During the Cold War, there was an implicit understanding—and a strong incentive—that the two superpowers would shore up weak states in their respective spheres of interest. After all, if one of the superpowers failed to do so, it would present an opportunity to the other, which encouraged Moscow and Washington to devote attention (and aid) to wobbly nation-states. After the Cold War, Moscow stopped playing that role, the UN proved incapable of stepping in and stepping up, and the United States tried to take up the slack. If nothing else, the post-Cold War era has taught us that fragile, failing and failed nation-states are a global problem that won’t go away. But what’s the best way for the United States and its closest allies to address this persistent problem? A review of recent history offers many lessons—some more helpful than others.

Triage
Faced with a man-made famine that had claimed hundreds of thousands of lives—and with a drumbeat of demands on CNN that “something must be done”—President George H.W. Bush dispatched 28,000 troops to Somalia at the closing hours of his presidency in December 1992. He assured the Somali people that America did not intend to “dictate political outcomes...We come only to feed the starving.” And he explained to the American people that “Some crises in the world cannot be resolved without American involvement…American involvement is often the catalyst for broader involvement in the community of nations.”

U.S. troops rapidly took control of vital ports, roads and the food shipments rival clans had used as a weapon. Within days, chaos turned to order. Within weeks, the manmade famine was over. In a sense, Bush 41 had triaged and then stabilized the near-death nation of Somalia.

The problem with triaging a patient is it only deals with the patient’s most immediate, life-threatening problems. Triage doesn’t deal with the root causes of the problem, the long-term issues, the organs that need to be repaired or rehabbed or removed, the pathologies that triggered intervention in the first place.

Reconstruction

That explains why triage often leads to larger, more ambitious intervention, which leads us to President Bill Clinton’s attempt to reconstruct Somalia in 1993. With Clinton’s blessing, the UN expanded what was a limited humanitarian mission into an ambitious reconstruction of Somalia’s government—exactly what Bush 41 promised America wouldn’t do. Worse, even as the complexity of the mission expanded, Clinton shrank the number of U.S. troops committed to the mission, from 28,000 to just 3,000.

Angered by the UN’s meddling, tribal leader Farah Aidid rallied thousands of Somalis against the very people who had saved their nation from mass-starvation. They began firing on UN personnel, prompting the peacekeepers to return fire and issue arrest warrants for Aidid and his henchmen. Clinton then sent hundreds of Special Operations personnel into Somalia to apprehend Aidid. A brutal, 60-day war ensued, culminating on October 3, 1993, in a day-long gun battle in the dusty alleys of Mogadishu. When the guns fell silent, 18 Americans and hundreds of Somalis were dead, triggering the beginning of the end of U.S. involvement in Somalia.

In stark contrast to the limited mission outlined by his predecessor, Clinton’s experiment in Somalia would end aimlessly and violently. The fragile order constructed by Bush 41 returned to chaos. American blood, treasure and prestige had changed nothing in Somalia.

Although the trigger for intervention was vastly different, President George W. Bush followed a similar logic in trying to reconstruct Iraq and Afghanistan. If sick societies spawned a day of terror (see the Taliban’s Afghanistan) and decades of cross-border aggression (see Saddam Hussein’s Iraq), Bush 43 concluded that only wholesale reconstruction of these failed states—only excising the unhealthy tissue and transplanting democracy into the region—could address the problem.

Importantly, U.S. intervention didn’t push Afghanistan and Iraq into failed-state status. By 2001, after six years of Taliban rule and decades of war, Afghanistan was universally considered a failed state. Likewise, Saddam Hussein’s Iraq was “a country wrecked and poisoned,” as the late Fouad Ajami observed. Iraq was not broken because outside powers intervened. Rather, outside powers intervened because Iraq was broken.

However, Bush 43 and Clinton learned that in places as broken as Somalia, Afghanistan and Iraq, “resurrection” may be a more appropriate term than “reconstruction.”

Rehab
Burned by Somalia, Clinton would take a different approach in the Balkans. Rather than trying to reconstruct the war-scarred Balkans, Clinton chose a less ambitious course of treatment: rehabilitation. Bosnia and Kosovo, unlike Somalia, were broken but not beyond repair. Not only were these proto-states in better shape (relatively speaking) than Somalia; the U.S. and Europe were willing to devote more resources to treatment.

The U.S.-led NATO force was in Bosnia for nine years, at which time NATO handed off its mission to an EU peacekeeping force, which remains in Bosnia today. U.S. and NATO peacekeepers are still in Kosovo 18 years after they arrived. The reason the U.S. and NATO made this long-term commitment underscores why the Balkans is a success story, why Somalia is not and perhaps why humanitarian interventions seldom do much more than triage: The U.S. and the rest of NATO concluded that their interests—not just their collective conscience—were impacted by what was happening on Western Europe’s doorstep. Moreover, not only were NATO’s Balkan interventions spurred by the warzone’s close proximity to Western Europe, they benefitted from close proximity to Western Europe.

Shock Treatment and Quarantine

Wanting to avoid another Iraq but pulled by conscience—and by Europe and CNN—into “doing something,” President Barack Obama reluctantly agreed to intervene in the failing state of Libya in 2011. Regrettably, unleashing airstrikes to decapitate the Libyan government without any plan for follow-on peacekeeping operations was akin to trying experimental shock treatments that would either kill or save the patient. Yes, NATO stopped Qaddafi from turning Benghazi into another Srebrenica; but in toppling Qaddafi, NATO destroyed any semblance of government in Libya. The result was anarchy.

With Iraq, Libya and the polls at home guiding his steps, Obama then steered clear of Syria’s civil war altogether, even as Assad used chemical weapons and indiscriminate bombing to cling to power. As if to quarantine Syria, Obama sent troops and aid to Jordan, Turkey and even Iraq. But that course of treatment proved futile, as the open wound Syria had become hemorrhaged across the Middle East, Africa and Europe.

Guidance

In short, the next time a fragile state falls into anarchy—and there will be a next time—President Donald Trump has several treatment options. Yet this less-than-sterling record helps explain why Trump has vowed, “the era of nation-building will be ended,” and has described “trying to topple various people”—we can infer from context he was talking about Saddam Hussein and Muammar Qaddafi—as “a tremendous disservice…to humanity.”

Trump may not be interested in failed states, but failed states are interested in him.

Why? First, fragile, failed and failing states open the door to a host of global ills—terrorism, piracy, drug-trafficking, human-trafficking, mass-migration, regional instability. These have a way of threatening U.S. interests. Second, mass-media and mass-suffering have a way of pulling on America’s heartstrings and overriding a president’s plans.

The result of these factors, as longtime diplomat James Dobbins concludes, is that “nation-building is the inescapable responsibility of the world’s only superpower.”

Nation-building operations, interventions in failed states and humanitarian missions are anything but post-Cold War aberrations. Recall that in 1898, the U.S. intervened in Cuba “to check the hopeless sacrifices of life by internecine conflicts” and stand up “a stable government, capable of maintaining order,” as President William McKinley explained. Washington deployed some 5,600 troops to quell disorder in China in 1912. Of the 300-plus U.S. military interventions since 1798 tallied by the Congressional Research Service, at least 35 were humanitarian interventions and at least 60 involved failed states.

There is no formula or matrix for when to intervene. But there are some guiding questions policymakers can ask before choosing a course of treatment for fragile, failing and failed states.

First, are U.S. interests in jeopardy? When U.S. interests will likely be impacted by a failing or failed state, intervention is not only sensible; it’s arguably necessary. Often, there is a confluence of U.S. interests and ideals. McKinley made such a case to the American people in Cuba; Clinton did likewise in Bosnia and Kosovo. Obama did not in Syria, even though he could and arguably should have: After all, the war in Syria shocked the conscience, threatened Turkey, Jordan and Israel (U.S. allies all), served as a magnet for Russia, Iran and Hezbollah (U.S. adversaries all), fueled jihadist groups, and destabilized the region.

William Burns, Michèle Flournoy and Nancy Lindborg argue that by “concentrating its efforts where its interests are greatest, where the stakes for regional order are most profound, and where, together with its partners, it can invest in prevention and resilience,” U.S. intervention is most effective.

Second, are policymakers ready for the long haul? “We’ve done these things quickly and we’ve done them well,” Dobbins explains, “but we’ve never done them quickly and well.” As Burns, Flournoy and Lindborg add, “It often takes years or even decades for a state to transcend fragility.” If policymakers are not willing to make such a commitment—and if a short-term commitment will only stave off a return to what triggered intervention in the first place—they should weigh the costs of not intervening at all.

Third, can anyone else help? If, to borrow Bush 41’s phrase, the U.S. can serve as a catalyst or enabler for broader international action to stop a humanitarian crisis, intervention is sensible. If not, the president should consider the costs—in political capital, national treasure and U.S. military personnel—of taking on a long-term humanitarian operation alone.

Fourth, are the American people on board? It is prudent to seek public support through Congress. Whether and how long that support will last is an entirely different matter, but obtaining it is helpful. Burns, Flournoy and Lindborg urge policymakers to “be straight with the American people…about both the limits of our means and also about the costs and consequences of inaction.”

That brings us to a fifth guiding question. Will words help or hinder? Obama said things like this: “When dictators commit atrocities, they depend upon the world to look the other way until those horrifying pictures fade from memory.” (That was after Assad’s gassing of Ghouta.) And this: “Awareness without action changes nothing…‘Never again’ is a challenge to us all…Too often, the world has failed to prevent the killing of innocents on a massive scale.” (That was after a year of butchery in Syria.) Words like this raise the expectation for intervention, while lowering the threshold for intervention. That’s a dangerous mix.

These guiding questions may be unsatisfying. But the exercise is a reminder that determining when and where to intervene is not a science. In a broken world full of broken nations, American policymakers must seek the counsel of the heart and the head, aim for the achievable, and choose the least-bad option.