims, and continued to expand into the South China Sea and seize disputed waters, islands, and shoals. At the same time, the United States appealed to China to respect the “legally binding” verdict yet had not even ratified the UN Convention on the Law of the Sea that it urged China to observe. From ignoring the International Court of Justice over the mining of Nicaraguan harbors in 1986 to bombing Serbia in 1999 without a UN mandate, the United States has infringed on the letter of international law when it has found that other interests or values were compelling. It exercised a vigilante’s privilege. So too did other major powers. For less powerful and emerging states, the writ of liberal order was often remote, as they “routed around” rules to pursue their interests. In this century, Africa, from the Great Lakes region to the Sudan, has seen millions butchered, displaced, and unavenged. The era may have involved greater degrees of “rule following” than earlier eras. But it was not “based” on the observation of rules, at least not for the major powers. To rebrand this fraught history of power politics as an era of rule-bound civility is perverse. There is nothing intrinsically wrong in calling for the conservation, or restoration, of an order on the basis that it represents something better. In this case, though, the nostalgia rests on delusion — about what the world was and what it can be.

The Burden of Liberal Nostalgia

The debate around “liberal order” is consequential. Memories of what went before will condition the ability of the United States and its allies to navigate their way ahead. At issue here is neither the legitimacy of American power in the world nor many of its benefits. If there was to be a superpower emerging from the rubble of world war in midcentury, we should be grateful it was the United States, given the totalitarian alternatives on offer. Under America’s aegis, there were islands of liberty where prosperous markets and democracies grew. U.S. internationalism rebuilt Western Europe and East Asia and successfully contained Soviet communism. The central issue is whether this created a wider “liberal” system, and whether the actual historic process of world ordering can even be achieved by liberal means. The answer, in both cases, is no. Even at its zenith, America did not have the power to reorder the world so fundamentally. Its more ambitious efforts to do so generated illiberal as well as liberal results. The limits on power and knowledge are too strong for any world ordering to be so straightforward and benign.

How do we best explain these contradictions? The most accurate answer is not that the United States cynically preached liberalism while practicing realpolitik. Liberalism is an authentic and powerful engine of U.S. diplomacy. It is a pillar of the American diplomatic mind. But it was never all-conquering. Long before Trump, large swaths of the globe (and of American statecraft) were unsubjugated by it. Today’s lamentations confuse two phenomena, liberal ideals and institutions on one hand and a hegemon’s world ordering on the other. To underwrite the order it promoted, the superpower exercised prerogatives in ways that cannot adequately be labelled “liberal,” implicitly claiming a privilege to do so. Except in atypical circumstances, large liberal projects require murky bargains. At worst they contain the seeds of their own unraveling, especially in the countries that become laboratories for the most doctrinaire attempts.

Endless recall of the “liberal order” is not only ahistorical. It is harmful. It damages the intellectual capacity to diagnose the failures of the recent past. It harms the effort to construct a workable design for the future. It impedes Washington from undertaking a needed reassessment of its grand strategy that has put the United States where it now is: struggling under the weight of spiraling debt, confronting multiplying foreign conflicts and domestic discord, and set on a collision course with rivals. Appeals to take up the burden, again, of spreading liberalism overseas presupposes the worldview of idealistic technocrats, confident in their capacity to reprogram the world despite growing evidence to the contrary. At a time when a sober reappraisal and some retrenchment is needed, both Trump and his critics undermine that task by peddling ahistorical reductionism. There is a better, non-Trumpian critique to be made of a failing foreign policy consensus, and on behalf of an alternative order based on a wiser combination of restraint, deterrence, and power sharing.

A review of the United States’ current grand strategic situation suggests that a clear-eyed stock taking is in order. Trump’s presidency doesn’t signify a general retrenchment of the United States and a retreat from international commitments. Under Trump, Washington’s growing commitments still exceed its power. The United States feels its capacity to impose order strained, even with the significant investments it already makes. As Richard Betts once suggested of the annual defense budget, half a trillion dollars is more than enough.⁶⁷ It is in the size of the policy ambition relative to capabilities, rather than merely the size of those capabilities, where the dangerous imbalance lies. Despite his threats to overturn the old order, the power of the foreign policy establishment and its habitual ideas have steered Trump to quickly conform to the fundamentals of traditional U.S. grand strategy.⁶⁸ He now aggressively reasserts U.S. primacy. If he poses a danger, it is not from abandonment but overreach. On its current course, the United States is prone to two forms of self-inflicted wounds: self-encirclement, whereby a state undermines its own security by provoking resistance and counterpower; and imperial overextension, whereby a state expands to the point where the costs outstrip the benefits.⁶⁹

The United States is accumulating record deficits and growing, unsustainable debts. According to the Congressional Budget Office, federal debt will reach 150 percent of GDP by 2047.⁷⁰ Because repayment obligations are the first, compulsory items in expenditure and because heavy fiscal burdens beyond a certain proportion of debt-to-GDP tend to choke economic growth,⁷¹ a growing debt load directly impedes the country’s ability to sustain its way of life alongside its extensive international commitments. U.S. grand strategy also gives Washington a proclivity to continuous wars that it chooses to fund through deficits. According to one estimate, U.S. wars from 2001 to 2016 had a budgetary cost of approximately $4.79 trillion, taking into account indirect costs such as interest on borrowing and through-life care for veterans.⁷² Those wars have led to further geopolitical crises and demand for further commitment. Conflict-induced anarchy in Iraq and Libya created footholds for the Islamic State and, by upsetting the balance of power in the Persian Gulf, opened the way to a Saudi-Iran cold war that now implicates the United States.

The Trump administration has not reversed this imbalance but aggravated it. It has significantly increased the defense budget, while significantly reducing taxes. It has embarked on a deficit-financed military buildup, a pattern that historically increases imbalances in the economy and triggers a “boom-bust” cycle, and where overreaching wars (like Iraq) and financial meltdowns (like the global financial crisis) are linked.⁷³ The final 2018 defense budget is expected to be 13 percent higher than that of 2017.⁷⁴ The United States’ grand strategy of primacy saddles it with defense and national security expenditures that amount to over 68 percent of discretionary spending, taking into account the base budget and overseas contingency operations and support for veterans affairs, homeland security, and the nuclear weapons program.⁷⁵

Meanwhile, the overall direction so far of President Trump’s foreign policy has been to multiply America’s security commitments and entanglements. The United States has implicated itself more deeply in the geopolitics of the Persian Gulf. Trump has intensified America’s confrontation with Iran by abandoning the multiparty settlement on Iran’s nuclear program. He has reinforced U.S. patronage of Saudi Arabia and the Gulf monarchies, while hardening Washington’s alignment with Israel by recognizing Jerusalem as its capital. By November 2017, Trump had increased the number of troops and civilians working for the Department of Defense in the Middle East by 33 percent.⁷⁶ At the time of writing, the status of America’s commitment in Syria is not clear, with the administration both promising to withdraw yet indicating it would stay to defeat the remnants of the Islamic State, and threatening to continue to punish Syria for chemical weapons use. He increased the U.S. commitment to the Afghanistan-Pakistan theater. Lastly, the Trump administration’s National SecurityStrategy openly acknowledges competition against “revisionist” powers Russia and China.⁷⁷ And its Nuclear PostureReview expands the conditions under which the United States would threaten nuclear use and plans an increased arsenal of low-yield nuclear bombs.⁷⁸ Escalating rivalries are the likely result. Not only is this imbalance between power and commitments financially difficult to service. It also makes the country harder to govern. Recurrent clashes over federal budgets and the increasing tradeoffs between consumption, investment, and defense lead to periods of paralysis. We see a dangerous interaction between domestic discord and foreign policy failure.

These deteriorating circumstances make it imperative for Washington to conduct a cold reassessment of its grand strategy. It needs to ask what works and what doesn’t, to rank its interests into a hierarchy and distinguish what is vital from what is desirable, to assess what is achievable, and what costs and sacrifices it can bear. The growing demand on already scarce resources, from the mounting costs of defense to the current and future burdens of entitlements, means that it will be difficult for the superpower to increase its extraction of resources from its population base. For a reassessment to be realistic, the country must be able to consider retrenchment, burden shifting, the accommodation of potential rivals, and the limitation of commitments. History suggests strategies that bring a state’s power and commitments into balance and that can successfully prevent overstretch, insolvency, or exhaustion.⁷⁹ To do this, decisionmakers can draw on an American tradition of prudential, realist thinking about aligning resources and goals. As Samuel P. Huntington summarized it, to address the gap between ambitions and capabilities, states can attempt

to redefine their interests and so reduce their commitments to a level which they can sustain with their existing capabilities; to reduce the threats to their interests through diplomacy; to enhance the contribution of allies to the protection of their interests; to increase their own resources, usually meaning larger military forces and military budgets; to substitute cheaper forms of power for more expensive ones, thus using the same resources to produce more power; to devise more effective strategies for the use of their capabilities, thereby securing also greater output in terms of power for the same input in terms of resources.⁸⁰

If, however, Washington is held to a fictitious and demanding historical standard, this exercise will be impossible. If “liberal order” visions prevail, it will be deemed immoral even to consider an alternative of restraint. A pernicious byproduct of such nostalgia is its reductionism, whereby traditionalists assert a false choice between primacy or “global leadership” on one hand and inward-looking isolation on the other. Accordingly, advocates of primacy brand today’s realists who call for retrenchment as Trumpian.⁸¹ By contrast, if Washington can be liberated from the burdensome historical fantasy that hegemonic nostalgists impose upon it, then it can gain a clearer-sighted appreciation of the choices now before it.

Notes:

1. As Anne Applebaum puts it, “Is This the End of the West as We Know It?,” Washington Post, March 4, 2016.
2. Patrick M. Stewart, “Trump and World Order: The Return of Self-Help,” Foreign Affairs 96, no. 2 (March/April 2017): 52-57.
3. James Kirchick, The End of Europe: Dictators, Demagogues, and the Coming Dark Age (New Haven, CT: Yale University Press, 2017).
4. “Is the Liberal Order in Peril?,” Foreign Affairs Online, undated, https://www.foreignaffairs.com/ask-the-experts/liberal-order-peril.
5. Naazneen Barma, Ely Ratner, and Steven Weber, “The Mythical Liberal Order,” National Interest 124 (March/April 2013): 56-67; see also Andrew Bacevich, “The Global Order Myth,” The American Conservative, June 15, 2017; Brahma Chellany, “Mirage of a Rules-Based Order,” Japan Times, July 25, 2016; and Michael Brendan Dougherty, “The Endless Hysteria about the Liberal World Order,” National Review, March 27, 2018.
6. Adam S. Posen, “The Post-American World Economy: Globalization in the Trump Era,” Foreign Affairs 97, no. 2 (March/April 2018): 28-38.
7. See Mackubin Thomas Owens, “A Balanced Force Structure to Achieve a Liberal World Order,” Orbis 50, no. 2 (Spring 2006): 307-25; Michael Mandelbaum, The Case for Goliath: How America Acts as the World’s Government in the 21st Century (New York: Public Affairs, 2005), pp. 34-35; Zbigniew Brzezinski, The Grand Chessboard: American Primacy and Its Geostrategic Imperatives (New York: Basic Books, 1997), pp. 30-57; Aaron Friedberg, “China’s Understanding of Global Order Shouldn’t Be Ours,” Foreign Policy Online, January 24, 2018; and Stephen Brooks and William Wohlforth, America Abroad: The United States’ Global Role in the 21st Century (New York: Oxford University Press, 2016), p. 4.
8. Richard N. Haass, “Liberal World Order: R.I.P,” Project Syndicate, March 21, 2018.
9. Aaron Friedberg, Twitter post, February 10, 2018, 3:47 p.m.
10. Jeremy Suri, “How Trump’s Executive Orders Could Set America Back 70 Years,” Atlantic, January 27, 2017.
11. See, for example, “Full Text: Donald Trump 2016 RNC Draft Speech Transcript,” Politico, July 21, 2016, https://www.politico.com/story/2016/07/full-transcript-donald-trump-nomination-acceptance-speech-at-rnc-225974.
12. See J. G. Ruggie, “International Regimes, Transactions, and Change: Embedded Liberalism in the Postwar Economic Order,” International Organization 36, no. 2 (Spring 1982): 380.
13. Amitav Acharya, The End of American World Order (Malden, MA: Polity, 2014), pp. 106-19; Henry Kissinger, World Order: Reflections on the Character of Nations and the Course of History (New York: Penguin, 2014), pp. 361-75; and Michael J. Mazarr, “The Once and Future Order: What Comes after Hegemony?,” Foreign Affairs 96, no. 1 (January/February 2017): 25-32.
14. Hal Brands and Colin Kahl, “Trump’s Grand Strategic Train Wreck,” Foreign Policy Online, January 31, 2017.
15. Jake Sullivan, “The World after Trump: How the System Can Endure,” Foreign Affairs 97, no. 2 (March/April 2018): 10.
16. Geir Lundestad, “Empire by Invitation? The United States and Western Europe, 1945-1952,” Journal of Peace Research 23, no. 3 (1986): 263-77.
17. The journal Foreign Affairs, the house organ of the foreign policy establishment, asked 32 experts whether the “liberal order is in peril,” and most agreed it is, with 26 respondents registering a “confidence level” of 7 out of 10.
18. G. John Ikenberry, “The Future of the Liberal World Order: Internationalism after America,” Foreign Affairs 90, no. 3 (May/June 2011): 56.
19. G. John Ikenberry, “The End of International Order?,” International Affairs 94, no. 1 (January 2018): 7-23.
20. A representative article in this regard is Martin Wolf, “Davos 2018: The Liberal International Order Is Sick,” Financial Times, January 23, 2018.
21. Robin Niblett, “Liberalism in Retreat: The Demise of a Dream,” Foreign Affairs 96, no. 1 (January/February 2017): 24.
22. Doug Stokes, “Trump, American Hegemony, and the Future of the Liberal International Order,” International Affairs 94, no. 1 (January 2018): 133-50.
23. For an honorable exception, see Jeff D. Colgan and Robert O. Keohane, “The Liberal Order Is Rigged,” Foreign Affairs 96, no. 3 (May/June 2017): 36-44.
24. Building Situations of Strength: A National Security Strategy for the United States, Brookings Institution, 2017, https://www.brookings.edu/research/building-situations-of-strength/.
25. Robert Kagan, “The Twilight of the Liberal World Order,” in Big Ideas for America, ed. Michael O’Hanlon (Washington: Brookings Institution Press, 2017), pp. 267-75.
26. Eliot A. Cohen, “Trump’s Lucky Year: Why the Chaos Can’t Last,” Foreign Affairs 97, no. 2 (March/April 2018): 2-9; Eliot A. Cohen, The Big Stick: The Limits of Soft Power and the Necessity of Military Force (New York: Basic Books, 2016); Robert J. Lieber, comments in “Roundtable 10-2,” International Security Studies Forum, October 27, 2017; Robert J. Leiber, Retreat and Its Consequences: American Foreign Policy and the Problem of World Order (Cambridge: Cambridge University Press, 2016); and Thomas Wright, “The Foreign Crises Awaiting Trump,” Atlantic, January 20, 2017.
27. Constance Duncombe and Tim Dunne, “After Liberal World Order,” International Affairs 94, no. 1 (January 2018): 25-42; and Tim Dunne and Trine Flockhart, eds., Liberal World Orders (New York: Oxford University Press, 2011).
28. Robert Cooper, “The New Liberal Imperialism,” Guardian, April 7, 2002.
29. Robert Cooper, The Breaking of Nations: Order and Chaos in the Twenty-First Century (New York: Atlantic Books, 2004), pp. 61-62.
30. Michael Ignatieff, “The American Empire: The Burden,” New York Times, January 5, 2003.
31. See Jane Mayer, “Outsourcing Torture,” New Yorker, February 14, 2005, pp. 106-23.
32. Thomas Wright, a proponent of liberal world order, also argued that the United States, especially during the Cold War, exercised informal imperial control: see Daniel Nexon and Thomas Wright, “What’s at Stake in the American Empire Debate?,” American Political Science Review 101, no. 2 (May 2007): 253-72; Eliot A. Cohen, “Calling Mr. X,” New Republic, January 19, 1998, pp. 17-19; Richard Haass, “Imperial America,” paper presented at the Atlanta Conference, November 11, 2000; and Max Boot, The Savage Wars of Peace: Small Wars and the Rise of American Power (New York: Basic Books, 2002).
33. Andrew Hurrell, On Global Order: Power, Values, and the Constitution of International Society (New York: Oxford University Press, 2007); and Daniel Deudney and G. John Ikenberry, “The Nature and Sources of Liberal International Order,” Review of International Studies 25, no. 2 (April 1999): 179-96.
34. Michael Anton, “America and the Liberal International Order,” American Affairs 1, no. 1 (Spring 2017): 113-25.
35. Edward Luce, “The New World Disorder,” Financial Times, June 24, 2017.
36. Michael Lind, “There’s No Such Thing as ‘the’ Liberal World Order,” National Interest, September 5, 2017.
37. Niall Ferguson, “The Myth of the Liberal International Order,” Global Times, January 11, 2018.
38. Leon Hollerman, “International Economic Controls in Occupied Japan,” Journal of Asian Studies 38, no. 4 (1979): 719.
39. As Pankaj Mishra argues in “The Rise of China and the Fall of the ‘Free Trade’ Myth,” New York Times Magazine, February 7, 2018.
40. Gowling WLG, “Global Protectionism: Are You Leaving Yourself Open?,” p. 6, https://gowlingwlg.com/en/insights-resources/topics/global-protectionism/overview.
41. Marc Jones, “World Has Racked Up 7,000 Protectionist Measures since Crisis: Study,” Reuters, November 14, 2017.
42. See, for example, Daniel Griswold, Stephen Slivinski, and Christopher A. Preble, “Ripe for Reform: Six Good Reasons to Reduce U.S. Farm Subsidies and Trade Barriers,” Cato Institute Trade Policy Analysis no. 30, September 14, 2005, https://www.cato.org/publications/trade-policy-analysis/ripe-reform-six-good-reasons-reduce-us-farm-subsidies-trade-barriers.
43. Sheldon L. Richman, “Ronald Reagan: Protectionist,” Free Market 6, no. 5 (1988).
44. See Maura R. O’Connor, “Subsidizing Starvation,” Foreign Policy Online, January 11, 2013.
45. “Bush Says Sacrificed Free-Market Principles to Save Economy,” CNN, December 17, 2008.
46. Mark P. Lagon and Bettany L. Fried, “China and Singapore: The Models Not to Follow,” Freedom House, August 24, 2016; and Freedom House, Freedom in the World 2017, pp. 3, 20-21.
47. Jeffrey Sachs,“Eastern Europe’s Economies,” The Economist, January 13, 1990.
48. See Lawrence R. Klein and Marshall Pomer, The New Russia: Transition Gone Awry (Palo Alto, CA: Stanford University Press, 2002); and David Stuckler, Lawrence King, and Martin McKee, “Mass Privatisation and the Post-Communist Mortality Crisis: A Cross-National Analysis,” Lancet 373, no. 9661 (January 2009): 399-407.
49. Joseph E. Stiglitz, Globalization and Its Discontents (New York: Penguin, 2002), pp. 144-45.
50. “Mass Murder and the Market,” The Economist, January 22, 2009.
51. See Stephen Cohen, Failed Crusade: America and the Tragedy of Post-Communist Russia (New York: W. W. Norton, 2000), p. 15.
52. For a strong articulation of this case, see Richard B. Frank, Downfall: The End of the Imperial Japanese Empire (New York: Penguin, 1999), pp. 349-61.
53. See Nuno Monteiro, “Unrest Assured: Why Unipolarity Is Not Peaceful,” International Security 36, no. 3 (Winter 2011/2012): 9-40; and Bruce Porter, “The Warfare State,” American Heritage 45, no. 4 (July/August 1994): 56-69.
54. See Mary L. Dudziak, War-Time: An Idea, Its History, Its Consequences (New York: Oxford University Press, 2012), pp. 95-133.
55. Benn Steil, “How to Win a Great Power Competition: Alliances, Aid, and Diplomacy in the Last Struggle for Global Influence,” Foreign Affairs Online, February 9, 2018.
56. Keith Kyle, Suez: Britain’s End of Empire in the Middle East (London: I. B. Tauris, 1991, 2011 edition), pp. 411-12.
57. The definitive book on the U.S. economic pressure on the United Kingdom and France during this period is Diane B. Kunz, The Economic Diplomacy of the Suez Crisis (Chapel Hill: University of North Carolina Press, 1991). On Eisenhower and the Suez Canal Crisis, see Evan Thomas, Ike’s Bluff: President Eisenhower’s Secret Battle to Save the World (New York: Little, Brown, 2012), pp. 219-34.
58. “Yemen 2017/2018,” Amnesty International, https://www.amnesty.org/en/countries/middle-east-and-north-africa/yemen/report-yemen/.
59. See “From Dictatorship to Democracy: The Role Ex-Nazis Played in Early West Germany,” Spiegel Online, March 6, 2012; and Stefan Wagstyl, “Postwar West German Ministry ‘Burdened’ by Ex-Nazis, Study Says,” Financial Times, October 10, 2016.
60. Cited in David Harvey, A Brief History of Neoliberalism (New York: Oxford University Press, 2005), p. 6.
61. Jeanne Morefield, Empires without Imperialism: Anglo-American Decline and the Politics of Deflection (New York: Oxford University Press, 2014), p. 242.
62. For this argument in greater depth, see Michael C. Desch, “America’s Liberal Illiberalism: The Ideological Origins of Overreaction in U.S. Foreign Policy,” International Security 32, no. 3 (Winter 2007/2008): 7-43.
63. David H. Petraeus, “America Must Stand Tall,” Politico, February 7, 2017.
64. On “signature strikes,” see Micah Zenko, Reforming U.S. Drone Strike Policies, Council on Foreign Relations Special Report no. 65, January 2013, pp. 12-14; on Petraeus’s advocacy of signature strikes, see Greg Miller, “CIA Seeks New Authority to Expand Yemen Drone Campaign,” Washington Post, April 18, 2012.
65. David Petraeus, interview by Jake Tapper, “Petraeus Explains How Jihadis Could Be Peeled Away to Fight ISIS — and Assad,” CNN Politics, September 1, 2015.
66. See Gene Gerzhoy, “Alliance Coercion and Nuclear Restraint: How the United States Thwarted West Germany’s Nuclear Ambitions,” International Security 39, no. 4 (Spring 2015): 91-129.
67. Richard Betts, “A Disciplined Defense: How to Regain Strategic Solvency,” Foreign Affairs 86, no. 6 (November/December 2007): 67-80.
68. See my article, Patrick Porter, “Why American Grand Strategy Has Not Changed: Power, Habit and the Foreign Policy Establishment,” International Security 42, no. 4 (Spring 2018): 9-46.
69. Jack Snyder, The Myths of Empire: Domestic Politics and International Ambition (Ithaca, NY: Cornell University Press, 1991), p. 60.
70. Congressional Budget Office, The 2017 Long-Term Budget Outlook (March 2017), p. 39, https://www.cbo.gov/publication/52480.
71. Manmohan S. Kumar and Jaejoon Woo, “Public Debt and Growth,” IMF Working Paper, WP/10/74, July 2010, http://www.imf.org/external/pubs/ft/wp/2010/wp10174.pdf; and Carmen M. Reinhart and Kenneth S. Rogoff, “Growth in a Time of Debt,” American Economic Review 100, no. 2 (May 2010): 573-78, https://scholar.harvard.edu/files/rogoff/files/growth_in_time_debt_aer.pdf; see also Salim Furth, “High Debt Is a Real Drag,” Heritage Foundation Issue Brief no. 3859, February 22, 2013.
72. Neta C. Crawford, U.S. Budgetary Costs of Wars through 2016: $4.79 Trillion and Counting, Watson Institute, Brown University, September 2016, http://watson.brown.edu/costsofwar/files/cow/imce/papers/2016/Costs%20of%20War%20through%202016%20FINAL%20final%20v2.pdf.
73. See Thomas Oatley, A Political Economy of American Hegemony: Buildups, Booms, and Busts (New York: Cambridge University Press, 2015).
74. Jennifer Rizzo and Brad Lendon, “Pentagon Asks for Major Budget Increase amid Threats from Russia, China, and North Korea,” CNN, February 13, 2018.
75. According to the estimate of the Watson Institute’s “Costs of War” Project: http://watson.brown.edu/costsofwar/military-spending-2018.
76. John Haltiwanger, “Trump’s Secret War? U.S. Military’s Presence in Middle East Has Grown 33 Percent in Past Four Months,” Newsweek, November 21, 2017.
77. White House, National Security Strategy of the United States of America (Washington: White House, December 2017), pp. 2, 25.
78. Office of the Secretary of Defense, Nuclear Posture Review (Washington: Department of Defense, February 2018), p. XII.
79. Paul K. MacDonald and Joseph M. Parent, “Graceful Decline? The Surprising Success of Great Power Retrenchment,” International Security 35, no. 4 (Spring 2011): 7-44; T. V. Paul, ed., Accommodating Rising Powers: Past, Present, and Future (New York: Cambridge University Press, 2016), p. 21.
80. Samuel P. Huntington, “Coping with the Lippmann Gap,” Foreign Affairs 66, no. 3 (May/June 1988): 453-77.
81. Hal Brands and Peter Feaver, “Saving Realism from the So-Called Realists,” Commentary, August 14, 2017.

Patrick Porter is professor of international security and strategy at the University of Birmingham.

Erik Goepner

Afghans have endured 40 years of uninterrupted war, and there is no plausible argument that war will soon end. In all the debate about troop surges or maintaining the status quo, two critical questions rarely get asked: Why have Afghans been at war for so long, and why can’t the United States and the international community end it? Some of the obvious answers include an incompetent Afghan government and security force, rebel sanctuaries in the mountains and in Pakistan, and the lucrative and illicit opium trade. Almost entirely ignored, however, is the role played by the decades of bone-jarring trauma experienced by Afghans.

Afghanistan has become a trauma state, stuck in a vicious cycle: war causes trauma, which drives more war, which in turn causes more trauma, and so on. Thanks to 40 years of uninterrupted war, Afghans suffer from extremely high rates of post-traumatic stress disorder and other mental illnesses, substance abuse, and diminished impulse control. Research shows that those negative effects make people more violent toward others. As a result, violence can become normalized as a legitimate means of problem solving and goal achievement, and that appears to have fueled Afghanistan’s endless war. Thus, Afghanistan will be difficult, if not impossible, to fix.

Trauma at this level imposes profound limits on America’s ability to effect enduring change in Afghanistan and other places. Accordingly, the United States should decrease its military footprint in the country and focus on efforts to incentivize a more effective and less corrupt Afghan government. More broadly, America should restrain its use of military force to those instances in which it is both effective and necessary, since sustained war in already traumatized states such as Afghanistan increases psychological damage and societal instability, making continued war more likely. Although it has become a common element of U.S. foreign policy, intervening with military force in another country’s civil war is almost never necessary to secure U.S. interests. When the United States does intervene, however, the population’s mental health status should be included in military planning and intelligence estimates as a relevant factor affecting the war and the likelihood of future stability.

Introduction

The central thesis of this analysis is that 40 years of war have fundamentally changed Afghans and made the country more prone to war in the future.¹ A coup in 1978 ushered in a civil war followed immediately by the Soviet invasion. By the time the Soviet Union left in 1989, 7 percent to 9 percent of the Afghan population had been killed, with the death count rising to a staggering one in five for working-age males.² Civil war resumed. Before the U.S. invasion in 2001, war in Afghanistan had already killed, wounded, or displaced half of the population.³ Then in late 2001, the United States invaded Afghanistan to destroy al Qaeda and dislodge the Taliban and later continued fighting to stabilize the country and establish a democratic government. As of 2018, Afghans remain mired in war, and the Taliban contest, influence, or control more territory than at any point since America initiated combat operations.⁴

U.S. efforts have been significant, yet American objectives remain largely unmet. Since October 2001, more than 2,000 Americans have been killed in Afghanistan at an estimated financial cost of $840 billion. Forty-one other countries have also contributed to the war in varying degrees.⁵

Seventeen years in, the United States remains torn between maintaining the status quo, surging military forces, or leaving the country altogether. The Trump administration has chosen to surge forces, but regardless of the path pursued, Americans can expect continued civil war involving the Taliban and other insurgent groups, as well as a corrupt, illiberal, and largely incompetent Afghan government. An end to the violence will happen only after one group finally monopolizes the use of force in Kabul and a sufficient number of provinces outside the capital, but even then there is a much higher than average probability that civil war will resume.

Unfortunately, neither the United States nor the international community can substantially improve Afghanistan’s situation. Instead, the future of the country rests primarily in the hands of Afghans who, to date, have largely been incapable of or uninterested in fundamentally changing conditions on the ground. A large number of policy analyses suggest otherwise: that a substantial and enduring U.S. presence will sufficiently improve the situation. However, those analyses typically ignore two critical questions: Why have Afghans been at war for so long? And why haven’t the United States and the international community ended the war after 16 years of trying?

The reasons for Afghanistan’s bleak future can be found in the answers to those two questions. Some of the more obvious explanations include the Afghan National Defense and Security Forces’ failure to stop the insurgency, the low opportunity cost of rebel recruitment, and insurgent sanctuary in the mountains and in Pakistan. Other likely causes include rebels motivated by grievances against their extremely corrupt government, as well as ethnolinguistic fractionalization between Pashtuns and others (e.g., Tajiks, Uzbeks, and Hazaras). Financial incentives likely motivate a number of insurgent groups too, as perpetual war perversely provides them an enduring income stream from the illicit opium trade that would otherwise be confined to traditional criminal elements if the conflict ended.

In addition to analyzing those areas, this policy analysis goes a step further and suggests an answer largely overlooked in the security studies literature — Afghanistan as trauma state. Simply put, Afghans have endured so much trauma that the society has fractured and now finds itself unable to function normally. Beyond Afghanistan, this analysis should inform future U.S. policies toward other states in the midst of civil war, such as Iraq, Syria, and Libya, which have histories of extreme trauma and are home to America’s current nemesis, Islamist-inspired terrorists.

This analysis begins with a brief review of the literature on the prevalence of civil war. Why are some countries, like Afghanistan, home to so much civil war while others never experience it? The next section provides a detailed answer to the more specific question, Why is there so much war in Afghanistan? The subsequent sections explore the reasons American and international efforts have failed to end the war in Afghanistan. The final section offers recommendations for U.S. efforts in Afghanistan now and for other high-trauma civil war states in the future.

Afghanistan’s Endless War

My experience in Khaki Khel in 2010 explains a lot.⁶ Our helicopters touched down just outside the village in this remote province in southern Afghanistan. The Afghan Army and police, along with an American military unit, had just conducted operations in and around the village, more to build confidence among the Afghan population than to kill or capture insurgents. At the conclusion of such operations, I would normally escort members of the Afghan government and medical community out to facilitate a dialogue between the government and their village constituents and to provide basic medical care. This time was different. The U.S. military unit made a mistake one evening when they fired off an illumination round. Instead of safely falling to earth after jettisoning its contents, the metal canister sliced through the bodies of two children asleep on the roof of their home to escape the summer heat inside. We had all come to pay our respects. Because of Afghan cultural considerations regarding gender roles, my female lieutenant’s mission for the day was to spend time with the grieving mother, apologize for the tragedy, and express our sympathies to her.

We made our way down the steep hill, away from the mud homes and into a field of poppies. A clearing opened up, and we took our seats to begin the shura (a traditional meeting of elders designed to share important information and potentially arrive at a consensus-based decision). The literacy rate varied between 1 percent and 10 percent in our province, but out here in such a remote place with no real access to a school or qualified teacher, probably less than 1 percent could read and write. They had no electricity, no cars, and no paved roads to connect them to other Afghans or business opportunities. As subsistence farmers, they lived harvest to harvest, and the droughts of the past few years had reduced crop yield and killed livestock. The gentle rebuke of other elders at an earlier time had taught me to never again say, “What a beautiful day” or, “Isn’t the weather nice?” These Afghans operated on a different step of Maslow’s hierarchy of needs. For them, the weather was the difference between being able to feed your family or going hungry. The poppy field loomed in the background, both an economic incentive for the impoverished residents and a reminder of why the insurgents were interested in this remote village.

In accordance with custom, the village elders spoke first. They hit on two of the talking points that we would hear at virtually every shura: Why are you here when the insurgents are over there (said while pointing in the direction of neighboring Pakistan) and how do you expect us to stand up to the Taliban when you and 42 other countries can’t defeat them? The second point was so specific that we had long ago concluded that the insurgents had actually told it to them. The number would vary slightly from village to village, but it always stayed between 40 and 49. We concluded that the insurgents used the talking point to intimidate the villagers into submission, along the lines of: “Don’t bother resisting us. If the United States and 42 other countries with all of their weapons and technology can’t beat us, you shouldn’t think you can either.”

As the elders spoke in turn, one lost his composure and became very emotional in this public setting, something our training had suggested Pashtun men avoided at all costs. He implored the Afghan and American security forces to fight the Taliban “down here” in the field, not up among the villagers’ homes. He said it several times and in different ways. The message was clear: tell the Taliban to fight you down here, so innocent villagers won’t be caught in the crossfire. I turned to my interpreter, an Afghan man in his mid-20s who had my complete trust and admiration, and asked if he thought the elder was joking. “No,” he replied, “he is serious.” When my interpreter and I talked later, we concluded that the elder did not know the insurgents often intentionally put civilians in harm’s way as a means to achieve their goals. We also concluded that the elder truly believed the insurgents would fight us at the location of our choosing. Of course, from the insurgents’ perspective, their survival required them to never meet us at a time or place of our choosing.

Khaki Khel illuminates both of the primary arguments for the prevalence of civil war: the opportunity for rebellion and the motivation to rebel. The opportunity for rebellion typically exists when the state has ineffective or nonexistent security forces, when recruiting rebels is easy, and when rebels can readily find sanctuary. Grievances among the population and rebel groups’ desire to financially profit from illicit activities have typically fueled the motivation to rebel.⁷ Afghan security forces rarely operate in Khaki Khel, so for practical purposes the security forces do not exist. And even when combined with the U.S. military, Afghan forces were ineffective, having accidentally killed members of the very village they had come to protect. Additionally, with the villagers living harvest to harvest, low opportunity costs for rebel recruitment persisted. Even modest payments from insurgents go a long way for the average Khaki Khel family. As for the motivation to rebel, the villagers certainly have a number of grievances to choose from: an incompetent government that cannot even provide them with security, a corrupt government rated worse than 96 percent of all governments in the world, and a government that unwittingly kills their fellow villagers.⁸ Finally, the poppy field serves as a visible reminder of the perverse role financial incentives may play in fueling the rebellion.

But Khaki Khel made it clear that existing theories were incomplete. Something else important was going on. Evidence for a new theory emerged: a vicious cycle of war causing extreme trauma, and trauma, in turn, causing more war.

Afghanistan, the War State

Afghanistan’s 40 straight years of war provide ample support for both of the main theories of civil war and all of their subarguments. Data from those 40 years also support the argument that as war begets trauma, trauma also perpetuates war, and the vicious cycle continues.

Opportunity for Rebellion. The opportunity for rebellion has long been a feature of Afghan life. A strong federal government has proved elusive. The state security force has been largely nonexistent and, where present, highly ineffective. Rebels enjoy safe haven in the extensive mountains and in neighboring Pakistan, while Afghanistan’s enduring poverty reduces the costs of recruiting new rebels.

Ineffective or Nonexistent Security Forces. For the past century, Afghanistan has had an ineffective state security force, with the possible exception of 1953 to 1963 during Mohammed Daoud Khan’s tenure as prime minister.⁹ The country’s five civil wars during the 20th century speak to the enduring incompetence of Afghan security forces.¹⁰ Observers have characterized them as “almost useless,” “tactically inept,” “in disarray,” and able to conduct “only limited defensive operations.”¹¹ Today, despite numbering more than 365,000, they barely control or influence half of the country’s districts.¹² This incompetence, in part, motivated the Soviet Union to invade in 1979 and to try to prop up the communist regime in Kabul. Today, security force ineffectiveness keeps American advisers and trainers there after 16 years of trying to professionalize the Afghan force.

Incompetent Afghan security forces also make the villagers’ lives more difficult. At shuras I attended, the senior Afghan government leader presented gifts, paid for by American taxpayers, to each of the elders in attendance. At slightly more than half of those shuras, I watched the elders politely refuse the gifts. Initially, it seemed to make no sense. Even if the elders hated their government, why refuse a free prayer rug or Koran? Years earlier, Osama bin Laden had reminded his followers of the value of trying to financially bankrupt their enemies. Taking the gifts with no strings attached appeared to be a good, albeit small, way to get back at America.

At one shura, however, my confusion was cleared up: everything came with strings attached. As we made our way into the village, only one elder greeted us. The Afghan deputy governor expressed his disbelief, “Where are all of your white beards?” The elder said the others were out in the fields working, as he pointed up into the nearby hills.

The deputy governor did not believe him. He noted that the Afghan Army and the Americans had been out there for three days, and they had certainly told the elders we were coming today.

The back-and-forth went on for 15 minutes until the elder finally admitted that the others were hiding in a nearby compound for fear of what the insurgents would later do to them. He agreed to fetch the other elders and allow the shura to proceed. As the shura began, one of the elders drove the point home. He asked us and his government not to come out to his village anymore. He said that three years before, the Afghan government and a previous American unit had come to the village and given them supplies to clean their irrigation system. A few days later, after the government and the Americans left, the insurgents came and destroyed much of their kareze (a traditional communal irrigation system that relies on tunneling to tap existing groundwater) and abused their elders in front of everyone, accusing them of working with the infidel government and the Americans. Two years before that, he said, the government and the Americans came and dug the villagers a well. But after they left, the insurgents returned, destroyed the villagers’ well, and humiliated the elders in front of the people. The elder acknowledged that the previous year the government and the Americans had respected their wishes and did not come out. He said that the villagers had put their meager monies together and bought a small farming machine. Unfortunately, the insurgents assumed the Afghan government or the Americans had bought it for them, so they destroyed it and again punished the elders for cooperating with the infidels.

He made his point clear enough. The transitory presence of the Afghan security forces put the elders and villagers in greater peril than if the security forces had never come. In both scenarios, the Taliban largely controlled the lives of these Afghans, but in the second scenario the Afghans suffered less.

Low Opportunity Costs for Rebel Recruitment. Afghanistan has typically had low opportunity costs for rebel recruitment, and persistent war has only reduced them further. As Figure 1 shows, for the last 40 years of the 20th century, Afghanistan’s gross domestic product (GDP) per capita averaged 40 percent of that of all states at war and one-sixth of the worldwide average.¹³ When America invaded in 2001, Afghans had an average annual income of $117.¹⁴ The Central Intelligence Agency’s 2017 World Factbook ranks Afghanistan 207th out of 230 for income per capita despite the billions poured in by the United States and other members of the international community.¹⁵ Afghanistan’s GDP per capita is 11 percent of the global average (and 2.8 percent of America’s). Comparatively, Afghanistan hit its high-water mark in 1950, when its GDP per capita reached 30 percent of the global average. Afghanistan’s GDP per capita has also grown at an inferior rate. Over the past 60 years, the worldwide average has grown 263 percent versus just 35 percent for Afghanistan.¹⁶

Figure 1: Gross domestic product per capita, 1960-1999 (in current U.S. dollars)

Source: Ibrahim Elbadawi and Nicholas Sambanis, “How Much War Will We See? Explaining the Prevalence of Civil War,” Journal of Conflict Resolution 46, no. 3 (2002): 307-34.

Afghans suffer from extremely high rates of illiteracy, and even after 16 years of effort from the international community the prospects for improvement remain bleak. The United Nations Educational, Scientific and Cultural Organization (UNESCO) estimates a 32 percent literacy rate for Afghans as of 2011, woefully below the global average of 85 percent.¹⁷ ONE, an international nongovernmental organization, currently rates Afghanistan the world’s fourth-worst country in education for girls.¹⁸ An uneducated workforce offers little hope for economic growth. Insurgency, therefore, remains an attractive source of income.

Two bright spots, however, emerged in Khaki Khel, where a meager 1 percent to 10 percent were literate. When we went on patrol, the children would gravitate to us with shouts of “Qalam, Mister. Qalam, Mister.” They wanted pens. Not money or food, but pens. My Afghan interpreter thought it indicated both an aspiration to be literate on the kids’ part and a desire to have something unique and prestigious. The young women in Alamat, our province’s capital city, represented the second bright spot. They wore their uniforms when walking to school, which made them a visible target for the insurgents. They could easily have put on burqas, but they chose to have everyone see them in their uniforms, a particularly bold step in the very conservative province.

Rare bright spots aside, decades of war have crushed Afghanistan’s economy and the human capital that would normally undergird it. GDP per capita has been set back, making insurgency a more lucrative source of income. However, the widespread trauma, the internally displaced persons and refugees, and the lack of education have severely affected Afghans’ capabilities over the long term.¹⁹ When countries remain at war for too long, waging war can become the citizens’ only marketable skill.²⁰

Rebel Sanctuary. Insurgents have benefited from sanctuary in neighboring Pakistan.²¹ The Terrorism Research and Analysis Consortium indicates that Taliban leaders have enjoyed safe haven in Pakistan since shortly after the United States initiated combat operations in Afghanistan back in 2001. The so-called Quetta Shura — named after the Pakistani city in which they enjoy refuge — even openly collects funds through various charity fronts in Quetta and other Pakistani cities.²² In 2009, President Barack Obama publicly called on Pakistan to “demonstrate its commitment to rooting out al Qaeda and the violent extremists within its borders.”²³ President Trump recently repeated a similar refrain when he announced a surge of forces back into Afghanistan: “We can no longer be silent about Pakistan’s safe havens for terrorist organizations, the Taliban, and other groups that pose a threat to the region and beyond.”²⁴ Presidential protestations aside, Pakistan has provided sanctuary to rebels for 16 years and counting.

Additionally, the Taliban, al Qaeda, and other insurgent groups have successfully sought refuge in Afghanistan’s mountains.²⁵ Osama bin Laden, for example, hid in the Spin Ghar mountain range before escaping during the Battle of Tora Bora in late 2001.²⁶ More recently, the Taliban occupied the same cave and tunnel complex only to be sent fleeing, ironically, by Islamic State fighters who then took up residence there.²⁷ Territorial sanctuaries continue to enhance insurgent viability.

Motivation to Rebel. Grievances. “We don’t want any more of your mercy.” The elder spoke those words during a shura between his village and the Afghan provincial government in the summer of 2010. His comments seemed directed primarily to his government and, to a lesser extent, me and the other representatives of the U.S. military in attendance. As my interpreter translated, I could not help but think, “With friends like these, who needs enemies?” The elder’s turn of phrase, however, held no hint of humor, only years of pent-up pain and anguish. And he was basically right. Legitimate (and some illegitimate) grievances filled his life and the lives of the villagers he represented. His government remained mired in the worst levels of corruption and delivered no goods or services to his village: no roads, schools, or agricultural assistance and, worst of all, no security. Insurgents came and went at will, forcing villagers to provide them with food and other logistical needs, while humiliating and beating elders to compel compliance. Additionally, the insurgents, typically young men in their 20s, upended the Afghan cultural norm of respect and deference to the aged.

In addition to legitimate and enduring grievances against their government, water rights and ethnic fractionalization also fuel lasting resentments. The Asia Foundation’s annual survey of Afghan sentiments routinely shows water availability to be one of the top concerns at the local level.²⁸ During my time there, frustrations over water access became evident in several ways. First, residents near a U.S. military base had lodged a complaint well before my arrival, a complaint that persisted after I had departed. They alleged that construction on the military base had unintentionally curtailed their water supply to almost nothing.

Second, elders petitioned us to install several wells in their village. My staff checked the records left by previous teams and noted that five wells had reportedly been dug in that village a few years earlier. When the team members shared that information with the elders, the elders fired back that one family and its extended members had monopolized those wells, leaving everyone else in the village to fend for themselves.

Ethnic fractionalization also provides fodder for enduring grievances. According to Barry Goodson, Middle East studies professor at the Army War College, absent the “preexisting ethnic tensions,” the civil war that began in 1978 would probably not have started so rapidly or spread so “vigorously.” Referring to the civil war that followed the Soviet departure, Goodson describes the “internecine fighting” among the different mujahideen groups. More broadly, Barnett Rubin of New York University notes the “powerful force” of ethnic division that pitted Pashtuns, Uzbeks, Tajiks, and Shiites against one another.²⁹

Financial Incentives. Financial incentives also appear to fuel Afghanistan’s enduring war, with rebel groups using the conflict to shield their unlawful activities. The illicit opium trade incentivizes war, as insurgent groups profit from their illegal activity in the midst of instability more easily than they would if an established rule of law existed and the government enforced it. The United Nations reports that Afghanistan’s opium production increased 87 percent in 2017 from the previous year. The report goes on to note that poppy cultivation occurs on 328,000 hectares of the country — the most in the 24 years of available data and exponentially higher than the 8,000 hectares under cultivation when the United States invaded in 2001.³⁰ For two decades now, opium has been the country’s “leading cash-generating economic activity,” accounting for an estimated one-third to one-half of total economic output.³¹

As noted by the U.S. Special Inspector General for Afghanistan Reconstruction, the billions of U.S. and international community dollars flooding into the country have inadvertently introduced “perverse incentives.”³² The artificial and unsustainable increase in the size of the economy encourages Afghans to enter political life for corrupt purposes and further incentivizes them to keep the war going lest Americans and their money leave. Afghan government officials have siphoned off an estimated 20 percent of each contract, while the insurgents typically require a payment as well to prevent them from destroying the new project.³³ The net result? More grievances against the government, increased viability for the insurgents, and more war.

Afghanistan, the Trauma State

The discussion so far is common to most civil wars. But to really understand Afghanistan’s case, policymakers should consider the role of trauma. Afghans have endured four consecutive decades of bone-jarring trauma that has changed them psychologically and physically. Those changes have ushered in harmful consequences not just for the traumatized individuals, but also for the population at large, which increase the likelihood that war will continue.

The American Psychiatric Association defines a traumatic stressor as any event that may “cause or threaten death, serious injury, or sexual violence to an individual, a close family member, or a close friend.”³⁴ The severest traumatic stressors include torture, rape, and war. Increased rates of trauma are associated with psychological and physical changes to individuals, which often profoundly affect them and those around them. Exposure to traumatic stressors frequently results in mental disorders, especially post-traumatic stress disorder (PTSD) and major depressive disorder (depression).³⁵ Populations that have endured mass conflict, high rates of torture, and significant displacement similar to Afghans have a reported prevalence rate of between 17 percent and 50 percent for PTSD. That compares with a substantially lower estimated global rate of only 5 percent.³⁶

Traumatic stressors also lead to physical changes. Exposure to trauma correlates with stunted growth in key brain areas.³⁷ These changes include lowered hippocampal volume, decreased corpus callosum size, and diminished activity in the basal ganglia.³⁸ Such physical changes often result in a lowered IQ, reduced impulse control, difficulty paying attention, memory impairment, diminished capacity to reason, inability to plan, and poor problem-solving skills.³⁹

Finally, traumatic events have a more pernicious effect when the events occur during adolescence. Two generations have now come of age in the Afghan trauma state.

Afghans have been, and continue to be, exposed to an extraordinary number of traumatic events, both in severity and frequency. Studies indicate that, on average, Afghan adults have experienced 7 traumatic events, while children have endured between 5.7 and 6.6. Those events include being bombed or shelled during war, being physically beaten by members of armed groups, domestic abuse, forced displacement, and the death of a loved one.⁴⁰ That compares with fewer than one to two events for European samples, one to three for U.S. adults, and an estimated 1.8 in a multicountry sample.⁴¹ With regard to specific traumatic events, approximately 52 percent of Afghans, for example, report having experienced some form of violent assault compared with just 4 percent who live in a developed European nation.⁴² These findings generally conform to the broader trauma literature that suggests conflict-affected poor countries are home to high rates of traumatic stressors and that more traumatic stressors result in increased rates of PTSD and depression.⁴³

Trauma, Mental Illness, and Other Negative Outcomes. Mark, my young Afghan interpreter, and I had just finished another depressing meeting at the Afghan governor’s office.⁴⁴ Various officials had taken turns mocking and swearing at one another, making bizarre claims, and arriving at exactly zero decisions. We were now walking the hundred yards back to our compound. Mark’s head hung down, typically a sign he was upset.

I asked him what he made of the meeting, wanting to hear his insights into what was really going on. He, better than anyone, could make sense of the dizzying complexities of Afghans interacting with one another. His reply caught me off-guard.

“We all have PTSD,” he said. “Don’t listen to me or any other Afghans. We don’t even know what we’re saying. All of this has made us crazy.”

I did not believe him. Mark always performed his duties superbly. Moreover, during the numerous rocket attacks we endured together and the constant threat of improvised explosive devices, he never displayed the slightest fear.

I also doubted that many Afghans had PTSD because of the training I had received before deploying. The lecturers and readings (incorrectly) dissuaded me from believing that PTSD could explain much, if anything, about Afghan behavior or the war. Deeply held religious beliefs, the power of the Pashtunwali honor code, and close families, we were told, largely inoculated Afghans from mental illness or any other undesirable outcome that could be caused by the acute trauma they had experienced.

I was wrong, and so were the readings and the training we received before going to Afghanistan. It turns out that study after study shows that significant numbers of Afghans do meet the criteria for PTSD, depression, and various anxiety disorders. Although on average only 5 percent of the world’s population will meet the criteria for PTSD at any point in their lives, an estimated 29 percent of Afghans meet the definition now. Studies suggest even higher depression rates of between 37 percent and 68 percent for Afghans.⁴⁵ Virtually all of the studies conclude that the more traumatic events a person is exposed to, the more severe the follow-on negative consequences for that person — psychologically and physically. For example, children who experience five or more traumatic events have a 300 percent increased risk of mental illness (as mentioned, Afghan children have endured, on average, between 5.7 and 6.6).⁴⁶

In their meta-analysis on trauma and mental health outcomes published in the Journal of the American Medical Association, Zachary Steel (et al.) observed that populations with very high reported rates of torture had a 46 percent prevalence rate for PTSD and 50 percent for depression. When respondents also came from countries with substantial amounts of political violence and terror, as measured by the Political Terror Scale, the estimated rate of PTSD rose to 54 percent.⁴⁷ Unfortunately, and as will be further outlined later, Afghanistan has met most of these criteria for the past 40 years.⁴⁸

Making matters worse, Afghans have no real opportunity to receive professional care. Researchers have reported that Afghanistan’s mental health services are “nonexistent,” that there is an “acute shortage” of qualified providers, and that the general situation is one in which “chronic men­tal illness has been left unattended in Afghani­stan for decades.”⁴⁹

Increased exposure to traumatic stressors causes an increase in mental illness, substance abuse, and diminished impulse control.⁵⁰ People meeting the criteria for a mental disorder are 2.7 times more likely to also meet the criteria for an alcohol or drug disorder, with substantially more succumbing to a drug disorder than to one involving alcohol.⁵¹ Although approximately 30 percent of those with mental illness will also be diagnosed with a substance abuse disorder at some point, the number rises to 50 percent for those with “severe” mental disorders.⁵² Experiencing traumatic stressors, particularly during childhood, decreases an individual’s impulse control.⁵³ Chronic traumatization, like Afghanistan’s for the past 40 years, intensifies the effect.⁵⁴ No surprise, then, that both PTSD and depression are associated with impulse control disorders.⁵⁵

Trauma and Violence. Colonel Naseri began berating Colonel Habib in front of their subordinates and their American counterparts in the operations center. Habib, the number-two police officer for the province, had angered Naseri, the chief of the provincial security directorate, by arresting one of Naseri’s men. The arrest took place after an investigation into the serial raping of an Afghan boy. The Afghan National Police had wanted to arrest their prime suspect, but the suspect’s brother — an agent who worked for Naseri — kept using his position to keep the police at bay. Finally, word had made its way to Habib. Fed up, he had the suspect and his brother arrested for obstructing the investigation. Naseri fumed in response. An intellectual, he chose this moment to publicly mock the uneducated Habib, who had spent most of his life at war. We, the American forces, loved Habib. He was one of the few brave men who consistently took the fight to the enemy, and the drug addiction that we surmised he had was understandable in a land where self-medication was about the only option.

As the barrage came his way, Habib could do little to match Naseri’s verbal skills. Eventually, a switch flipped and Habib unholstered his handgun. There, in the Afghan equivalent of a war room, Habib aimed his weapon at Naseri. Fortunately, a nearby American officer rushed in between the two men and stood in front of the loaded firearm. Unable to communicate in Pashto, he spoke the only English Habib understood and accompanied it with hand gestures, “It’s OK, Habib. It’s OK.” Habib holstered his weapon. The next day, all of the government buildings had paper signs posted with a picture of a handgun surrounded by a red circle with a red line running through it: no guns allowed. The Afghan general in charge of the operations center also banned Habib from the premises for 30 days.

The mental health literature indicates that people with mental illness, substance abuse issues, and diminished impulse control commit more acts of violence against others, all else being equal.⁵⁶ Studies suggest that Afghans suffer from atypically high rates of all three. By themselves, those factors would predict that Afghanistan should be home to higher rates of violence than other countries.

Mental health experts characterize the Afghan population as being “greatly affected by psychological distress.”⁵⁷ The United Nations notes the “widespread” use of opium within the country and a problem-drug-use rate twice the global average and climbing.⁵⁸

Afghans also experience (and mete out) extremely high rates of domestic violence.⁵⁹ International and national human rights organizations, such as the UN, Amnesty International, and the Afghan Independent Human Rights Commission, assess the problem of physical abuse of Afghan women as “desperate.”⁶⁰ Global Rights described the violence as “so prevalent and so pervasive that practically every Afghan woman will experience it in her lifetime.”⁶¹

Likewise, academics have characterized the violence as “pervasive and socially tolerated.”⁶² In a study of more than 4,000 women currently living within the country, 39 percent responded that their husbands had hit them within the past year. Eighty-seven percent reported they experienced at least one form of physical, sexual, or psychological violence, and a substantial majority said they had experienced multiple forms of violence.⁶³ Nearly a quarter of the time, women identified their mother-in-law as the primary abuser. Domestic violence has become so normalized that “many women noted satisfactory marital relationships while simultaneously reporting experiences of violence in the home.”⁶⁴ By comparison, global estimates of lifetime physical abuse rates by an intimate partner range from 10 percent to 50 percent.⁶⁵

Research on Afghan children indicates that 35 percent have experienced physical violence in the past month at home, and 77 percent have experienced or observed at least one lifetime episode of violence in the home. The children, aged 7 to 15, reported an average of 4.3 lifetime violent episodes within the home. Nearly a third reported witnessing their fathers beat their mothers, and 60 percent said their mothers had beaten them (vs. 42 percent who reported their fathers had beaten them).⁶⁶

During my time in Afghanistan in 2010, nonwar violence occurred frequently and in ways that seemed excessive. One of the district chiefs — similar to a U.S. mayor — dispatched his bodyguard to establish an illegal checkpoint and shake down motorists, particularly those transporting goods for sale. The practice of illegal checkpoints was common enough that businessmen had created workarounds. In this instance, the entrepreneur had illegally hired Afghan National Police officers to guard his convoy of wares en route to Pakistan. As the loaded trucks rumbled up to the illegal checkpoint, the district chief’s bodyguard and cronies motioned for the convoy to stop. It did not. The bodyguard brandished his weapon, and the police officers responded by making it clear they were police. A firefight erupted.

Police officers assaulted one another. Even senior-ranking officials would occasionally strike their peers during arguments. At a shura, I watched as the district chief publicly smacked a police officer with all his might. The incident occurred as final preparations were being made just before the start of the meeting. Neither my interpreter nor I could determine a particular reason for the physical violence beyond the stress often associated with putting on large events attended by dignitaries. In 50 years, I have never seen that behavior in America, but within a year I saw it in Afghanistan.

The “green-on-blue” attacks — as incidents in which Afghan forces attack U.S.-led coalition forces came to be called — offer another potential example of the connection between trauma and violence. Those events, in which an Afghan security force member attempts to kill his American or coalition counterpart, began occurring more frequently in 2011 when 35 coalition forces were killed and 34 wounded over the course of 16 incidents.⁶⁷ The U.S. government estimated that 40 percent of the incidents resulted from “stress of various kinds,” while the Taliban, disguised as Afghan security forces, committed only 10 percent of the attacks.⁶⁸ In 2012, the commander of the International Security Assistance Force (and senior coalition military officer in the country) concluded that the majority of attacks stemmed from “disagreements and animosities” and “personal grievance, social difficulties.”⁶⁹ Insurgent infiltration, impersonation, and coercion accounted for only 25 percent of the attacks.⁷⁰ In response, military leaders implemented a “guardian angel” program, requiring an armed coalition member to protect other coalition forces any time they interact with their Afghan counterparts (e.g., advising, assisting, and training them).⁷¹ The guardian angel program continues as of this writing.⁷²

This perfect storm has likely made Afghans more violent and has helped legitimize violence as an acceptable option for problem solving and goal achievement in daily life. Ever-present domestic violence within the home and the war that permeates the entire society should therefore come as no surprise.

Trauma and Rebellion. Trauma’s effects on individuals and societies help explain civil war prevalence by providing an explanation for why some populations willingly resort to violence against their government while others do not. The civil war literature treats the threshold separating motivated citizens who will not employ lethal force against their government from those who will as a constant, despite, for instance, the obvious variation in violence rates across countries not at war. Research also indicates that the willful taking of human life is rare, even among military members. As a result, militaries provide substantial training to their recruits to ensure that they will actually kill in combat.⁷³ Trauma’s aggravating effect on violence norms may be a cause of civil war: as a population’s exposure to trauma increases, the risk of civil war also increases. Figure 2 diagrams the potential effect of trauma.

Figure 2 : Trauma and civil war: Hurt people hurt people

Source: Author’s representation.

Trauma can increase in three ways. First, traumatic stressors can become more severe. Victims of deliberate attacks, for instance, tend to suffer more symptoms than survivors of natural disasters. The severest traumas include torture, rape, and war. Second, the amount of trauma may increase over time. Third, the victim does not have time to heal. The trauma either continues or is relieved by only short intervals before the next traumatic exposure.⁷⁴

An additional personal anecdote serves to illustrate the normality of torture and political violence in Afghanistan. One morning as Mark and I arrived at the provincial governor’s compound to attend a staff meeting, we saw the deputy governor standing outside on the patio waiting for us. After responding to my greeting in my limited Pashto, the deputy governor directed his comments to my interpreter. He asked if I had heard about the young girl and her father who had died at Checkpoint 7. I said no and asked for some details. The deputy governor said the child and her father had been caught between the Taliban and the Afghan National Police. The Taliban ambushed the police checkpoint, and during the gun battle both father and daughter had been wounded. After the fighting had died down, the police commander refused to let any villagers into the area.

“During the night,” the deputy governor said, “she was rolling around in a small space like this.” He held his hands apart as Mark translated his words.

I could not miss the horror suggested by his literal translation. The child spent her final hours writhing in pain alongside her father, who could not save her, as the two slowly bled to death.

The deputy governor went on to confirm that the Afghan National Police commander did not help the child and her father and would not let the villagers help them either. He said the police commander reported it might be a Taliban trap and that the area remained unsafe, but he readily conceded that the police officer had lied.

The deputy governor reckoned that the police officer wanted to send the people a message: if you let the Taliban stage an attack from your village, expect no help from us. So he gave them no help, and they died. The deputy governor concluded by saying that the officer had been jailed but would be released in a few days and then reassigned to the headquarters.⁷⁵

People with mental illness, substance abuse issues, or diminished impulse control behave, on average, more violently than individuals without those conditions, and the highest risk for violence comes from individuals with both mental and substance abuse disorders.⁷⁶ The Diagnostic and Statistical Manual for Mental Disorders includes a chapter on disruptive, impulse-control, and conduct disorders that involve “problems in the self-control of emotions and behaviors.”⁷⁷ The negative effects of these disorders include behaviors that “violate the rights of others (e.g., aggression, destruction of property) and/or that bring the individual into significant conflict with societal norms or authority figures.” Experiencing traumatic stressors such as neighborhood violence, physical or sexual abuse, and harsh or neglectful parenting increases the probability of having one of these disorders.⁷⁸

To put this in context, most individuals will not kill their fellow human beings. Writing on the logic of violence within civil war, Oxford’s Stathis Kalyvas observes, “Most are repelled by the prospect of acting violently, and so they will not.”⁷⁹ Dave Grossman, a psychology professor and Army Ranger, notes that even military personnel go out of their way to avoid killing while in combat, driving the U.S. military to implement significant training efforts to ensure that they do kill.⁸⁰

Current explanations for civil war typically treat this high threshold for deadly force as a constant. This is noteworthy because norms for measures of violence, such as gun violence and murder rates, vary dramatically across countries. The International Homicide Statistics database from the UN Office on Drugs and Crime, for instance, shows the United States with a homicide rate four to eight times greater than similar countries (e.g., Canada, Australia, Germany, France, and the United Kingdom).⁸¹

Explaining societal violence after civil conflict has ended, Chrissie Steenkamp, a peacebuilding scholar, refers to a “culture of violence” in which society adopts “the norms and values that underpin the sustained use of violence.” Countries with high rates of trauma can eradicate previous norms and values and usher in new ones that “sustain the use of violence.”⁸² Political scientists Roos Haer and Tobias Bohmelt advance a similar argument for child soldiers, emphasizing the effects of trauma and the influence of learning by observation and imitation during war. In such cases, violence becomes normalized as a technique to solve problems and achieve goals in the postwar environment.⁸³

The trauma argument provides insight into why some populations are more violent than others and how this occurs (i.e., by lowering the lethal-force threshold). A traumatized society will become more violent than a nontraumatized one, all else being equal. Likewise, an equal amount of grievance or greed in a traumatized society should cause more civil war than would occur in a society without severe prior trauma.

Measuring Afghanistan’s Trauma. To assess the potential impact of trauma, it is important to measure it. Afghans have endured a sickening number of the severest traumatic events such as torture, rape, and war over the past 40 years. In all measures, they not only have suffered more trauma than the average global citizen but also have been afflicted at even higher rates than those confronting other countries at war. (Most measures have been normalized and presented in a 0-to-100 format to allow comparison across the different trauma measures. See the appendix for more information.)

Torture. Afghans have suffered extremely high rates of torture for the past 40 years (Figure 3).⁸⁴ The CIRI Human Rights Data Project (named for head researchers David L. Cingranelli and David L. Richards) made an assessment that the government “frequently” tortures its citizens — the highest possible rating — each year almost without exception.⁸⁵ The data cover 1981 through 2011. Over that period, Afghanistan received the worst score 23 times, the middle score twice, and no score for six years because of government collapse or foreign occupation.⁸⁶ In the history of the CIRI Human Rights Data Project, Afghanistan has never received the lowest (best) possible torture score.

Figure 3: Amount of torture, 1978-2014

Sources: CIRI Human Rights Data Project, http://www.humanrightsdata.com/; Political Terror Scale, www.politicalterrorscale.org.
Note: For more information on how the scores were derived, see the appendix.

The Political Terror Scale corroborates the CIRI data. The scale measures the amount of political terror by country from 1976 to 2016. Political terror is defined as “violations of basic human rights to the physical integrity of the person by agents of the state within the territorial boundaries of the state in question,” and the scale includes torture as an example of the violations governments can commit.⁸⁷ Scores range from 1 to 5, with 1 defined, in part, as “torture is rare or exceptional” and 5 as “the terrors of Level 4 have been extended to the whole population.”⁸⁸ Afghanistan has averaged a score of 4.6 over the 40 years. To put this into perspective, a score of 4 indicates “torture [is] a common part of life” and a score of 5 suggests terror affects the whole population.⁸⁹

Rape. Of all traumatic stressors, rape has the highest conditional rate for PTSD. Nearly half of women and two-thirds of men who have been raped will, at some later point, meet the criteria for the disorder.⁹⁰ The number of rapes that a population experiences typically increases during times of war.⁹¹ Having been at war for so long, Afghanistan will presumably have higher rates of rape than if the country had enjoyed peace during the same period.

In its 1995 report on Afghanistan, Amnesty International found that all warring factions “committed rape and other forms of torture,” particularly of women and children. The report also stated that women and girls throughout the country “live in constant fear of being raped by armed guards.”⁹² An interpreter I worked with, an Afghan man who lived in Kabul during the 1990s, described events that matched Amnesty International’s assessment. He said that when battle lines shifted too quickly, families did not have enough time to flee. He recalled watching in horror as several of his schoolgirl friends committed suicide to avoid being raped.

Some surveys indicate very low rates of rape in Afghanistan.⁹³ The low reporting, however, is likely a response to strong cultural taboos rather than an accurate estimate of the situation. For example, some scholars have excluded survey questions that explicitly ask about sexual violence to avoid gathering inaccurate data. Instead, they use less detailed wording or ask questions that deemphasize the sexual aspects and instead focus more broadly on the violence.⁹⁴

As Figure 4 shows, more rapes occur in Afghanistan than in the average country that experiences war, and substantially more occur in Afghanistan than the global average. Afghanistan’s score suggests rape is more than “a problem” but not “widespread.” (See the appendix for more information on this measure.)

Figure 4: Amount of rape, 1990-2014

Source: Author’s calculations, with data derived from U.S. Department of State’s annual Country Reports on Human Rights Practices.

War. As the following exchange in 2010 between a young captain on my team, an Afghan district governor, and a village elder conveys, the war has clouded people’s thinking and made them less predictable.

The elder shared his frustrations with both his government and the American forces: “We cannot come closer to you. We have no security. The Afghan forces and ISAF [International Security Assistance Force] come occasionally and only stay for a short time. When they leave, the Taliban come in and hurt us because they think we are cooperating with you,” he explained.

The American captain asked if the elder would be interested in overseeing a local self-defense force comprising men from his village. They would be armed to protect themselves and the other villagers and paid for providing security. The elder called the proposal ridiculous and said the Taliban would kill him and all his men. The captain gently probed the elder’s assumptions. He asked how many Taliban come into the village at a time. Ten to 20, the elder responded. He then inquired about the number of village men the elder could arm. Two hundred and fifty, the elder replied.

The captain asked the elder why he believed that the men from his village could not be armed to protect themselves and their village. The elder simply replied, “Because the Taliban will kill us.”⁹⁵

The elder’s fear of the Taliban stemmed, in part, from the fact that neither his government nor the ISAF provided sufficient security for those in his village. His villagers were not alone. I estimated that 80 percent of the population in our province had no reasonable assurance of security on a round-the-clock basis. Instead, they lived in the midst of war. One day the Americans could bring out members of the Afghan government or security forces and the next day the Taliban might ride into the village. Security did not exist in any meaningful way, and from the villagers’ perspective uncertainty remained a constant.

The following three figures depict different ways to measure the amount of war trauma. Each figure compares Afghanistan with the average experienced by countries that had a war during any portion of the past 40 years and the average for all states in the international system. The different measures acknowledge that all wars are not created equal, with trauma levels varying substantially across them. Although Afghanistan has been at war for each of the past 40 years, the average length of time that other countries were at war during that period was a much briefer 11.8 years.

The war in Afghanistan directly affects more of the country than a typical war does, and it has done so consistently over time. As Figure 5 indicates, in each of the 30 years for which data are available, the war in Afghanistan affected more of the country than the average for all states at war. Whereas Afghanistan tends to receive the highest rating for virtually all years (i.e., the war affects more than half of the country), for the average war state only a quarter of the country is affected.

Figure 5: Amount of war (area magnitude), 1978-2007

Source: Political Instability Task Force, State Failure Problem Set. Data and codebook are available at Center for Systemic Peace, “Integrated Network for Societal Conflict Research (INSCR) Data Page,” http://www.systemicpeace.org/inscrdata.html.
Note: For more information on how the scores were derived, see the appendix.

Afghans have also suffered more war fatalities per capita than the average war state in each year for which data are available (Figure 6). Throughout the Soviet occupation and civil war that followed, Afghanistan received the maximum score possible. Only when the Taliban took power did battle deaths decline a bit, and they have remained near this level ever since (except for the first few years after the U.S. invasion, when they temporarily declined even further).

Figure 6: Amount of war (battle deaths), 1978-2014

Sources: Uppsala Conflict Data Program Battle-Related Deaths Dataset; Peace Research Institute Oslo Battle Deaths Dataset; Political Instability Task Force, State Failure Problem Set (data and codebooks are available at the Uppsala Conflict Data Program (UCDP), “Downloads,” http://ucdp.uu.se/downloads/; Peace Research Institute Oslo (PRIO), “Battle Deaths Data,” https://www.prio.org/Data/Armed-Conflict/Battle-Deaths/; and Center for Systemic Peace, “Integrated Network for Societal Conflict Research (INSCR) Data Page,” http://www.systemicpeace.org/inscrdata.html).
Note: For more information on how the scores were derived, see the appendix.

Time helps heal all wounds, as mental health professionals well know.⁹⁶ Afghans, though, have endured 40 years of war trauma without any years of peace during which healing could begin. In contrast, the average country that experienced war has benefited from 28 years of peace during that same period (Figure 7). For the populations in those states, some degree of healing could take place during the periods of peace that elude Afghans. And, as expected, the average country in the international system has overwhelmingly enjoyed periods of peace rather than war.

Figure 7: Amount of war (years of peace), 1978-2017

Source: Meredith Reid Sarkees and Frank Wayman, Resort to War: 1816-2007 (Washington: CQ Press, 2010).

Natural Disasters. Adding insult to injury, Afghans have even suffered more natural disasters than the populations of most other countries (Figure 8). Between 1978 and 2014, Afghanistan averaged a disaster score 20 percent higher than the average for all states that experienced a war at some point during that period. Interestingly, both Afghanistan and the broader group of war states had substantially higher proportions of their populations directly affected by disasters compared with the global average. Afghanistan’s average score is 84 percent higher than the global average, and the average score of war states overall is 53 percent higher. Since 1978, for example, disasters have left 42 percent of the Afghan population in need of “immediate assistance” compared with just 10 percent of Americans.⁹⁷

Figure 8: Amount of natural disasters, 1978-2014

Source: Data and explanatory notes are available at Centre for Research on the Epidemiology of Disasters (CRED), “Emergency Events Database,” http://www.emdat.be/.
Note: For more information on how the scores were derived, see the appendix.

Refugees and Internally Displaced Persons. Afghanistan has suffered horrific rates of refugees and internally displaced persons (IDPs). In 1990, for instance, more than half of the population qualified as refugees or IDPs. And in 7 of the past 40 years, more than 40 percent of Afghans found themselves fleeing for their lives. On average, one-fifth of the population have been refugees or IDPs every year since 1978. As with virtually every previous trauma measure, Afghans again rank first. Their average country score is 40 percentage points higher than the average for all countries that had a war and 60 percentage points higher than the global average (Figure 9).⁹⁸ No wonder, then, that Larry Goodson, professor at the U.S. Army War College, referred to the Afghan refugee population as the “largest in the world,” and Louis Dupree, an authority on Afghanistan, called the massive dislocation “migratory genocide.”⁹⁹

Figure 9: Levels of refugees and internally displaced persons, 1978-2014

Sources: United Nations High Commissioner for Refugees; Center for Systemic Peace, Forcibly Displaced Persons dataset.
Note: Refugee and IDP numbers were converted into a per capita proportion for each country-year. The data were then normalized via z-score and finally transformed into a percentage from zero to 100.
Note: For more information on how the scores were derived, see the appendix.

For the past 40 years, Afghans have suffered trauma rates beyond those of other countries at war. Even worse, they endured that trauma without interruption, while the populations of the average war state faced high levels for only a fraction of the time. As a result, Afghanistan’s civil war problem has become even more intractable. The trauma has inadvertently helped normalize the use of violence as a means for goal achievement and problem resolution. It also makes the Afghan security forces less effective because they recruit from a population beset with mental illness. Finally, excessive exposure to torture, rape, and war motivates Afghans to continue rebelling because all of those trauma victims hold some individual or group responsible for their pain.

Why External Intervention Fails

Since initiating combat operations against al Qaeda and the Taliban in October 2001, America has deployed nearly three million military members and more than 2,000 Americans have lost their lives in Afghanistan at an estimated financial cost of $840 billion. Forty-one other countries have contributed to the Afghan war in varying degrees too.¹⁰⁰

These gargantuan efforts have achieved very little. In 2001, just before the United States sent them fleeing, the Taliban controlled or contested 90 percent of the country with an estimated force of 35,000. Today, the Afghan government barely controls or influences half the country, despite having a security force 10 times larger than the Taliban’s, not to mention the benefit of 16 years of American combat power, expertise, and money.¹⁰¹ Nearly 17 years after the United States initiated Operation Enduring Freedom, Afghans are not even safe, much less free. Freedom House currently gives the country its lowest possible rating — “not free” — the same rating it gave Afghanistan in 2001.¹⁰² The assessment also remains abysmal for corruption, with Transparency International ranking Afghanistan below 96 percent of all other countries in the international system. More broadly, one of America’s main goals in invading Afghanistan — to destroy and defeat al Qaeda and other terror groups with global reach — appears further out of reach now than when the war on terror began. Back then, the State Department’s list included only al Qaeda and 12 other similarly motivated groups, comprising an estimated 32,000 fighters. Now the number of groups has mushroomed to 44, and their adherents have swelled to an estimated 110,000.¹⁰³

It is likely that the war continues because, despite the efforts of the United States and the international community, the drivers of endless war in Afghanistan have not been addressed. No doubt the U.S. surge in 2010 temporarily depleted the Taliban’s ranks, but when the time came for Afghan security forces to take responsibility for their nation’s security, they showed themselves incapable or unwilling. Regardless of the reasons, after 16 years of being trained and equipped by the world’s mightiest military, Afghan security forces — more than 350,000 strong — continue to cede ground to a much smaller, shoddily equipped insurgency. Although GDP per capita has increased, Afghans remain in the bottom 10 percent for the world, which keeps rebel recruitment costs low.¹⁰⁴ And despite the rhetoric and assurances, insurgents still find refuge in Pakistan (as well as in Afghanistan’s vast mountain ranges).

The motivation for rebellion also remains high. Afghans have every right to be angry with their government. Outside observers rate Afghans as not free and the government at near rock bottom for corruption, and they judge that the government has failed to provide its citizens with security and basic goods or services. Ethnic fractionalization runs rampant, intensified by the ongoing trauma and erosion of trust that accompanies war. Hundreds of billions of dollars that poured in from America and the international community unwittingly incentivized corruption and drug smuggling within the government and insurgent groups, often blurring the lines between friend and foe. Moreover, the lucrative opium trade motivates insurgent groups to keep the war going so they can conduct their illicit activities with little government interference.

Finally, the decades of bone-jarring trauma — including that from the U.S. invasion and follow-on military operations — have fundamentally changed Afghans. Violence has become normalized. The threshold at which aggrieved citizens will use deadly force against members of their own government has been lowered, making continued war more likely. The mental health care needs of Afghans likely outstrip the capabilities of nongovernmental organizations and exceed the time commitments that any Western liberal democracy could reasonably make. Even if and when widespread trauma finally subsides to a comparatively normal level, Afghans will likely be unable to self-govern peaceably and stably until the negative effects dissipate over time.

Recommendations for U.S. Policymakers

Withdraw American military forces from Afghanistan. Little or no correlation appears to exist between American efforts in Afghanistan and the ability or willingness of Afghans to fundamentally change the situation on the ground. Each year U.S. leaders say that gains are being made and that next year will be different, yet it never is. American blood and treasure should not be spent on a mission that only makes sense if the years of evidence are ignored. Additionally, America’s reputation abroad will continue to suffer as long as the country supports an Afghan government that ranks at the bottom on freedom and at the top on corruption. Moreover, the use of military force in Afghanistan and other Muslim-majority states has hardened anti-American sentiments. Survey data indicate that more citizens in a number of Muslim-majority states agree than disagree with the statement, “The US presence in the region justifies attacks against the US everywhere.”¹⁰⁵ Those countries include the likes of Jordan, Kuwait, and Iraq. And finally, the Taliban threat does not necessitate a continued American military presence in Afghanistan.

Since 2001, the United States has been conducting a social experiment in Afghanistan (and the broader Muslim world) with its employment of military force and simultaneous attempts to establish democratic governments. Early in the war on terror, President George W. Bush spoke of a “global democratic revolution” led by the United States, a revolution that has, to date, failed.¹⁰⁶ America’s political leaders sought to address the underlying causes of terrorism and usher in a sustained period of peace. Quite the opposite has happened. A decade and a half later, Afghans remain “not free” and their government continues its horrible record on civil rights and political liberties.¹⁰⁷ Elsewhere in the Islamic world, more civil wars now rage and terror activity has increased substantially. Instead, future U.S. military interventions should occur only when a vital national interest is at stake.

Incorporate Population’s Mental Health Status Into Military Planning And Intelligence Estimates. The U.S. military has adapted significantly during the war on terror. Intelligence estimates that once focused on the physical terrain and enemy capabilities now analyze the “human terrain” — the psychological, cultural, and behavioral attributes of the populations American forces seek to protect.¹⁰⁸ Military members have learned the languages, customs, and histories of the countries in which they fight. However, when fighting insurgencies, more is needed.

Just as the Department of Defense now recognizes the significant effects that PTSD and other mental health problems can have on its own troops after the war, U.S. military planners and national security policymakers should account for a foreign nation’s mental health status before intervening in its civil war.

For instance, had planners and policymakers analyzed the Afghan population before embarking on a decade and a half of nation building, the analysis would have cast significant doubt on the prospects for peace and the ability of Afghans to implement a functioning democracy — likely the most challenging form of government. Similarly, before getting further involved in any number of ongoing conflicts, such as those in Yemen, Syria, and Somalia, planners and policymakers would be well served by estimating how much trauma the populations have already endured, what that trauma has done to their mental health, and the extent to which violence norms have changed and endless war has become the new normal.

If military planners and national security policymakers continue to ignore the impact of trauma and a population’s mental health status, then they will fail to account for important factors that affect the very war outcome they seek to control.

Incentivize a more effective, less corrupt Afghan government. Finally, if policymakers demand a continued American presence in Afghanistan, then the focus should switch from military force to economic and diplomatic power. To date, American dollars have unwittingly fueled corruption with perverse incentives for Afghan government officials. To keep the money flowing, most Afghan government officials likely believe they need U.S. troops to remain, and that requires continued conflict. However, America could turn its approach to financially supporting Afghanistan upside down. Future funding could be tied to improvements in Afghan governance (e.g., provision of goods and services to its citizens, a reduction in corruption, and more political freedoms). Conceivably, then, the amount of aid flowing in would increase as the situation improves, until leveling off and eventually declining as Afghan self-sufficiency is restored. American diplomatic efforts should shore up enduring agreements from the international community, preferably through the United Nations, to both decrease costs to the American taxpayer and bolster the perception that Afghanistan is an international (rather than U.S.) mission. Even this approach, though, is fraught with risk, as substantial scholarship has found little or no relationship between aid and economic growth or basic human development indicators.¹⁰⁹

Conclusion

The United States invaded Afghanistan in 2001 to destroy al Qaeda, remove the Taliban from power, and ensure that the country would not become a sanctuary for transnational terrorists again. Sixteen years later, those objectives are largely unmet. Al Qaeda has not been defeated, and the number of other Islamist-inspired terrorist groups has proliferated. The Taliban no longer constitute the national government, but they do control, influence, or contest almost half of Afghan districts, while the nominally democratic government ranks at or near the bottom of all states in capacity, transparency, and freedom. Additionally, terror groups like ISIS appear to be increasingly active within the country. U.S. efforts have largely failed and will continue to fail because of the dysfunctional features of a society that only Afghans can fix.

Instead of focusing on changes around the margins, the United States should take a step back and ask why Afghanistan has been at war for 40 years and why no one has been able to end it. The opportunity for war continues to exist, grievances remain at elevated levels, and two generations have come of age within the Afghan trauma state. As a result, America should decrease its military footprint and focus on efforts to incentivize a more capable, less corrupt Afghan government.

Beyond Afghanistan, America should restrain its use of military force to those instances vital to U.S. interests, since sustained war in already traumatized states like Afghanistan increases psychological trauma and societal instability, making victory unlikely. Finally, in those rare instances when the United States finds it necessary to apply military force in other countries’ civil wars, the population’s mental health status should be included in military planning and intelligence estimates because it will certainly affect the conduct of the war.

Appendix

The Trauma Index

To create the composite trauma index, it is first necessary to ensure all components of the index are comparable, that is, measured on the same scale. Measures for torture, war, rape, and other trauma are standardized using one of two methods — min-max or z-score. Each score is then converted to a scale of zero to 100, in which zero represents the least trauma and 100 represents the most trauma for each index.

For example, the trauma measure includes two variables: the CIRI Human Rights Data Project (torture) and the Political Terror Scale. The range of possible scores is 0-2 and 1-5, respectively. To combine these two variables into the single trauma index, each variable is converted to the same scale of zero to 100. First, since a CIRI score of zero represents the most amount of torture, the index is flipped such that a score of 2 now indicates the highest degree of torture and zero the lowest. Next, the score is converted to a 0-to-100 scale by multiplying the score by 50, in which a score of 2 on the CIRI index becomes 100, 1 becomes 50, and so on. Finally, the Political Terror Scale is mapped from the 1-to-5 scale by multiplying by 20, in which a score of 1 becomes 20, 4 becomes 75, and so on. Now that the two indexes lie on the same 0-to-100 scale, the composite trauma index is created by taking the simple average of the two scores by adding each score and dividing by two (i.e., the number of variables). (The complete codebook is available upon request.)

To estimate the prevalence of rape, I developed a measure based on Dara Kay Cohen and Ragnhild Nordås’s previous work on rape and sexual violence during military conflicts.¹¹⁰ I compiled a dataset using annual Country Reports on Human Rights Practices produced by the U.S. Department of State (1990-2015). The methodology and scoring criteria followed Cohen and Nordås, who used an ordinal scale from zero to 3.¹¹¹ Assessments of rape as “massive,” “systematic,” or a “tactic to [punish, terrorize, etc.]” are scored a 3. “Common,” “widespread,” and “serious problem” receive a 2. Characterizations of rape such as “some reports” or “a problem” were scored 1. Countries not meeting any of the above criteria received a zero. The qualitative assessments contained in the annual State Department reports were derived from a number of sources, including government data, assessments from nongovernmental organizations, and news accounts. The min-max method was used to normalize the data.

The term “rape” occurs more frequently over time in the State Department reports. Reviewing the reports in reverse chronological order, it appeared that the frequency of “rape” began to decrease substantially from 1998 to 1997. I then conducted a statistical analysis to investigate whether a significant change in frequency occurred over time. The chi-square results indicated a significant and positive relationship between year and use of “rape”: a more recent year corresponded with an increase in use of the term. As a result, I added three more possible scores for all years before 1997. If the report indicated that sexual violence was “widespread” or a “serious problem,” I recorded a 0.75. If the report characterized sexual violence as “a problem,” then the country received a score of 0.5. Finally, if the report did not include any of the above conditions but did describe domestic violence as “widespread” or a “serious problem,” the country would be scored 0.25 because domestic violence, sexual violence, and rape share a strong association.¹¹²

The actual number of reported rapes, when available, was not considered for several reasons. First, a sample of countries that included both quantitative and qualitative measures of rape was analyzed and the results indicated no statistically significant relationship, at p-value < 0.10, between the characterization of rape and the number of reported rapes in a country. Second, rape is considered an underreported crime but different country reports provide substantially different estimates regarding the degree of underreporting, and some country reports make no mention of underreporting at all. Third, analyzing reported numbers would bias the scoring against those countries that have police forces that citizens trust to report crime to, as well as those states with government bureaucracies that maintain and publicize crime statistics.

For the war variable measuring area magnitude, the data come from the Political Instability Task Force State Failure Problem Set. Scores range from zero (lowest area magnitude score for a country at war) to 4 (greatest area magnitude). A score of zero indicates that “less than one-tenth of the country and no significant cities” are affected, and a 4 means the war affects “more than one-half of the country.” I normalized scores such that a 4 corresponds to 100 points, a 3 to 80 points, and so on, with zero corresponding to 20 points. In cases where no war occurred (and the task force, therefore, provided no score), I assigned the respective country-year zero points.¹¹³

Notes

The author thanks Nikolaus Pittore for his valuable research.

¹. Meredith Reid Sarkees and Frank Wayman, Resort to War: 1816-2007 (Washington: CQ Press, 2010).

². Noor Ahmad Khalidi, “Afghanistan: Demographic Consequences of War, 1978-1987,” Central Asian Survey 10, no. 3 (1991): 101, 106, 108.

³. Larry Goodson, Afghanistan’s Endless War: State Failure, Regional Politics, and the Rise of the Taliban (Seattle: University of Washington Press, 2001), p. 5.

⁴. Bill Roggio and Alexandra Gutowski, “LWJ Map Assessment: Taliban Controls or Contests 45% of Afghan Districts,” FDD’s Long War Journal, September 26, 2017, www.longwarjournal.org/archives/2017/09/lwj-map-assessment-taliban-controls-or-contests-45-of-afghan-districts.php.

⁵. iCasualties.org, “U.S. Fatalities in and around Afghanistan,” Operation Enduring Freedom, http://icasualties.org/oef/; Jeanne Sahadi, “The Financial Cost of 16 Years in Afghanistan,” CNN Money, August 22, 2017, http://money.cnn.com/2017/08/21/news/economy/war-costs-afghanistan/index.html; ISAF Troop Contributing Nations,” October 1, 2009, www.nato.int/ISAF/structure/nations/index.html;Special Inspector General for Afghanistan Reconstruction, “Quarterly Report to the United States Congress,” July 30, 2014, p. 5, www.sigar.mil/pdf/quarterlyreports/2014-07-30qr.pdf”>pdf/quarterlyreports/2014-07-30qr.pdf; and Watson Institute, “US Veterans and Families,” Costs of War website, http://watson.brown.edu/costsofwar/costs/human/veterans. Other cost estimates are even higher. See, for example, Neta Crawford, “US Budgetary Costs of Wars through 2016: $4.79 Trillion and Counting; Summary of Costs of the US Wars in Iraq, Syria, Afghanistan and Pakistan and Homeland Security,” Watson Institute, September 2016, http://watson.brown.edu/costsofwar/files/cow/imce/papers/2016/Costs%20of%20War%20through%202016%20FINAL%20final%20v2.pdf; and Linda Bilmes, “The Financial Legacy of Iraq and Afghanistan: How Wartime Spending Decisions Will Constrain Future National Security Budgets,” Faculty Research Working Paper no. RWP13-006, Harvard Kennedy School, March 2013, https://research.hks.harvard.edu/publications/workingpapers/citation.aspx?PubId=8956.

⁶. All names of Afghans and Afghan locations have been changed.

⁷. The civil war literature refers to the desire for financial profit as the “greed” or “rebellion-as-business” approach to civil war.

⁸. Transparency International, Corruption Perception Index 2016, www.transparency.org/news/feature/ corruption_perceptions_index_2016.

⁹. Martin Ewans, Afghanistan: A Short History of Its People and Politics (New York: Harper Perennial, 2002), pp. 122, 136, 153, 158-59.

¹⁰. Sarkees and Wayman, Resort to War.

¹¹. Amin Saikal, A. G. Ravan Farhadi, and Kirill Nourzhanov, Modern Afghanistan: A History of Struggle and Survival (London: I. B. Tauris, 2012), p. 193; and Angelo Rasanayagam, Afghanistan: A Modern History; Monarchy, Despotism or Democracy? The Problems of Governance in the Muslim Tradition (London: I. B. Tauris, 2005), p. 114.

¹². Special Inspector General for Afghanistan Reconstruction, “Quarterly Report,” pp. 88, 101, 104.

¹³. Averages derived from data provided in Ibrahim Elbadawi and Nicholas Sambanis, “How Much War Will We See? Explaining the Prevalence of Civil War,” Journal of Conflict Resolution 46, no. 3 (2002): 307-34.

¹⁴. World Bank, “GDP per Capita (Current US$),” https://data.worldbank.org/indicator/NY.GDP.PCAP.CD?locations=AF.

¹⁵. Megan McCloskey et al., “Behold: How the US Blew $17 Billion in Afghanistan,” ProPublica, December 18, 2015, www.pri.org/stories/2015-12-18/behold-american-taxpayer-what-happened-nearly-half-billion-your-dollars; and Central Intelligence Agency, “Country Comparison: GDP per Capita (PPP),” The World Factbook, www.cia.gov/library/publications/the-world-factbook/rankorder/2004rank.html.

¹⁶. Maddison Historical Statistics, Groningen Growth and Development Centre, 2013, www.ggdc.net/maddison/maddison-project/home.html.

¹⁷. Percentages are based on those age 15 and older. UNESCO Institute for Statistics, “Data for the Sustainable Development Goals,” http://uis.unesco.org; World Bank, “Literacy Rate, Adult Total (% of People Ages 15 and Above),” https://data.worldbank.org/indicator/SE.ADT.LITR.ZS.

¹⁸. “The Toughest Places for a Girl to Get an Education,” https://s3.amazonaws.com/one.org/pdfs/IDG_2017_media_briefing_en_web.pdf.

¹⁹. William Maley, The Afghanistan Wars (New York: Palgrave, 2002), p. 154.

²⁰. Barnett R. Rubin et al., “Building a New Afghanistan: The Value of Success, the Cost of Failure,” Center on International Cooperation, New York University, March 2004, pp. 23-24.

²¹. Malou Innocent and Ted Galen Carpenter, “Escaping the ‘Graveyard of Empires’: A Strategy to Exit Afghanistan,” Cato Institute White Paper, 2009, pp. 6-7; Barrack Obama, “Remarks by the President on a New Strategy for Afghanistan and Pakistan,” March 27, 2009, https://obamawhitehouse.archives.gov/the-press-office/remarks-president-a-new-strategy-afghanistan-and-pakistan; and Cato Institute, Cato Handbook for Policymakers (Washington: Cato Institute, 2008), p. 534.

²². Terrorism Research and Analysis Consortium, “Quetta Shura Taliban,” www.trackingterrorism.org/group/quetta-shura-taliban-qst.

²³. Obama, “New Strategy for Afghanistan and Pakistan.”

²⁴. Donald Trump, “Remarks by President Trump on the Strategy in Afghanistan and South Asia,” August 21, 2017, www.whitehouse.gov/the-press-office/2017/08/21/remarks-president-trump-strategy-afghanistan-and-south-asia.

²⁵. Obama, “New Strategy for Afghanistan and Pakistan.”

²⁶. John Swift, “Battle of Tora Bora,” Encyclopaedia BritannicaOnline, 2007, www.britannica.com/event/Battle-of-Tora-Bora.

²⁷. Rod Nordland and Fahim Abed, “ISIS Captures Tora Bora, Once Bin Laden’s Afghan Fortress,” New York Times, June 14, 2017, www.nytimes.com/2017/06/14/world/asia/isis-captures-tora-bora-afghanistan.html?_r=0.

²⁸. Mohammad Osman Tariq, Najla Ayoubi, and Fazel Rabi Haqbeen, Afghanistan in 2011: A Survey of the Afghan People (Kabul: Asia Foundation, 2011), pp. 25-26.

²⁹. Barnett R. Rubin, The Fragmentation of Afghanistan: State Formation and Collapse in the International System, 2nd ed. (New Haven, CT: Yale University Press, 2002), p. 273; and Goodson, Afghanistan’s Endless War, pp. 57, 69.

³⁰. “Afghanistan Opium Production Jumps 87 Per Cent to Record Level — UN Survey,” UN News, November 15, 2017, https://news.un.org/en/story/2017/11/636182-afghanistan-opium-production-jumps-87-cent-record-level-un-survey.

³¹. Vanda Felbab-Brown, “High and Low Politics in Afghanistan: The Terrorism-Drugs Nexus and What Can Be Done about It,” Ahora (translated and reprinted at Brookings.Edu), April 29, 2016, www.brookings.edu/articles/high-and-low-politics-in-afghanistan-“>afghanistan-the-terrorism-drugs-nexus-and-what-can-be-done-about-it/”>the-terrorism-drugs-nexus-and-what-can-be-done-about-it/.

³². Special Inspector General for Afghanistan Reconstruction, “Corruption in Conflict: Lessons from the U.S. Experience in Afghanistan,” September 2016, p. 19, www.sigar.mil/pdf/LessonsLearned/SIGAR-16-58-LL.pdf.

³³. Special Inspector General for Afghanistan Reconstruction, “Corruption in Conflict,” p. 20.

³⁴. American Psychiatric Association, Diagnostic and Statistical Manual of Mental Disorders, 5th ed. (DSM-5) (Washington: American Psychiatric Association, 2013), p. 830; Rangaswamy Srinivasa Murthy et al., The World Health Report 2001 — Mental Health: New Understanding, New Hope, World Health Organization, 2001, www.who.int/whr/2001/en/; Rangaswamy Srinivasa Murthy, “Mass Violence and Mental Health: Recent Epidemiological Findings,” International Review of Psychiatry 19, no. 3 (2007): 183-92; Zachary Steel et al., “Association of Torture and Other Potentially Traumatic Events with Mental Health Outcomes among Populations Exposed to Mass Conflict and Displacement: A Systematic Review and Meta-Analysis,” JAMA 302, no. 5 (2009): 537-49; and Niloofar Afari et al., “Psychological Trauma and Functional Somatic Syndromes: A Systematic Review and Meta-Analysis,” Psychosomatic Medicine 76, no. 1 (2014): 2-11.

³⁵. American Psychiatric Association, DSM-5; and Steel et al., “Association of Torture with Mental Health Outcomes.”

³⁶. Steel et al., “Association of Torture with Mental Health Outcomes”; Aziz Yasan et al., “Trauma Type, Gender, and Risk of PTSD in a Region within an Area of Conflict,” Journal of Traumatic Stress 22, no. 6 (2009): 663-66; and American Psychiatric Association, Diagnostic and Statistical Manual of Mental Disorders, 4th ed., text revision (DSM-IV-TR) (Washington: American Psychiatric Association, 2000).

³⁷. Deborah A. Weber and Cecil R. Reynolds, “Clinical Perspectives on Neurobiological Effects of Psychological Trauma,” Neuropsychology Review 14, no. 2 (2004): 115-29; Pia Pechtel and Diego A. Pizzagalli, “Effects of Early Life Stress on Cognitive and Affective Function: An Integrated Review of Human Literature,” Psychopharmacology 214, no. 1 (2011): 55-70; and Anke Karl et al., “A Meta-Analysis of Structural Brain Abnormalities in PTSD,” Neuroscience & Biobehavioral Reviews 30, no. 7 (2006): 1004-31.

³⁸. Karl et al., “A Meta-Analysis of Structural Brain Abnormalities in PTSD”; Fu Lye Woon, Shabnam Sood, and Dawson W. Hedges, “Hippocampal Volume Deficits Associated with Exposure to Psychological Trauma and Post-Traumatic Stress Disorder in Adults: A Meta-Analysis,” Progress in Neuro-Psychopharmacology and Biological Psychiatry 34, no. 7 (2010): 1181-88; Pechtel and Pizzagalli, “Effects of Early Life Stress on Cognitive and Affective Function,” p. 7.

³⁹. Chris R. Brewin et al., “Memory for Emotionally Neutral Information in Post-Traumatic Stress Disorder: A Meta-Analytic Investigation,” Journal of Abnormal Psychology 116, no. 3 (2007): 448-63; and Pechtel and Pizzagalli, “Effects of Early Life Stress on Cognitive and Affective Function.”

⁴⁰. Annette A. M. Gerritsen et al., “Physical and Mental Health of Afghan, Iranian and Somali Asylum Seekers and Refugees Living in the Netherlands,” Social Psychiatry and Psychiatric Epidemiology 41, no. 1 (2006): 168-69; and Israel Bronstein, Paul Montgomery, and Stephanie Dobrowolski, “PTSD in Asylum-Seeking Male Adolescents from Afghanistan: PTSD amongst Afghan UASC,” Journal of Traumatic Stress 25, no. 5 (2012): 551-57.

⁴¹. Dean G. Kilpatrick et al., “National Estimates of Exposure to Traumatic Events and PTSD Prevalence Using DSM-IV and DSM-5 Criteria: DSM-5 PTSD Prevalence,” Journal of Traumatic Stress 26, no. 5 (2013): 541; U. Hepp et al., “Prevalence of Exposure to Potentially Traumatic Events and PTSD: The Zurich Cohort Study,” European Archives of Psychiatry and Clinical Neuroscience 256, no. 3 (2006): 153-54; A. Perkonigg et al., “Traumatic Events and Post-Traumatic Stress Disorder in the Community: Prevalence, Risk Factors and Comorbidity,” Acta Psychiatrica Scandinavica 101, no. 1 (2000): 46-59, https://doi.org/10.1034/j.1600-0447.2000.101001046.x; Naomi Breslau, “Traumatic Events and Post-Traumatic Stress Disorder in an Urban Population of Young Adults,” Archives of General Psychiatry 48, no. 3 (1991): 217; Kate M. Scott et al., “Associations between Lifetime Traumatic Events and Subsequent Chronic Physical Conditions: A Cross-National, Cross-Sectional Study,” PLOS ONE 8, no. 11 (2013): 3; and Giel-Jan de Vries and Miranda Olff, “The Lifetime Prevalence of Traumatic Events and Post-Traumatic Stress Disorder in the Netherlands,” Journal of Traumatic Stress 22, no. 4 (2009): 261.

⁴². Hepp et al., “Zurich Cohort Study,” p. 154; Barbara Lopes Cardozo, “Mental Health, Social Functioning, and Disability in Postwar Afghanistan,” JAMA 292, no. 5 (2004): 579; and Willem F. Scholte, Miranda Olff, and Peter Ventevogel, “Mental Health Symptoms following War and Repression in Eastern Afghanistan,” JAMA 292, no. 5 (2004): 589.

⁴³. Qais Alemi et al., “Psychological Distress in Afghan Refugees: A Mixed-Method Systematic Review,” Journal of Immigrant and Minority Health 16, no. 6 (2014): 1256; Kilpatrick et al., “National Estimates of Exposure to Traumatic Events,” p. 545; and Derrick Silove, “The Psychosocial Effects of Torture, Mass Human Rights Violations, and Refugee Trauma: Toward an Integrated Conceptual Framework,” Journal of Nervous and Mental Disease 187, no. 4 (1999): 200-207.

⁴⁴. He was an Afghan man in his early 20s who served as an interpreter for American forces. Like most interpreters, he did not use his real name in an effort to protect himself and his family.

⁴⁵. Catherine Panter-Brick et al., “Violence, Suffering, and Mental Health in Afghanistan: A School-Based Survey,” Lancet 374, no. 9692 (2009): 814; Hillary Wildt et al., “War Trauma, Psychological Distress, and Coping among Afghan Civilians Seeking Primary Health Care,” International Perspectives in Psychology: Research, Practice, Consultation 6, no. 2 (2017): 82; Gerritsen et al., “Physical and Mental Health of Afghan, Iranian and Somali Asylum Seekers and Refugees,” p. 8; Scholte, Olff, and Ventevogel, “Mental Health Symptoms following War and Repression,” p. 590; Cardozo, “Mental Health, Social Functioning, and Disability in Postwar Afghanistan,” p. 575; Bronstein, Montgomery, and Dobrowolski, “PTSD in Asylum-Seeking Male Adolescents from Afghanistan,” p. 551; and American Psychiatric Association, DSM-5.

⁴⁶. Panter-Brick et al., “Violence, Suffering, and Mental Health in Afghanistan,” p. 813.

⁴⁷. Steel et al., “Association of Torture with Mental Health Outcomes,” pp. 540-41, 543, 545. Steel et al. used 40 percent as the cut-off point. If 40 percent or more of a population reported having been tortured, they were binned in the highest category, 20 percent to 39 percent in the middle category, and less than 19 percent in the lowest. For more information on the Political Terror Scale, which is run by scholars from the University of North Carolina Asheville, see “Documentation: Coding Rules,” www.politicalterrorscale.org/Data/Documentation.html.

⁴⁸. Data and documentation are available from the CIRI Human Rights Data Project, www.humanrightsdata.com/p/data-documentation.html.

⁴⁹. Cardozo, “Mental Health, Social Functioning, and Disability in Postwar Afghanistan,” p. 576; and Panter-Brick et al., “Violence, Suffering, and Mental Health in Afghanistan,” p. 813.

⁵⁰. Steel et al., “Association of Torture with Mental Health Outcomes”; Darrel A. Regier, Mary Farmer, and Donald Rae, “Comorbidity of Mental Disorders with Alcohol and Other Drug Abuse: Results from the Epidemiologic Catchment Area (ECA) Study,” JAMA 264, no. 19 (1990): 2514; and Timothy Weaver et al., “Comorbidity of Substance Misuse and Mental Illness in Community Mental Health and Substance Misuse Services,” British Journal of Psychiatry 183, no. 4 (2003): 304-13.

⁵¹. Regier, Farmer, and Rae, “Comorbidity of Mental Disorders,” p. 2514.

⁵². Regier, Farmer, and Rae, “Comorbidity of Mental Disorders, p. 2517; and Robert E. Drake et al., “A Review of Treatments for People with Severe Mental Illnesses and Co-Occurring Substance Use Disorders,” Psychiatric Rehabilitation Journal 27, no. 4 (2004): 361.

⁵³. John Fairbank, Frank Putnam, and William Harris, “The Prevalence and Impact of Child Traumatic Stress,” in Handbook of PTSD: Science and Practice (New York: Guilford Press, 2007), p. 240; Emil F. Coccaro, “Intermittent Explosive Disorder as a Disorder of Impulsive Aggression for DSM-5,” American Journal of Psychiatry 169, no. 6 (2012): 584; and Bessel A. van der Kolk, “The Neurobiology of Childhood Trauma and Abuse,” Child and Adolescent Psychiatric Clinics of North America 12, no. 2 (2003): 293, 298-99.

⁵⁴. Van der Kolk, “Neurobiology of Childhood Trauma and Abuse.”

⁵⁵. Michel Lejoyeux et al., “Impulse Control Disorders and Depression,” Journal of Nervous and Mental Disease 190, no. 5 (2002): 310, 314; M. Kotler et al., “Anger, Impulsivity, Social Support, and Suicide Risk in Patients with Post-Traumatic Stress Disorder,” Journal of Nervous and Mental Disease 189, no. 3 (2001): 162-67; and Byron Good and Devon E. Hinton, “Introduction. Culture, Trauma, and PTSD,” in Culture and PTSD: Trauma in Global and Historical Perspective, eds. Devon E. Hinton and Byron Good (Philadelphia: University of Pennsylvania Press, 2016), pp. 3-49.

⁵⁶. Dianne M. Tice, Ellen Bratslavsky, and Roy F. Baumeister, “Emotional Distress Regulation Takes Precedence over Impulse Control: If You Feel Bad, Do It!” Journal of Personality and Social Psychology 80, no. 1 (2001): 53, https://doi.org/10.1037//0022-3514.80.1.53; Menahem Krakowski, “Violence and Serotonin: Influence of Impulse Control, Affect Regulation, and Social Functioning,” Journal of Neuropsychiatry and Clinical Neurosciences 15, no. 3 (2003): 294-95, 300; Richard Van Dorn, Jan Volavka, and Norman Johnson, “Mental Disorder and Violence: Is There a Relationship beyond Substance Use?,” Social Psychiatry and Psychiatric Epidemiology 47, no. 3 (2011): 487; and Michael A. Norko and Madelon V. Baranoski, “The Prediction of Violence; Detection of Dangerousness,” Brief Treatment and Crisis Intervention 8, no. 1 (2008): 76.

⁵⁷. Alemi et al., “Psychological Distress in Afghan Refugees,” p. 1256.

⁵⁸. United Nations Office on Drugs and Crime, “Drug Use in Afghanistan: 2009 Survey. Executive Summary,” pp. 3, 6, www.unodc.org/documents/data-and-analysis/Studies/Afghan-Drug-Survey-2009-Executive-Summary-web.pdf.

⁵⁹. Claudia Catani et al., “War Trauma, Child Labor, and Family Violence: Life Adversities and PTSD in a Sample of School Children in Kabul,” Journal of Traumatic Stress 22, no. 3 (2009): 163-71; Wildt et al., “War Trauma, Psychological Distress, and Coping among Afghan Civilians.”

⁶⁰. Jennifer Hatfield, Wilfreda E. Thurston, and Sadiqa Basiri, “Women’s Participation in Domestic Violence Health Policy Development: Afghanistan Component,” Women’s Domestic Violence Health Project, 2008, p. 3, www.ucalgary.ca/wethurston/files/wethurston/Report_WDVHPAfghanistan.pdf.

⁶¹. Diya Nijhowne and Lauryn Oates, “Living with Violence: A National Report on Domestic Abuse in Afghanistan,” Global Rights, 2008, p. 1, www.humanitarianresponse.info/system/files/documents/files/Living%20with%20Violence%20-%20A%20National%20Report%20on%20Domestic%20Violence%20in%20Afghanistan%2C%20Global%20Rights%2C%202008.pdf.

⁶². Kenneth E. Miller et al., “Daily Stressors, War Experiences, and Mental Health in Afghanistan,” Transcultural Psychiatry 45, no. 4 (2008): 615, https://doi.org/10.1177/1363461508100785.

⁶³. Nijhowne and Oates, “Living with Violence,” p. 1.

⁶⁴. Nijhowne and Oates, “Living with Violence,” p. 2.

⁶⁵. World Health Organization, “Violence against Women,” Fact Sheet no. 239, 2000, http://collections.infocollections.org/ukedu/en/d/Js0517e/.

⁶⁶. Catani et al., “War Trauma, Child Labor, and Family Violence,” pp. 166-67.

⁶⁷. Bill Roggio and Lisa Lundquist, “Green-on-Blue Attacks in Afghanistan: The Data,” FDD’s Long War Journal, August 23, 2012, www.longwarjournal.org/archives/2012/08/green-on-blue_attack.php.

⁶⁸. David A. Arenas, “Afghanistan: ‘Green-on-Blue Attacks,’” United States Marine Corps Command and Staff College, 2013, www.dtic.mil/dtic/tr/fulltext/u2/a601559.pdf.

⁶⁹. Thom Shanker, “General Notes Taliban Coercion in Some Attacks on Troops,” New York Times, August 23, 2012, www.nytimes.com/2012/08/24/world/asia/general-notes-taliban-coercion-in-some-attacks-on-troops-in-afghanistan.html.

⁷⁰. Tom Bowman, “U.S. Faces Growing ‘Insider Attacks’ in Afghanistan,” NPR, August 23, 2012, www.npr.org/2012/08/23/159926492/us-confronts-growing-insider-attacks-in-afghanistan.

⁷¹. John Reed, “Guardian Angels in Afghanistan,” Foreign Policy, August 14, 2012, http://foreignpolicy.com/2012/08/14/guardian-angels-in-afghanistan/.

⁷². Grace Geiger, “This Time It’s Different: Collective Training Advisors in Afghanistan,” Army.Mil, January 10, 2017,www.army.mil/article/180327/this_time_its_different_collective_training_advisors_in_afghanistan.

⁷³. Stathis N. Kalyvas, The Logic of Violence in Civil War (Cambridge: Cambridge University Press, 2006), p. 14; and Dave Grossman, On Killing: The Psychological Cost of Learning to Kill in War and Society (New York: Open Road Media, 2014), Kindle edition.

⁷⁴. Steel et al., “Association of Torture with Mental Health Outcomes”; Regier, Farmer, and Rae, “Comorbidity of Mental Disorders”; and Weaver et al., “Comorbidity of Substance Misuse and Mental Illness.”

⁷⁵. All of my experiences suggested that the Afghan government never held anyone accountable for criminal wrongdoing, only for ulterior motives (e.g., to shift the power balance). Instead, officials would make symbolic gestures, like putting someone in jail for two days, in an apparent effort to mollify their ISAF counterparts.

⁷⁶. Norko and Baranoski, “Prediction of Violence,” p. 76; and Richard Van Dorn et al., “Mental Disorder and Violence,” p. 487. Debate exists over the role of socioeconomic status within mental illness and violence. For example, is socioeconomic status independent of mental illness or endogenous to it, and do issues of reverse causality exist between the two? See, for example, John Monahan and Henry Steadman, “Crime and Mental Disorder: An Epidemiological Approach,” in Crime and Justice: An Annual Review of Research, eds. Michael Tonry and Norval Morris, vol. 4 (Chicago: University of Chicago Press, 1983), pp. 145-89.

⁷⁷. American Psychiatric Association, DSM-5.

⁷⁸. American Psychiatric Association, DSM-5.

⁷⁹. Stathis N. Kalyvas, Logic of Violence in Civil War, p. 14.

⁸⁰. Grossman, On Killing.

⁸¹. Data are available from the World Bank, http://data.worldbank.org/indicator/VC.IHR.PSRC.P5.

⁸². Chrissie Steenkamp, “The Legacy of War: Conceptualizing a ‘Culture of Violence’ to Explain Violence after Peace Accords,” Round Table 94, no. 379 (2005): 254-55, 264.

⁸³. Roos Haer and Tobias Böhmelt, “Child Soldiers as Time Bombs? Adolescents’ Participation in Rebel Groups and the Recurrence of Armed Conflict,” European Journal of International Relations 22, no. 2 (2016): 414.

⁸⁴. Torture is defined as “the deliberate … infliction of physical or mental suffering … to force another person to yield information … or for any other reason.” “Declaration of Tokyo: Guidelines for Physicians concerning Torture and Other Cruel, Inhuman or Degrading Treatment or Punishment in Relation to Detention and Imprisonment,” World Medical Association, 1975.

⁸⁵. The CIRI Human Rights Data Project uses different wording for torture without materially changing the definition: “Torture refers to the purposeful inflicting of extreme pain, whether mental or physical.” However, for the CIRI measure, torture is assessed only if done by “government officials or by private individuals at the instigation of government officials.” See “CIRI Short Variable Descriptions,” p. 3, www.humanrightsdata.com/p/data-documentation.html.

⁸⁶. David Cingranelli and David Richards, “The Cingranelli-Richards (CIRI) Human Rights Data Project Coding Manual Version 5.20.14,” 2014, www.humanrightsdata.com/p/data-documentation.html; and David Cingranelli, David Richards, and K. Chad Clay, “The CIRI Human Rights Dataset,” 2014, www.humanrightsdata.com.

⁸⁷. Mark Gibney et al., “Political Terror Scale 1976-2012,” 2014, p. 1, www.politicalterrorscale.org/.

⁸⁸. Peter Haschke, “The Political Terror Scale (PTS) Codebook,” 2017, p. 4, www.politicalterrorscale.org.

⁸⁹. Haschke, “PTS Codebook,” p. 4; and Mark Gibney et al., “Political Terror Scale 1976-2016,” 2017, www.politicalterrorscale.org.

⁹⁰. Ronald C. Kessler, “Post-Traumatic Stress Disorder in the National Comorbidity Survey,” Archives of General Psychiatry 52, no. 12 (1995): 1052-53, https://doi.org/10.1001/archpsyc.1995.03950240066012.

⁹¹. Dara Kay Cohen, “Explaining Rape during Civil War: Cross-National Evidence (1980-2009),” American Political Science Review 107, no. 3 (2013): 461-77; and Dara Kay Cohen and Ragnhild Nordås, “Sexual Violence in Armed Conflict: Introducing the SVAC Dataset, 1989-2009,” Journal of Peace Research 51, no. 3 (2014): 418-28, https://doi.org/10.1177/0022343314523028.

⁹². Amnesty International, Afghanistan: International Responsibility for Human Rights Disaster (London: Amnesty International, 1995), p. 47.

⁹³. Scholte, Olff, and Ventevogel, “Mental Health Symptoms following War and Repression”; and Cardozo, “Mental Health, Social Functioning, and Disability in Postwar Afghanistan.”

⁹⁴. Catherine Panter-Brick et al., “Trauma Memories, Mental Health, and Resilience: A Prospective Study of Afghan Youth,” Journal of Child Psychology and Psychiatry 56, no. 7 (2015): 815; and Catani et al., “War Trauma, Child Labor, and Family Violence,” p. 165.

⁹⁵. Reprinted from Erik W. Goepner, “Battered Spouse Syndrome: How to Better Understand Afghan Behavior,” Military Review 92, no. 1 (2012): 59-66.

⁹⁶. Steel et al., “Association of Torture with Mental Health Outcomes”; and Good and Hinton, “Introduction. Culture, Trauma, and PTSD,” p. 9.

⁹⁷. Université catholique de Louvain and Debarati Guha-Sapir, “EM-DAT: The Emergency Events Database,” www.emdat.be.

⁹⁸. Refugee and IDP numbers were converted into a per capita proportion for each country-year. The data were then normalized via z-score and finally transformed into a percentage from zero to 100.

⁹⁹. Goodson, Afghanistan’s Endless War, p. 61; Alemi et al., “Psychological Distress in Afghan Refugees,” p. 1247; and Louis Dupree, “Afghanistan in 1983: And Still No Solution,” Asian Survey 24, no. 2 (1984): 234, https://doi.org/10.2307/2644442.

¹⁰⁰. Data are available from the Watson Institute, http://watson.brown.edu/costsofwar/costs/human/veterans; iCasualties.org, http://icasualties.org/oef/; Sahadi, “Financial Costs,” http://money.cnn.com/2017/08/21/news/economy/war-costs-afghanistan/index.html; and ISAF Troop Contributing Nations, www.nato.int/isaf/structure/nations/index.html.

¹⁰¹. Bill Roggio and Alexandra Gutowski, “LWJ Map Assessment”; and Erik Goepner, “Afghanistan’s Biggest Obstacle Is Its Government,” Orange County Register, November 1, 2017, www.ocregister.com/2017/11/01/afghanistans-biggest-obstacle-is-its-government/.

¹⁰². Freedom House, “Freedom in the World 2001” and “Freedom in the World 2017,” https://freedomhouse.org/report-types/freedom-world.

¹⁰³. A. Trevor Thrall and Erik Goepner, “Step Back: Lessons for U.S. Foreign Policy from the Failed War on Terror,” Cato Institute Policy Analysis no. 814, June 26, 2017, p. 9, www.cato.org/publications/policy-analysis/step-back-lessons-us-foreign-policy-failed-war-terror.

¹⁰⁴. Central Intelligence Agency, “Country Comparison: GDP per Capita (PPP).”

¹⁰⁵. Erik Goepner and A. Trevor Thrall, “Time to Step Back from the War on Terror,” War on the Rocks, October 26, 2017, https://warontherocks.com/2017/10/time-to-step-back-from-the-war-on-terror/.

¹⁰⁶. Fred Barbash, “Bush: Iraq Part of ‘Global Democratic Revolution,’” Washington Post, November 6, 2003.

¹⁰⁷. Freedom House, “Freedom in the World 2001” and “Freedom in the World 2017.”

¹⁰⁸. Andrew Garfield, “Understanding the Human Terrain: Key to Success in Afghanistan,” Small Wars Journal, July 16, 2010, http://smallwarsjournal.com/blog/understanding-the-human-terrain-key-to-success-in-afghanistan; and Joint Chiefs of Staff, Counterinsurgency, Joint Publication no. 3-24 (Washington: Department of Defense, 2013), pp. IV-4.

¹⁰⁹. Ian Vásquez, “The New Approach to Foreign Aid: Is the Enthusiasm Warranted?” Cato Institute Foreign Policy Briefing no. 79, September 17, 2003, p. 2, https://object.cato.org/sites/cato.org/files/pubs/pdf/fpb79.pdf; William Easterly et al., “Aid, Policies, and Growth: Comment,” American Economic Review 94, no. 3 (2004): 774-80; and Claudia Williamson, “Exploring the Failure of Foreign Aid: The Role of Incentives and Information,” Review of Austrian Economics 23, no. 1 (2010): 17-33.

¹¹⁰. Cohen, “Explaining Rape during Civil War”; Cohen and Nordås, “Sexual Violence in Armed Conflict.”

¹¹¹. See the Sexual Violence in Armed Conflict Dataset, www.sexualviolencedata.org/dataset/.

¹¹². Patricia Tjaden and Nancy Thoennes, “Extent, Nature, and Consequences of Intimate Partner Violence,” U.S. National Institute of Justice, July 2000, https://stacks.cdc.gov/view/cdc/21858; Pamela Powell and Marilyn Smith, “Domestic Violence: An Overview,” Fact Sheet no. 11-76, University of Nevada, 1989, www.unce.unr.edu/publications/files/cy/2011/fs1176.pdf; and Stephanie Holt, Helen Buckley, and Sadhbh Whelan, “The Impact of Exposure to Domestic Violence on Children and Young People: A Review of the Literature,” Child Abuse & Neglect 32, no. 8 (2008): 800, https://doi.org/10.1016/j.chiabu.2008.02.004.

¹¹³. Codebook and data are available from the Center for Systemic Peace, “Integrated Network for Societal Conflict Research (INSCR) Data Page,” www.systemicpeace.org/inscrdata.html.

Antonio Rodiles and Erik Jennische

On April 19, Raúl Castro stepped down from his self-ascribed title as president of Cuba and transferred the post to his deputy, Miguel Díaz-Canel. For the first time since 1959, neither of the Castro brothers, Raúl or the late Fidel, supposedly rules the island. The handover of power to a new generation — Díaz-Canel is 57 years old — and changes to some political rules, such as the introduction of term limits, have fueled hopes that in the midterm a democratic opening might be in the cards. However, this so-called transition is just a mirage.

The Castro family remains firmly in control of the government and the military. Raúl has kept his posts as secretary general of the Communist Party of Cuba (PCC) and commander-in-chief of the Cuban Revolutionary Armed Forces (FAR). Meanwhile, his son, Alejandro Castro Espín, is at the center of a new power structure that Raúl carefully put in place in recent years, one in which the military elite and the second tier of the Communist Party leadership are in charge.

Fifty-two-year-old Castro Espín is currently a colonel in the Interior Ministry. He is the coordinator of the intelligence and counterintelligence services, which makes him one of the most powerful figures in Cuba. He was also the head of the National Defense and Security Commission, a recently disbanded advisory body to Raúl Castro that many perceived to be a “parallel government.” Rumor has it that Raúl’s goal is to place Castro Espín as secretary general of the PCC by 2021, which would make him the effective ruler of Cuba.¹

Díaz-Canel — although nominally the president — will not wield real power. He himself confirmed this in his inaugural speech to the National Assembly, in which he stated that “Raúl Castro Ruz, as First Secretary of the Communist Party of Cuba, will make the most important decisions for the present and the future of the nation.”²

The idea that a democratic transition is underway in Cuba is further belied by two well-documented developments: the increased crackdown on dissidents and groups in civil society, and the regime’s backtracking on the timid economic policy changes Raúl Castro implemented when he came to power in 2006.

A GRIM OUTLOOK FOR CIVIL LIBERTIES

Despite the hopes stirred by the diplomatic rapprochement between Washington and Havana more than three years ago, it is now clear that the Cuban regime does not intend to change the fundamental nature of its Stalinist political system. In fact, there is evidence that the dictatorship has increased its repression of dissidents and civil society.

The number of arbitrary detentions for political reasons reached 9,940 in 2016, exceeding that of any previous year since 2010,³ and detentions have since remained high.⁴ Since Barack Obama visited the island in March 2016, the Cuban political police have made it even more difficult to demonstrate. Police officers are now placed outside the houses of dozens of activists during the weekends to prevent those activists from participating in the Sunday marches organized by the Damas de Blanco (Ladies in White) and #TodosMarchamos. The police also make it nearly impossible for other human rights defenders and political activists to meet. Hostility against the families of activists has increased as well. The number of political prisoners has doubled to 140 over the last couple of years.⁵

Even though in 2013 the regime lifted — although not entirely — the requirement for ordinary Cubans to get permission to travel abroad, democracy activists’ ability to travel is severely limited. Dozens of Cubans have been arrested by the security police days before traveling, on their way to the airport, or even at the airport, and have thus been prevented from participating in seminars and conferences abroad organized by international human rights organizations. Limits on travel to the island also persist for Cubans living abroad: individuals who have criticized the dictatorship or been politically active against it are not allowed to visit.

In April 2016, the PCC gathered for its Seventh Congress. In his opening speech, Raúl Castro made it clear that there would be no reforms that could threaten the “unity of the majority of the people behind the Party” or “cause instability and insecurity.”⁶ Referring to international demands for a multiparty system, Castro clarified that such a system would occur “neither today, nor ever” and warned that “if one day they succeed in fragmenting us, it would be the beginning of the end of our fatherland, of the Revolution, socialism, and national independence.” Foreign minister Bruno Rodríguez claimed in his speech that Barack Obama’s visit had been “an attack on our conception, on our history, on our culture, and on our symbols.”⁷

The objective of the PCC’s Seventh Congress was to discuss two documents. The first describes the principles and theories of the economic and social model of the Cuban government. It states that the PCC is “the superior leading force of the society and the State.”⁸ The second document outlines the “Vision of the Nation” for 2030. Cuba should be “sovereign, independent, socialist, democratic, prosperous, and sustainable,” and to achieve this, the document deems it necessary to have an “efficient and socialist government.”⁹

Two guiding principles of that vision — national defense against aggression and national security — reinforce the regime’s current defense doctrine, which states that the regime’s institutions, political and mass organizations, and the rest of the population will participate in confronting the activities of “the enemy.” The militarization of society and the inclusion of ordinary citizens in the surveillance system have been two of the most effective strategies to curb self-organizing in Cuban civil society.

There is scant mention of reform in those documents. Freedom of expression or assembly or multiparty democracy cannot be part of the regime’s narrative. When the PCC declares those principles and that vision for the coming 15 years, it does not have in mind anything other than continuing in exactly the same way as it has for the last six decades.

The only reform within the political system announced at the PCC Seventh Congress concerned the age of individuals entering the highest positions of the party — the central committee, the secretariat, and the political bureau — and how long they will be allowed to hold their positions. In the future, nobody above the age of 60 will enter those bodies or serve for more than two five-year terms. With this amendment, Raúl Castro wanted to rejuvenate the apparatus and create a new network of loyalists for his son and inheritor, Alejandro. Castro also promised that those changes would be included in the constitution and proposed a constitutional reform and subsequent referendum, a process that would “ratify the irrevocable nature of the political and social system.”¹⁰

BACKTRACKING ON TIMID POLICY CHANGES

The consolidation of power in the hands of the Castro family does not seem to be the main topic of concern for most Cuba watchers. Instead, they focus on the promise of the economic program adopted by the PCC in 2011, aimed at creating a small-business sector that could generate employment and improve services. Unfortunately, those policies have been too timid to bring about meaningful change to the Cuban economy, and the regime is now backtracking on some of them.

The number of independent microbusinesses grew rapidly between 2010 and 2014, but that growth has significantly decelerated since.¹¹ In recent months, the regime has announced new restrictions on the private sector because of complaints about, among other things, “excess accumulation of wealth.”¹² It also stopped handing out licenses for small businesses, saying it needs to reevaluate the legal framework around these businesses and combat corruption related to them. In July 2017, Raúl Castro openly criticized the dynamics of the microbusinesses. A leaked video recently showed Díaz-Canel saying that the regime sees entrepreneurs as capitalist instruments who can destroy the revolution. As The Economist’s Bello column rightly points out, “the government wants a market economy without capitalists or businesses that thrive and grow.”¹³

The regime continues to exert absolute control over the legal labor market, retaining — or confiscating — around 95 percent of the hard-currency earnings of all Cubans working in the formal dollar economy.¹⁴ These profits are then invested in the state’s repressive machinery and in the personal coffers of the Communist Party leadership. This modern-day system of slavery will not lead to the empowerment of Cuban workers or to the advancement of their rights.

Moreover, the concentration of economic power in the hands of the FAR has accelerated since 2014. The FAR own at least 57 companies and half of the retail businesses in Cuba, along with car fleets, gas stations, and supermarkets — all of which are key sectors of the economy.¹⁵ They also control at least 40 percent of the foreign capital in the country through their holding company, Grupo de Administración Empresarial Sociedad Anónima (GAESA). This means that foreign investors in Cuba must establish direct relations with GAESA and its CEO, Luis Alberto Rodríguez López-Callejas, Raúl Castro’s son-in-law.

PROSPECTS FOR CHANGE

The lack of human rights and democracy is the essence of Cuba’s totalitarian political and economic system. The legacy of the Castro brothers includes not only executions, imprisonments, assassinations, torture, beatings, harassment, and intimidation but also a constitutional and legal framework that legalizes repression and promotes widespread violations of human rights by the authorities.

It is not realistic to expect that the Cuban regime will embrace democracy and the rule of law any time soon. A real transition in Cuba must involve the immediate release of political prisoners, the restitution of all fundamental rights and freedoms, the complete dismantling of the dictatorship, and the celebration of free, multiparty, and competitive elections — in other words, the construction of a functioning democracy.

NOTES

¹ Roberto Álvarez Quiñonez, “¿Necesita Alejandro Castro, el hijo consentido, ser presidente?” Diario Las Américas, January 22, 2018, https://www.diariolasamericas.com/america-latina/necesitaalejandro-castro-el-hijo-consentido-ser-presidente-n4141838.

²“Primer discurso completo de Díaz-Canel como presidente de Cuba,” Pulso de los Pueblos, April 19, 2018, http://pulsodelospueblos.com/primer-discurso-completo-de-diaz-canel-como-presidentede-cuba/.

³Deutsche Welle, “Cuba Arbitrary Arrests Soared in 2016, Dissidents Say,” January 6, 2017, http://www.dw.com/en/cubaarbitrary-arrests-soared-in-2016-dissidents-say/a-37033489.

⁴ Data f rom Defenders’ Databas e, https://database.civilrightsdefenders.org/.

⁵ Figures from the Cuban Commission on Human Rights and National Reconciliation (CCDHRN), http://ccdhrn.org/.

⁶ Raúl Castro Ruz, “Informe Central al VII Congreso del Partido Comunista Cuba,” Cuba Debate, April 17, 2016, http://www.cubadebate.cu/noticias/2016/04/17/informe-centralal-vii-congreso-del-partido-comunista-cuba/#.Wt9DkBPwZhE.

⁷“Raúl Castro y canciller cubano arremeten contra visita de Obama,” CBS News, April 18, 2016, https://www.cbsnews.com/news/raul-castro-y-canciller-cubano-arremeten-contra-visita-de-obama/.

⁸ See “Documentos del 7mo. Congreso del Partido aprobados por el III Pleno del Comité Central del PCC el 18 de mayo de 2017 y respaldados por la Asamblea Nacional del Poder Popular el 1 de junio de 2017,” http://www.granma.cu/file/pdf/gaceta/%C3%BAltimo%20PDF%2032.pdf.

⁹“Documentos del 7mo. Congreso del Partido aprobados por el III Pleno del Comité Central del PCC el 18 de mayo de 2017 y respaldados por la Asamblea Nacional del Poder Popular el 1 de junio de 2017.”

¹⁰ Article 5 of the Cuban Constitution already states that “The Communist Party of Cuba … is the highest leading force of society and of the state, which organizes and guides the common effort toward the goals of the construction of socialism and the progress toward a communist society.”

¹¹ Redacción, “Licencias para trabajo privado en Cuba: las que sí, las que no,” On Cuba, August 3, 2017, https://oncubamagazine. com/economia-negocios/licencias-trabajo-privado-cuba-las-quesi-las-que-no/.

¹² Reuters, “Cuba Tightens Regulations on Nascent Private Sector,” December 21, 2017, https://www.reuters.com/article/us-cuba-economy/cuba-tightens-regulations-on-nascent-privatesector-idUSKBN1EF318.

¹³ Bello, “A Year Without Fidel.” The Economist, December 9, 2017, p. 38.

¹⁴“Testimony of the International Group for Corporate Social Responsibility in Cuba before the Inter-American Commission on Human Rights,” October 15, 2005, p. 17, http://www.cubastudygroup.org/index.cfm/files/serve?File_id=5411e1c1-7883-4015-9d22-2d0ecc99318b.

¹⁵ Michael Smith, “Want toMarkets Do Business in Cuba? Prepare to Partner with the General.” Bloomberg Markets, September 30, 2015, https://www.bloomberg.com/news/features/2015-09-30/want-toinvest-in-cuba-meet-your-partner-castro-s-son-in-law.

Colin Grabow, Inu Manak, and Daniel J. Ikenson

For nearly 100 years, a federal law known as the Jones Act has restricted water transportation of cargo between U.S. ports to ships that are U.S.-owned, U.S.-crewed, U.S.-registered, and U.S.-built. Justified on national security grounds as a means to bolster the U.S. maritime industry, the unsurprising result of this law has been to impose significant costs on the U.S. economy while providing few of the promised benefits.

This paper provides an overview of the Jones Act by examining its history and the various burdens it imposes on consumers and businesses alike. While the law’s most direct consequence is to raise transportation costs, which are passed down through supply chains and ultimately reflected in higher retail prices, it generates enormous collateral damage through excessive wear and tear on the country’s infrastructure, time wasted in traffic congestion, and the accumulated health and environmental toll caused by unnecessary carbon emissions and hazardous material spills from trucks and trains. Meanwhile, closer scrutiny finds the law’s national security justification to be unmoored from modern military and technological realities.

This paper examines how such an archaic, burdensome law has been able to withstand scrutiny and persist for almost a century. It turns out that, as in so many other cases of rent seeking, there is an asymmetry of motivations among those who benefit from the Jones Act’s protections and the vastly greater number who bear its costs. The protected domestic shipbuilding industry has a captive market from which it benefits handsomely and seeks to preserve by promoting fallacious arguments about the law’s necessity to national security, while the vast costs are dispersed across the economy in the form of higher prices, inefficiencies, and forgone opportunities that few people can even tie to the cause. That so many federal agencies and congressional committees have at least partial jurisdiction over different facets of the Jones Act also helps to explain its longevity. Lastly, this paper presents a series of options for reforming this archaic law and reducing its costly burdens.

Terence Kealey

In 2015 the Department of Health and Human Services and the Department of Agriculture released the latest iteration of their dietary advice, Dietary Guidelines for Americans 2015-2020. Upon receiving it, Congress, citing concerns over scientific integrity, commissioned the National Academy of Medicine to review the process of generating those guidelines. In its commission, Congress asked the National Academy of Medicine for full transparency, lack of bias, and the inclusion of all latest available research, however challenging.

By so asking, Congress was suggesting that the federal government’s dietary recommendations — and in particular its long-standing demonization of fats and its praise for carbohydrates — were suspect.

The story starts on January 14, 1977, when the Senate Select Committee on Nutrition and Human Needs published its Dietary Goals for the United States, which, for the first time, attacked overeating. Previously, the Committee had worried about undernutrition, but by the late 1970s it worried that the epidemic of heart attacks could be attributed to an excessive intake of saturated fats. It therefore recommended that Americans eat carbohydrates instead.

Unbeknownst to the vast majority of Americans, however, the theory that replacing saturated fats with carbohydrates would lower the risk of heart attacks was unproven and disputed. Moreover, the government’s dietary advice led Americans to indulge in the widespread consumption of trans unsaturated fats, which are themselves dangerous. Further, this advice coincided with — and probably contributed to — the subsequent epidemics of obesity and type 2 diabetes.

Today, most nongovernmental dietary advice focuses on the benefits of plant-based fats and a Mediterranean diet, and while that, too, may be only a work in progress, it is much better than the paradigm that was disseminated by the government during the 1970s. Yet the government still propagates the oversimplified idea that fats are bad and carbohydrates are good.

In fact, the federal government may be institutionally incapable of providing wise dietary advice, as Thomas Jefferson warned us in his 1787 Notes on the State of Virginia: “Was the government to prescribe to us our medicine and diet, our bodies would be in such keeping as our souls are now.”

Introduction

The federal government’s agencies have been issuing dietary advice for more than a century, but before 1977 they limited themselves largely to addressing malnutrition among the poor. The first two-thirds of the 20th century had witnessed the early triumph of nutrition research when, in a world still concerned with malnourishment, the discipline had helped oversee the discovery of vitamins and the elaboration of the basic principles of metabolic biochemistry. In 1968, the Senate created the Select Committee on Nutrition and Human Needs. At the instigation of its first chairman, Sen. George McGovern (D-SD) — who was to achieve his greatest prominence when he ran against President Richard Nixon — the committee focused initially on the problems of undernutrition. But a decade later, on January 14, 1977, when it published its Dietary Goals for the United States, the committee launched an attack on the apparent problems of overconsumption. In his foreword to the Goals McGovern wrote:

This is the first comprehensive statement by any branch of the federal government on risk factors in the American diet.

Too much fat … [is] … linked directly to heart disease, cancer, obesity and stroke.

… six out of the ten leading causes of death in the United States [heart disease, cancer, vascular disease, diabetes, arteriosclerosis and cirrhosis of the liver] have been linked to our diet.¹

The committee reported unanimously, and in his own foreword Sen. Charles Percy (R-IL) wrote, “without government … commitment to good nutrition, the American people will continue to eat themselves to poor health.”² Consequently, the committee explained, “We as a government … have an obligation to provide practical guides to the individual consumer as well as to set national dietary goals for the country.” Accordingly, Americans were urged to:

1. Increase carbohydrate consumption to account for 55 to 60 percent of energy (caloric) intake.
2. Reduce overall fat consumption from approximately 40 to 30 percent of energy intake.
3. Reduce saturated fat consumption to account for about 10 percent of total energy intake.
4. Reduce cholesterol consumption to about 300 mg a day.³

The committee had, in short, officially launched the anti-fat, high-carbohydrate campaign that was to dominate the world of nutrition until recently and which still reigns in official circles.

Why Did the Senate Select Committee Launch an Attack on Fats?

The problem was heart attacks. These had seemingly come out of nowhere and by 1968, at the height of the epidemic, accounted for more than a third (37 percent) of all deaths in the United States. By contrast, all cancers accounted for only a sixth (17 percent) of all deaths, and strokes only a tenth (10 percent). Accidents, at 6 percent, were the fourth most common cause of death.⁴ The sudden epidemic of heart attacks was profoundly alarming, especially as it seemed to target otherwise-healthy people at the peak of their performance.

Some physicians argued the epidemic was illusory, the result of better diagnosis and an aging population; yet that argument, although not trivial, was to be disproved. In 1966, for example, Leon Michaels, a Canadian physician, showed that the absence of evidence for heart attacks before the 20th century was indeed evidence for their absence: on comparing the characteristic chest pain of heart attacks with the characteristic pain and symptoms of migraine and gout, he showed that, whereas frequent descriptions of migraine and gout can be found in medical texts from all eras, stretching back to Greek and Roman times, angina and heart attacks started to be described with any frequency only in the 20th century.⁵ He thus concluded that the death rates from those diseases had increased up to 200 fold between 1901 and 1962, and although that rate of increase cannot be known with certainty, it is now accepted that the rise in the incidence of heart attacks during the 20th century was real (see Figure 1).

Figure 1: Death rates per 100,000 people from atherosclerotic heart disease and stroke

Sources: See Daniel T. Lackland et al., “Factors Influencing the Decline in Stroke Mortality: A Statement from the American Heart Association/American Stroke Association,” Stroke 45, no. 1 (December 5, 2013): 315-53, doi:10.1161/01.str.0000437068.30550; and Leon Michaels, “Aetiology of Coronary Heart Disease: An Historical Approach,” British Heart Journal 28, no. 2 (March 1966): 258-64.
Note: The data in solid lines come from the joint 2013 statement of the American Heart and Stroke Associations. Data on atherosclerotic heart disease were not collected before 1950, so the dotted lines are extrapolated from the estimates in the Michaels article.

Faced with such an epidemic, some commentators argued it was surely reasonable for the U.S. government to address it. In 1974, for example, Marc LaLonde, Canada’s Minister of National Health and Welfare, had published a working paper, A New Perspective on the Health of Canadians, intended to prescribe the diet of the Canadian people (in which he suggested that Canadians should eat less fat — in particular less saturated fat and less cholesterol — and more carbohydrates) and some commentators argued it was surely not unreasonable for the U.S. Senate Select Committee on Nutrition and Human Needs also to take a position.

It might indeed not have been unreasonable — had the committee’s position been a wise one. Yet the federal government may be institutionally incapable of providing wise dietary advice.

Why the Demonization of Fat?

The senators on the Select Committee were not, of course, nutritionists, so they took their lead from the scientists, and from one scientist in particular: Ancel Keys (1904-2004), professor of physiology at the University of Minnesota. It was Keys who launched the modern dietary era in 1953 when he published his paper, “Atherosclerosis: A Problem in Newer Public Health.”⁶

By 1953 the Food and Agriculture Organization of the United Nations had collected dietary information on 22 countries. Keys selected data from six countries (Australia, Canada, England and Wales, Italy, Japan, and the United States) to generate his renowned graph, in which he showed that as the percentage of fat in the diet rose, ranging from Japan (7 percent fat in the diet) through to the United States (40 percent fat in the diet), so the death rates from heart disease of males aged 55-59 years rose accordingly, from 0.5 per thousand in Japan to just under 7 per thousand in the United States.

In 1953, as today, heart disease was understood essentially as the consequence of atherosclerosis. But whereas today we understand atherosclerosis to be an inflammation of the arteries, in 1953 it was seen more as a hardening (technically an arteriosclerosis) caused by the deposition of cholesterol within the arterial walls. Keys proposed that as people ate too much fat (in particular too much cholesterol) the blood vessels silted up with cholesterol, leading the heart to become diseased as its own arteries, being narrowed, failed to provide enough oxygen or nourishment. That process in turn would lead to a heart attack or myocardial infarction as the blood clotted over the cholesterol-filled artery and thus killed the patient. Here was Keys’s model:

Initial Criticism of Keys’s Model. It is often forgotten that, from its moment of conception, Keys’s model was heavily criticized: in 1955, for example, the World Health Organization (WHO) convened a small seminar of international experts, who proceeded comprehensively to demolish it (one of his colleagues remembered the WHO meeting as “the pivotal moment in Keys’s life. He got up from being knocked around and said ‘I’ll show these guys.’”).⁷ Among the critics were Jacob Yerushalmy and Herman Hilleboe, from the University of California, Berkeley, and the New York State Commission of Health, respectively, who attacked the model both at the seminar and shortly afterwards in a paper.⁸

At the 1955 seminar Keys had claimed “there is a remarkable relationship between the death rate from degenerative heart disease and the proportion of fat calories in the national diet,” which was supportable. But then he claimed “no other variable … shows anything like such a consistent relationship,” which wasn’t supportable.⁹ When Yerushalmy and Hilleboe reexamined the Food and Agriculture Organization data, they showed the association between animal protein and heart disease was stronger than between fat and heart disease. And the association between the total consumption of calories and heart disease was stronger yet. Meanwhile, the strongest determinant of calorie and meat intake seemed to be GDP per capita.

On the other hand, consumption of vegetables seemed to protect against heart disease, even though vegetables contain fat, protein, and carbohydrates. To further complicate matters, it appeared that the greater the consumption of animal fat and protein, the lower the death rates from every other condition except for heart disease.

So, the model that best accounted for the empirical facts was:

As Yerushalmy and Hilleboe pointed out at the 1955 WHO seminar, and as they expanded in their 1957 paper, the data thus suggested the citizens of poor countries (who largely ate vegetables, including starchy vegetables such as maize/corn, rice, and potatoes) didn’t die much of heart disease (but they were vulnerable to other diseases); while the citizens of rich countries (who ate a lot of meat, which includes much fat) died largely of heart disease (but were protected from other causes of death).

And to confirm that GDP per capita seemed central to the development of heart attacks, Yerushalmy and Hilleboe noted that atherosclerosis remitted when people reverted to a pre-Western lifestyle: during the Second World War and its aftermath, many parts of Europe had been reduced to meager diets, and those parts of Europe saw their rates of heart disease fall; but when normal food supplies were restored, heart disease returned. Heart disease, in short, seemed to be a consequence of a Western diet (which was in turn a consequence of Western wealth), but no single element of that diet could be identified as especially responsible.

The discussions at the 1955 WHO seminar were prescient, because the delegates saw ischemic heart disease not as the consequence of a single cause (cholesterol) but as the consequence of many complex, still-to-be-elucidated causes. More of those possible causes were soon being identified, and when in 1957 professor of physiology John Yudkin of the University of London reexamined the Food and Agriculture Organization data, he found:

a moderate but by no means excellent relationship between fat consumption and coronary mortality… . A better relationship turned out to exist between sugar consumption and coronary mortality in a variety of countries.¹⁰

Yudkin therefore proposed that sugar, not fat, was bad for the heart, and in his book Pure, White and Deadly he wrote of “good nutritious foods like meat and cheese and milk.”¹¹ Also in 1957, in a paper entitled “Dietary Control of Serum Lipids in Relation to Atherosclerosis” published in the Journal of the American Medical Association, Pete Ahrens (1915-2000) of the Rockefeller Institute of Medical Research — the doyen of fat biochemistry — who had long recognized carbohydrates as cardiac killers, was protesting that “unproved hypotheses are enthusiastically proclaimed as facts.”¹² (Ahrens was later to greet the Senate Select Committee on Nutrition and Human Needs’ recommendations as treating people as if they were “a homogenous group of Sprague-Dawley rats.”¹³ ) George V. Mann, a University of Vanderbilt biochemist, was yet another researcher who had shown that Keys’s demonization of animal fats did not fit the cardiac facts.¹⁴

Meanwhile, in 1970, Richard Doll, the epidemiologist who had earlier reported that cigarettes caused lung cancer, found, “It is cigarette smoking … which is implicated in the aetiology and manifestation of myocardial infarction.”¹⁵ Figure 2 illustrates that one nondietary phenomena that tracks the incidence of atherosclerotic heart deaths is cigarette smoking.

Figure 2: Male peptic ulcer death rates and per capita cigarette consumption

Source: Alexander Mercer, Infections, Chronic Disease, and the Epidemiological Transition: A New Perspective (New York: University of Rochester Press, 2014), p. 184. The data on cigarette consumption come from Centers for Disease Control and Prevention, Achievements in Public Health, 1900-1999: Tobacco Use, United States, 1900-1999. Morbidity and Mortality Weekly Report (Washington: CDC, 1999): 986-93.

Nor did the list of possible causes stop there, and further causes, including a lack of exercise and excessive stress, were soon identified.

By the 1970s, therefore, the criticisms of the 1955 WHO seminarians had been vindicated, and the only sure model compatible with the data was that found in Figure 3.

Figure 3: Proposed causes of atherosclerosis, circa 1970s

Source: These data have been extracted by the author from the contemporary literature.

But which of the many possible factors was responsible for the epidemic of heart attacks could not be isolated.

So why in 1977 did the Senate committee back Keys’s dietary fat hypothesis against all the other possible causes of atherosclerosis? Well, Keys had generated a model: whereas no one could easily suggest how calories, protein, cigarettes, sugar, a lack of exercise, or a lack of vegetables could provoke atherosclerosis, Keys could suggest how fat could do it.

Keys’s First Model for Atherosclerosis. Keys’s first set of points noted that heart attacks were caused by atherosclerosis; atherosclerotic plaques were full of cholesterol; and patients who suffered heart attacks had elevated blood levels of cholesterol. Keys recommended avoiding cholesterol, noting that rabbits fed high levels of cholesterol develop atherosclerosis.¹⁶

Under this model:

By 1955, though, Keys had appreciated that dietary cholesterol is not a human danger: our livers synthesize most of our cholesterol, and when we ingest it in our food, our livers simply reduce their creation of cholesterol.¹⁷ This is not true of all animals, particularly not of herbivores such as rabbits, which — because plants are low in cholesterol — do not normally handle significant amounts of it. So herbivores, when fed high cholesterol in the laboratory, respond with raised cholesterol in the blood: their livers don’t know any better. But we are omnivores, our livers are not naive, and when fed high cholesterol we do not respond with raised levels of blood cholesterol. For us:

Asymmetrical Science

Although by 1955, within two years of originally proposing it, Keys had abandoned the dietary cholesterol hypothesis, for another 60 years the federal government continued to warn against consuming cholesterol-rich foods. It was only in 2015 that its Dietary Guidelines Advisory Committee classified high-cholesterol foods such as eggs, shrimp, and lobster as safe to eat: “cholesterol is not a nutrient of concern for overconsumption.”¹⁸

This 60-year delay shows how asymmetrical the official science of nutrition can be: a federal agency can label a foodstuff dangerous based on a suggestion, yet demand the most rigorous proof before reversing its advice. The Harvard professor of epidemiology and nutrition Walter Willett, commenting on the asymmetry in a related area of government nutrition advice, described it as “Scandalous. They say ‘You really need a high level of proof to change the recommendations,’ which is ironic, because they never had a high level of proof to set them.”¹⁹

And it was Keys himself who championed asymmetry in dietary advice when he wrote in 1957 that nobody had “adequate evidence to state that there is not a causal relationship between dietary fat and the tendency to develop atherosclerosis in man” (i.e., he could condemn fat on the basis of a hypothesis only, yet it could only be classified as safe after exhaustive study).²⁰ So Keys not only launched the cholesterol/fat paradigm on inadequate evidence, he also biased the debate in its favor.

Further, it may seem safer to advise abstention from a particular food than to clear one as safe. Yet this abstention may itself be dangerous, because abstention can have unintended consequences: other foodstuffs must be consumed instead. The Select Committee’s major scientific adviser was Mark Hegsted of Harvard, who wrote in his introductory statement to the Goals of 1977:

The question to be asked, therefore, is not why we should change our diet but why not? What are the risks associated with eating less meat, less fat, less saturated fat, less cholesterol, less sugar, less salts, and more fruits, vegetables, unsaturated fat and cereal products — especially whole grain cereals?²¹

The answer to this question, namely that it would lead to the eating of more carbohydrates and more trans fats, both of which really are dangerous, would soon emerge. (Trans fats are chemically synthesized unsaturated fatty acids associated with increased risk of coronary heart disease.)

Meanwhile, the 60 years of official misinformation has taken its toll: a 2015 survey by Credit Suisse Foundation (a social research charity) found 54 percent of doctors falsely believed eating cholesterol-rich food raises blood levels of cholesterol and damages the heart. In the words of the survey, “This is a clear example of the level of misinformation that exists among doctors.”²²

Keys’s Second Model

Keys always saw dietary cholesterol as only one of two problematical factors, and by 1955 he was presenting his second set of facts, noting that eating fat, especially saturated or animal fat, raised the blood levels of cholesterol. His recommendation was that we should avoid eating saturated and animal fats.

From this, he generated his second model:

This, of course, was the model the Select Committee was to eventually to endorse in 1977. But as we have seen, on its being presented to the WHO delegates back in 1955, they had immediately been skeptical, noting:

the evidence is circumstantial … no conclusions of etiological [causative] relationships should be attempted unless the factor is found to be related to the disease by evidence [are] from entirely different types of investigations.²³

To counter this skepticism, Keys had launched his famous Seven Countries study, published in 1970, in which he personally examined what people were eating in Finland, Greece, Italy, Japan, the Netherlands, the United States, and Yugoslavia. He refined his surveys to look at saturated (i.e., animal) fats rather than total fats in the diet, and yet again he found a strong association between (saturated) fat ingestion and deaths from heart disease.²⁴

But the Seven Countries study wasn’t the “entirely different type of investigation” for which the WHO experts had called, as it still generated data that were, in the WHO experts’ words, only “circumstantial.”²⁵ Indeed, when one of Keys’s colleagues, Alessandro Menotti, recharacterized the foodstuffs in the Seven Countries study, he found “sweets” (sugar-rich products, cakes, and other confectioneries) in the diet correlated more strongly with coronary mortality than did “animal food” (butter, meat, eggs, margarine, lard, milk, and cheese). Even Keys’s own program of work, therefore, suggested it might be carbohydrates, not fats, that killed people.²⁶

Bad Science

The Select Committee’s gravest offense wasn’t going beyond the verifiable science, but actually flouting it. There is in epidemiology a “hierarchy of evidence”: some data are recognized to be more credible than others, and in particular randomized controlled trials (i.e., experiments) are recognized to produce harder data than observations (which may report only associations, and which may, in turn, mislead). By 1977 (when the Select Committee published its report) no fewer than six randomized trials had been performed on a total of 2,467 males (of whom 423 died from heart problems during the trials) in which their total and saturated fat intakes were reduced by placing them on low-fat diets. As predicted by the Keys hypothesis, the subjects on reduced dietary fat showed a fall in their circulating blood levels of cholesterol. But, contrary to the hypothesis, their mortality rates did not fall. In their devastating review of the six trials, freelance nutritionist and author Zoë Harcombe and her colleagues in Wales and Kansas City concluded, “Dietary recommendations were introduced for 220 million US citizens … in the absence of supporting evidence from randomized controlled trials.”²⁷

That lack of hard evidence was recognized at the time, and in 1977 the American Medical Association responded to the Goals with the statement:

The evidence for assuming that benefits [are] to be derived from the adoption of such universal dietary goals … is not conclusive, and there is potential for harmful effects.²⁸

Although the committee had, in its Goals, acknowledged the incomplete state of the science of the day, it had also written with approval:

Marc LaLonde, Canada’s Minister of National Health and Welfare, said: “Even such a simple question as whether one should severely limit his consumption of butter and eggs can be a matter of endless scientific debate … [so] it would be easy for health educators and promoters to sit on their hands… . But many of Canada’s health problems are sufficiently pressing that action has to be taken even if all the scientific evidence is not in.”²⁹

On being challenged on the incompleteness of the science, Senator McGovern said “Senators do not have the luxury that the research scientist does of waiting until every last shred of evidence is in,” which is the opposite of the truth: research scientists are at leisure — and are perhaps even obligated — to explore every possible hypothesis, but senators should not issue advice until every last shred of evidence is in, because they may otherwise issue misleading or even dangerous advice.³⁰ As they did in 1977.

We see here a second reason why official dietary advice may be institutionally biased, because officialdom may be under pressure to issue it prematurely, sometimes even by decades. Moreover, this advice may be based on models rather than on hard facts.

Why Did the Randomized Controlled Clinical Trials Not Confirm the Dietary Saturated Fat Model?

The Keys model was two-staged:

Confusingly, both stages of the model were true. Mark Hegsted, the head of the Department of Nutrition at Harvard and the Select Committee’s major adviser, had shown that when humans ate saturated fat their circulating blood levels of cholesterol did indeed rise. Equally, two future Nobel laureates from Texas, Michael Brown and Joseph Goldstein, showed — in certain inherited diseases of metabolism — that high blood levels of cholesterol can indeed cause heart attacks. So, the committee put two and two together and supposed:

thus reminding us of H. L. Mencken’s aphorism that for every complex problem there is a solution that is clear, simple, and wrong. As we saw above, when no fewer than six randomized controlled clinical trials had tested the complete model by withdrawing saturated fat from the diet of vulnerable men, their blood levels of cholesterol fell but their heart death rates did not. Why not?

It transpires there are at least three different types of circulating blood cholesterol. One type, so-called HDL (high density lipoprotein) is positively healthful, as it draws cholesterol out of the arteries. Another type, lLDL (large low density lipoprotein), is largely neutral, while a third type, sLDL (small low density lipoprotein) is the one that kills, as it (and its oxidized forms) tend to lodge in the arteries and precipitate the inflammation we know as atherosclerosis. Moreover, saturated fat in the diet tends to raise the circulating levels of the essentially neutral lLDLs, while carbohydrate in the diet tends to raise the circulating levels of the dangerous sLDLs. Therefore, it is carbohydrate, not fat, in the diet that raises the dangerous type of cholesterol, although the mechanisms remain to be fully elucidated.

So it is no surprise the six randomized controlled clinical trials failed to find a fall in the rate of heart death rates following the reduction of saturated fat in the diet of vulnerable men. This is because the compensatory rise in carbohydrate intake was dangerous. To Mark Hegsted’s question in his introductory statement to the Goals— “What are the risks associated with eating less meat, less fat, less saturated fat, less cholesterol?” — we can now reply that if, in consequence, people were to follow his advice and eat more carbohydrates and more trans fats in compensation, the risks are of precipitating early death from atherosclerosis. Irony of ironies.

The American People Were Dutiful

Nonetheless the American people did as they were advised — not just by the government, but also by the mass media, which reinforced the government’s message.³¹ A survey performed jointly by the National Heart, Lung, and Blood Institute and the Food and Drug Administration showed that, by 1986, 72 percent of adults “believed that reducing high blood levels of cholesterol would have a large effect on heart disease.”³² Consequently, between 1960 and 2000, their per capita consumption of saturated fatty acids fell from 55 to 46 grams per day, and their per capita consumption of cholesterol fell from 465 to 410 milligrams per day. Meanwhile, their per capita consumption of carbohydrates rose from 380 to 510 grams per day and consumption of fiber rose from 18 to 26 grams per day (Figure 4).³³

Figure 4: Obesity and the consumption of different foods in America, 1960-2000

Source: The data come from Shi-Sheng Zhou et al., “B-Vitamin Consumption and the Prevalence of Diabetes and Obesity among US Adults: Population Based Ecological Study,” BMC Public Health 10 (2010): 746, https://doi.org/10.1186/1471-2458-10-746.

Americans also increased their intake of unsaturated vegetable-derived fats chemically modified as trans fats (as in replacing butter with margarine) because it was believed that saturated rather than unsaturated fat was dangerous. So whereas in 1911 per capita consumption of butter was 19 pounds a year to 1 pound of margarine, by 1976 butter consumption had fallen to 4 pounds a year while margarine’s had risen to 12.³⁴ We now know that trans fats lower healthful HDLs, raise the dangerous sLDLs, and are inflammatory. Consequently, in the words of a recent authoritative review, they “contribute significantly to increased risk of coronary heart disease events.”³⁵

The Select Committee’s demonization of saturated animal fat can only have damaged the health of Americans who followed its advice by causing them to increase their consumption of trans fats.

Keys and the Select Committee Vindicated? Nonetheless, on following the Select Committee’s advice, an odd thing happened to the American people: their rates of heart disease and stroke fell, really quite dramatically (see Figure 1). So, does Figure 1 vindicate the Select Committee’s advice?

Before invoking Figure 1 as vindication of the Select Committee’s advice, consider Figure 2, which illustrates the course of another 20th-century epidemic, that of peptic ulcers. It is often now forgotten, but the 20th century witnessed an epidemic of peptic ulcers that mirrored that of heart attacks. Since nobody has suggested saturated fat is the cause of peptic ulcers, this must warn us not to confuse correlation with causation: a disease may reflect the incidence of heart attacks without that incidence proving that saturated fat is responsible.

Moreover, the stroke data in Figure 1 are incompatible with the Keys hypothesis. Strokes are caused by the same atherosclerotic process as heart attacks, yet their epidemiology looks different: namely, they peak before the 20th century, and during the first half of the 20th century they appear to decline even in the face of the high-fat Western diet. During the second half of the 20th century they continue to decline, even as the Western diet becomes lower-fat and higher-carbohydrate. Figure 1, indeed, shows that variations in dietary composition cannot explain the epidemiology of strokes.

Strokes and the Metabolic Syndrome

So, how can we explain the epidemiology of strokes? Well, the World Health Organization reports that in 2015, the latest year for which we have data, the two commonest causes of death among low-income economies were chest infections and diarrhea (i.e., bacterial and viral infections of the lungs and guts). But stroke was third.³⁶

Strokes are a disease of poverty, and the British epidemiologist David Barker best explained the association of strokes with poverty when he showed — unexpectedly — a correlation between maternal mortality rates and strokes.³⁷ In his words:

If you want to know how much … stroke there is in any city, any town, any rural village, do not count the hamburger outlets, the tobacconists, the playgrounds. Ask instead how many mothers died in childbirth several years ago.³⁸

Maternal mortality rates are highest where mothers are poor and ill-fed and where, in consequence, their fetuses are malnourished. And a malnourished fetus is a fetus under metabolic pressure, so it must make a choice: which organs will it protect? Will it allow all its organs to be malnourished equally, or will it protect some organs at the expense of others? It appears a malnourished fetus chooses to protect its brain (these responses are little different from an adult’s: when adult mammals are starved, most of their organs shrink, the exceptions being the brain and, in the case of male mice, the testicles).³⁹

So, on being starved, the fetus will deprive its other organs of nourishment, and in consequence it will grow into a short adult; and to help achieve that smallness, it will induce its muscles and other major organs to become resistant to insulin. The role of insulin is to direct the glucose we absorb from our food into our muscles and other major organs, so when we become insulin-resistant, their glucose uptake will be suppressed, thus sparing it for the brain.⁴⁰ And, unexpectedly perhaps, it transpires that insulin resistance contributes to atherosclerosis.

Malnourished or poor mothers, therefore, produce children who are prone to developing strokes. But well-nourished or rich mothers do not. Hence the slope of the line in Figure 1: as mothers in the West have grown increasingly well-nourished, so their babies have been ever-less prone to developing strokes.

Understanding the Incidence of Heart Attacks in the 20th Century: Summarizing What We Know

From the divergences between the two lines in Figure 1 we can see that the causes of strokes and heart attacks must be different. Although both are caused by atherosclerosis, this condition must affect the arteries of the brain and heart in subtly different ways. All, therefore, we can currently state with certainty is that atherosclerosis is an inflammation of the arteries of which there are many possible causes, including the insulin-resistance of the metabolic syndrome and raised levels of sLDL, but also including smoking, stress, hypertension, diabetes, and aging, as well as a range of inflammatory diseases including arthritis, lupus, chronic infections, and inflammations of unknown cause. Another inflammatory disease that can apparently cause heart attacks is peptic ulceration, which is caused primarily by infection with Helicobacter pylori, because it transpires there is an association between H. pylori infection and atherosclerosis.⁴¹ Importantly, therefore, we still cannot know with certainty what accounted for the epidemiology of heart deaths in the 20th century, which must weaken the confidence with which we can pronounce on any putative causative factor.

Shattering the Cholesterol Story

Fifty years after Keys had captured the Senate Select Committee’s imagination, the failings of the cholesterol hypothesis had become so obvious that in 2007 Gary Taubes, a science journalist, could publish a book Good Calories, Bad Calories, which became a bestseller. In the book he claimed that carbohydrates in general, and sugar in particular, were the dietary hazards; natural fats were healthful.⁴²

As we have seen above, there was already a long scientific tradition led by such professors as Yudkin, Mann, and Ahrens arguing for carbohydrates/sugar, not fats, being the cardiac killers, but Taubes also had a populist predecessor, Robert Atkins (1930-2003), who in 1972 had published Dr. Atkins’ Diet Revolution. Atkins was a New York cardiologist who found that, for slimming, a diet low in carbohydrates and high in fat and protein worked.⁴³

While Yudkin, Mann, and Ahrens had focused on sugar/carbohydrates as the cause of heart attacks, Atkins had focused on sugar/carbohydrates as the cause of obesity. And Taubes continued in Atkins’s wake, noting that the epidemic of obesity (and type 2 diabetes) accelerated when fat in food was being replaced by carbohydrate (see Figure 4). Thus, between 1960 and 2000 the incidence of obesity more than doubled, from 13.4 percent of the population to 30.9 percent, while the incidence of type 2 diabetes rose even more markedly, from 2.6 percent to 6 percent of the population. Since 2000 it has continued to rise, and it reaches nearly 10 percent today.⁴⁴

Yet we should eschew easy myth-making. Thus Jennie Brand-Miller at the University of Sydney, Australia, has shown that in Australia and the United Kingdom, sugar consumption fell between 1980 and 2003 even as obesity rose, which suggests that obesity cannot be attributed to any single nutrient.⁴⁵ Moreover, a recent massive study of the literature (data on 68.5 million people) found that while overweight (body mass index [BMI] over 25) and obesity (BMI over 30) are — as is widely known — associated with cardiovascular and other diseases, the impact may be less important than is widely feared, and the authors of the study noted that “the rate of this increase has been attenuated owing to decreases in underlying rates of death from cardiovascular disease.”⁴⁶ Indeed, life expectancies in the United States have continued to rise (as they have continued to rise among industrialized countries since 1840) by three months for every year lived, or six hours for every day lived, which is truly extraordinary.⁴⁷ Moreover, the so-called “obesity paradox” reveals how, under some circumstances, being overweight can apparently be healthful, and we should remember that as early as 1955 the WHO symposium noted that the Western diet had dual effects in both stimulating and damaging our health.⁴⁸ We are still trying to understand these effects, and it must be premature to confidently dictate our diet.

There is nonetheless currently a scientific consensus on health and diet, which has been summarized by the American Heart Association:

• Dietary saturated fat does increase the rate of cardiovascular disease significantly (Keys got that right), but
• replacing it with refined carbohydrates and sugars does not reduce the rate of cardiovascular disease (Keys got that wrong, whereas his critics, including Yudkin, Mann, and Ahrens, got that right), but
• replacing saturated (animal) fat with unsaturated (vegetable) fat does reduce the rate of cardiovascular disease significantly (as effectively as does treatment with statins) as long as those unsaturated fats are not trans fats.⁴⁹

But this consensus has not yet deeply penetrated the public debate, and while most popular commentators now follow the Taubes anti-carbohydrate/anti-sugar story, the federal government and leading medical authorities still follow the traditional anti-fat story.⁵⁰ This divergence of opinion is both unnecessary and unhelpful.

A Nutritional Note

Reducing food to its constituent chemicals such as carbohydrate or fat is now increasingly criticized as “nutritionism,” because such reductionism may mislead by ignoring the complex and generally unknown interactions between the different chemicals in food.⁵¹ So, for example, the data may mean that meat is dangerous not because of its fat content but because of its protein or haem content (haem being the iron-containing chemical in meat that gives it a red color).⁵² Equally, a Harvard group has identified that plant-based diets that are rich in sugars, starch, or refined carbohydrates may be unhealthful.⁵³

The current American Heart Association advice, therefore, seeks to avoid the nutritionist error by recommending a “Mediterranean” diet (rich in olive oil, vegetables, fruit, nuts, and legumes such as peas, beans, lentils, and chickpeas; moderate in fish, poultry, alcohol, and wholegrain cereals; and low in red meat, processed meat, and sweet foods such as cakes or jams). The similar DASH (Dietary Approaches to Stop Hypertension) diet is also recommended by the association.⁵⁴

A Biochemical Note

Taubes suggested that, calorie for calorie, carbohydrates in the diet may promote obesity more than fat does because they stimulate the secretion of insulin, which in turn stimulates adipocytes (fat cells) to store fat: just as a pubertal girl puts on weight around her hips and at her breasts because of the local actions on fat cells by certain female hormones (not because she’s suddenly started to overeat), so someone who swaps fats for carbohydrates may start to put on weight because of the generalized effect on fat cells of insulin. Further, insulin drives down blood sugar levels, which in turn promotes the secretion of hunger hormones such as ghrelin, which therefore stimulates further eating.⁵⁵

Conclusion

The central question remains: Why, in 1977, did the Senate Select Committee on Nutrition and Human Needs publish its Dietary Goals for the United States when so many credible authorities, including Yudkin, Atkins, Ahrens, Mann, and the American Medical Association, had anticipated, at least in part, today’s understanding of carbohydrates and other saturated fat substitutes as dangerous?

One problem is that scientists are much less scientific than is popularly supposed. John Ioannidis of Stanford University has shown in his 2005 paper “Why Most Published Research Findings Are False” (which has been cited nearly 5,000 times) that the poor application of statistics allows most published research findings to indeed be false, while Brian Nosek of the University of Virginia recently reported — again in large part because of the poor application of statistics — that fewer than half of published studies in psychology can be reproduced.⁵⁶ Although Nosek’s findings have been challenged, it is now commonplace to describe a “crisis of reproducibility” in science, and mainstream scholars now write articles with such titles as “Saving Science.”⁵⁷ Too many research papers, in short, cannot be trusted. Why not?

There is a perverse reason that scientists use poor statistics: career progression. In a paper entitled “The Natural Selection of Bad Science,” Paul Smaldino and Richard McElreath of the University of California, Merced, and the Max Planck Institute, Leipzig, found that scientists select “methods of analysis … to further publication rather than discovery.”⁵⁸ Smaldino and McElreath report how entire scientific disciplines — despite isolated protests from whistleblowers — have, for more than half a century, selected statistical methods precisely because they will yield publishable rather than true results. The popular view is that scientists are falsifiers, but in practice they are generally verifiers, and they will use statistics to extract data that support their hypotheses. Keys, for example, was not a dishonest man, he was merely a typical scientist who had formulated a theory, which — by using poor statistics — he was able over the course of a long career and many publications to appear to verify.

Aggravating the problem of poor science is that research operates a version of public-choice theory: in his 1965 Logic of Collective Action, Mancur Olsen showed how small interest groups can capture public policy, and publicly funded science is no exception. Once armed with power over government-funded research grants, access to peer-reviewed journals, and appointments to university positions, Keys and his fellow elite researchers could enforce their paradigm on the whole field. Which was why the successful paradigm-shifter emerged not as a mainstream scientist but as a journalist — Taubes — who was spared the pressure to conform. And this public-choice aspect of science was aggravated by the Senate Select Committee’s endorsement of the fat paradigm, which fueled it, literally, with public money.

The federal government’s failings were further aggravated by lobbying. When McGovern’s committee reported, the reaction from the meat, egg, and other food lobbyists was so vitriolic that the committee was forced to hold additional public hearings. Following those, a second edition of Dietary Goals for the United States was hurriedly released in late 1977, retracting some of its strongest earlier claims. For example, the committee added this sentence: “science [cannot] at this time insure that an altered diet [will] provide improved protection from killer diseases such as heart disease.”⁵⁹ As it happened, the meat and egg lobbyists were not wrong in all their objections, but they influenced the committee not because of their superior science, but because of their electoral power.

Because of the controversy, the reputation of the Select Committee suffered, and its mandate was allowed to lapse. The issuance of further federal dietary advice was then charged to the Department of Health and Human Services (HHS) and the Department of Agriculture (USDA), which have since jointly published, every five years, their Dietary Guidelines for Americans (the most recent was in 2015), which have, however, only reinforced the anti-fat, pro-carbohydrate message of the original 1977 Dietary Goals. This message was to be popularized in the USDA’s Food Guide Pyramid (1992), MyPyramid (2005), and MyPlate (2011).

The USDA was so empowered specifically to help protect agricultural interests, which inevitably has distorted its advice. Marion Nestle, for example, professor of nutrition at New York University, reports that when she was hired to edit the 1988 Surgeon General’s Report on Nutrition and Health,

My first day on the job, I was given the rules: no matter what the research indicated, the report could not recommend “eat less meat” as a way to reduce the intake of saturated fat, nor could it suggest restrictions on intake of any other category of food.⁶⁰

Indeed, Marion Nestle’s book Food Politics:How the Food Industry Influences Nutrition and Health, was first published in 2002; it is now in its third edition, and it has been cited no fewer than 2,250 times, yet the food industry still influences the federal government’s advice. Thus the New York Times for January 18, 2016, reported that the early drafts of the 2015-2020 Guidelines confirmed that red and processed meats increase the risks of developing bowel and other cancers, but — following the lobbying of Congress by the National Cattlemen’s Beef Association — the reference was removed from the final version.⁶¹

The problem of poor science is aggravated in food because of the vast weight of papers (containing highly selective information and highly selective statistics) that are published in distinguished journals by food companies themselves. A recent survey showed, for example, that reviews of the literature authored by scientists with financial links to the sugar industry were five times less likely to conclude that sugar aggravated obesity or weight gain than did reviews authored by independent scientists.⁶² Where a field of research such as food science can be dominated by producers’ research, therefore, it can be doubly difficult to determine the various sources of bias. And the bias can be secret: we now know, for example, that Mark Hegsted, the Senate committee’s chief scientific adviser, was being secretly paid by the sugar producers to condemn fat and exonerate sugar and carbohydrates.⁶³

Governments may be institutionally incapable of providing disinterested advice for at least four reasons. First, the scientists themselves may be divided, and by choosing one argument over another, the government may be making a mistake. Second, by abusing the precautionary principle, the government may be biasing its advice away from objectivity to risk-avoidance long before all the actual risks have been calculated. Third, because of public pressure, it may offer premature advice. And fourth, its advice will be distorted by lobbying.

Congress has lost patience with official dietary advice, so it commissioned the National Academy of Medicine to review “the entire process used” to generate the official guidelines, although unfortunately the review was limited only to methodology and not results.⁶⁴ But why should the federal government issue health advice at all? The general public, sadly, will give greater credence to federal government pronouncements than the science can bear, and it would surely be healthier to promote a free market in research ideas. There are a large number of scientific institutions qualified to give advice on food, and it would surely be healthier for the public to be exposed directly to their disagreements rather than for the federal government to proselytize an apparent consensus that, in reality, is only partial and selective.

The tragedy is that food science is, because of the power and money of its commercial sponsors, deeply flawed, yet government — by inserting its own biases — has only amplified that science’s faults. Society does need a truly independent, truly high-powered entity to interrogate food science’s output, but that will have to be sought among the ranks of people like Gary Taubes, who have made a career of probing the biases of science.

There is a tradition of politicians involving themselves in science. From William Jennings Bryan’s attack on evolutionary theory in the Scopes Trial, to Al Gore and Donald Trump distorting modern climate science, politicians inevitably politicize science and almost always get it wrong; Senator McGovern’s actions were only one example of the many that have damaged the credibility of both science and politics. More than 200 years ago Thomas Jefferson warned politicians particularly not to engage with dietary science, and we need to reheed his warning.⁶⁵

Notes

¹ Select Committee on Nutrition and Human Needs, United States Senate, Dietary Goals for the United States (Washington: Government Printing Office, 1977), p. 1, https://archive.org/details/CAT10527234.

²Dietary Goals for the United States, p. v.

³Dietary Goals for the United States, p. 12.

⁴ U.S. Department of Health, Education, and Welfare, Vital Statistics of the United States, 1968: Volume II-Mortality, Part A (Rockville, MD: Public Health Service, 1972), pp. 1-6, www.cdc.gov/nchs/data/vsus/mort68_2a.pdf.

⁵ Leon Michaels, “Aetiology of Coronary Artery Disease: An Historical Approach,” British Heart Journal 28, no. 2 (1966): 258-64, doi:10.1136/hrt.28.2.258.

⁶ Ancel Keys, “Atherosclerosis: A Problem in Newer Public Health,” Journal of the Mount Sinai Hospital 2 (July and August 1953): 118-39.

⁷ Nina Teicholz, Big Fat Surprise (London: Scribe Publications, 2014), p. 36.

⁸ Study Group on Atherosclerosis and Ischaemic Heart Disease, Proceedings 117 (Geneva: World Health Organization, 1957).

⁹ Ancel Keys and Joseph T. Anderson, Symposium on Atherosclerosis: Proceedings (Washington: National Academy of Sciences-National Research Council, 1954). Quoted in Jacob Yerushalmy and Herman E. Hilleboe, “Fat in the Diet and Mortality from Heart Disease: A Methodological Note,” New York State Journal of Medicine 57, no. 14 (1957): 2343-54.

¹⁰ John Yudkin, Pure, White and Deadly, 2nd ed. (London: Penguin, 2012), p. 86; and John Yudkin, “Diet and Coronary Thrombosis: Hypothesis and Fact,” Lancet II (1957): 155.

¹¹ Yudkin, Pure, White and Deadly, p. 63.

¹² Edward H. Ahrenset et al., “Dietary Control of Serum Lipids in Relation to Atherosclerosis,” Journal of the American Medical Association 164, no. 17 (August 24, 1957): 1905-11, doi:10.1001/jama.1957.62980170017007d.

¹³ Carmel McCoubrey, “Edward Ahrens Cholesterol Researcher, Is Dead at 85,” New York Times, December 16, 2000, http://www.nytimes.com/2000/12/16/us/edward-ahrens-cholestrol-researcher-is-dead-at-85.html.

¹⁴ George V. Mann, “Diet-Heart: End of an Era,” New England Journal of Medicine 297, no. 12 (October 1977): 644-50, doi:10.1056/nejm197709222971206.

¹⁵ A. E. Bennett et al., “Sugar Consumption and Cigarette Smoking,” Lancet 295, no. 7655 (May 16, 1970): 1011-14, doi:10.1016/s0140-6736(70)91147-5.

¹⁶ Gerald Finking and Hartmut Hanke, “Nikolaj Nikolajewitsch Anitschkow (1885-1964) Established the Cholesterol-Fed Rabbit as a Model for Atherosclerosis Research,” Atherosclerosis 135, no. 1 (1997): 1-7, doi:10.1016/s0021-9150(97)00161-5.

¹⁷ Ancel Keys et al., “Effects of Diet on Blood Lipids in Man, Particularly Cholesterol and Lipoproteins,” Clinical Chemistry 1, no. 1 (1955): 34-52, http://clinchem.aaccjnls.org/content/clinchem/1/1/34.full.pdf.

¹⁸ U.S. Department of Agriculture and Department of Health and Human Services, Scientific Report of the 2015 Dietary Guidelines Advisory Committee (Washington: USDA and HHS, 2015), p. 17, https://health.gov/dietaryguidelines/2015-scientific-report/PDFs/Scientific-Report-of-the-2015-Dietary-Guidelines-Advisory-Committee.pdf.

¹⁹ Gary Taubes, “The Soft Science of Dietary Fat,” Science 291, no. 5513 (March 30, 2001): 2536-45, doi:10.1126/science.291.5513.2536.

²⁰ Ancel Keys, “Epidemiologic Aspects of Coronary Artery Disease,” Journal of Chronic Diseases 6, nos. 4-5 (1957): 552-59.

²¹Dietary Goals for the United States, p. 3.

²² Manuel Hörl, “Fat: The New Health Paradigm,” Credit Suisse, September 22, 2015, https://www.credit-suisse.com/corporate/en/articles/news-and-expertise/fat-the-new-health-paradigm-201509.html.

²³ Study Group on Atherosclerosis and Ischaemic Heart Disease, Proceedings 117.

²⁴ Ancel Keys, ed., Seven Countries: A Multivariate Analysis of Death and Coronary Heart Disease (Cambridge, MA: Harvard University Press, 1980). Quoted in Teicholz, Big Fat Surprise, p. 38.

²⁵ Study Group on Atherosclerosis and Ischaemic Heart Disease, Proceedings 117.

²⁶ Alessandro Menotti et al., “Food Intake Patterns and 25-Year Mortality from Coronary Heart Disease: Cross-Cultural Correlations in the Seven Countries Study,” European Journal of Epidemiology 15, no. 6 (July 1999): 507-15.

²⁷ Zoë Harcombe et al., “Evidence from Randomised Controlled Trials Did Not Support the Introduction of Dietary Fat Guidelines in 1977 and 1983: A Systematic Review and Meta-Analysis,” Open Heart 2, no. 1 (2015): 1-6, doi:10.1136/openhrt-2014-000196.

²⁸“Dietary Goals for the United States: Statement of the American Medical Association to the Select Committee on Nutrition and Human Needs, United States Senate,” Rhode Island Medical Journal 60, no. 12 (1977): 576-81.

²⁹Dietary Goals for the United States, p. 10.

³⁰ Quoted in Harcombe et al., “Evidence from Randomised Controlled Trials,” pp. 1-6.

³¹ For an account of the mass media’s support for the government’s message, see Teicholz, Big Fat Surprise, pp. 47-53.

³² Beth Schucker, “Change in Public Perspective on Cholesterol and Heart Disease,” Journal of the American Medical Association 258, no. 24 (December 1987): 3527, doi:10.1001/jama.1987.03400240059024.

³³ Shi-Sheng Zhou et al., “B-Vitamin Consumption and the Prevalence of Diabetes and Obesity among US Adults: Population Based Ecological Study,” BMC Public Health 10, no. 1 (December 2, 2010), doi:10.1186/1471-2458-10-746.

³⁴ Data from the USDA are reproduced in Roberto A. Ferdman, “The Generational Battle of Butter vs. Margarine,” Washington Post, June 17, 2014, https://www.washingtonpost.com/news/wonk/wp/2014/06/17/the-generational-battle-of-butter-vs-margarine/?utm_term=.b9b8f2a12ea5.

³⁵ D. Mozaffarian, A. Aro, and W. C. Willett, “Health Effects of Trans-Fatty Acids: Experimental and Observational Evidence,” European Journal of Clinical Nutrition 63, no. S2 (May 2009): S5-S21, doi:10.1038/sj.ejcn.1602973.

³⁶ World Health Organization, “The Top 10 Causes of Death,” http://www.who.int/mediacentre/factsheets/fs310/en/index1.html.

³⁷ C. N. Hales and D. Barker, “Type 2 (non-insulin-dependent) Diabetes Mellitus: The Thrifty Phenotype Hypothesis,” International Journal of Epidemiology 42, no. 5 (October 2013): 1215-22, doi:10.1093/ije/dyt133.

³⁸ Quoted in Caroline Fall and Clive Osmond, “David Barker,” Sight and Life 27 (2013): 64-66.

³⁹ Richard Weindruch and Rajindar S. Sohal, “Caloric Intake and Aging,” New England Journal of Medicine 337, no. 14 (October 2, 1997): 986-94, doi:10.1056/nejm199710023371407.

⁴⁰ P. L. Hofman, “Insulin Resistance in Short Children with Intrauterine Growth Retardation,” Journal of Clinical Endocrinology & Metabolism 82, no. 2 (February 1997): 402-06, doi:10.1210/jc.82.2.402.

⁴¹ John N. Duggan and Anne E. Duggan, “The Possible Causes of the Pandemic of Peptic Ulcer in the Late 19th and Early 20th Century,” Medical Journal of Australia 185, no. 11/12 (December 2006): 667-69; and John R. Brooks and Angelo J. Eraklis, “Factors Affecting the Mortality from Peptic Ulcer,” New England Journal of Medicine 271, no. 16 (October 15, 1964): 803-9, doi:10.1056/nejm196410152711601.

⁴² Gary Taubes, Good Calories, Bad Calories: Challenging the Conventional Wisdom on Diet, Weight Control, and Disease (New York: Alfred A. Knopf, 2007).

⁴³ Edgar S. Gordon et al., “A New Concept in the Treatment of Obesity,” Journal of the American Medical Association 186, no. 1 (October 5, 1963): 50-60, doi:10.1001/jama.1963.63710010013014.

⁴⁴ Robert J. Kuczmarski, “Increasing Prevalence of Overweight among US Adults,” Journal of the American Medical Association 272, no. 3 (July 20, 1994): 205, doi:10.1001/jama.1994.03520030047027; Katherine M. Flegal, “Prevalence and Trends in Obesity among US Adults, 1999-2000,” Journal of the American Medical Association 288, no. 14 (October 9, 2002): 1723, doi:10.1001/jama.288.14.1723; and Centers for Disease Control and Prevention, National Diabetes Statistics Report, 2017 (Atlanta: Centers for Disease Control and Prevention, 2017).

⁴⁵ Alan W. Barclay and Jennie Brand-Miller, “The Australian Paradox: A Substantial Decline in Sugar Intake over the Same Timeframe that Overweight and Obesity Have Increased,” Nutrition 3, no. 4 (April 3, 2011): 491-504.

⁴⁶ GBD 2015 Obesity Collaborators, “Health Effects of Overweight and Obesity in 195 Countries over 25 Years,” New England Journal of Medicine 377, no. 1 (July 6, 2017): 13-27, doi:10.1056/nejmoa1614362.

⁴⁷ National Institute on Aging, Living Longer 2015, https://www.nia.nih.gov/research/publication/global-health-and-aging/living-longer.

⁴⁸ Katherine M. Flegal et al., “Association of All-Cause Mortality with Overweight and Obesity Using Standard Body Mass Index Categories,” Journal of the American Medical Association 309, no. 1 (January 2, 2013): 71-82, doi:10.1001/jama.2012.113905.

⁴⁹ Frank M. Sacks et al., “Dietary Fats and Cardiovascular Disease: A Presidential Advisory from the American Heart Association,” Circulation 136, no. 3 (June 15, 2017), doi:10.1161/cir.0000000000000510.

⁵⁰ U.S. Department of Health and Human Services and U.S. Department of Agriculture, 2015-2020 Dietary Guidelines for Americans, 8th ed. (Washington: HHS, 2015), https://health.gov/dietaryguidelines/2015/guidelines/.

⁵¹ Michael Pollan, In Defense of Food: An Eater’s Manifesto (New York: Penguin Books, 2008).

⁵² Thomas C. Campbell, “A Plant-Based Diet and Animal Protein: Questioning Dietary Fat and Considering Animal Protein as the Main Cause of Heart Disease,” Journal of Geriatric Cardiology 14, no. 5 (May 2017): 331-37, doi:10.11909/j.issn.1671-5411.2017.05.011.

⁵³ Ambika Satija et al., “Healthful and Unhealthful Plant-Based Diets and the Risk of Coronary Heart Disease in U.S. Adults,” Journal of the American College of Cardiology 70, no. 4 (July 25, 2017): 411-22, doi:10.1016/j.jacc.2017.05.047.

⁵⁴ Mayo Clinic, “DASH Diet: Healthy Eating to Lower Your Blood Pressure,” April 8, 2016, http://www.mayoclinic.org/healthy-lifestyle/nutrition-and-healthy-eating/in-depth/dash-diet/art-20048456?pg=1.

⁵⁵ Tomomi Shiiya et al., “Plasma Ghrelin Levels in Lean and Obese Humans and the Effect of Glucose on Ghrelin Secretion,” Journal of Clinical Endocrinology & Metabolism 87, no. 1 (January 2002): 240-44, doi:10.1210/jcem.87.1.8129.

⁵⁶ John P. A. Ioannidis, “Why Most Published Research Findings Are False,” PLoS Medicine 2, no. 8 (August 30, 2005), doi:10.1371/journal.pmed.0020124; and Open Science Collaboration, “Estimating the Reproducibility of Psychological Science,” Science 349, no. 6251 (August 28, 2015), doi:10.1126/science.aac4716.

⁵⁷ Daniel T. Gilbert et al., “Comment on ‘Estimating the Reproducibility of Psychological Science,’” Science 351 (March 4, 2016): 1037; and Daniel Sarewitz, “Saving Science,” The New Atlantis (Spring/Summer 2016): 4-40, https://www.thenewatlantis.com/publications/saving-science.

⁵⁸ Paul E. Smaldino and Richard McElreath, “The Natural Selection of Bad Science,” Royal Society Open Science 3, no. 9 (September 21, 2016), doi:10.1098/rsos.160384.

⁵⁹ Quoted in Gerald M. Oppenheimer and I. Daniel Benrubi, “McGovern’s Senate Select Committee on Nutrition and Human Needs Versus the Meat Industry on the Diet-Heart Question (1976-1977),” American Journal of Public Health 104, no. 1 (January 2014): 59-69, doi:10.2105/ajph.2013.301464.

⁶⁰ Marion Nestle, Food Politics: How the Food Industry Influences Nutrition and Health (Berkeley: University of California Press, 2013), p. 3.

⁶¹ Jane E. Brody, “What’s New in the Dietary Guidelines,” New York Times, January 18, 2016, well.blog.nytimes.com/2016/0/18what’s-new-in-the-dietary-guidelines/r_r=0.

⁶² Maira Bes-Rastrollo et al., “Financial Conflicts of Interest and Reporting Bias Regarding the Association between Sugar-Sweetened Beverages and Weight Gain: A Systematic Review of Systematic Reviews,” PLoS Medicine 10, no. 12 (December 31, 2013): 1-9, doi:10.1371/journal.pmed.1001578.

⁶³ Cristin E. Kearns, Laura A. Schmidt, and Stanton A. Glantz, “Sugar Industry and Coronary Heart Disease Research,” JAMA Internal Medicine 176, no. 11 (November 1, 2016): 1680-85, doi:10.1001/jamainternmed.2016.5394.

⁶⁴ Peter Whoriskey, “Congress: We Need to Review the Dietary Guidelines for Americans,” Washington Post, December 18, 2015.

⁶⁵ Thomas Jefferson, Notes on the State of Virginia, ed. Frank Shuffelton (New York: Penguin, 1999), p. 165.

Jeffrey Miron

In the past several years, the national movement to end drug prohibition has accelerated. Nine states and Washington, DC, have legalized recreational marijuana, with at least three more states (Connecticut, Michigan, and Ohio) likely to vote on legalization by the end of 2018. Dozens of others have decriminalized the substance or permitted it for medicinal use. Moreover, amid the nation’s ongoing opioid crisis, some advocates and politicians are calling to decriminalize drugs more broadly and rethink our approach to drug enforcement.

Drug legalization affects various social outcomes. In the debate over marijuana legalization, academics and the media tend to focus on how legalization affects public health and criminal justice outcomes. But policymakers and scholars should also consider the fiscal effects of drug liberalization. Legalization can reduce government spending, which saves resources for other uses, and it generates tax revenue that transfers income from drug producers and consumers to public coffers.

Drawing on the most recent available data, this bulletin estimates the fiscal windfall that would be achieved through drug legalization. All told, drug legalization could generate up to $106.7 billion in annual budgetary gains for federal, state, and local governments. Those gains would come from two primary sources: decreases in drug enforcement spending and increases in tax revenue. This bulletin estimates that state and local governments spend $29 billion on drug prohibition annually, while the federal government spends an additional $18 billion. Meanwhile, full drug legalization would yield $19 billion in state and local tax revenue and $39 billion in federal tax revenue.

In addition, this bulletin briefly examines the budgetary effects of state marijuana legalizations that have already taken place in Colorado, Oregon, and Washington. This study finds that, so far, legalization in those states has generated more tax revenue than previously forecast but generated essentially no reductions in criminal justice expenditure. The bulletin offers possible explanations for those findings.

Estimates for Expenditures and Tax Revenue

Jeffrey Miron and Katherine Waldock previously estimated government anti-drug expenditures and potential tax revenue from legalization in a 2010 Cato Institute study.¹ That study examined the reduction in criminal justice spending ($41.3 billion per year) and the increase in tax revenue ($46.7 billion per year) that would result from legalizing all drugs in the United States at the federal level and in all states.

This study draws on more recent data and presents exciting new conclusions. These updates are useful because a number of states have legalized marijuana since the 2010 report; this provides some evidence on the validity of the 2010 estimates.² Moreover, additional states are considering marijuana legalization going forward, and updated estimates might be relevant to the legalization debates in those states.

This report uses the same analytic framework as the 2010 Cato study and therefore omits some details.³ The underlying data in this update are the latest numbers available and are sourced from 2015 or 2016 unless otherwise noted. Tables 1-6 in this report update the key information from Tables 2-7 in the 2010 report.

Table 1 shows estimated state and local criminal justice expenditures related to drug prohibition in 2016.⁴ In nearly all categories, estimated expenditure has risen, albeit slightly. For all drugs, the estimate is $29.4 billion; for marijuana, $6.0 billion; for heroin and cocaine, $12.8 billion; for synthetic drugs, $4.9 billion; and for all other drugs, $5.6 billion. Adjusted for inflation, these figures represent only about a 3 percent increase since 2010.

Table 2 provides a state-by-state breakdown of state and local expenditure on drug prohibition in 2016.

Table 3 presents estimated federal spending on drug prohibition enforcement for 2015 (in 2016 dollars). Real federal spending has risen by 4 percent since 2008. This report attributes that change to growth in the U.S. population.

Table 4 presents updated estimates of the tax revenue that federal, state, and local governments could collect if drugs were legal. Compared with the original study, these figures suggest that overall federal and state tax revenue would be higher than previously estimated after accounting for inflation; the federal government would take in $39.2 billion in federal tax revenues as a result of legalization today, compared with $31.2 billion in 2008. In real terms, this represents growth of about 12 percent. States would collect $19.6 billion today, compared with $15.6 billion in 2008. Beyond the effects of population growth and inflation, this upward trend reflects increasing use of marijuana, cocaine, and heroin-and therefore increasing consumer spending and potential tax revenue associated with those substances.

Table 1: State and local expenditures attributable to drug prohibition, billions of dollars, 2016

Source: Author’s calculations

Table 2: State and local expenditures attributable to drug prohibition, millions of dollars, 2016

Source: Author’s calculations.

Table 3: Federal expenditures attributable to drug prohibition, billions of dollars, 2015 (in 2016 dollars)

Source: Author’s calculations.

Table 4: State and federal tax revenues from drug legalization, billions of dollars, 2016

Source: Author’s calculations.

Table 5 estimates the tax revenue generated in each state by allocating the estimates from Table 4 to each state on the basis of population.

Table 6 summarizes the updated estimates for expenditure savings and additional revenues tied to drug legalization. Three aspects stand out. First, the total effect of drug legalization on government budgets would be approximately $106.7 billion in combined savings and additional revenue. In real terms, that marks an 8 percent increase from the estimates in the 2010 Cato study. Second, as in the previous report, nearly 60 percent of budgetary gains would come from legalizing heroin and cocaine. Third, the fiscal benefits of drug legalization would be roughly evenly shared between the states and the federal government.

Table 5: State tax revenues from drug legalization, distributed by population, millions of dollars, 2016

Source: Author’s calculations.

Table 6: Summary of expenditure savings and additional revenues from drug legalization, billions of dollars, 2016

Source: Author’s calculations.

Tax Revenue Comparisons

This section compares estimates from the 2010 Cato study with observed tax revenues following state-level marijuana legalization in a handful of U.S. states. Comparing those predictions to actual outcomes sheds light on the accuracy of the 2010 study.

Table 7 presents marijuana-related tax revenue from Colorado, Oregon, and Washington, the three states that had legalized and implemented recreational marijuana commerce at the time of writing. Although Colorado and Washington officially legalized marijuana in 2012 and Oregon followed suit in 2014, marijuana sales did not commence until regulatory frameworks were established. In each state, that process took several years. For example, Colorado did not begin issuing licenses to sell retail marijuana until the end of 2014. Table 7 also lists the tax revenue projections from the original report.

Washington collected nearly $70 million in marijuana tax revenues during the first year of legalization, almost exactly the estimate in the 2010 report once adjusted for inflation. In fiscal year 2016, however, Washington collected nearly triple that amount, and in fiscal year 2017 tax revenues reached nearly $320 million. Oregon collected only $20.6 million in fiscal year 2016, about half the 2010 estimate, but it collected $70.3 million in fiscal year 2017, well above the 2010 estimate. In Colorado, marijuana tax revenues have risen from $67.6 million in calendar year 2014 to $247.4 million in calendar year 2017. Even adjusting for inflation, those figures far outstrip the 2010 estimates as well as the updated estimates presented in this paper.

The discrepancy between the 2010 estimates and experience so far reflects, in part, an unexpectedly high amount of marijuana tourism in these three states; initial reports suggest that out-of-state marijuana shoppers account for a significant fraction of tax revenue. For example, a 2015 survey of adult tourists in Colorado found that 23 percent identified legal marijuana as a reason they traveled to the state.⁵

Alternatively, the discrepancy between the 2010 estimates and experience so far may indicate that some of the assumptions behind the original estimates were incorrect. For example, the 2010 report assumed that marijuana prices would fall by 50 percent in states that legalized; however, pricing data analyzed in a 2016 Cato Institute study by Angela Dills, Sietse Goffard, and Jeffrey Miron suggest that marijuana prices have not dropped that much.⁶ If true, this assertion would help explain the higher-than-expected tax revenue.

The implications of these initial data are therefore unclear. One caveat is that tax revenues are still in flux because of the recent nature of existing legalizations. Revenues may continue to increase over time as more stores open or if demand increases as a result of greater cultural acceptance of marijuana. This is a plausible explanation for the observed growth in tax revenue; Colorado, Oregon, and Washington have all seen gradually rising levels of marijuana use, according to survey data. But revenues in existing legalization states may also moderate if other states or the federal government legalize marijuana.⁷ Another consideration is that a nontrivial share of tax revenue in Colorado, Oregon, and Washington has been generated from collection of one-time application and licensing fees. To date, Colorado has collected $57.3 million in marijuana licensing and application fees, or about 8 percent of total state marijuana revenues since legalization.⁸ As recreational marijuana becomes a more established industry, states will likely see a decline in the number of new entrants and therefore a decline in licensing revenue.

Table 7: State tax revenue from marijuana, 2016 dollars

Sources: Colorado Department of Revenue; Washington State Liquor and Cannabis Board; Oregon Department of Revenue.

Further, if marijuana were legalized at the federal level, it would likely be taxed at both the state and federal level, similarly to how cigarettes are currently taxed. The addition of a federal tax would increase the price and drive down demand. States would then see less revenue as users reduced use in response to the price change. As California prepares to tax recreational marijuana sales statewide with a tax rate of 45 percent, we will be able, in the future, to observe consumer responses to widely varying tax rates.

Criminal Justice Expenditures in States with Marijuana Legalization

This section examines the effect of marijuana legalization on state-level criminal justice expenditures. Unlike tax revenues, no direct indicator shows how legalization affects state spending. Nevertheless, no evidence to date suggests that legalization generates a sharp decline in police, judicial, or correctional expenditures.

Table 8 shows marijuana arrests in Oregon and Washington and the percentage of total drug arrests and total statewide arrests attributed to marijuana offenses. Colorado does not publish comparable data for marijuana-specific offenses, so the table reports total arrests for any drug-related violation. The shaded box indicates the year in which marijuana legalization measures were passed.

Despite the sharp decline in marijuana arrests, criminal justice expenditures in Colorado, Oregon, and Washington have risen slightly, as shown in Table 9. One possible explanation is that marijuana offenses accounted for a small share of arrests and prosecutions even before legalization. In Table 2, the criminal justice expenditure attributable to marijuana represented only 15 percent of total expenditure in these three states. Another possible explanation is that states are shifting resources toward other types of drug and nondrug crimes.

Table 8: Marijuana arrests

Sources: Oregon Criminal Justice Information Services, https://www.oregon.gov/osp/CJIS/Pages/annual_reports.aspx; Washington Association of Sheriffs and Police Chiefs, http://www.waspc.org/crime-statistics-reports; Federal Bureau of Investigations Uniform Crime Reports.

Table 9: State and local expenditures by function (thousands of dollars)

Source: United States Census Bureau, http://www.census.gov/govs/local.

Conclusion

At both the federal and state levels, government budgets would benefit enormously from drug legalization policies. This report estimates that $47.9 billion is spent annually on drug prohibition enforcement, whereas $58.8 billion could potentially be raised in tax revenue. Combined, these figures suggest that the fiscal windfall of drug legalization could be as high as $107 billion. Moreover, although media outlets and policymakers mostly focus on marijuana, the majority of budgetary gains would likely come from legalizing heroin and cocaine. In addition to providing new estimates of the revenue and expenditure implications of drug legalization, this report suggests two conclusions based on experience to date from state-level legalizations.

First, the tax revenues generated by legalization might significantly exceed the estimates in the 2010 Cato study. At the same time, several factors might moderate the tax revenues if legalization occurs widely at the state level or at the federal level.

Second, the reductions in criminal justice expenditures from legalization are likely to be modest in practice, even if the number of drug arrests falls substantially. Early experience suggests that governments will reallocate rather than reduce those expenditures. That reallocation may be beneficial, but it does not have a direct effect on the budget deficit.

Notes

¹ Jeffrey A. Miron and Katherine Waldock, “The Budgetary Impact of Ending Drug Prohibition,” Cato Institute white paper, September 27, 2010.

² Since the 2010 report, 30 states and Washington, DC, have decriminalized, medicalized, or fully legalized marijuana. Colorado (2012); Washington (2012); Alaska (2014); Oregon (2014); Washington, DC (2014); California (2016); Maine (2016); Massachusetts (2016); Nevada (2016); and Vermont (2018) have legalized recreational marijuana consumption. Connecticut (2011), Michigan (2012), Maryland (2014), Missouri (2014), Delaware (2015), Illinois (2016), and New Hampshire (2017) have decriminalized marijuana use. Arizona (2010), New Jersey (2010), Minnesota (2014), New York (2014), Utah (2014), Georgia (2015), Louisiana (2015), North Carolina (2015), Texas (2015), Arkansas (2016), Florida (2016), North Dakota (2016), Ohio (2016), and Pennsylvania (2016) have enacted medical marijuana laws.

³ An appendix available on request contains details of all the calculations.

⁴ This report ignores the difference made in the 2010 report between gross and net expenditures, which equals the value of asset seizures and fines collected in the process of prohibition enforcement. As noted in the 2010 report, those amounts are too small to have a meaningful effect on the overall estimates.

⁵ Jason Blevins, “Only 4% of Colorado Tourists Came for the Legal Weed in 2015, Survey Says,” Denver Post, July 20, 2016.

⁶ Angela Dills, Sietse Goffard, and Jeffrey Miron, “Dose of Reality: The Effect of State Marijuana Legalizations,” Cato Institute Policy Analysis no. 799, September 16, 2016.

⁷ Federal taxation would make it harder for states to impose significant taxation without recreating the black market. So even though federal legalization would generate federal revenue, it might imply lower state revenue in some cases.

⁸ Colorado Department of Revenue, “Marijuana Tax Data,” https://www.colorado.gov/pacific/revenue/colorado-marijuana-tax-data.

Ryan Bourne

Federal, state, and local governments seek to assist poor households financially using transfers, minimum wage laws, and subsidies for important goods and services. This “income-based” approach to alleviating poverty aims both to raise household incomes directly and to shift the cost of items, such as food, housing, or health care, to taxpayers. Most contemporary ideas to help the poor sit firmly within this paradigm.

A “cost-based” approach would instead reform existing government interventions that raise living costs for the poor. Shelter, food, transport, and apparel and footwear alone account for 59 percent of spending by the average household in the bottom 20 percent of the income distribution, and government policies raise prices in all those sectors. Local land-use and zoning regulations constrain housing supply, which raises housing costs and deters labor mobility. State child-care staffing regulations reduce the number of infant centers in poor areas, increasing prices and reducing the payoff to work. The federal sugar program, milk-marketing orders, and ethanol mandates make grocery shopping more expensive. Federal fuel-standard regulations and state-level automobile dealership laws increase the cost of driving. Protectionist tariffs raise clothing and footwear prices, and state occupational licensing creates barriers to entry that raise the price of many services, from hair braiding to dentistry, while reducing labor-market opportunities.

Using cautious assumptions, I estimate that these interventions, combined, cost typical low-income households between $830 and $3,500 per year directly through higher prices. Pro-market reforms in these areas could significantly reduce living costs for the poor, while also improving labor mobility and job matching. With the federal budget deficit growing and demands for radical labor-market policies proliferating, such an agenda would represent an economically efficient means of improving the well-being of the poor without requiring more government spending or intervention.

Introduction

American government at the federal, state, and local levels delivers policies intended to help households on low incomes. Total annual expenditure on U.S. anti-poverty programs is estimated to exceed $1 trillion per year.¹ Governments redistribute income, provide benefits-in-kind, and subsidize the provision of certain services on the basis of need. They also pass mandates and regulations, such as minimum wage laws or limits on drug prices.

Though liberals and conservatives have different theories about the causes of poverty, the dominant paradigm for alleviating it rests on “income-based” approaches.² Policies attempt to raise the incomes of the poor directly through cash transfers, tax breaks, and minimum wage laws or to raise the poor’s disposable income indirectly by shifting expenditure on goods and services to taxpayers through programs such as Medicaid.

This income-based approach underpins contemporary policy ideas. Senate Democrats advocate a $15 per hour federal minimum wage.³ Their 2016 presidential candidates, Hillary Clinton and Sen. Bernie Sanders (I-VT), proposed new universal preschool and child-care programs.⁴ More recently, Senator Sanders backed a federal jobs guarantee designed to ensure a labor market wage floor.⁵

Universal basic income and negative income taxes, regularly touted as more efficient and freedom-enhancing means of income redistribution, nevertheless remain firmly in the income-based school of poverty alleviation.⁶“Reform conservatives” have likewise long advocated for increasing the generosity of the earned income and child tax credits.⁷ Even conservatives who want to see less generous redistributive programs agree that income growth is important to reduce poverty.⁸ They want to improve incentives and broaden economic growth so that the poor can earn their own way out of poverty.

Income is important to well-being. But focusing on earnings and transfers overlooks another way to help the less fortunate: reforming existing government policies that raise the prices of basic goods and services and thereby hurt the poor through higher living costs.

In markets where low-income households spend significant amounts — on housing, childcare, food, transport, clothing, and services regulated through occupational licensing — interventions designed to achieve other objectives restrict supply and in turn raise prices. Since these goods are relative necessities, these interventions impose disproportionate burdens on the poor. They are left with less disposable income, heightening calls for further taxpayer-funded redistribution or government interventions to counteract the effects of the policy.

This paper sets out nine policy areas across all levels of government that, combined, directly raise spending for typical households in the bottom 20 percent of the income distribution by anywhere from $830 to $3,500 per year. This list is hardly comprehensive; to avoid subjective judgments about the effect on prices relative to other objectives, this analysis focuses exclusively on anti-competitive interventions and regulations that both raise prices and reduce overall economic efficiency.⁹ A “cost-based” approach to poverty alleviation through reform in these areas could therefore provide a significant financial boost to low-income households.

For too long, scholars on the left and right have thought about alleviating poverty as something that should occur after market-based activity has taken place. But removing misguided regulatory interventions would reduce poverty while expanding markets, simultaneously reducing the cost of living for low-income families and growing the economy. Even on cautious assumptions, the indicative numbers outlined here suggest that reform in these areas could be a powerful tool against poverty and should take precedence over new programs, regulations, and interventions.

Why a pro-market agenda for those on low incomes?

The dominant “income-based” approach to helping the poor can directly alleviate financial hardship. Accounting for federal cash benefits, tax credits, and benefits-in-kind, the Center on Budget and Policy Priorities estimates that the U.S. poverty rate fell from 18.9 to 10.9 percent between 1964 and 2011, as redistributive spending increased substantially.¹⁰ Bruce Meyer and Derek Wu recently concluded that five of the six programs they examined — Social Security, Supplemental Security Income, Temporary Assistance for Needy Families, housing assistance, and food stamps — help reduce measured poverty substantially.¹¹

Of course, these types of analyses fail to model a counterfactual world in which extensive government redistribution does not exist. With lower tax burdens, civil society institutions and charities would surely offer more generous support for those in need. Without extensive welfare and entitlement programs, worker and household behavior in the long run would be very different. The real net effect of government redistribution on the financial position of poor households is uncertain and theoretically ambiguous.

But it would be unsurprising if government transfers and benefits-in-kind raised disposable incomes for some recipients above what they could obtain from market-based activity and civil society assistance, particularly in the short run. Minimum wage hikes likewise raise incomes for workers from poor households with low pay rates who are lucky enough to keep their jobs and hours (though minimum wages are not a well-targeted poverty reduction tool generally).¹²

Yet, even accepting that the “income-based” approach raises income levels for many today does not mean that further expanding this approach is the best way to help the poor going forward. Consider the following:

• Diminishing returns. Economists Bruce Meyer and James Sullivan have estimated that less than one-third of the reduction in the after-tax income poverty rate seen between 1960 and 2010 took place after 1972, with no progress at all after 2000, despite massive spending increases.¹³ This is consistent with redistribution exhibiting diminishing returns as a poverty- reduction tool.
• The fiscal environment. The federal deficit is projected to rise to 5.1 percent of GDP by 2022.¹⁴ This adds to an unsustainable long-term federal debt outlook, driven primarily by projected increases in Social Security and Medicare spending as the population ages.¹⁵ Increasing spending to further reduce poverty would, absent tax increases, worsen the structural deficit and make an unsustainable fiscal outlook worse.
• Negative consequences of more redistribution. Substantial additional redistribution would eventually require raising taxes. This would depress the level of GDP by raising marginal tax rates, at a time when future potential economic growth rates are already expected to be low.¹⁶ The means-tested nature of redistributive transfers means that increasing their generosity also results in either steeper withdrawal rates for recipients or more people being drawn into the system. The higher effective marginal tax rates both these outcomes generate would further erode work incentives.
• Negative consequences of minimum wage hikes. Set conservatively, minimum wages have modest effects on overall employment, with the burden falling heavily on those with low levels of labor market attachment (such as teenagers).¹⁷ Recent evidence from Seattle suggests much larger disemployment effects occur when minimum wages are increased from an already high level.¹⁸ Huge minimum wage hikes would therefore bring significant risks at a time when the labor market is looking increasingly healthy, with the unemployment rate at just 4 percent.¹⁹
• Redistribution is vulnerable to changing sentiment. Attitudes to welfare can be volatile, and preferences for redistribution replaced by narratives about “moochers” and “welfare queens.” The experience of other countries suggests politicians find it easier to cut working-age welfare expenditure in times of fiscal crisis than other major spending categories.²⁰

You do not have to believe existing anti-poverty programs have failed in order to acknowledge these unintended consequences, diminishing returns, and need for taxpayer goodwill.

A “cost-based” agenda focused on removing damaging government interventions, in contrast, would not require additional government spending. By making essential goods cheaper, such a policy may reduce spending levels by lowering the political demands for redistributive transfers. If delivered through reforms to policies that currently undermine economic efficiency, it would also raise GDP and market-obtained incomes without the risks of unemployment from minimum wage hikes or the need for higher marginal tax rates. A beneficial side effect might also be restored faith in the market economy to deliver affordable goods and services, resulting in a political environment more conducive to pro-growth reforms in other sectors.

None of this means a pro-market cost-of-living agenda would be easy to deliver. Powerful supporters of existing interventions will resist such change. Zoning and land-use planning reforms often run counter to the interests of existing homeowners and will be opposed by coalitions of NIMBYs.²¹ Professionals with occupational licenses will argue that licensing improves service quality. Industries that benefit from extensive government protection, such as dairy and sugar farmers, textiles producers, and automobile dealerships, will petition state and federal politicians to protect their own interests. The bureaucracies that implement these programs and regulations also have a vested interest in ensuring their continuation.

Yet, these interventions currently come at a high cost to the poor. A pro-market “cost-based” reform agenda to reduce prices of essential goods and services should be considered an important tool in an effective and enduring “first do no harm” approach to reducing poverty.

Where might a pro-market agenda have a big effect?

The Bureau of Labor Statistics Consumer Expenditure Survey shows the average amount spent by households across the income distribution on categories of goods and services. Table 1 shows households in the bottom 20 percent of the income distribution tend to spend a much higher proportion than the rest of the population on “essential” goods and services. Shelter, food, transport, and apparel together account for 59 percent of the $25,318 spent by the average household in the poorest income quintile, compared with 50.9 percent for the average household across the whole population and 46.5 percent for the average household in the richest quintile.

Table 1: Expenditure by category

Source: Data from Bureau of Labor Statistics, Consumer Expenditure Survey, https://www.bls.gov/cex/tables.htm#avgexp.

This masks substantial differences by household composition and region. The average single-parent family spends proportionately more on apparel than do two-parent families. Households in some major U.S. cities spend much more on shelter. In San Francisco, even the average household apportions as much as 28.7 percent of spending to shelter, and similarly high figures are seen in New York (26.5 percent), Boston (25.2 percent), Los Angeles (24.2 percent), and Miami (24.0 percent).²² Families with young children where both parents are employed face very costly child-care bills too. In Washington, D.C., Child Care Aware estimates an average annual cost of formal infant care of $23,089.²³

Without data disaggregated by region, household composition, and income level, one cannot reach firm conclusions about the financial costs of existing policies to individual families. Nevertheless, this high-level analysis shows markets where a meaningful anti-poverty agenda will have the biggest effect. Housing and child-care costs are likely to be particularly significant for those households in major metropolitan cities or with young children.

The remaining analysis highlights current government policies that drive up the cost of housing, childcare, food, transport, apparel and footwear, and services with occupational licensing requirements, and estimates their likely cost to poorer households.²⁴

Shelter

The single largest expenditure for most families is shelter (rent or the cost of owner-occupied housing). It makes up 25.2 percent of total spending for the average household in the poorest quintile, and 21.8 percent for the average single-parent household. Since the poorest quintile includes many older and poorer households with low incomes, spending as a proportion of income is higher still. The Pew Foundation estimates households in the bottom third of the income distribution spent 40 percent of their income on housing in 2014, while renters spent nearly half.²⁵

The United States has relatively cheaper housing overall than other major developed English-speaking countries. But prices and rents are extraordinarily high in certain metropolitan areas. Demographia’s median multiple index (median house price divided by median income) is over 9 in Los Angeles and San Francisco, and just below 6 for Seattle and New York (see Table 2).²⁶ Thirty overall housing markets and 13 major metropolitan markets are defined as “severely unaffordable,” meaning they have median multiples of 5.1 or over. But even these are quite broad markets, including suburban areas on the outskirts of cities. The online housing marketplace Zumper estimates that the median one-bedroom rental price in March 2018 was $3,400 per month in San Francisco; $2,900 in New York; $2,450 in San Jose; $2,300 in Boston; and $2,220 in Washington, D.C.²⁷

Table 2: Severely unaffordable housing markets

Source: Demographia.com, 14th Annual Demographia International Housing Affordability Survey, 2018, January 22, 2018, http://demographia.com/dhi.pdf
Note: The median multiple is the result of dividing median house price by median household income.

High housing costs have major consequences for the poor, both in direct financial terms and, indirectly, in terms of labor mobility and job match. They encourage families to live in smaller apartments and condominiums, to commute greater distances to jobs, and can even act as a prohibitive financial barrier to taking up employment opportunities in certain cities.

Regulatory restraints at the local-government level have a significant effect on housing affordability. Land-use planning and zoning laws — including urban growth boundaries, minimum lot sizes, density and height restrictions, and design requirements — raise the costs associated with providing new housing, restricting the potential supply and making it less responsive to changes in demand. The result of the latter is structurally higher prices as incomes rise and the population grows.

Because of the vast, complex, and differentiated nature of regulations across the country, it is difficult to measure and compare the permissiveness toward development across regions, but economists have used two techniques to measure the effects of regulations.

Some estimate an implied “regulatory tax” as the deviation between new house prices and marginal building costs. Using this method, Ed Glaeser, Joseph Gyourko, and Raven Saks estimated that Manhattan condominium prices were 50 percent higher in the early 2000s than under a free development regime.²⁸ For single-family homes across the country, their estimates show regulatory costs much higher in some areas than others — being indistinguishable from zero in cities such as Baltimore and Houston, but as high as 53 percent in the San Francisco Bay Area, 34 percent in Los Angeles, 22 percent in Washington, D.C., and 19 percent in Boston. Work by the Cato Institute’s Vanessa Brown Calder has subsequently found that regulatory burdens have intensified in many areas since the Glaeser et al. article appeared. We would therefore expect these implied regulatory taxes to be higher in many cities today.²⁹

Other economists estimate the effect of land-use regulations on prices and rents econometrically. Results from these studies, again, consistently suggest that tighter regulatory constraints drive higher housing costs. A 1996 paper by Stephen Malpezzi examining metropolitan markets found that increasing regulation by one standard deviation from average lowered construction by 11 percent and raised house prices by 22 percent.³⁰ A more recent assessment found that a similar one-standard-deviation increase reduced construction by a larger 17 percent, with twice the upward effect — 34 percent — on housing prices.³¹ A study of cities in Florida also found that restricting growth through farm preservation and open-space zoning made housing more expensive, with the most pronounced effects on the price of smaller houses.³²

Anti-development regulations have regressive effects. Poorer households are more likely to rent (61 percent of households in the bottom quintile and 66 percent of single-parent households rent, compared with just 38 percent for the population as a whole). An increase in housing costs has unambiguously negative consequences for renters. Poor households also tend to spend relatively more on housing, are more likely to value lower housing costs over improved amenities, and are more susceptible to being locked out of rich, productive cities and the economic opportunities they bring. This can have a big macroeconomic impact. Chang-Tai Hsieh and Enrico Moretti estimate that lowering the level of housing regulation to the median level across all U.S. cities for New York, San Francisco, and San Jose alone would raise long-term U.S. GDP by nearly 9 percent.³³

The negative consequences of land-use and zoning laws can also result in policies that exacerbate these regressive effects further. Local rent control laws, for example, are notionally justified as attempts to keep rents affordable, but binding controls deter investment in the rentable stock and encourage existing landlords to convert units to noncontrolled tenure types or to be more discerning about tenants. A recent study on the expansion of rent control in San Francisco in 1994 shows how this hurts the poor.³⁴ Landlords converted some properties to owner-occupied apartments and condos better suited to higher-income families. The overall supply of new housing fell too, increasing market rents by over 5 percent. Rent control both increased the cost of rental accommodation and intensified gentrification.

Federal taxpayers foot the bill for these mistakes, with relatively more housing aid flowing to states with restrictive zoning and land-use rules.³⁵ Treating the symptoms in this way helps entrench unnecessarily restrictive regulations. Subsidies ease the pressure on local governments to address the cause of high housing costs.

How much do existing regulations raise house prices or rents for households in the poorest 20 percent of the income distribution? It depends on where they live. Residents in many rural areas face no real housing cost increases. But estimates of regulatory taxes for major metropolitan areas by Glaeser et al. imply that average annual housing costs in New York are $2,060 more than in a competitive housing market; $3,200 in Boston; $5,230 in Los Angeles; $3,939 in D.C.; and a whopping $11,500 in San Francisco.³⁶

Some degree of regulatory tax in major cities might be appropriate given the externalities associated with new building, not least congestion. In cities such as San Francisco, the income distribution is very different from the national average too, meaning that there are fewer poor people residing in the city who would benefit directly from liberalization (though this is partly the result of high housing costs).³⁷ On the flip side, the Glaeser et al. estimates apply to the housing markets of nearly 20 years ago; since then the regulatory burden has intensified. New York as a whole has an income distribution similar to the overall U.S. population. Even using Glaeser’s older regulatory tax estimate implies that the poorest 20 percent there currently pay $1,044 per year more for shelter than they would under a permissive development regime.³⁸ These calculations would be much higher still for several cities in California.

Calculating an average effect for poor households across the country is difficult. Salim Furth has estimated that the average household’s annual housing costs increase by $1,700 as a result of land-use regulation. This implies housing costs for the poorest fifth are about $1,000 higher than they need be annually, given relative differences in spending on shelter. A similar result arises using Calder’s alternative measure of land-use regulation. Making the assumption that those states with above-average regulatory burdens were able to reduce these to the average of the rest of the country implies annual savings of $1,075 per year for poor households. But given that poorer households are more likely to live in rural areas, those figures may somewhat overestimate the effect.

Nevertheless, the direct cost of land-use planning and zoning regulations on low-income households could reasonably be anywhere between $0 and around $2,000 per year in the long term, depending on location. The broader economic costs are much greater still, given the secondary effect of poorer families finding it more difficult to move to areas with high-paying jobs. For single-parent households the range would be even wider, with regulatory costs up to around $3,500 or more for wealthier single-parent households in California’s most restrictive cities. Land-use and zoning liberalization could, in the long term, reduce housing costs significantly and greatly increase economic opportunities.

Childcare

Childcare is expensive. The average annual cost of infant-center care varies from a low of $5,178 in Mississippi to a high of $23,089 in D.C. (25.7 percent and 114.5 percent of the federal poverty income level, respectively).³⁹ Even accounting for income variance by state, care costs for an infant average 89.1 percent of median single-parent family income in D.C.; 70.9 percent in Massachusetts; and 57.0 percent in New York. Even in cheaper states, these costs average 27.2 percent in Mississippi; 28.9 percent in Louisiana; and 30.0 percent in Alabama.⁴⁰ For a family with two young children, the cost burden can be extremely heavy.

State governments control child-care policy, and variation exists in terms of assistance for poorer families.⁴¹ Overall, though, U.S. out-of-pocket costs for a typical single parent working full time are higher than any other OECD (Organisation for Economic Co-operation and Development) country.⁴² Not only are U.S. market prices higher than average, but parents receive less in the way of taxpayer subsidies.

These high prices can have negative consequences for poor families. Poorer single mothers are sensitive to child-care prices when making decisions about entering the labor market.⁴³ Mothers from poorer families, or those with low levels of educational attainment, are least likely to be working.⁴⁴ Previous data from the U.S. Census Bureau also show that poorer families are price sensitive in the type of care they choose. Children with employed mothers living in poverty are more than twice as likely to be cared for by an unlicensed relative.⁴⁵

More mothers of young children are choosing to work (in 1975, 28.3 percent of mothers with children under the age of 3 and 33.2 percent of mothers with children under the age of 6 were employed, compared with 59.4 and 61.5 percent, respectively, in 2016), making high child-care costs a salient political issue. Pressure is building for governments to help poor families with these costs.⁴⁶

There are good reasons why prices for formal childcare are high. It is a labor-intensive personalized service entailing the care of children, whom parents tend to value highly. A strong correlation between areas with high child-care costs and costs as a proportion of income suggests childcare is strongly “income-elastic” too — richer people want to spend relatively more on it.

Yet economic evidence suggests child-care prices are also driven higher by state-level regulations. Input requirements designed to improve care “quality,” including staff-qualification requirements and minimum staff-to-child ratios, significantly raise prices, with little evidence that they achieve other objectives.

These regulations are particularly regressive but get justified on “market failure” grounds. Parents are supposedly unable to observe accurately the quality of care in the sector, or underestimate the social benefits arising from “high-quality” childcare, necessitating minimum quality standards.

But these theoretical arguments are not robust and ignore the market context.⁴⁷ Most importantly, the regulations cannot ensure quality directly, not least because the true child-care market includes much more than formal infant-center care. If regulations affect prices, they can be the cause of substitution away from formal centers into more informal arrangements, the quality of which varies greatly.

Suppose a new regulation requires an increase in the staff-child ratio or child-care workers to achieve higher qualification levels. The former could increase quality by increasing staff interactions with individual children and the latter by making caregivers better trained to interact with the child in ways that foster development. The combination of the regulations may satisfy some parents that their children will be well cared for, and this “quality assurance” effect may raise overall demand for formal care.

Yet, raising the staff-child ratio has the effect of restricting the revenue-raising potential of each worker or of raising staffing requirements for a given number of children. These increased costs reduce the supply of formal care, thus increasing prices, and could lead parents to choose less costly alternatives. If centers compensate by paying staff lower wages to avoid this, the industry may attract lower-quality workers. Child-care providers likewise may respond to the cost increase arising from higher government certification requirements on caregivers by hiring cheaper, lower-quality support staff or purchasing lower-quality equipment. The effects of both regulations on the quality and use of childcare are therefore theoretically ambiguous.

Empirical work suggests that staff-child ratio regulation increases child-care prices substantially. Diana Thomas and Devon Gorry analyze variation in prices and staff-child ratios across states, estimating that loosening the requirement by one child across all age groups (regulations tend to vary by child age) reduces prices by between 9 and 20 percent.⁴⁸ This supports an older result from Randal Heeb and Rebecca Kilburn, who found that reducing the number of children per staff member by two would raise the price of childcare by 12 percent.⁴⁹

The poor suffer disproportionately from these higher prices. Thomas and Gorry show that a small but measurable number of mothers stop working altogether. These are more likely to be low-income people for whom the payoff for moving into work is smaller. Joseph Hotz and Mo Xiao, using a panel dataset across three census periods with extensive child-care center data, data on home care by state, and a host of control variables, find that tightening the staff-child ratio by one child reduces the number of child-care centers by 9.2 to 10.8 percent, without increasing employment levels at other centers.⁵⁰ This reduced supply occurs exclusively in lower-income areas and leads to substitution to home daycare. Importantly, there is no evidence that increasing the stringency of this regulation improves quality. It simply reduces accessibility to formal care for the poor, making it more expensive and leading to substitution toward other care settings.

Staff qualification requirements also appear to have a big effect on prices. Thomas and Gorry find the requirement for lead teachers to have a high school diploma increases prices by between 25 and 46 percent. Hotz and Xiao likewise find that increasing the average required years of education of center directors by one year reduces the number of child-care centers in the average market by between 3.2 and 3.8 percent. Again, this effect manifests itself overwhelmingly in low-income areas, with quality improvements (proxied here by accreditation for the center) occurring in high-income areas.

Like housing, childcare is a sector where government regulations restrict the supply of the service to the financial detriment of the poor. For those on the margins of the labor market, child-care regulations can reduce the payoff to work. In return for these higher costs, there is little evidence that they yield much improvement in child-care quality. In fact, higher prices appear to cause demand substitution to potentially lower-quality childcare settings. (Though in the case of childcare, there is also a question about what “quality” actually means.)

Despite this evidence, some city and state governments continue undeterred. The D.C. government has passed regulations requiring teachers at child-care centers and caregivers at home-based centers to have associate degrees in early childhood education and assistant caregivers to obtain new child development associate certificates.⁵¹ Even if these do raise the quality of care, the requirements will further constrict supply — which is presumably why the District has delayed implementation and is engaged in new attempts to subsidize provision.⁵²

Deregulation of staffing requirements could therefore significantly reduce prices to the benefit of the poor, who tend to put much less weight on the “quality” desired by richer families and regulators. The current costs of these regulations to low-income families are significant. The cautious end of Thomas and Gorry’s estimates suggests that even modest relaxation of staff-to-child ratios by one child at all age groups alone could reduce average child-care prices by $466 per year in Mississippi and $2,078 per year in Washington, D.C.⁵³

Eliminating statutory regulations on child-care staffing entirely could reduce the cost of care even more significantly. Market mechanisms in the form of accreditation or certification agencies will arise if significant numbers of parents put a high premium on certain staffing structures and outcomes. Many major European countries already do not bother with mandated staff-to-child ratios, for example, seemingly with few ill effects.⁵⁴

But extensive deregulation might be a leap too far for state policymakers. For the purposes of examining the cost of child-care regulations for a typical family with a young child in the poorest quintile, then, I assume that the “cost” of regulation equates roughly to the potential gains from a modest relaxation in the staff-child ratio, as outlined above. The net benefits to poor households of more extensive deregulation would be much larger. The broader economic benefits are greater still since lower prices allow more low-income family members to fulfill their labor market preferences.⁵⁵

Food

The average household in the poorest 20 percent of the income distribution spent $3,682 on food in 2016 (15.4 percent of total spending and the highest proportion of any income group). Single-parent households spend proportionately more than other household types. Yet, the federal government makes groceries more expensive through such policies as milk-marketing orders, sugar programs, and ethanol mandates.

Milk-Marketing Orders. The federal government operates a byzantine system of marketing orders, price and income supports, and trade barriers in dairy markets.⁵⁶

Federal milk-marketing orders set monthly minimum prices that dairy processors must pay dairy farmers in 10 regions. These account for around 60 percent of total production, with another fifth of the remaining 40 percent from California, which operates similar schemes at the state level.⁵⁷ The marketing orders set regional prices for fluid milk and use complex formulae to determine nationwide prices for three other classes (soft manufactured products such as ice cream, hard cheese and cream cheese, and butter and dry milk).

The Milk Support Program supplements this with guarantees that the government will purchase any amount of cheese, butter, and dry milk from processors at a set minimum price. In order to ensure that these prices are not undercut by foreign producers, import barriers then insulate domestic dairy producers from competition through tariff rate quotas.

With modern storage techniques, there appears to be little need for this regional balkanization of the sector. The marketing orders and price supports stymie entrepreneurship and producers’ ability to provide low-cost milk to regions with higher milk prices. Import barriers further raise product prices and distort economic activity toward the dairy sector rather than allowing resources to be used most efficiently. The main economic effect of all this is higher average prices borne by consumers.

The best evidence of these price effects comes from Chouinard et al., who review the existing literature on what would happen if milk-marketing orders and associated supports were abolished.⁵⁸ All studies suggest retail fluid milk prices would fall by between 15 and 20 percent. The effect on manufactured milk and processed dairy products is less clear. Averaging previous studies suggests butter and ice cream prices would fall by 3 percent and 1 percent respectively while, counterintuitively, fresh cream, coffee additives, and yogurt prices would increase by 1.3 percent and cheese prices would rise by 0.5 percent.⁵⁹

The average poorest quintile household spent $246 per year on dairy products in 2016 ($97 on milk and fresh cream and $149 on other manufactured products).⁶⁰ But this masks significant variation. The average single-parent family spends $353 on dairy products.⁶¹ African Americans, who are more likely to be in poverty, experience high rates of lactose intolerance, so spending on dairy products for poor whites is likely to be higher than the average figures suggest. The Consumer Expenditure Survey is not sufficiently disaggregated to calculate the cost of these programs to the average household in the poorest quintile. We need information about how consumers would react to price changes and how far they will substitute away from some dairy products to others to calculate net savings.

The Chouinard et al. model seeks to do just that. It suggests that “lower income families [would] benefit more than wealthier families” from eliminating federal milk-marketing orders, meaning the regulations currently are very regressive. They also find that families with young children benefit far more than the childless.

Their results suggest annual savings for white families with annual incomes of $10,000 from the abolition of federal milk-marketing orders of $44; $38 for those with incomes at $20,000; and $33 for those with incomes at $30,000. They estimate the regulatory burden for black families to be somewhere between a third and a half of this, meaning that the average black family would save somewhere between $14 and $22 per year.

Taking into account these differential racial burdens, the average household in the bottom quintile faces a current regulatory cost of approximately $38 per year from dairy interventions, and the average single-parent family a higher cost of $54 per year.⁶²

Sugar intervention. In similar fashion, the U.S. federal government also effectively cartelizes the sugar market. As Cato scholar Colin Grabow has explained in detail, the U.S. Department of Agriculture (USDA) facilitates loans to sugar processors using raw sugarcane and refined beet sugar as collateral, effectively creating a floor for the domestic sugar price.⁶³ To ensure that loans will likely be repaid, it then restricts the supply of domestic sugar through allotment quantities, raises demand by making purchases, and limits the amount of sugar that can be imported either without tariffs or with low tariffs.

Unsurprisingly, these anti-competitive actions, which restrict supply and inflate demand, raise domestic sugar prices substantially. Data from the USDA show that in March 2018 the U.S. raw sugar price was 24.73 cents per pound, almost double the world price of 12.83 cents.⁶⁴ Not only do consumers pay higher retail prices for sugar but they also pay more for manufactured foods that contain sugar as an ingredient.

Economic analysis of the consumer cost of the program has examined the aggregate effect on consumers. Economist Michael Wohlgenant has suggested that the burden amounts to $2.4 billion per year, or an average of around $19 per household.⁶⁵ A 2017 paper by John Beghin and Amani Elobeid estimated the loss to consumer welfare from the sugar program at between $2.4 billion and $4 billion in 2009 dollars.⁶⁶ Adjusted for inflation, that is equivalent to $2.8 billion to $4.7 billion today.

This suggests that the program costs between $22 and $36 per year for the average household.⁶⁷ Determining a more precise figure for low-income households is fraught with difficulty. On the one hand, poorer households tend to contain fewer people, and on the other, some evidence suggests overall sugar consumption is highest among those with the lowest incomes.⁶⁸

In the absence of more complete evidence, I take the midpoint of these household estimates and assume that the costs of the program are spread evenly across individuals, such that the average poor household (with 1.6 members) is $18 per year worse off as a result of the policy, and the average single-parent family (with 2.9 members) is $33 worse off.

Renewable Fuel Standard. The federal Renewable Fuel Standard (RFS) mandates quotas for the amounts of biofuels blended into transportation fuel sold. The origins of the standard are the Energy Policy Act of 2005, which amended the Clean Air Act, and the Energy Independence and Security Act of 2007, which expanded the ethanol mandates.

These regulations raise food prices for consumers. The increased demand for corn for biofuels raises the prices of corn and corn-based products directly. This higher price then raises production costs for meat and dairy products, since corn is used as animal feed. Dedicating agricultural land to growing corn also restricts available land supply for other crops, such as soybeans, raising their price too.

A 2009 Congressional Budget Office (CBO) analysis found that demands for ethanol subsequently raised total food spending by between 0.8 and 1 percent.⁶⁹ This corroborates the work of Richard Perrin, who estimated that growth in demand for ethanol raised overall food prices by 1 to 2 percent in 2008.⁷⁰ But how much consumers would benefit financially from the repeal of the RFS at any given time depends on the oil price.

In 2014, when oil prices were much higher than today, CBO analysis suggested “suppliers would probably find it cost-effective to use a roughly 10 percent blend of corn ethanol in gasoline in 2017 even in the absence of the RFS,” meaning that total food spending would fall only very slightly were the RFS repealed (by 0.1 percent).⁷¹ Today, however, oil prices are significantly lower, meaning that there is a bigger incentive to use relatively more oil in gas production.

Given that the price of oil today falls between the levels seen in 2008 and 2009, I assume that the RFS currently raises food prices by 1 percent. Assuming that food spending is price inelastic, the direct cost of this policy can be estimated at about $39 per year for the average family in the lowest income quintile, or $58 for the average single-parent family.⁷²

Transport

The average household in the bottom 20 percent of the income distribution spent $3,767 on transport in 2016 (15 percent of total spending). The vast majority was on private vehicles: $1,332 on vehicle purchases; $902 on gasoline and motor oil; and $1,308 on other vehicle expenses.⁷³ Just $225 was spent, on average, on public transportation.

Averages mask the real experience of families, of course. Whereas 9 percent of all households do not have a vehicle,⁷⁴ this increases to 20 percent for households in poverty and 11 percent for households with incomes at 100 to 200 percent of the federal poverty level. Spending on motor vehicles for vehicle-owning households is therefore higher than the figures above suggest.⁷⁵

Two government regulations increase motoring costs: Corporate Average Fuel Economy Standards at the federal level and dealership franchise laws at the state level. These not only increase transport costs for the poor directly, but also make it more difficult for poor families to have physical accessibility to jobs, health care, training, and childcare.⁷⁶

Corporate Average Fuel Economy Standards. First created in 1975, Corporate Average Fuel Economy Standards (CAFE) sought to increase the fuel economy of cars and trucks to limit dependence on foreign oil. It was originally thought that consumers undervalued fuel savings from more efficient vehicles, though recent research suggests fears over consumer short sightedness were overstated.⁷⁷ Now CAFE standards are justified as a tool to reduce carbon emissions.

The regulations require manufacturers to achieve a sales-weighted fuel economy average for car and light-truck fleets. Their stringency has increased since they were tied to a vehicle’s physical footprint beginning in 2012. President Barack Obama had agreed to raise the standards significantly from 2022 through 2025, to 60 miles per gallon for small cars and 46 for large cars, and 50 miles per gallon for small trucks and 30 for large trucks. But President Trump has outlined plans to relax these rules.⁷⁸ More recently, the administration proposed freezing the standards entirely at 2020 levels and preventing states (particularly California) from unilaterally imposing stricter regulations.⁷⁹

CAFE standards increase costs to consumers overall, although the effects are not uniform across vehicles. Meeting fuel economy standards requires high fixed-cost investments in technological improvements by manufacturers. But to hit the sales-weighted averages, manufacturers have to adjust prices to incentivize purchases. Evidence suggests consumers prefer larger, more powerful vehicles. Firms therefore have to offer discounts for smaller, more fuel-efficient models, cross-subsidized by higher prices for larger vehicles. Increased prices for new cars lead to higher prices in the used car market too, as consumers substitute toward older models of the larger vehicles they tend to prefer.

It is beyond the scope of this paper to assess the merits of curbing carbon emissions. Economists are, in any case, doubtful this type of policy will have large effects on emissions, not least because making vehicles more expensive leads consumers to keep older, less fuel-efficient cars on the road longer and incentivizes owners of the more efficient cars to drive more.

If one is worried about the externality of carbon emissions, theory and evidence show that CAFE standards have more regressive effects than an equivalent gas tax for such a goal.⁸⁰ CAFE standards are roughly equivalent to a tax on the gasoline used per mile of travel. The difference between the consumption of rich and poor on this metric is lower than the total gas consumed. This makes CAFE standards more regressive than a simple gas tax. Additionally, larger vehicle categories also face less stringent standards. The fact that vehicle size tends to increase with income exacerbates the regressive impact of the standards.

Economists Lucas Davis and Christopher Knittel estimate the implicit tax from CAFE standards in 2012 to be around $180 per vehicle for those in the poorest income quintile.⁸¹ Adjusted for inflation, that’s more like $194 today.⁸² The standards have become much more stringent since then, suggesting the effect today would be far larger.

Other academic studies find larger effects for broader long-term consumer welfare losses (which include the welfare costs of substituting away from preferred vehicles). In today’s prices, for example, Mark Jacobsen estimates a long-run consumer surplus loss of $226 for every one-mile-per-gallon standard increase for people with incomes below $25,000.⁸³ Assuming that this effect was linear (we might assume the marginal cost increases with the standard), this implies that the tightened standards seen between 2011 and 2018 caused consumer welfare losses of more than $2,230 per vehicle. If the standards planned by President Obama were implemented through 2025, this loss would more than double.⁸⁴ This corresponds closely to figures from much older studies. David Greene found that for every one-mile-per-gallon increase in vehicle fuel economy, the average per-vehicle cost was from $225 to $450 in today’s prices, and figures of these magnitudes have been corroborated in a broader review of the literature.⁸⁵

Given that the average used car price is now around $20,000, this suggests the ratchet in standards since 2011 is likely to account for over 10 percent of the price of a used vehicle.⁸⁶ A 10 percent reduction in vehicle prices would save the average poorer household $133 and the average single-parent family $307 annually.⁸⁷

This appears to be a reasonable estimate. A 2015 paper by Mark Jacobsen and Arthur van Benthem estimated that the standards enacted in 2012 would cause used vehicle prices to rise by $103 per year relative to the old standards enacted in 2007.⁸⁸ An earlier paper by David Austin and Terry Dinan found an annual cost per vehicle per year of $153 for an increase in the standards of just under 4 miles per gallon.⁸⁹ More recently, the Reason Foundation’s Julian Morris estimated that the average price of a new pickup truck has risen by 25 percent since 2013, overwhelmingly because of the CAFE standards. That works out to a net cost of about $100 per year.⁹⁰

While a range exists, all these estimates suggest CAFE standards increase new and used vehicle prices for consumers. Theory and evidence suggests the effects are regressive. While manufacturers are unlikely to undo technological changes that have delivered improved fuel efficiency, President Trump’s planned policy of capping standards at 2020 levels would deliver significant annual savings for purchasers of vehicles relative to the trajectory planned by President Obama.

Dealership Franchise Laws. Every state has laws governing the economic relationships of car manufacturers with new car dealers. These require dealers to be licensed and can also incorporate restrictions on when franchise relationships can be terminated, canceled, or transferred, restrictions on opening new dealerships in existing market areas, and requirements that manufacturers buy back vehicles or other accessories when a dealership franchise is terminated.⁹¹

The most prominent effect of these laws is the restriction of direct sales by manufacturers. But the broad effect of all of them is to insulate dealerships from competition and prevent manufacturers from optimizing their inventory and distribution to best match the demands and preferences of consumers.

“Good cause” regulations, for example, mean manufacturers can only terminate a franchised dealership for a set of enumerated reasons, often not including efficiency. Manufacturers can face penalties and charges if they terminate dealerships because of demand patterns. Though states often allow termination for noncompliance of a franchise agreement, even then the manufacturer faces the burden of proof in showing that they have acted in good faith, the termination is reasonable, and they have given notice with an opportunity for the franchisee to deal with the issue at hand.

Plenty of states have laws that protect existing franchisees from “encroachment” too. Manufacturers must show the need for a new dealership if it is proposed within the same relevant market area as an existing one. Protection of exclusive territories creates effective monopoly power for dealers, raising profits, when manufacturers might prefer to increase the quantity of sales.

These regulations were justified in the early 20th century as correcting for asymmetric information between the franchiser (the manufacturer) and the franchisee (the dealer) that led to manufacturers exploiting dealers. But today, calls for auto dealership laws are based on the supposed “social benefits” of dealerships, including their roles in the community and as sponsors of local events. Because such claims could be made about all local businesses, they do not provide robust “market failure” justifications for the interventions.

These regulations raise consumer prices, though the magnitudes of the effect are disputed. A paper exploring data from 1972 suggested that new car prices were raised by around 9 percent.⁹² A report for the Federal Trade Commission in 1986 found an average price increase of just over 6 percent across all car types.⁹³ In 2001, the Consumer Federation of America summarized the existing literature, concluding that these laws raised new automobile prices by between 6 and 8 percent.⁹⁴ This was subsequently questioned by the National Automobile Dealers Association, which concluded that the true effect was much lower, at 2.2 percent.⁹⁵ But papers focusing on other countries have found effects similar to the 2001 study.⁹⁶

Unfortunately, little modern evidence exists on this subject, and it is beyond the scope of this paper to develop new calculations. The internet may have helped to reduce some of the burden on consumers, and though the price effects will induce substitution to the used car market, we perhaps would not expect prices to be affected to the same extent. Given the best estimates from older work, the average regulatory cost on a household in the bottom quintile is likely to be around $61 annually, or $140 for the average single-parent family.⁹⁷

Apparel and footwear

In 2016, the average household in the bottom income quintile spent $860 on apparel and footwear, or 3.4 percent of overall spending — the highest proportion of any income quintile.⁹⁸ The average single-parent household put 4.5 percent of total expenditure toward these goods.⁹⁹ The poor spend a disproportionate amount on clothing and footwear, and family structures most likely to be recipients of means-tested welfare programs (single-parent households) spend most of all.

Yet the federal government makes clothing and footwear more expensive through import tariffs, which are often higher than those imposed on other goods. The United States raised $33.1 billion overall in tariff revenue in 2017, but $14 billion of that came from tariffs on apparel and footwear alone. These items account for 4.6 percent of the value of U.S. imports, but 42 percent of duties paid. The average effective tariff rate for U.S. imports overall is just over 1.4 percent. Rates for apparel and footwear are 13.7 percent and 11.3 percent, respectively.¹⁰⁰

The Cato Institute’s Daniel Ikenson has examined the evolution of clothing and textile protectionism.¹⁰¹ He concludes that such high tariffs do not exist to protect domestic apparel manufacturing. Data from the U.S. Trade Representative estimated that 91 percent of manufactured apparel goods and 96.5 percent of footwear are imported despite the tariffs.¹⁰² In February 2018, just 116,400 people were employed in domestic apparel manufacturing, a collapse from 939,000 in January 1990.¹⁰³

Why then are such highly regressive tariffs imposed? The answer appears to be the lobbying efforts of the capital-intensive U.S. textile industry. Textiles are the major input for labor-intensive apparel production, which largely occurs overseas. To quote Ikenson:

The U.S. textile industry insists on preserving those tariffs as leverage to compel foreign apparel producers to purchase their inputs. Preferential access [to U.S. markets] is conditioned on use of U.S. textiles. The high rates of duty apply, generally, to all “normal trade relations” partners. But those duties are much lower or excused entirely for trade agreement partners, provided that the finished garment comprises of textiles made in countries that are signatories to the agreement.¹⁰⁴

U.S. consumers pay the price of this protectionism, and poorer consumers especially. In fact, protectionism is doubly regressive. Not only do poorer households spend relatively more on clothes and footwear, but Edward Gresser’s work has shown how often luxury clothes and shoes face lower tariff rates than inexpensive products.¹⁰⁵

Consider Table 3 (an updated version of Gresser’s work) below. Where duties are applicable, a pure cashmere sweater import incurs a 4 percent tariff, a wool sweater a 16 percent tariff, and an acrylic sweater a whopping 32 percent. Men’s silk shirts would see a 0.9 percent tariff, cotton shirts a 19.7 percent tariff, and cheaper polyester shirts a 32 percent tariff. Leather dress shoes have an 8.5 percent tariff, whereas cheap sneakers would see a 43 percent tariff. Windbreakers, leggings, tank tops, and other clothes made cheaply from synthetic fabrics face a 32 percent tariff if sourced from countries that the United States does not have a free-trade agreement with.

Table 3: Regressive tariffs (percentage on various goods)

Source: United States International Trade Commission, Tariff Database, https://dataweb.usitc.gov/scripts/tariff.asp
Note: The codes needed to find tariff rates for these products are: 64035960, 64029142, 64029160, 61109010/61101210, 61103030, 61059040, 61051000, 61052020, 71131921, 71131110, 71131120, 63012000, 63013000.

Assuming poorer households tend to buy cheaper products, these differential tariffs have perniciously regressive effects. (And not just for clothes; as Table 3 shows, similar trends are seen for consumer goods such as handbags, necklaces, and blankets).

It is difficult to calculate the true overall cost of these tariffs to poor families. That would require detailed information on the effect on domestic substitute goods prices, knowledge of products bought by poor families and their propensity to import in the absence of protectionism.

Nevertheless, we can develop cautious lower-bound estimates of the financial cost. The average household in the poorest income quintile spends $655 on apparel and $206 on footwear per year. Assuming the import propensities for the population as a whole apply to poorer people implies $595 of apparel spending and $199 of footwear spending is on imported goods. Taking average effective tariff rates for apparel and footwear for this spending (13.7 and 11.3 percent) implies a combined direct tariff cost of $92 per year for the average household in the poorest income quintile, or $204 per year for the average single-parent household.¹⁰⁶

These figures underestimate the true burden, though, because they only represent the direct cost from current spending on imported goods. They assume tariffs do not raise domestically produced goods prices, though in reality the anti-competitive effect of the tariffs would be expected to raise prices here too. The calculation also assumes the same effective tariff rates for apparel and footwear apply for the poorest households as for the whole population, but we have seen that products that the poor are more likely to buy tend to face higher tariff rates. Consumer welfare losses from tariffs are higher than the implied savings here, of course, since tariffs make consumers less willing to buy imported products that they would otherwise prefer.

Nevertheless, these figures correspond well to calculations by Jason Furman, Katheryn Russ, and Jay Shambaugh that provide an estimate of the overall tariff burden (all goods, not just apparel and footwear) of around $100 and $238 per year for poorer and single-parent households, respectively.¹⁰⁷

Occupational Licensing

State governments regulate numerous occupations through education, training, or test requirements, creating barriers to entry to practicing a trade. This restricts the supply of providers within the state and discourages movement of professionals across state lines, raising the price of services.

Licensing gets justified on grounds of imperfect information between buyer and seller, particularly when harm could result from low-quality service. This argument is most forcefully made about medical professionals, where it is argued that “quack” practitioners might do substantial harm to patients. Yet restrictions on entry come with tradeoffs, including higher prices and deterring talented people from entering a profession. Ideally, one must weigh up any benefits of reduced quackery against these supply-restricting consequences.

Other sectors commonly licensed include hair braiding, barbers, and sign-language interpreters, where any costs associated with low-quality providers are likely much smaller. In these cases, consumers are best placed to judge a price-quality bundle, and intermediate institutions such as online rating sites can provide information about the nature and quality of service. Markets may even deliver certification mechanisms for safety- or quality-sensitive customers. Instead, licensure boards are often dominated by existing providers with a vested interest against competition. The arguments that licensure corrects for some “market failure” are therefore increasingly difficult to justify. Yet 25 to 30 percent of Americans now work in licensed occupations.¹⁰⁸

A plethora of research has found licensure raises wages in licensed sectors (relative to no licensure or no certification). A recent study found that “having a license when it is not required has no influence on wage determination, but, when it is required, licensing raises wages by 7.5 percent,” even controlling for a host of worker and occupational characteristics.¹⁰⁹

Whether this raises prices depends on whether consumers would demand equally robust entry barriers in the form of certification absent government intervention. Without licensure constraints, prices of services are likely to be lower, unless state governments provide economies of scale in license provision relative to private certificates. However, in many cases consumers are unlikely to demand substitute certification at all, and so the wage premium would evaporate.

Research on individual markets confirms this intuition. Relaxing licensing laws to allow nurse practitioners to perform tasks without medical doctor supervision was found to reduce well-child exam prices by between 3 and 16 percent.¹¹⁰ Delicensing of funeral servicing providers in Colorado lowered funeral prices significantly.¹¹¹ Older papers estimated that dental assistant and hygienist licensing raised prices of dental visits by between 7 and 11 percent, and optician licensing the price of eye care by between 5 and 13 percent.¹¹²

Two attempts have been made to estimate the aggregate costs of occupational licensing to consumers. Morris Kleiner, Alan Krueger, and Alex Mas estimated a $203 billion annual cost, or $1,567 per household.¹¹³ The Heritage Foundation’s Salim Furth estimates a lower figure, with a cost to the average household of $1,033 per year.¹¹⁴

Ideally, we would produce a more accurate estimate using detailed data of the cost of licensure by sector mapped against spending patterns for poor households. We know, for example, that poorer households spend relatively more on health care than richer households but also that richer households spend more on other grooming services affected by the licensure premium. Poorer households are likely to be more price sensitive and less discerning about “quality.” One must bear in mind, too, that because licensing restricts people from practicing certain occupations, the potential labor supply is increased in nonlicensed sectors, putting downward pressure on labor costs and hence prices in other industries.

Bearing these caveats in mind, I assume that the total spending ratio of poorer and single-parent households to the average household is the same as the ratio of spending on licensed services between the groups. Using the household costs of licensure from Kleiner and Furth implies an average annual cost to poorer households of between $450 and $690 per year, and between $760 to $1,160 per year for the average single-parent household.¹¹⁵ Again, this is a lower bound to the true economic costs for poorer people, not least because occupational entry barriers prevent individuals from taking up new and better job opportunities because of the time and financial costs of meeting the licensure requirements.

Conclusion

This paper has demonstrated how government interventions raise the cost of living for poor and single-parent households.

Debates on policies to help the poor tend to focus on redistribution, tax breaks, minimum wage hikes, and government-provided services. But liberalizing reform in the markets outlined above could improve the financial well-being of less well-off households without new government expenditure or risky labor-market interventions. A “cost-based” approach to poverty alleviation should be considered a key tool in helping the less well-off.

Table 4 summarizes the estimated direct costs to poorer and single-parent households of existing interventions. These are based on extremely cautious assumptions and likely understate the true financial impacts. The ranges are wide, reflecting differences in household location and composition. However, a reasonable central range for poorer households would be a lower bound cost of $830 per year for a household with no children living in a rural area, up to $3,500 for a poor family living in an expensive city such as New York with a young child in full-time infant care.¹¹⁶

Table 4: Summary of costs of interventions on the poor and single parents (dollars per year)

Source: Author’s calculations applied largely through the Bureau of Labor Statistics Consumer Expenditure Survey, https://www.bls.gov/cex/.
Note: Housing costs range from areas with no effective regulatory tax through to the implied regulatory tax in expensive cities (adjusted to relative spending by poor and single-parent households). Child-care costs range from $0 for those with no young children in infant care to the implied savings from relaxing staff-child ratios by one child in the most expensive region for childcare (D.C.). Milk-marketing order costs calculated using information from Chouinard et al. Sugar program costs calculated applying per person midpoint cost of Beghin and Elobeid cost estimate, multiplied by average household numbers for poor and single-parent families. Renewable food standard cost based on 1 percent uplift in food prices applied to food spending levels. Corporate average fuel economy standards cost calculated as 10 percent increase in vehicle prices applied to vehicle purchase spending levels. Dealership franchise laws cost assumes new vehicle prices are increased by 6 percent and used car prices by 4 percent as a result of the regulations. Tariff costs estimated on current spending levels by poor and single-parent households, based on whole population import propensities and average effective tariff rates on apparel and footwear. Occupational licensing costs based on the average cost of licensure per household estimates from Kleiner et al. and Furth adjusted for spending levels of the poor and single parents. For more information on the assumptions behind these calculations, see the discussion in relevant sections of the paper.

The figures here relate only to the direct effects of these policies on prices and so ignore the broader effect on productivity and market incomes. For land-use and zoning laws, child-care regulations, policies that increase driving costs, and occupational licensure these secondary effects could be very large indeed. Liberalization could improve labor mobility, willingness to move into the labor market, and job options available to the unemployed and existing low-paid workers.

The calculations are cautious for other reasons. Estimates of the cost of sugar interventions and tariffs on clothing and footwear assume that the poor have the same consumption habits as the broader population, though there are reasonable grounds to suspect the cost borne by them is even higher. The cost of child-care regulations is pegged at the savings from very modest relaxation of existing staffing regulations, rather than full repeal, which could deliver huge price reductions. This analysis also ignores regressive interventions in other areas where the poor spend significant amounts, especially health care and utilities.

Individual households will face very different costs depending on exactly where they live, whether they have young children, their means of transportation, and other spending tastes and preferences. This paper has shown, though, that a “cost-based” reform agenda could deliver major financial savings for poor families.

A concerted anti-poverty agenda across all levels of government overturning these damaging policies could have political benefits too. The lower cost of living would lessen political demands for government to redistribute income. The aspirations of the “living wage” campaign would be much more likely to be achieved, but through lower living costs rather than demands for states or cities to raise minimum wages. Better financial outcomes for the poor through market activity might lead to greater support for economic liberalization in other sectors.

Reform would be politically challenging. But with the federal finances suffering from large and growing imbalances and widespread concern about the future of labor markets, now is an opportune time for a new approach to assist those on low incomes. For too long an obsessive focus on the role of government transfers and minimum wage laws in alleviat­ing poverty has blinded campaigners and poli­ticians to areas where existing policies raise living costs. We should aspire to undo this damage, rather than doubling down on a more interventionist agenda that, in part, seeks to treat the symptoms of current mistakes.

Notes

¹. Michael Tanner, “The American Welfare State: How We Spend Nearly $1 Trillion a Year Fighting Poverty-and Fail,” Cato Institute Policy Analysis no. 694, April 11, 2012, https://www.cato.org/publications/policy-analysis/american-welfare-state-how-we-spend-nearly-%241-trillion-year-fighting-poverty-fail.

². See discussion of theories of causes of poverty in Michael Tanner, The Inclusive Economy: How to Bring Wealth to America’s Poor, Cato Institute, forthcoming, December 2018.

³. Sean Higgins, “Democrats Officially Introduce $15 Minimum Wage Bill,” Washington Examiner, May 25, 2017.

⁴. See “A Living Wage,” BernieSanders.com, https://berniesanders.com/issues/a-living-wage/; and “Early Childhood Education,” HillaryClinton.com, https://www.hillaryclinton.com/issues/early-childhood-education/.

⁵. Ryan Bourne, “A Jobs Guaranteed Economic Disaster,” Cato at Liberty, Cato Institute, April 24, 2018, https://www.cato.org/blog/jobs-guaranteed-economic-disaster.

⁶. Dylan Matthews, “Hillary Clinton Almost Ran for President on a Universal Basic Income,” Vox.com, September 12, 2017; see also Charles Murray, “A Guaranteed Income for Every American,” Wall Street Journal, June 3, 2016.

⁷. Ramesh Ponnuru, “Tax Relief for Parents,” Statement before the Senate Committee on Finance on “Individual Tax Reform,” September 14, 2017, http://www.aei.org/publication/tax-relief-for-parents/; and Sen. Mike Lee, “Sens. Lee and Rubio to Introduce Child Tax Credit Refundability Amendment,” November 29, 2017, https://www.lee.senate.gov/public/index.cfm/2017/11/sens-lee-and-rubio-to-introduce-child-tax-credit-refundability-amendment.

⁸. “A Better Way: Our Vision for a Confident America,” Poverty, Opportunity, and Upward Mobility, June 7, 2016, https://abetterway.speaker.gov/_assets/pdf/ABetterWay-Poverty-PolicyPaper.pdf.

⁹. For example, it is widely acknowledged that policies that seek to ameliorate climate change by raising the cost of carbon emissions hit the poor hard. But the objective of these policies is to internalize the social costs of carbon, which theoretically raises overall economic efficiency. Though one can debate whether existing policies do this well, or whether they are needlessly regressive, this paper avoids issues where these kinds of judgments are required, instead focusing on areas where market failure arguments are weak and the policies have clear regressive effects.

¹⁰. Arloc Sherman, “Official Poverty Measure Masks Gains Made over Last 50 Years,” Center for Budget and Policy Priorities, September 13, 2013, https://www.cbpp.org/research/official-poverty-measure-masks-gains-made-over-last-50-years.

¹¹. Bruce D. Meyer and Derek Wu, “The Poverty Reduction of Social Security and Means-Tested Transfers,” NBER Working Paper no. 24567, May 2018, http://www.nber.org/papers/w24567.

¹². David Neumark, “Reducing Poverty via Minimum Wages, Alternatives,” Federal Reserve Bank of San Francisco Economic Letter no. 2015-38, December 28, 2015, https://www.frbsf.org/economic-research/publications/economic-letter/2015/december/reducing-poverty-via-minimum-wages-tax-credit/.

¹³. Bruce Meyer and James Sullivan, “Winning the War: Poverty from the Great Society to the Great Recession,” Brookings Papers on Economic Activity, Fall 2012, https://www.brookings.edu/wp-content/uploads/2012/09/2012b_Meyer.pdf.

¹⁴. Congressional Budget Office, The Budget and EconomicOutlook: 2018 to 2028, April 2018.

¹⁵. Congressional Budget Office, The 2017 Long-Term BudgetOutlook, March 2017.

¹⁶. See Karel Martens and José L. Montiel Olea, “Marginal Tax Rates and Income: New Times Series Evidence,” NBER Working Paper no. 19171, September 2017, http://www.nber.org/papers/w19171; and Robert J. Barro and Charles J. Redlick, “Macroeconomic Effects from Government Purchases and Taxes,” NBER Working Paper no. 15369, December 2011, http://www.nber.org/papers/w15369. Concerning the latter, Barro states in the Wall Street Journal that “My research with Charles Redlick, published in 2011 by the Quarterly Journal of Economics, suggests that cutting the average marginal tax rate for individuals by 1 percentage point increases gross domestic product by 0.5% over the next two years.” Robert J. Barro, “Tax Reform Will Pay Growth Dividends,” Wall Street Journal, January 4, 2018.

¹⁷. See David Neumark and Cortnie Shupe, “Declining Teen Employment: Minimum Wages, Other Explanations, and Implications for Human Capital Investment,” Mercatus Center Working Paper, February 7, 2018, https://www.mercatus.org/publications/declining-teen-employment-minimum-wage-human-capital-investment.

¹⁸. See summary of the literature in Ryan Bourne, “A Seattle Game-Changer?,” Regulation 40, no. 4 (Winter 2017-2018): 8-11, https://object.cato.org/sites/cato.org/files/serials/files/regulation/2017/12/regulation-v40n4-6.pdf.

¹⁹. Neil Irwin, “The Unemployment Rate Rose for the Best Possible Reason,” New York Times, July 6, 2018.

²⁰. In Britain, working-age welfare bore a hugely disproportionate share of the deficit reduction measures seen following the general election during 2010.

²¹. Andrew Hall and Jesse Yoder, “Does Homeownership Influence Political Behavior? Evidence from Administrative Data,” Department of Political Science, Stanford University, August 7, 2018.

²². Bureau of Labor Statistics, Consumer Expenditure Survey, Metropolitan Statistical Area Tables, https://www.bls.gov/cex/tables.htm#MSA.

²³. Child Care Aware of America, “2017 Appendices: Parents and the High Cost of Child Care,” http://usa.childcareaware.org/costofcare.

²⁴. The costs to the average poor household of anti-competitive, regressive regulations would be higher still if we also examined some utilities and health care interventions; but that is beyond the scope of this paper.

²⁵. Pew Charitable Trusts, “Household Expenditures and Income,” March 2016. The poorest quintile are more likely to be renters: 61 percent of households in the bottom quintile and 66 percent of single-parent households rent, compared to just 15 percent of households in the top income quintile; see Bureau of Labor Statistics, Consumer Expenditure Survey 2016, Table 1101, Quintiles of income before taxes, and Table 1502, Composition of consumer unit.

²⁶. Demographia, “14th Annual Demographia International Housing Affordability Survey: 2018,” January 22, 2018, http://demographia.com/dhi.pdf. It is worth noting that Demographia limits its analysis to metropolitan areas with populations of 1 million people or more.

²⁷. Crystal Chen, “Zumper National Rent Report: March 2018,” Zumper.com, February 28, 2018.

²⁸. Edward L. Glaeser, Joseph Gyourko, and Raven Saks, “Why Is Manhattan So Expensive? Regulation and the Rise in Housing Prices,” Journal of Law and Economics 48, no. 2 (October 2005): 331-69.

²⁹. Vanessa Brown Calder, “Zoning, Land-Use Planning, and Housing Affordability,” Cato Institute Policy Analysis no. 823, October 18, 2017, https://www.cato.org/publications/policy-analysis/zoning-land-use-planning-housing-affordability.

³⁰. Stephen Malpezzi, “Housing Prices, Externalities, and Regulation in U.S. Metropolitan Areas,” Journal of HousingResearch 7, no. 2 (1996): 209-41.

³¹. Raven Saks, “Job Creation and Housing Construction: Constraints on Metropolitan Area Employment Growth,” Finance and Economics Discussion Series, Divisions of Research and Statistics and Monetary Affairs, Federal Reserve Board, Working Paper no. 2005-49, September 22, 2005.

³². Keith R. Ihlanfeldt, “The Effect of Land Use Regulation on Housing and Land Prices,” Journal of Urban Economics 61, no. 3 (May 2007).

³³. Chang-Tai Hsieh and Enrico Moretti, “Why Do Cities Matter? Local Growth and Aggregate Growth,” Econometrics Laboratory, University of California, Berkeley, Working Paper, April 2015.

³⁴. Rebecca Diamond, Timothy McQuade, and Franklin Qian, “The Effects of Rent Control Expansion on Tenants, Landlords, and Inequality: Evidence from San Francisco,” NBER Working Paper no. 24181, January 2018, http://www.nber.org/papers/w24181.

³⁵. Vanessa Brown Calder, “Zoning, Land-Use Planning, and Housing Affordability.”

³⁶. Bureau of Labor Statistics, Consumer Expenditure Survey 2016, Tables 3004, 3024, and 3033.

³⁷. Jordan Weissman, “So You’re Rich for an American. Does That Make You Rich for New York?” Slate Moneybox, August 29, 2014.

³⁸. Calculation by author subtracting Glaeser et al. estimate of implied regulatory tax for New York (12.2 percent) from average New York household expenditure on shelter of $16,882 in 2016 (taken from Consumer Expenditure Survey 2016, Table 3004). This implies regulation raises average household expenditure by $1,836 per year. Given those in the poorest quintile, on average, spend 57 percent of average household expenditure on shelter across the whole of the United States (Consumer Expenditure Survey 2016, Table 1101), this suggests these households pay $1,044 more for shelter than they would if the regulatory tax was zero.

³⁹. Child Care Aware of America, “2017 Appendices: Parents and the High Cost of Child Care,” Appendix I: 2016 Average Annual Cost of Full-Time Center-Based Child Care by State, http://usa.childcareaware.org/costofcare.

⁴⁰. Child Care Aware of America, “2017 Appendices: Parents and the High Cost of Child Care,” Appendix III: 2016 Ranking of Least-Affordable Center-Based Infant Care, http://usa.childcareaware.org/costofcare.

⁴¹. Dawn Lee, “State Child Care Assistance Programs,” Single Mother Guide, January 10, 2018, https://singlemotherguide.com/state-child-care-assistance/.

⁴². OECD, Society at a Glance 2016: OECD Social Indicators (Paris: OECD Publishing, 2016), Figure 1.14. Childcare costs are around 15% of net family income across the OECD. Out-of-pocket childcare costs for a single parent: full-time care at a typical childcare center, http://dx.doi.org/10.1787/9789264261488-en.

⁴³. Rachel Connelly and Jean Kimmel, “The Effect of Child Care Costs on the Employment and Welfare Recipiency of Single Mothers,” Southern Economic Journal 69, no. 3 (January 2003): 498-519, https://www.jstor.org/stable/1061691?seq=1#page_scan_tab_contents.

⁴⁴. Allison Linn, “Opt Out or Left Out? The Economics of Stay-at-Home Moms,” NBC News, May 12, 2013.

⁴⁵. Lynda Laughlin, “Who’s Minding the Kids? Childcare Arrangements: Spring 2011.” U.S. Census Bureau Current Population Reports no. P70-135, April 2013. Among children with employed mothers, those living below the poverty line were more than twice as likely to be cared for by an unlicensed relative (20.7 percent vs. 9 percent).

⁴⁶. Author’s calculations based on data from U.S. Bureau of Labor Statistics, “Women in the Labor Force: A Databook,” BLS Report no. 1071, November 2017, https://www.bls.gov/opub/reports/womens-databook/2017/pdf/home.pdf.

⁴⁷. Ryan Bourne and Len Shackleton, “Getting the State out of Preschool and Childcare,” Institute of Economic Affairs, February 6, 2017, https://iea.org.uk/publications/getting-the-state-out-of-pre-school-childcare/.

⁴⁸. Diana Thomas and Devon Gorry, “Regulation and the Cost of Child Care,” Mercatus Center Working Paper, August 17, 2015.

⁴⁹. Randal Heeb and M. Rebecca Kilburn, “The Effects of State Regulations on Childcare Prices and Choices,” RAND Labor and Population Working Paper no. WR-137-NICHD, January 2004, https://www.rand.org/content/dam/rand/pubs/working_papers/2004/RAND_WR137.pdf.

⁵⁰. V. Joseph Hotz and Mo Xiao, “The Impact of Regulations on the Supply and Quality of Care in Child Care Markets” American Economic Review 101, no. 5 (August 2011): 1775-1805.

⁵¹. Nicholas Clairmont, “D.C.’s Misguided Attempt to Regulate Daycare,” Atlantic, July 11, 2017.

⁵². Martin Austermuhle, “D.C. Delays New Degree Requirements for Childcare Workers,” WAMU, November 17, 2017.

⁵³. Author’s calculation applying Thomas and Gorry’s lower bound estimate of price fall from loosening staff-child ratios by one across the board (9 percent) to the average annual cost of the most expensive care in the country (D.C. at $23,089 per year) and the cheapest (Mississippi at $5,178). This implies prices would fall by around $2,078 in DC and $466 in Mississippi, assuming a linear relationship.

⁵⁴. Ryan Bourne and Len Shackleton, “Getting the State out of Preschool and Childcare.”

⁵⁵. Though not analyzed here, it is likely that the cost of childcare is also pushed up by other, broader government policies which have damaging economic consequences. Zoning and land-use ordinances quite often prohibit new facilities in residential areas, and home daycare has to comply with requirements on lot size, parking, and architectural regulations. Barriers to immigration from poorer countries also restrict the supply of lower-cost child-care workers.

⁵⁶. Chris Edwards, “Milk Madness,” Cato Institute Tax and Budget Bulletin no. 47, July 2007, https://www.cato.org/publications/tax-budget-bulletin/milk-madness.

⁵⁷. Owen Townsend, Kenneth Burdine, and Tyler Mark, “The History and Class Pricing of Federal Milk-Marketing Orders,” University of Kentucky Department of Agricultural Economics, Agricultural Economics Extension Series Number: 2017-13, https://www.uky.edu/Ag/AgEcon/pubs/extfedmilkord06.pdf.

⁵⁸. Hayley Chouinard et al., “Milk Marketing Order Winners and Losers,” Applied Economic Perspectives and Policy 32, no. 1 (March 2010): 59-76.

⁵⁹. The averaged data are calculated by Chouinard et al., “Milk Marketing Order Winners and Losers.”

⁶⁰. Bureau of Labor Statistics, Consumer Expenditure Survey 2016, Table 1101, Quintiles of income before taxes.

⁶¹. Bureau of Labor Statistics, Consumer Expenditure Survey 2016, Table 1502, Composition of consumer unit: Annual expenditure means, shares, standard errors, and coefficients of variation.

⁶². Calculated by author. Based on a regulatory cost of $44 for a white household earning $10,000 and $38 for a household earning $20,000, a straight-line extrapolation suggests the cost to a household with average income for the quintile ($11,363) would be $43. Assuming this cost holds for all nonblack households, and that the burden is one-third of this for black households, one can use the proportion of black to nonblack households in the bottom quintile (20:80) to calculate the overall average regulatory burden for the quintile (80% x $43 + 20% x $14) = $37.51. Given that we know the ratio of spending by the average single-parent household to the average household in the poorest quintile on dairy is 1.43, the average regulatory cost for single-parent households can be estimated $37.51 x 1.43 = $54.31.

⁶³. Colin Grabow, “Candy-Coated Cartel: Time to Kill the U.S. Sugar Program,” Cato Institute Policy Analysis no. 837, April 10, 2018, https://www.cato.org/publications/policy-analysis/candy-coated-cartel-time-kill-us-sugar-program.

⁶⁴. U.S. Department of Agriculture Economic Research Service, “Sugar and Sweeteners,” https://www.ers.usda.gov/topics/crops/sugar-sweeteners/.

⁶⁵. Michael K. Wohlgenant, “Sweets for the Sweet: The Costly Benefits of the U.S. Sugar Program,” American Enterprise Institute, July 12, 2011.

⁶⁶. John C. Beghin and Amani Elobeid, “Analysis of the US Sugar Program,” Agricultural Policy in Disarray: Reforming The Farm Bill, American Enterprise Institute, November 2017, http://www.aei.org/publication/analysis-of-the-us-sugar-program/.

⁶⁷. This number is calculated by dividing the $2.8 billion and $4.7 billion range of consumer cost by 129.55 million household units, as per Table 1101 of the Consumer Expenditure Survey.

⁶⁸. Elyse S. Powell, Lindsey P. Smith-Taillie, and Barry M. Popkin, “Added Sugars Intake across the Distribution of US Children and Adult Consumers: 1977-2012,” Journal of theAcademyofNutritionand Dietetics 116, no. 10 (October 2016): 1543-50.

⁶⁹. Congressional Budget Office, “The Impact of Ethanol Use on Food Prices and Greenhouse-Gas Emissions,” April 2009, https://www.cbo.gov/publication/41173, pp. 8-10.

⁷⁰. Richard K. Perrin, “Ethanol and Food Prices-Preliminary Assessment,” Faculty Publications: Agricultural Economics 49, University of Nebraska-Lincoln, May 9, 2008.

⁷¹. Congressional Budget Office, “The Renewable Fuel Standard: Issues for 2014 and Beyond,” June 2014, https://www.cbo.gov/publication/45477.

⁷². Calculations by author based on food expenditure data from Table 1101, Quintiles of income before taxes, and Table 1502, Composition of consumer unit, from Consumer Expenditure Survey.

⁷³. Bureau of Labor Statistics, Consumer Expenditure Survey 2016, Table 1101, Quintiles of income before taxes.

⁷⁴. U.S. Census Bureau, 2012-2016 American Community Survey 5-Year Estimates, DP04 Selected Housing Characteristics.

⁷⁵. Daniel C. Vock, “More Poorer Residents Are Driving Cars, Presenting New Issues for Transit Agencies,” Governing.com, April 9, 2018, http://www.governing.com/topics/transportation-infrastructure/gov-car-ownership-poverty.html.

⁷⁶. Lisa M. Brabo et al., “Driving Out of Poverty in Private Automobiles,” Journal of Poverty 7, no. 1-2 (2003): 183-96, https://www.researchgate.net/publication/239804081_Driving_Out_of_Poverty_in_Private_Automobiles.

⁷⁷. For a summary of recent research, see Peter Van Doren, “Regulation without Results,” U.S. News and World Report, March 27, 2017.

⁷⁸. Timothy Puko, Mike Spector, and Chester Dawson, “EPA Will Ease Vehicle-Emissions Standards,” Wall Street Journal, April 2, 2018.

⁷⁹. Timothy Puko, “Trump Administration May Eliminate Increases in Fuel-Economy Standards,” Wall Street Journal, April 27, 2018.

⁸⁰. For a full exposition of this argument, read Arik Levinson, “Energy Efficiency Standards Are More Regressive Than Energy Taxes: Theory and Evidence,” May 8, 2018, http://faculty.georgetown.edu/aml6/pdfs&zips/RegressiveMandates.pdf. Empirical evidence at Lucas W. Davis and Christopher R. Knittel, “Are Fuel Economy Standards Regressive?” NBER Working Paper no. 22925, December 2016, http://www.nber.org/papers/w22925 and Mark R. Jacobsen, “Evaluating US Fuel Economy Standards in a Model with Producer and Household Heterogeneity,” AmericanEconomicJournal 5, no. 2 (May 2013): 148-87, https://www.aeaweb.org/articles?id=10.1257/pol.5.2.148.

⁸¹. Davis and Knittel, “Are Fuel Economy Standards Regressive?”

⁸². Adjusted by Consumer Price Index for All Urban Consumers: All Items. Federal Reserve Economic Data, Federal Reserve Bank of St. Louis, https://fred.stlouisfed.org/series/CPIAUCSL.

⁸³. Mark R. Jacobsen, “Evaluating US Fuel Economy Standards in a Model with Producer and Household Heterogeneity.”

⁸⁴. Calculations averaging standards in 2011, 2018, and 2025, assuming two-thirds of vehicles are ordinary cars and one-third light trucks. This would see average miles-per-gallon standards increase from 28.3 in 2011 to 38.3 in 2018, and then further to 50 by 2025.

⁸⁵. David L. Greene, “Short-Run Pricing Strategies to Increase Corporate Average Fuel Economy,” Economic Inquiry 29, no. 1 (January 1991): 101-14. Paper shows for every one-mile-per-gallon increase in vehicle fuel economy, the average per-vehicle cost would be within the range of $100 to $200 in constant 1985-dollar terms, adjusted to 2018 prices. For a more detailed review of the literature see Thomas Klier and Joshua Linn, “Corporate Average Fuel Economy Standards and the Market for New Vehicles,” AnnualReviewofResourceEconomics 3, no. 1 (October 2011): 445-62.

⁸⁶. Rob Looker, “Average New- and Used-Car Prices, and the Advantages of Flexible Financing,” RoadLoans.com, March 1, 2018, https://roadloans.com/blog/average-car-price.

⁸⁷. Net outlay on vehicle purchases averages $1,331 per year for the average household in the poorest quintile in 2016.

⁸⁸. Mark R. Jacobsen and Arthur A. van Bentham, “Vehicle Scrappage and Gasoline Policy,” NBER Working Paper no. 19055, May 2013, http://www.nber.org/papers/w19055.

⁸⁹. David Austin and Terry Dinan, “Clearing the Air: The Costs and Consequences of Higher CAFE Standards and Increased Gasoline Taxes,” Journal of Environmental Economics andManagement 50, no. 3 (November 2005): 562-82, https://www.sciencedirect.com/science/article/pii/S0095069605000550.

⁹⁰. Julian Morris, “The Effect of Corporate Average Fuel Economy Standards on Consumers,” Reason Foundation Policy Brief, April 1, 2018, https://reason.org/policy-brief/the-effect-of-corporate-average-fuel-economy-standards-on-consumers/.

⁹¹. For a more comprehensive examination of these issues, see Francine Lafontaine and Fiona Scott Morton, “State Franchise Laws, Dealer Terminations, and the Auto Crisis,” Journal ofEconomicPerspectives 24, no. 3 (Summer 2010): 233-50, StateFranchiseLawsDealerTerminationsandtheAutoCrisis.pdf”>http://faculty.som.yale.edu/FionaScottMorton/documents/StateFranchise”>StateFranchiseLawsDealerTerminationsandtheAutoCrisis.pdf.

⁹². Richard L. Smith II, “Franchise Regulation: An Economic Analysis of State Restrictions on Automobile Distribution,” Journalof Law and Economics 25, no. 1 (April 1982): 150.

⁹³. Robert P. Rogers, “The Effect of State Entry Regulation on Retail Automobile Markets,” Bureau of Economics Staff Report to the Federal Trade Commission, January 1986.

⁹⁴. Mark Cooper, “A Roadblock on the Information Superhighway: Anticompetitive Restrictions on Automotive Markets,” Consumer Federation of America, February 2001, https://consumerfed.org/pdfs/internetautosales.pdf.

⁹⁵. Brian Shaffer, “An Assessment of Franchise Laws and Internet Auto Sales,” National Automobile Dealers Association Public Affairs, August 2001.

⁹⁶. See, for example, Frank Mathewson and Ralph Winter, “The Economic Effects of Automobile Dealer Regulation,” Annalesd’Économie et de Statistique, no. 15/16 (July-December 1989): 409-26, http://www.jstor.org/stable/20075766; and Gerald Bodisch, “Economic Effects of State Bans on Direct Manufacturer Sales to Car Buyers,” U.S. Department of Justice Economic Analysis Group, May 2009, http://www.justice.gov/atr/economic-effects-state-bans-direct-manufacturer-sales-car-buyers.

⁹⁷. Assumes new vehicle prices are increased by 6 percent and used car prices by 4 percent as a result of the regulations. Average spending on new cars and trucks is $462 and $843 on used for those in the lowest quintile. This implies the cost of the regulations for the average household in the poorest quintile is $27.72 and $33.72 on new and used vehicles, respectively, summing to $61.44; for the average single-parent household these figures are $55.74 and $84.60, summing to $140.34.

⁹⁸. Bureau of Labor Statistics, Consumer Expenditure Survey 2016, Table 1101, Quintiles of income before taxes.

⁹⁹. Bureau of Labor Statistics, Consumer Expenditure Survey 2016, Table 1502, Composition of consumer unit.

¹⁰⁰. U.S. International Trade Commission, “Interactive Tariff and Trade DataWeb,” at http://dataweb.usitc.gov. Data for imports for consumption, and effective rates calculated using “customs value” and “calculated duties” for 2017.

¹⁰¹. See, for example, Daniel Ikenson, “Threadbare Excuses: The Textile Industry’s Campaign to Preserve Import Restraints,” Cato Institute Trade Policy Analysis no. 25, October 15, 2003, https://www.cato.org/publications/trade-policy-analysis/threadbare-excuses-textile-industrys-campaign-preserve-import-restraints; and Ikenson, “Cutting the Cord: Textile Trade Policy Needs Tough Love,” Cato Institute Free Trade Bulletin no. 15, July 23, 2013, https://www.cato.org/publications/free-trade-bulletin/cutting-cord-textile-trade-policy-needs-tough-love.

¹⁰². Jason Furman, Katheryn Russ, and Jay Shambaugh, “U.S. Tariffs Are an Arbitrary and Regressive Tax,” Vox.com, CEPR’s Policy Portal, January 12 2017, https://voxeu.org/article/us-tariffs-are-arbitrary-and-regressive-tax. See in particular Technical Appendix to the CEX-HS Crosswalk and Matched Effective Tariffs.

¹⁰³. Federal Reserve Economic Data, Federal Reserve Bank of St. Louis, All Employees: Nondurable Goods: Textile Mills, https://fred.stlouisfed.org/series/CEU3231300001.

¹⁰⁴. Daniel Ikenson, “Washington’s Coddling of U.S. Textile Industry Is Hurting Shoppers,” Forbes, July 23, 2013.

¹⁰⁵. Edward Gresser, “Toughest on the Poor: America’s Flawed Tariff System,” Foreign Affairs, November/December 2002.

¹⁰⁶. The average single-parent household spends $1,402 on apparel and $512 on footwear per year. Assuming the import propensities for the population as a whole apply to single-parent households, this would mean $1,274 of apparel spending and $494 of footwear spending is on imported goods. Taking average effective tariff rates for apparel and footwear for this spending (13.7 and 11.3 percent) implies a direct cost of protectionism of $204 per year.

¹⁰⁷. Jason Furman, Katheryn Russ, and Jay Shambaugh, “U.S. Tariffs Are an Arbitrary and Regressive Tax.”

¹⁰⁸. National Conference of State Legislatures, “The National Occupational Licensing Database”; Morris M. Kleiner, “Reforming Occupational Licensing Policies,” Brookings Institution Discussion Paper No. 2015-01, March 2015; and Maury Gittleman, Mark Klee, and Morris Kleiner, “Analyzing the Labor Market Outcomes of Occupational Licensing,” IndustrialRelations 57, no.1: 57-100.

¹⁰⁹. Maury Gittleman, Mark Klee, and Morris Kleiner, “Analyzing the Labor Market Outcomes of Occupational Licensing.”

¹¹⁰. Morris M. Kleiner et al., “Relaxing Occupational Licensing Requirements: Analyzing Wages and Prices for a Medical Service,” NBER Working Paper no. 19906, February 2014.

¹¹¹. Brandon Pizzola and Alex Tarrabok, “Occupational Licensing Causes a Wage Premium: Evidence from a Natural Experiment in Colorado’s Funeral Services Industry,” International Review of Law and Economics 50, issue C, 50-59, 2017.

¹¹². Nellie J. Liang and Jonathan D. Ogur, “Restrictions on Dental Auxiliaries: An Economic Policy Analysis,” Federal Trade Commission, 1987; and Deborah Haas-Wilson, “The Effect of Commercial Practice Restrictions: The Case of Optometry,” Journal of Law & Economics 29, no. 1 (1986): 165-86.

¹¹³. Morris Kleiner, Alan B. Krueger, and Alex Mas, “A Proposal to Encourage States to Rationalize Occupational Licensing Practices,” A Proposal to the Brookings Institution Hamilton Project, April 2011, https://www.hhh.umn.edu/file/9441/download.

¹¹⁴. Salim Furth, “Costly Mistakes: How Bad Policies Raise the Cost of Living,” Backgrounder No. 3081, Heritage Foundation, November 23 2015.

¹¹⁵. Author’s calculations based on the average cost of licensure per household estimates from Kleiner et al. and Furth and adjusting for spending levels of the poor and single parents using data from the Consumer Expenditure Survey, Tables 1101 and 1502.

¹¹⁶. Author’s calculations using estimates outlined in Table 4. The lower estimate assumes $0 regulatory cost for housing and childcare and the lower end of the range presented for the cost of occupational licensure. The higher estimate assumes a $1,044 cost of housing regulation; a $1,350 cost of child-care regulation (using calculations for New York State from Child Care Aware of America); and the top end of the occupational licensure cost range.

Chris Edwards

The Tax Cuts and Jobs Act of 2017 was the largest overhaul of the federal income tax in decades. The law changed deductions, exemptions, and tax rates for individuals, while reducing taxes on businesses.

More than 86 percent of middle- and higher-income households received an individual tax cut.¹ Most lower-income households do not pay income taxes, but many of them received increased benefits from refundable credits. The average benefit across all households in 2018 is $1,260.²

This report looks at changes to individual income taxes, particularly the state and local tax (SALT) deduction. The 2017 tax law cut individual tax rates and roughly doubled standard deductions, but it also imposed a $10,000 cap per return on SALT deductions. Those changes are expected to reduce the number of households that deduct state and local income, sales, and property taxes from 42 million in 2017 to 17 million in 2018.³

Millions of households will feel a larger bite from state and local taxes and will thus become more sensitive to tax differences between the states. The tax law may prompt an outflow of mainly higher-earning households from higher-tax states to lower-tax states.

Even before the new tax law, a substantial number of Americans were moving from higher-tax to lower-tax states. Looking at migration flows between the states in 2016, almost 600,000 people with aggregate income of $33 billion moved, on net, from the 25 highest-tax states to the 25 lowest-tax states in that single year.

Interstate migration flows are influenced by many factors, including retirement, job opportunities, housing costs, and climate. Experts disagree about how large a role taxes play in migration, but that role will certainly be increased by the new tax law.

The raw data suggest that taxes do influence migration. Of the 25 highest-tax states, 24 of them had net out-migration in 2016. Of the 25 lowest-tax states, 17 had net in-migration. The largest out-migration is from high-tax New York, whereas the largest in-migration is to low-tax Florida. Florida is enjoying an influx of wealthy entrepreneurs and retirees looking for a tax climate that boasts no income tax or estate tax.

The following sections discuss changes to the SALT deduction and examine trends in interstate migration. Then, the report looks at the relationship between taxes and migration. The out-migration of high earners is a serious threat to high-tax states because those individuals pay a large share of state income taxes, invest in new businesses and generate jobs, and are heavily engaged in philanthropy.

In this new era of intensified tax competition, state policymakers should rethink their tax codes with an eye toward retaining and attracting residents. They should improve the efficiency of government services to give taxpayers more value for their money. And they should reduce regulations on individuals and businesses, given that Americans are migrating, on net, to states that provide more economic freedom.

State and Local Tax Deductions

Before the passage of the 2017 tax law, individuals had no direct limit on the amount of state and local taxes they could deduct on their federal returns.⁴ The deduction was available to households that itemized deductions, which in 2017 was 27 percent of tax filers. The other 73 percent took the standard deduction.⁵ Of households that itemized, 93 percent took the SALT deduction.

The effect of the SALT deduction was to soften the blow of state and local income, property, and sales taxes. For example, taxpayers in New York in the 33 percent federal bracket in 2017 who paid $30,000 in state and local income and property taxes could reduce their federal taxes by $10,000. The federal government essentially gave them a rebate of that amount.

Meanwhile, high-income taxpayers in California subject to the 39.6 percent federal tax rate and a state rate of 13.3 percent effectively faced just an 8.0 percent state rate because of federal deductibility. The higher the household’s income, the larger the effective federal subsidy.

The 2017 tax law changed that. Most higher-income taxpayers will now face the full brunt of state and local taxes-the full $30,000 cost in New York and the full 13.3 percent rate in California.

Before the law change, federal deductibility subsidized high-tax states and encouraged them to load their taxes onto higher earners. The SALT deduction induced “shifting of the jurisdiction’s tax burden to those individuals best positioned to receive the federal tax subsidy,” which were high earners because they were generally the ones who itemized.⁶

Taxes are the “price” residents pay for state and local services such as police and schools, but the SALT deduction effectively reduced that price, thus inducing residents to demand too much spending. The Congressional Budget Office noted of the SALT deduction, “Because of the subsidy, too many of those services may be supplied, and state and local governments may be bigger as a result.”⁷

The prior SALT deduction mainly benefited higher earners. Before the law change, 91 percent of the benefit went to households with incomes above $100,000.⁸ The deduction favored higher-income and higher-tax states over other states. In California, 96 percent of state and local deductions that exceeded $10,000 were taken by households with incomes above $100,000.⁹

The new SALT limit is a long-needed reform. Leading up to the Tax Reform Act of 1986, the Reagan administration proposed eliminating the deduction, with President Reagan arguing, “Perhaps if the high-tax states didn’t have this federal crutch to prop up their big spending, they might have to cut taxes to stay competitive.”¹⁰ The 1986 law did eliminate the deductibility of sales taxes, but Congress added back that deduction in 2004.

The 2017 law capped the SALT deduction at $10,000 per year, for both single and married tax filers. It also nearly doubled standard deductions. Those changes will reduce the number taking the SALT deduction from 42 million in 2017 to 17 million in 2018.¹¹ For people who continue to take it, the average benefit will be less than half as large. The $10,000 cap is not adjusted for inflation.

The full weight of state and local taxes will now be felt by an additional 25 million households. The states where SALT deductions were the largest relative to incomes were New York, New Jersey, Connecticut, California, Maryland, Oregon, Rhode Island, Massachusetts, and Minnesota.¹² Those are generally high-tax states.

Governments in high-tax states are worried that the SALT reform will induce additional high-earning taxpayers to move out. If they move in substantial numbers, it would be a blow to state budgets. In New York, the top 1 percent of highest earners pay 41 percent of state income taxes; in New Jersey, the share is 37 percent.¹³ In California, the top 1 percent pay a remarkable 50 percent of state income taxes.¹⁴

Under the new federal law, some states are becoming more cautious about raising taxes. New Jersey’s legislature passed bills to raise taxes on millionaires five times under former governor Chris Christie. Christie vetoed them. But now that New Jersey has a governor eager to raise taxes on millionaires, the legislature has shied away. State Senate President Steve Sweeney (D), who previously supported higher taxes, earlier this year cautioned, “This state is taxed out” and the federal tax law “changed the game for us.”¹⁵ But in July, Sweeney caved, agreeing with the governor to raise the top individual income tax rate from 8.97 percent to 10.75 percent.

To shield themselves from the SALT changes, some states are considering converting part of their mainly nondeductible income taxes into deductible employer payroll taxes. Other states are trying to convert state tax payments into charitable contributions, which continue to be fully deductible. New York has enacted limited versions of those mechanisms, but the schemes may not be effective or pass muster under federal tax law. The U.S. Treasury has issued a notice saying that it will be imposing regulations.

A more productive response to the federal tax changes would be for high-tax states to reduce their tax rates so that people have less incentive to migrate. State policymakers should use federal tax reform as an opportunity to rethink their state budgets to ensure that residents receive high-value services at minimum cost.

Will more people in higher-tax states move to lower-tax states under the new federal tax law? The next section looks at current migration trends to provide some clues.

Trends in Interstate Migration

This report examines interstate migration data produced by the Internal Revenue Service (IRS).¹⁶ The IRS flags when a tax filer’s address changes and has built a database aggregating moves in and out of every county and state. The database includes each tax filer’s adjusted gross income (AGI).

The IRS data show that 2.8 percent of tax filers-essentially households-moved between states in 2016. Tax-filing households may be either singles or married couples, with or without children. An average household comprises 2.1 people in the IRS data.¹⁷

The IRS data do not include households that do not file tax returns, so it misses about 13 percent of the population.¹⁸ However, the data are quite precise because they are not based on survey data, as are migration data produced by the U.S. Census Bureau.

Experts agree on the basics of interstate migration.¹⁹ The migration rate has dropped since the 1980s, although the IRS data show a smaller drop than the census data.²⁰ The internal migration rate is higher in the United States than in most other high-income countries. Migration rates decline with age. Renters are more likely to move than homeowners. Singles are more likely to move than married couples. Migration is somewhat pro-cyclical.

Table 1, column 1, shows the number of net domestic in-migrants (in-migrants less out-migrants) for each state in 2016, based on the IRS data. All data in this report exclude international immigration and emigration.

New York lost 218,937 households to other states in 2016, gained 142,722 households from other states, and thus had a net loss of 76,215.

The other states with the largest net migration losses were Illinois (41,965 households), New Jersey (25,941), California (25,913), Pennsylvania (19,516), Massachusetts (14,549), Ohio (13,254), Connecticut (12,254), Maryland (12,068), and Michigan (10,325).

Where did those domestic migrants go? The largest net inflows were to Florida (95,072), Washington state (30,480), North Carolina (25,601), Colorado (24,672), Arizona (24,211), Oregon (21,729), Texas (19,414), South Carolina (18,519), Georgia (17,798), and Nevada (14,236).

For nearly all states, the 2016 migration flows represent extended trends. For 48 of the 50 states, the net direction of migration (in or out) in 2016 matched the direction of the total net flows over the past five years. The two states that were different were oil producers Oklahoma and North Dakota, which used to have net in-migration but now have net out-migration. Also note that Texas had abnormally low in-migration in 2016-its net inflows over the past five years have been larger than Florida’s.

For many states, recent trends extend back decades. The largest net migration losers between 1993 and 2010 were New York, California, Illinois, Michigan, and New Jersey. The largest net migration winners over that period were Florida, Arizona, Texas, North Carolina, and Georgia.²¹

Table 1, column 2, shows the ratio of gross in-migration to gross out-migration in 2016. States losing population have ratios of less than 1.0. States gaining population have ratios of more than 1.0. New York’s ratio is 0.65, meaning that for every 100 households that left, only 65 moved in. Florida’s ratio is 1.45, meaning that 145 households moved in for every 100 that left.

Table 1: Interstate migration flows, 2016

Source: Author’s calculations based on Internal Revenue Service data.

The IRS database includes AGI, so we can see how much aggregate income is migrating between the states. The AGI is the income reported in the first year a household is at a new address. Strictly speaking, AGI does not migrate; people do. An individual’s income may be higher or lower after moving to a new state. Nonetheless, saying that income is “migrating” is rough shorthand for saying that the earning power of households is moving between states.²²

In 2016, households with $227 billion of income moved between states. Looking at Table 1, column 3, New York lost a net $8.4 billion in income to other states in 2016, whereas Illinois lost $4.8 billion. Florida gained $17.2 billion.

The ratios in column 4 are similar to those in column 2, except that they are ratios of income, not households. Florida’s household ratio is 1.45, and its income ratio is 2.46. The larger income ratio means that in-migrants to Florida have much higher incomes than out-migrants do. Other states with a much higher income ratio than household ratio include Idaho, Montana, South Carolina, Utah, and Wyoming. Those states are attracting large numbers of high-income in-migrants compared with their out-migrants.

When the column 4 ratio is lower than the column 3 ratio, it means that the state has net out-migration particularly of higher-income households. Connecticut is a good example.

Table 2 shows net migration ratios for households headed by persons age 65 and older and for households with incomes of more than $200,000. Some of the places that both seniors and high earners are leaving, on net, are Alaska, California, Connecticut, Illinois, Maryland, New Jersey, New York, North Dakota, Pennsylvania, and West Virginia. Some of the places that the two groups are moving to, on net, are Arizona, Colorado, Florida, Idaho, Montana, Nevada, North Carolina, Oregon, South Carolina, Tennessee, Utah, and Washington State.

Table 2: Ratios of in-migrants to out-migrants, seniors and high earners, 2016

Source: Author’s calculations based on Internal Revenue Service data.

Why Do People Move?

An annual Census Bureau survey asks people who move any distance the main reason for their decision out of 19 choices. The most popular choices in 2017 were “wanted new or better home” (16.0 percent), “to establish own household” (11.5 percent), “other family reason” (11.3 percent), “new job or job transfer” (9.9 percent), and “wanted cheaper housing” (8.3 percent).²³

The Census Bureau does not ask movers about taxes. But some of the 19 choices may reflect the influence of taxes. For example, people moving for housing reasons may consider the level of property taxes since those taxes are a standard item listed on housing sale notices. Similarly, people moving for new jobs may consider the effect of income taxes if they are, for example, moving between a high-tax state such as California and a state with no income tax such as Nevada.

A national survey by Bankrate found that taxes play a substantial role in retirement location decisions.²⁴ Based on their survey, Bankrate weighted the location choice factors as follows: cost of living (20 percent), taxes (20 percent), health care quality (15 percent), weather (15 percent), crime (10 percent), cultural vitality (10 percent), and well-being (10 percent). Bankrate found that 47 percent of Americans would consider moving when they retire.

Figure 1 shows that many people are moving from northern states to southern states. Liberal analysts typically attribute that fact to people wanting to live in warmer states. Conservative analysts typically attribute it to people wanting to live in lower-tax states. Looking at the Census Bureau survey data for interstate moves only, of the 19 choices, only 2.2 percent chose “change of climate” as the move reason.²⁵ That is surprisingly low. Apparently, there is more to the popularity of many southern states than just higher temperatures.

Figure 1: Interstate migration flows, ratios of in-migrants to out-migrants, 2016

Source: Author’s calculations based on Internal Revenue Service data.

Americans Are Moving to Lower-Tax States

Americans are moving from higher-tax states to lower-tax states in substantial numbers. That is clear from the raw migration data discussed here.

State and local tax revenues averaged 10.1 percent of personal income in the nation in 2015, according to the Census Bureau.²⁶ Sales taxes were 3.5 percent, property taxes were 3.1 percent, individual income taxes were 2.4 percent, and corporate and other taxes were 1.1 percent. Sales taxes include general sales taxes and selective sales taxes on products such as gasoline, alcohol, and cigarettes.

Which of those taxes might influence individual migration decisions?

Polls have asked Americans their “most disliked” taxes.²⁷ The most disliked state and local tax has long been the property tax. After that, Americans dislike sales taxes and individual income taxes. Other polls have asked which taxes are the least “fair.” Property and various selective sales taxes are often the top responses, followed by individual income and general sales taxes.

Thus, sales, property, and individual income taxes likely have the most influence on migration decisions, as they are the largest state-local taxes and the most disliked. Corporate taxes are less disliked in polls, which is not surprising because they are less visible to the public.

Table 3 shows data for the combined net migration flows for the 25 highest-tax and 25 lowest-tax states. Taxes are measured as state and local sales, property, and individual income taxes as a percentage of state personal income. In 2016, 578,269 people moved from the highest-tax states and the District of Columbia to the lower-tax states, on net. Of the 25 highest-tax states, 24 had net out-migration. Of the 25 lowest-tax states, 17 had net in-migration.

Table 3: Net migration from high-tax to low-tax states, 2016

Source: Author’s calculations based on Internal Revenue Service data.

The tax gap between the 25 highest and 25 lowest states may not seem large at 2.2 percentage points of income. But many of the largest migration flows are between the states with the very highest and very lowest taxes. The largest outflow state, New York, has a tax burden by this measure of 13.0 percent, whereas the largest inflow state, Florida, has a tax burden of 6.6 percent.

Figure 2 shows the relationship between tax levels and migration ratios. The migration ratios are from Table 1, column 2. The tax variable is the average state and local sales, property, and individual income taxes as a percentage of personal income.²⁸

Figure 2: Tax levels and net migration ratios, 2016

Source: Author’s calculations based on data from the Internal Revenue Service and U.S. Census Bureau.

The figure shows a clear negative relationship between tax levels and migration. On the left, states have lower taxes and net in-migration (a ratio greater than 1.0). On the right, states have higher taxes and net out-migration (a ratio less than 1.0).

There were 17 states that had net in-migration in 2016 (a ratio of more than 1.0). Of those, 17 had a tax burden of less than 8.5 percent.

Of the 26 states with a tax burden of 8.5 percent or greater, 25 of them had net out-migration. The only high-tax state with in-migration was Maine. (The District of Columbia had a migration ratio of 1.0.)

Figure 2 shows a fitted regression line. A simple regression of the migration ratio on the tax variable produces a highly statistically significant fit. The F-statistic (12.1) and t-statistic (3.5) are significant above the 99 percent level. State tax levels and net migration flows are highly correlated.

Here are some patterns in the interstate migration flows:²⁹

• The Northeast. New Hampshire enjoyed net in-migration in four of the past five years of IRS data (2012 to 2016). It is a low-tax state with no personal income tax or general sales tax. Nearby, higher-tax Massachusetts, Rhode Island, and Vermont suffered net out-migration all five years. New Hampshire enjoys net in-migration from all three of those states.
• The Midwest. South Dakota has enjoyed net in-migration in four of the past five years. By contrast, its higher-tax neighbors Nebraska, Iowa, and Minnesota had net out-migration all five years, and each had a migration deficit with South Dakota. South Dakota is one of the lowest-tax states and has no income tax.
• The Southeast. Kentucky has suffered net out-migration in each of the past five years, whereas Tennessee has enjoyed net in-migration every year, including from its neighbor. Kentucky is a relatively high-tax state, whereas Tennessee is one of the lowest-tax states and has no personal income tax.
• The West. The three largest destinations for California out-migrants in 2016 were Texas, Washington State, and Nevada-all low-tax states with no income taxes. California has a large migration deficit with all three states.

One interesting pattern that affects high-tax states across the nation is that the net migration ratio gets worse for older age groups. For example, California’s migration ratio for people age 26-34 is 0.92, but the ratio for age 55-64 is 0.60.³⁰ Connecticut, Illinois, Maryland, Massachusetts, Minnesota, New Jersey, New York, and Ohio show this same pattern. It appears that older people with higher incomes and higher taxes are even less willing to move to such states than young people with lower incomes and lower taxes.

Taxes are more likely to influence moves when interstate differences are large-and the differences between the highest- and lowest-tax states are large. The District of Columbia government produces an annual study comparing state and local taxes on hypothetical households at various income levels in the largest city in each state.³¹ The study includes sales, property, individual income, and automobile taxes.

Table 4 highlights some of the results. Families earning $75,000 a year could save about $5,000 a year by moving from a high-tax city to a low-tax city. Families earning $150,000 could save about $10,000 with such a move.³² Those differences would seem to be large enough to influence some people to move.

Table 4: Household taxes for the largest city in each state
State and local sales, property, individual income, and automobile taxes, 2016

Source: Government of the District of Columbia, “Tax Rates and Tax Burdens in the District of Columbia: A Nationwide Comparison, 2016,” December 2017.

People who are thinking about moving can easily learn about state tax differences. Two recent Kiplinger.com stories were “Best States to Move to in 2018 for Lower Taxes” and “The 10 Most Tax-Friendly States in the U.S.”³³ Another resource is WalletHub’s “2018’s Tax Burden by State,” which compares income, sales, property, and excise taxes.³⁴

Of course, many other factors influence interstate migration, and those factors are complex and sometimes interrelated. If high taxes in a state buy high-quality services such as good schools, then those services will draw migrants willing to put up with the higher taxes. That said, no clear relationship exists between tax levels and the quality of government services across the states.³⁵ In some states, tax revenues may be used efficiently to pay for quality services that residents and in-migrants want. In other states, tax revenues may be dissipated on high-cost bureaucracies or misallocated to activities that most people do not want.

Presumably, most people consider a combination of factors when moving. A recent CNBC article (“Californians Fed Up with Housing Costs and Taxes Are Fleeing State in Big Numbers”) suggests that high costs for housing, taxes, and gasoline were all pushing people out of California.³⁶ Comparable apartments cost twice as much in Los Angeles as in Las Vegas and Phoenix, and gasoline is a dollar a gallon more expensive in California than in Texas, partly because of taxes.³⁷ Although California is viewed as having perhaps the nicest weather in the nation, it has suffered from domestic out-migration for many years.

Economists have used regression analysis to determine the specific factors that drive interstate migration. However, the research has not come to any clear answers with respect to taxes-numerous studies have found substantial migration effects, but some have not.

Since the 1970s, economist Richard Cebula has been publishing statistical studies that attribute interstate migration flows to a range of factors, including state economic growth, housing costs, taxes, climate, crime rates, and public school spending.³⁸ Cebula has consistently found that income and property tax burdens are statistically significant in explaining interstate migration.

A 2012 study by Cebula and Usha Nair-Reichert, for example, performed regression analyses on the determinants of interstate migration between 2000 and 2008.³⁹ Controlling for state employment growth, unemployment, January temperatures, education spending, and the cost of living, they found that income and property taxes per capita were significant in explaining migration. Their results, they say, confirm the Tiebout hypothesis that people “vote with their feet” and move to jurisdictions offering better fiscal bargains.⁴⁰

Numerous other statistical studies have found that state and local taxes affect interstate migration, including those by Mark Gius; Yu Hsing; Robert Preuhs; Karen Conway Smith and Andrew Houtenville; David Clark and William Hunter; Antony Davies and John Pulito; and Roger Cohen, Andrew Lai, and Charles Steindel.⁴¹

New Jersey increased its top individual income tax rate from 6.37 percent to 8.97 percent in 2004. A statistical study by economists in the state’s Department of Treasury found that the hike induced net out-migration of 80 or more millionaires a year.⁴² That would be a modest effect, but the drain would add up over time if sustained.

With federal deductibility, the 2004 New Jersey increase of 2.6 percentage points was equivalent to an effective increase of 1.6 percentage points. By comparison, the 2017 tax law ended deductibility for most taxpayers, thus increasing the effective top New Jersey income tax rate by 3.6 percentage points.⁴³ Then in 2018, New Jersey hiked its top individual income tax rate from 8.97 percent to 10.75 percent, so we should expect larger outflows in coming years than after 2004.

Economists Cristobal Young and Charles Varner found a smaller effect from the 2004 New Jersey tax increase, and Young and coauthors on other studies have found that taxes have only small effects on the interstate migration of millionaires.⁴⁴ The authors argue that millionaires are not particularly footloose because they are “embedded” in their communities. That is, they often have social and business dealings in their states that make moving difficult.

In sum, numerous statistical studies have found that taxes affect interstate migration, but some studies have contrary findings. One reason for the mixed findings could be the mechanism of “capitalization.” A tax increase in one jurisdiction may cause an initial out-migration to other jurisdictions. That flow will reduce property values in the tax-increasing jurisdiction and raise property values in other jurisdictions. Those property value changes will ultimately stem the migration flow as the economy enters a new equilibrium. If differences in state taxes are mainly capitalized, then the related migration flows will be mitigated.

Wage adjustments may also offset the migration effects of taxes.⁴⁵ An income tax increase in a state may cause individuals to out-migrate over time. As that happens, gross wages would rise in the tax-increasing state relative to other states, and that would eventually stem the outflow.

A wide range of policy and amenity differences between states-not just taxes-may be capitalized in asset prices or offset by gross wage differentials. To the extent that those market adjustments occur, migration would decrease.

Nonetheless, migration does happen, as we have seen. State policies, individual preferences, and other factors are always changing, and incentives apparently change enough each year for more than 2 percent of U.S. households to move to a different state.

A Closer Look at High Earners

Households with incomes of more than $200,000 were 5 percent of all interstate movers in 2016, but they accounted for 36 percent of the income of all interstate movers.

The 2017 tax law changed migration incentives for this group. The large gap between New York City’s 12.7 percent top income tax rate and Florida’s 0 percent has been laid bare, as has the gap between California’s 13.3 percent top rate and the 0 percent rate in Texas, Nevada, and Washington State.⁴⁶

IRS data for 2016 show that the highest interstate migration rates are for households with the very lowest incomes, but that is because most of those households consist of young and often single people. When you look within each particular age category, the migration rate is much greater for high earners than it is for middle-income earners. For example, for people between ages 45 and 54, the migration rate for households with incomes between $50,000 and $200,000 was 1.4 percent, but the rate for households with incomes above $200,000 was 2.1 percent. So the highest earners are relatively mobile.

Another important fact about high earners is that they are more responsive to tax rates in general than are other individuals. Empirical academic studies generally agree that high earners respond more in their working, entrepreneurial, investing, and avoidance activities than do other people.⁴⁷ So we would expect them to be responsive to interstate tax differences.

High earners are often entrepreneurs, and they may move their businesses and related jobs with them when they migrate. Very wealthy entrepreneurs have been gravitating to Florida, which has no income tax or estate tax, as these examples illustrate:

• New Jersey’s richest person, David Tepper, moved with his hedge fund business Appaloosa Management to Florida in 2016. In a single move, the government of New Jersey lost as much as $100 million a year in income taxes, as Tepper sometimes makes more than $1 billion a year in income.⁴⁸ The state also lost the high-paying jobs that Tepper’s business creates.
• Electronic stock trading entrepreneur Thomas Peterffy moved from Connecticut to Florida in 2016. Peterffy’s firm Interactive Brokers is the largest electronic broker, and he has a net worth of about $20 billion.⁴⁹ Taxes were reported to have been a factor in his move.⁵⁰
• Investor and executive C. Dean Metropoulos left Connecticut for Palm Beach in 2014.⁵¹ Metropoulos has a controlling interest in Hostess Brands. He has a net worth of more than $2 billion earned from buying and turning around dozens of companies.
• Hedge fund manager Paul Tudor Jones moved from Connecticut to Palm Beach in 2015. Jones has a net worth or more than $4 billion and has earned up to $600 million a year.⁵² Connecticut lost up to about $30 million a year in annual income taxes from his move.
• Founder of Paychex business services firm Thomas Golisano moved from New York to Florida for lower taxes in 2009. The Associated Press reported: “The 67-year-old philanthropist from the Rochester area has long criticized the state’s government and high taxes… . [H]e decided to change his residency after lawmakers increased taxes on wealthy New Yorkers in the new state budget. Golisano says moving to a state without a personal income tax will save him $13,800 a day.”⁵³ That amounts to $5 million a year.
• Real estate investor Barry Sternlicht, who runs Starwood Capital Group, moved from Connecticut to Florida in 2016. He cited tax savings for his decision and said: “There’s a massive exodus from Connecticut… . As of July 1 … I’ve become a resident of Florida.”⁵⁴ When asked if he left because of high taxes, he responded, “Yeah.”
• Edward Lampert moved with his hedge fund ESL Investments from Connecticut to Miami in 2012. Lampert was worth about $3 billion at the time. A state representative and friend of Lampert’s said, “ESL’s departure … represents the loss of [not only] wonderful people and philanthropy, but also a large amount of state tax revenue.”⁵⁵

High-tax states are losing not just the income taxes paid by wealthy entrepreneurs moving out, but sometimes the income taxes paid by the people they employ. The head of Palm Beach County’s Business Development Board says good weather and low taxes have drawn 60 or 70 private equity and hedge fund firms to her city in the past few years.⁵⁶

Florida cities are aggressively courting wealthy individuals and their businesses from high-tax states. Florida gains not just the initial jobs that move but also incremental jobs added as relocated businesses grow. And Florida’s low taxes are drawing not just U.S. financial firms but also international firms that might otherwise have located in the Northeast.⁵⁷

Recent articles suggest that the new federal tax law is accelerating moves to Florida. A June Wall Street Journal article quoted a Palm Beach real estate expert as saying there is a boom in wealthy buyers looking to purchase to gain tax residency.⁵⁸ And a June Bloomberg story said: “Florida’s long-running effort to lure Wall Street hotshots is gaining traction thanks to a provision in the federal tax law that hits residents of high-tax states… . That’s because Florida doesn’t have a state income tax and its property taxes are relatively low, whereas the tri-state area has among the highest property taxes in the country… . ‘SALT has been the No. 1 theme when we speak with finance companies’ about relocating or opening branch offices in Miami, said Nitin Motwani, a lead developer of a $2 billion skyscraper in Miami.”⁵⁹

Some entrepreneurs have moved their financial firms out of high-tax California. Hedge fund Universa Investments moved its headquarters from California to Florida in 2014. The firm’s founder, Mark Spitznagel, “cited Miami’s favorable tax policies, emerging financial sector and access to Latin American and European investors as primary reasons for the move… . ‘Florida’s business-friendly policies, which are so different from California’s, offer the perfect environment for us as we expand,’ he said.”⁶⁰

Billionaire investor Ken Fisher moved with his investment firm Fisher Investments from California to the state of Washington in 2011. Fisher has a net worth of more than $3 billion, and his company employs more than 2,000 people. Before the move, Fisher had expressed his frustration with California’s high taxes, and he was looking for a lower-tax location not just for himself but also for his employees.⁶¹ Washington has no income tax.

As a zero-income-tax state, Tennessee is also booming. The Wall Street Journal reported in May: “AllianceBernstein Holding LP plans to relocate its headquarters, chief executive and most of its New York staff to Nashville, Tenn., in an attempt to cut costs… . In a memo to employees, AllianceBernstein cited lower state, city and property taxes compared with the New York metropolitan area among the reasons for the relocation. Nashville’s affordable cost of living, shorter commutes and ability to draw talent were other factors.”⁶² The company has about 3,400 employees, and it considered 30 different cities in its search for a new home.

Wealthy professional athletes are responsive to state tax differences. After California hiked its tax rate on millionaires in 2012, golfer Phil Mickelson said that he had to take “drastic action” because of the hike.⁶³ Mickelson moved from California to Texas in 2014. He earns about $60 million a year, so the move would have saved him millions of dollars a year.⁶⁴ In commenting on Mickelson’s tax situation, golfer Tiger Woods said he himself moved from California to Florida in the 1990s to reduce his tax burden.⁶⁵

Concern exists about the effect of the 2017 tax law on professional sports.⁶⁶ Even before the law, statistical research found that teams in low-tax cities outperform teams in high-tax cities because it is easier to recruit top players to the former.⁶⁷ With the tax law, the United States will become a little more like Europe, where large tax differences between countries drive wealthy athletes, entertainers, and other millionaires to Switzerland.⁶⁸

In their book An Inquiry into the Nature and Causes of the Wealth of States, Arthur Laffer and coauthors present data from 1993 to 2010 suggesting that the nine states without income taxes (Alaska, Florida, Nevada, New Hampshire, South Dakota, Tennessee, Texas, Washington, and Wyoming) are a particularly strong draw for high earners.⁶⁹

Recent IRS data support that view. In 2016, 166,000 interstate migrant households had incomes of more than $200,000, and their aggregate income was $86 billion. In that group, households with income totaling $32 billion moved to the nine states without income taxes. Thus, among those high-income migrants, 37 percent of the income moved to 18 percent of the states.

In 2016, the average in-migration ratio for the nine states with no income tax was 1.13 (Table 1, column 2). But the average ratio for those states in the over $200,000 group is much higher at 1.41 (Table 2, column 3). Thus, the zero-income-tax states are a net migration draw, but they are a particularly strong draw for high earners. The one exception is Alaska.

The average income level within the over $200,000 group can indicate the presence of very high earners. In 2016, the average income of all interstate migrants in that top group was $518,000. But the averages for in-migrants to Wyoming and Florida were $897,000 and $849,000, respectively, which were by far the highest averages among the states. Wyoming and Florida do not have income taxes or estate taxes, and they draw very-high-income migrants.

Some analysts say that high earners hesitate to move out of high-tax states because that is where their business and social relationships are. But New York residents who are sick of paying high taxes can establish a new permanent residence in Florida and continue to undertake business and social activities in New York for part of the year. Generally speaking, they can spend up to 183 days in New York without being a tax resident of that state. They need to carefully document their time and activities in Florida to withstand a challenge by New York tax authorities, but that is a common tax-reduction strategy by high earners.⁷⁰

When state policymakers think about the taxation of high earners, they should recognize that their states lose more than just income taxes when high earners leave. Many wealthy businesspeople are angel and venture investors. They plow their wealth back into young, growing companies, often in the region where they live. If a state loses wealthy individuals, it may also lose startup and entrepreneurial activities down the road.

The wealthy also make large contributions to health, education, and cultural charities in their states. You can see that pattern in a recent Philanthropia.com report that profiles the largest givers in each state.⁷¹ Phil Knight, cofounder of Nike in Oregon, gave $500 million to the Oregon Health and Science University. Phillip Frost, a Florida pharmaceutical billionaire, has given hundreds of millions of dollars to Miami-area health facilities, universities, and art and science museums. Jon Huntsman, founder of a chemical company, gave $175 million to the Huntsman Cancer Center in Utah. Darla Moore, a partner in the investment firm Rainwater Inc., has donated tens of millions of dollars to education and arts institutions in her state of South Carolina.

Individuals gave $280 billion to charity in 2016, and foundations gave $58 billion.⁷² The top 1 percent of earners make one-third of the nation’s charitable contributions.⁷³ The wealthiest 1.4 percent of Americans give 86 percent of all charitable donations made at death.⁷⁴ If high-tax states lose wealthy individuals to lower-tax states, they will likely lose a share of their state’s philanthropy.

A New Jersey report prompted by the possible loss of charitable giving because of taxpayer flight noted: “Wealthy households contribute disproportionately more to charitable causes both from their household assets and from their foundations, trusts, and donor advised funds. Our analysis indicates that in recent years wealth has been leaving New Jersey in larger amounts than wealth has been entering the state due to household migration.”⁷⁵

In sum, the wealthy are important to the states they reside in for many reasons. They pay a large share of state income taxes; they run businesses and create jobs; they invest in growth companies; and they engage in philanthropy. Time will tell how large the migration effect will be from the 2017 tax law. But even a modest increase in moves by top earners could be a substantial blow to high-tax states.

A Closer Look at Retirees

People age 65 and older accounted for 10 percent of interstate movers and 13 percent of the income of movers in 2016. States are increasingly putting out a welcome mat for this group by reducing taxes on retirement income and estates. The thinking is that seniors have substantial wealth to spend and they impose little cost on governments for services such as public schools.

Of the 41 states that impose broad-based income taxes, 36 provide special breaks for pension income.⁷⁶ A few states offer full exemptions, but most have partial exemptions with a dollar cap. More states offer breaks for public pension income than for private pension income. But Illinois, for example, exempts nearly all public and private pension income, including income from 401(k) plans.⁷⁷ The National Conference of State Legislatures says that states use such breaks as “an economic development tool by attracting retired people to, or retaining them in, a state.”⁷⁸

Nearly all the states with an income tax used it to fully tax pension income. But beginning in the 1970s, states began adopting pension tax exemptions, often in regional patterns. They adopted them when neighboring states did so as a “weapon of policy competition,” said Karen Smith Conway and Jonathan Rork.⁷⁹ Over time, exemption amounts have increased as states have raised the competitive bar to attract retirees.

A parallel trend has been the reduction of estate and inheritance taxes, as states have competed to attract wealthy retirees. All 50 states used to impose one or both of those “death taxes,” but today just 17 states and the District of Columbia do so.⁸⁰ The remaining death-tax states are in the Northeast (Connecticut, Maine, Maryland, Massachusetts, New Jersey, New York, Pennsylvania, Rhode Island, and Vermont), the Northwest (Oregon and Washington), and the Midwest (Illinois, Iowa, Minnesota, and Nebraska). Kentucky and Hawaii also impose them.

Before 2005, the federal government provided individuals a credit against state death taxes up to a fixed amount, so it made sense for states to adopt a death tax up to that level. In earlier decades, many states imposed excess death taxes beyond the federal credit amount. In the 1960s, many states imposed taxes on estates of as little as $10,000.⁸¹ But since the 1970s, a competitive trend for states has been to reduce these excess death taxes.⁸²

Then in 2005, the repeal of the federal credit accelerated the downsizing of death taxes. Since then, competition to attract wealthy retirees has prompted most states to completely repeal these taxes.⁸³ Many of the states that have retained them have increased the exemption amounts.

Clearly, state policymakers believe that retirees are responsive to tax differences. Retirees are often footloose, and they can consult sources such as Kiplinger’s “10 Most Tax-Friendly States for Retirees, 2017” and SmartAsset’s “Most Tax Friendly Places for Retirees.”⁸⁴ The latter is an interactive site with detailed tax information down to the county level, and it provides a “retirement tax friendliness index” for each location.

On its website, AARP discusses “Which States Provide the Best Tax Breaks for Retirees?”⁸⁵ The organization is a vigorous lobbyist for reducing state taxes on retirement income, and it informs its 38 million members about state tax differences.⁸⁶

Do taxes influence interstate moves by seniors? Statistical studies generally find that they do, but the results are mixed.

A 1992 study by David Clark and William Hunter finds that high inheritance and estate taxes deterred state in-migration.⁸⁷ A 2004 study by Jon Bakija and Joel Slemrod finds that “high state inheritance and estate taxes and sales taxes have statistically significant, but modest, negative impacts on the number of federal estate tax returns filed in a state.”⁸⁸ In 2001 and 2003 studies, Karen Smith Conway and Andrew Houtenville find that low income, property, and death taxes attract elderly migration.⁸⁹ But in 2006 and 2012 studies, Smith Conway and Jonathan Rork find that death taxes and pension tax breaks have little, if any, effect on migration.⁹⁰

However, the simple patterns of senior migration suggest that taxes do influence location decisions. Consider Kiplinger’s “State-by-State Guide to Taxes on Retirees,” which classifies each state as “most tax-friendly,” “tax-friendly,” “mixed,” “not tax-friendly,” and “least tax-friendly.”⁹¹ The classification considers income taxes, sales taxes, property taxes, motor vehicle taxes, and taxes on pension income and Social Security benefits.

Those categories can be compared with the elderly migration ratios in column 1 of Table 2. For the 20 states that are “most tax-friendly” or “tax-friendly,” the average ratio is 1.11, meaning net in-migration. For the 20 “not” and “least” tax-friendly, the ratio is 0.90, meaning net out-migration. Thus, seniors are gravitating toward states that have the best retirement tax climates.

To get an idea of where the very wealthiest retirees are moving, we can look at federal estate tax returns. In 2016, federal estate taxes were potentially payable above exemption amounts of $5.45 million for singles and $10.9 million for married couples. Florida has 6 percent of the U.S. population, but its residents paid 17 percent of federal estate taxes in 2016.⁹² That figure is up from 12 percent 20 years ago. Meanwhile, the estate tax share in four high-tax northeastern states has fallen, as more of the wealthiest elderly Americans are living elsewhere.⁹³

What Should States Do?

The passage of the 2017 federal tax law has heightened tax competition between the states. The capping of federal deductibility has increased the state and local tax bite on millions of households in high-tax states. At the same time, today people can easily find information to compare state tax burdens.

Some analysts say that people are migrating from northern states to southern states for the warmer climate, not lower taxes.⁹⁴ Yet, as noted, only 2 percent of interstate movers say climate is the main reason for their move. The next few years under the new tax law should give us a clearer view.

More importantly, policymakers in northern states cannot do anything about the warm and sunny climate in southern states. But they can do something about taxes, housing costs, school quality, and other standard-of-living factors affected by government policy. By providing government services more efficiently at lower cost, states can both attract in-migrants and benefit current residents alike.

To grow, states need in-migrants of all types, not just retirees and the wealthy. Urban economist Edward Glaeser argues that attracting younger, educated people with modest incomes is perhaps more important than attracting older, wealthier people.⁹⁵ He says that cities with young and brainy populations are best able to generate growth in our dynamic economy.

To attract all types of people and investment, states should create simple, neutral, low-rate tax codes. They should reduce regulations that dissuade entrepreneurship, and they should pursue land-use reforms to keep housing prices down.⁹⁶ States should also inject more competition into public schooling to improve quality and attract families.

State policymakers should focus on increasing individual freedom. The Cato Institute publishes Freedom in the 50 States, which scores the states on 175 variables covering fiscal, regulatory, and personal freedom.⁹⁷ The variables generally measure the ability of individuals to act without government restraint.

In the 2016 edition of the report, William Ruger and Jason Sorens find that interstate migration flows are strongly correlated with state freedom scores after controlling for climate, the cost of living, and other factors. Numerous studies have confirmed a link between economic freedom and interstate migration.⁹⁸ Americans are gravitating toward states offering greater individual freedom.

Those results are not surprising. Historically, the United States has been a huge draw for international migrants seeking economic and personal freedom. Freedom is a migration draw not just because it has intrinsic value but also because it fosters innovation and growth, which in turn attracts businesses, entrepreneurs, and job seekers.

Some U.S. states have been losing people from out-migration for decades. The new federal tax law should be a wake-up call for such laggard states to improve their tax codes, slim down their governments, and allow residents more economic and personal freedom.

Notes

¹. Frank Sammartino, Philip Stallworth, and David Weiner, “The Effect of the TCJA Individual Income Tax Provisions across Income Groups and across the States,” Tax Policy Center, March 28, 2018.

². Frank Sammartino, Philip Stallworth, and David Weiner, “The Effect of the TCJA Individual Income Tax Provisions across Income Groups and across the States,” Tax Policy Center, March 28, 2018.

³. Joint Committee on Taxation, “Tables Related to the Federal Tax System as in Effect 2017 through 2026,” JCX-32R-18, April 24, 2018, Table 7.

⁴. However, SALT deductions were limited for taxpayers subject to the alternative minimum tax and those subject to the “Pease” limitation on itemized deductions.

⁵. Joint Committee on Taxation, “The Taxation of Individuals and Families,” JCX-41-17, September 12, 2017, Tables 2 and A-1.

⁶. Kirk J. Stark, “Fiscal Federalism and Tax Progressivity: Should the Federal Income Tax Encourage State and Local Redistribution?,” University of California Law Review 51, no. 1 (2004): 1418.

⁷. Congressional Budget Office, “The Deductibility of State and Local Taxes,” February 2008, p. 7.

⁸. Joint Committee on Taxation, “Tables Related to the Federal Tax System as in Effect 2017 through 2026,” JCX-32R-18, April 24, 2018, Table 7.

⁹. State of California, Franchise Tax Board, “Preliminary Report on Specific Provisions of the Federal Tax Cuts and Jobs Act,” March 20, 2018, Table 3.

¹⁰. Ronald Reagan, “Remarks and a Question-and-Answer Session with Economic Editors during a White House Briefing on Tax Reform,” June 7, 1985,http://www.presidency.ucsb.edu/ws/index.php?pid=38740.

¹¹. Joint Committee on Taxation, “Tables Related to the Federal Tax System as in Effect 2017 through 2026,” JCX-32R-18, April 24, 2018, Table 7.

¹². Jared Walczak, “The State and Local Tax Deduction: A Primer,” Tax Foundation, March 2017.

¹³. State of New Jersey, Department of Treasury, “Statistics of Income,” January 12, 2018, Tables B and 3.1b. Also see New York State, Department of Taxation and Finance, “Analysis of 2014 Personal Income Tax Returns,” Table 21. New York data are for 2014; New Jersey data are for 2015.

¹⁴. Justin Garosi and Jason Sisney, “Top 1 Percent Pays Half of State Income Taxes,” California Legislative Analyst’s Office, December 4, 2014.

¹⁵. Jeff Stein, “N.J. Backs Down on Millionaire’s Tax,” Washington Post, May 29, 2018.

¹⁶. IRS migration data are at https://www.irs.gov/statistics/soi-tax-stats-migration-data. And see Kevin Pierce, “SOI Migration Data: A New Approach,” Statistics of Income Bulletin, Internal Revenue Service, Summer 2015.

¹⁷. The IRS data include the number of tax exemptions, which measures the number of people in each household or tax filing unit.

¹⁸. Raven Molloy, Christopher L. Smith, and Abigail Wozniak, “Internal Migration in the United States,” Journal of EconomicPerspectives 25, no. 2 (Spring 2011): 173-96.

¹⁹. Raven Molloy, Christopher L. Smith, and Abigail Wozniak, “Internal Migration in the United States,” Journal ofEconomicPerspectives 25, no. 2 (Spring 2011): 173-96. And see Greg Kaplan and Sam Schulhofer-Wohl, “Understanding the Long-Run Decline in Interstate Migration,” Federal Reserve Bank of Minneapolis, October 2015.

²⁰. U.S. Census Bureau, “CPS Historical Migration/Geographic Mobility Tables,” Table A-1. The Census Bureau reported that 1.7 percent of Americans moved between states in 2017.

²¹. Arthur B. Laffer et al., An Inquiry into the Nature and Causes of the Wealth of States (New York: Wiley, 2014), Table 2.7. See also Tax Foundation, “State to State Migration Data,” https://interactive.taxfoundation.org/migration.

²². The website www.howmoneywalks.com maps the IRS AGI data for the period from 1993 to 2016.

²³. U.S. Census Bureau, “CPS Historical Migration/Geographic Mobility Tables,” Table A-5.

²⁴. Taylor Tepper, “These Are the Best and Worst States for Retirement,” Bankrate.com, July 12, 2018.

²⁵. Tabulation from the U.S. Census Bureau, Current Population Survey, Annual Social and Economic Supplement, 2017, https://cps.ipums.org/cps/index.shtml.

²⁶. U.S. Census Bureau, “State and Local Government Finance,” https://www.census.gov/govs/local/.

²⁷. Karlyn Bowman, “Public Opinion on Taxes: 1937 to Today,” American Enterprise Institute, April 2016.

²⁸. The state tax ratios for 2015 are Alabama (7.2 percent), Alaska (4.9 percent), Arizona (8.2 percent), Arkansas (9.0 percent), California (9.6 percent), Colorado (8.1 percent), Connecticut (10.2 percent), Delaware (5.7 percent), District of Columbia (11.8 percent), Florida (6.6 percent), Georgia (8.1 percent), Hawaii (11.6 percent), Idaho (7.9 percent), Illinois (10.1 percent), Indiana (8.6 percent), Iowa (9.3 percent), Kansas (8.5 percent), Kentucky (8.8 percent), Louisiana (8.4 percent), Maine (11.0 percent), Maryland (9.5 percent), Massachusetts (9.0 percent), Michigan (8.5 percent), Minnesota (10.4 percent), Mississippi (9.3 percent), Missouri (7.9 percent), Montana (7.6 percent), Nebraska (9.2 percent), Nevada (8.2 percent), New Hampshire (7.1 percent), New Jersey (10.0 percent), New Mexico (8.9 percent), New York (13.0 percent), North Carolina (8.3 percent), North Dakota (8.7 percent), Ohio (9.5 percent), Oklahoma (7.2 percent), Oregon (8.4 percent), Pennsylvania (8.7 percent), Rhode Island (10.2 percent), South Carolina (7.9 percent), South Dakota (7.2 percent), Tennessee (6.5 percent), Texas (8.2 percent), Utah (8.4 percent), Vermont (10.9 percent), Virginia (7.8 percent), Washington (8.1 percent), West Virginia (9.4 percent), Wisconsin (9.3 percent), Wyoming (8.0 percent).

²⁹. State-to-state flows are available on the IRS “U.S. Population Migration Data” webpage, https://www.irs.gov/statistics/soi-tax-stats-migration-data.

³⁰. Detailed data by age are available on the IRS “U.S. Population Migration Data” webpage, https://www.irs.gov/statistics/soi-tax-stats-migration-data. The age is the age of the primary taxpayer, which the IRS derives by matching Social Security numbers on tax returns to data from the Social Security Administration.

³¹. Government of the District of Columbia, “Tax Rates and Tax Burdens in the District of Columbia: A Nationwide Comparison, 2016,” December 2017.

³². Note that Bridgeport and Newark have extraordinarily high property taxes, and that some of the other high-tax cities, such as Detroit and Philadelphia, impose city income taxes on top of state income taxes.

³³. “Best States to Move to in 2018 for Lower Taxes,” Kiplinger.com, November 21, 2017. See also Sandra Block and David Muhlbaum, “The 10 Most Tax-Friendly States in the U.S.,” Kiplinger.com, October 4, 2017.

³⁴. Adam McCann, “2018’s Tax Burden by State,” WalletHub, April 9, 2018.

³⁵. Laffer et al., An Inquiry into the Nature and Causes of the Wealth of States (New York: Wiley, 2014), pp. 229-43. The authors look at the relationship between state and local taxes, spending, and quality of services for schools, police, corrections, highways, and other services. They find, for example, that taxpayers in higher-tax California get less bang for their buck than taxpayers in lower-tax Texas. One reason is the excess compensation of employees in California government. Another reason is the excess regulation in California, which, for example, results in a mile of highway costing twice as much to build as in Texas.

³⁶. Jeff Daniels, “Californians Fed Up with Housing Costs and Taxes Are Fleeing State in Big Numbers,” CNBC.com, March 20, 2018.

³⁷. American Automobile Association, “Gas Prices,” https://gasprices.aaa.com.

³⁸. Richard J. Cebula, “Migration and the Tiebout-Tullock Hypothesis Revisited,” Review of Regional Studies 32 (2002): 87-96; Richard J. Cebula and Usha Nair-Reichert, “Migration and Public Policies: A Further Empirical Analysis,” Journal ofEconomics andFinance 36, no. 1 (2012): 238-48; Richard J. Cebula, Usha Nair-Reichert, and Christopher K. Coombs, “Total State In-Migration Rates and Public Policy in the United States,” Regional Studies,Regional Science 1, no. 1 (2014): 101-15; Richard J. Cebula and Gigi M. Alexander, “Determinants of Net Interstate Migration, 2000-2004,” Journal of Regional Analysis and Policy 36, no. 2 (2006): 116-23; Richard J. Cebula, “The Impacts of State Government Tax and Spending Policies on Domestic Migration in the United States,” Journal of Economics 42, no. 2 (2016): 1-22.

³⁹. Richard J. Cebula and Usha Nair-Reichert, “Migration and Public Policies: A Further Empirical Analysis,” Journal ofEconomics and Finance 36, no. 1 (2012): 238-48.

⁴⁰. The Tiebout hypothesis is named for Charles Tiebout. See Charles Tiebout, “A Pure Theory of Local Expenditures,” Journal of Political Economy 64 (1956): 416-24.

⁴¹. Yu Hsing, “A Note on Interstate Migration and Tax Burdens: New Evidence,” Journal of Applied Business Research 12 (Winter 1995/1996): 12-14; Karen Smith Conway and Andrew J. Houtenville, “Elderly Migration and State Fiscal Policy: Evidence from the 1990 Census Migration Flows,” National Tax Journal 54 (March 2001): 103-23; Karen Smith Conway and Andrew J. Houtenville, “Out with the Old, In with the Old: A Closer Look at Younger versus Older Elderly Migration,” Social ScienceQuarterly 84 (May 2003): 309-28; Mark Gius, “The Effect of Income Taxes on Interstate Migration: An Analysis by Age and Race,” Annals of Regional Science 46 (February 2011): 205-18; David E. Clark and William J. Hunter, “The Impact of Economic Opportunity, Amenities and Fiscal Factors on Age-Specific Migration Rates,” Journal of RegionalScience 32 (August 1992): 349-65; Robert R. Preuhs, “State Policy Components of Interstate Migration in the United States,” Political ResearchQuarterly 52 (September 1999): 527-47; Roger Cohen, Andrew Lai, and Charles Steindel, “Tax Flight Has Tangible Effects on Income Tax Revenue,” State Tax Notes, February 20, 2012; Antony Davies and John Pulito, “Tax Rates and Migration,” Mercatus Center at George Mason University, August 2011.

⁴². Roger S. Cohen, Andrew E. Lai, and Charles Steindel, “A Replication of ‘Millionaire Migration and State Taxation of Top Incomes,’” Public Finance Review 43, no. 2 (2015): 206-25.

⁴³. Assuming that the taxpayer was in the top federal bracket of 39.6 percent, the effective rate prior to 2018 was 5.42 percent (8.97-8.97*0.396).

⁴⁴. Cristobal Young and Charles Varner, “Millionaire Migration and State Taxation of Top Incomes,” National Tax Journal 64 (2011): 255-84; Cristobal Young et al., “Millionaire Migration and Taxation of the Elite,” American Sociological Review 81, no. 3 (2016): 421-66; Charles Varner, Cristobal Young, and Allen Prohofsky, “Millionaire Migration in California: Administrative Data for Three Waves of Tax Reform,” Stanford Center on Poverty and Inequality, July 2018. Young and Varner’s studies are critiqued in Antony Davies and John Pulito, “Tax Rates and Migration,” Mercatus Center at George Mason University, August 2011.

⁴⁵. Sally Wallace, “The Effect of State Income Tax Structure on Interstate Migration,” Georgia State University, Fiscal Research Program, December 2002. See also Martin Feldstein and Marian Vaillant Wrobel, “Can State Taxes Redistribute Income?” Journal of Public Economics 68, no. 3 (1994): 369-96; and Seth H. Giertz and Mehmet S. Tosun, “Migration Elasticities, Fiscal Federalism, and the Ability of States to Redistribute Income,” National Tax Journal 65, no. 4 (December 2012): 1069-92.

⁴⁶. New York City’s top rate is 3.876 percent, which is on top of New York State’s top rate of 8.82 percent.

⁴⁷. For example, see Emmanuel Saez, Joel Slemrod, and Seth H. Giertz, “The Elasticity of Taxable Income with Respect to Marginal Tax Rates: A Critical Review,” Journal ofEconomicLiterature 50, no. 1 (2012): 3-50; Martin Feldstein, “The Effect of Marginal Tax Rates on Taxable Income: A Panel Study of the 1986 Tax Reform Act,” Journal of Political Economy 103, no. 3 (1995): 551-72; Robert Carroll et al., “Entrepreneurs, Income Taxes, and Investment,” National Bureau of Economic Research Working Paper no. 6374, January 1998; Jonathan Gruber and Emmanuel Saez, “The Elasticity of Taxable Income: Evidence and Implications,” Journal of Public Economics 84, no. 1 (2002): 1-32; Gerald Auten and David Joulfaian, “The Taxable Income Elasticity of High-Income Taxpayers: Evidence from a Long Panel,” May 2009; Robert Carroll, “Do Taxpayers Really Respond to Changes in Tax Rates,” Working Paper no. 78, U.S. Treasury, Office of Tax Analysis, November 1998; Bradley Heim, “The Effect of Recent Tax Changes on Taxable Income: Evidence from a New Panel of Tax Returns,” Journal of Policy Analysis andManagement 9, no. 1 (2009): 147-63; and Aspen Gorry, R. Glenn Hubbard, and Aparna Mathur, “The Elasticity of Taxable Income in the Presence of Intertemporal Income Shifting,” National Bureau of Economic Research Working Paper no. 24531, April 2018. A number of studies on the responsiveness of high earners are summarized in Alan Reynolds, “Of Course 70% Tax Rates Are Counterproductive,” Wall Street Journal, May 7, 2012.