In 2003, David Petraeus, then a division commander in Iraq, famously asked “tell me how this ends?” in reference to the conflict just starting there. It was a good question then, and it’s a good question now. The war against the Islamic State gets a lot of attention, much of it focused on the immediate: Is the war going better or worse this month than last month? Is the Islamic State gaining ground or losing it? Are U.S. air strikes killing more Islamic State leaders or fewer? But these things only matter if they contribute to an ultimate end to the conflict on terms the United States can live with. Will they?

In fact, we have a lot of evidence on wars like this and how they typically end. But it’s not a very encouraging story. The Islamic State threat is likely to persist, in one form or another, for a long time. In the meantime, we’re going to be stuck with a policy that amounts to containment and damage limitation, whose shortcomings will frustrate many Americans.

Civil wars of the kind in which the U.S. conflict with the Islamic State is embedded are notoriously hard to terminate and typically drag on for years. Datasets vary slightly, but most put the median duration of such conflicts at seven to 10 years; and an important minority drag on for a generation or more.

When they do end, it’s rarely because an empowered, victorious army marches into the enemy capital, pulls down the flag, and governs a newly stable society. Civil wars like today’s conflict in Syria and Iraq are often complex, multi-sided proxy conflicts in which a variety of local combatants have ties to outside backers who fund, equip, train, and advise allies’ forces. This outside support enables fighters to weather setbacks and hang on in the face of military adversity. Outside backers usually have geopolitical reasons of their own to support local proxies, and for most such backers, a stable postwar state under rivals’ influence often looks worse than continued chaos—so outsiders also usually have incentives to keep the pot boiling by supporting guerrilla resistance if their proxies lose outright control of territory. But with outside backers on all sides injecting money and arms as needed to prevent a stable consolidation of power by rivals, the result can be a long, grinding stalemate wherein no one can establish durable control over a stabilized country (think of the Democratic Republic of Congo).

Instead, it’s often mutual exhaustion that finally ends such wars. War burns capital and destroys wealth. Even the most resilient combatants eventually run their war-supporting economies so far into the ground that they simply cannot continue. And even their outside backers eventually burn through enough of their own wealth that they lose enthusiasm for the project. When this happens, bargaining space for negotiated settlements opens up, and peace talks can finally end the fighting. The civil war in Angola following the country’s independence, for example, lasted 19 years and only ended in 1994, following the cessation of external support and the exhaustion of both rebel and government forces. (Sadly, fighting resumed in 1998, but without external support for the combatants, it was short-lived; a durable settlement appears to have been reached in 2002, following almost 30 years of intermittent fighting.)

Real exhaustion of this kind could take a long time in Syria and Iraq, and leave a lot of profound damage in its wake. To date, Iran, Russia, Saudi Arabia, the United Arab Emirates, and of course the United States, among others, have all supported proxies in Syria and Iraq. This list includes countries with some very deep pockets. Iran and Saudi Arabia, in particular, each see the war in Syria and Iraq as essential to prevent the other from seizing a geopolitical advantage in a regional conflict for survival. Russia has been willing to deploy its own troops to prevent its ally, Syrian President Bashar al-Assad, from falling. Local allies of all these parties have suffered reversals but have often been bailed out, given the stakes their backers see in the war; for any major combatant to be defeated will thus require a degree of war-weariness not yet apparent from any of the major outside benefactors.

What, then, of the U.S. war against the Islamic State? ISIS is more self-reliant than many contestants in this conflict, funding its war effort mainly by taxing economic activity in areas it controls. The Islamic State lacks the kind of outside benefactors that fuel Assad’s or Jabhat al-Nusra’s or the Iraqi government’s war efforts. And as the group’s ongoing incompetent governance destroys the economy in areas it controls, ISIS will likely run out of money sooner than many of the war’s other combatants. Indeed, it may already be showing signs of this: In late November 2015, the group slashed its fighters’ salaries in half, reporting in Western media suggests its revenues have plummeted, and the group has reportedly tried to create cash through exchange-rate manipulations (a doomed strategy for small economies, as former finance ministers from Mexico, Thailand, Argentina, and other developing economies could attest).

The real problem, however, runs much deeper than just the Islamic State. The war in Iraq and Syria pits a host of groups against one another, many of whom are almost as dangerous to U.S. interests as the Islamic State, and many of whom seek the same status of jihadi vanguard against the West that ISIS now enjoys. ISIS itself achieved this position at the expense of the al-Qaeda-affiliated Jabhat al-Nusra, with whom it competes actively for recruits and resources. Even if some U.S.-allied proxy succeeds in conquering the nominal Islamic State capital of Raqqa and pulling down the ISIS flag from its citadel, this is unlikely to end the war, stabilize Syria, or even remove the threat of jihadi terrorism against the United States from Syrian (or Iraqi) soil—it would just open the war to its next phase, in which Islamic State’s rivals compete for the status that group had enjoyed before them. In the absence of a mutual willingness to end the war and accept some new model of representative governance, even great progress against ISIS does not realize U.S. interests in the conflict—which are to end the terrorism threat, the humanitarian crisis, and the danger to regional stability, not merely to occupy the city of Raqqa. A mutual willingness to end the war on the part of most or all of today’s warring factions, however, seems a long way off.

In this context, real U.S. leverage to bring about a real end to the war—and actual realization of American interests in that war—is distinctly limited. None of the proposals popular in today’s Washington debate offer any meaningful prospect of achieving this. According to U.S. military doctrine, to defeat even an insurgency (much less a proto-state like ISIS) and stabilize a threatened population requires something like 20 counterinsurgents for every thousand civilians. That means 50,000-100,000 well-trained troops would be needed to hold the area now under Islamic State control (depending on how much of the population has fled), much less the rest of Syria. No one is now proposing a realistic plan to accomplish anything close to this—whether such a force comprises American troops, Iraqis, Kurds, Saudis, Turks, or anyone else. In the absence of this, bombing raids or offensives from Iraqi or Kurdish allies can accelerate the rate at which ISIS burns through its capital and perhaps hasten the day when the Islamic State is replaced by the next militant group in the queue—but limited efforts of this kind cannot end the war.

In practical terms, the answer to Petraeus’s famous question is thus relatively insensitive to U.S. policy. If this war plays out the way so many others have, its end will come not through an allied offensive to conquer a capital city but through the mutual exhaustion of multiple actors with multiple, often wealthy outside benefactors. This will eventually happen—but it will likely take many years yet. U.S. efforts won’t change these fundamentals much absent a major stabilization and nation-building effort that few Americans now support, or some diplomatic breakthrough that assuages Iranian, Russian, and Saudi long-run security concerns. And that leaves Americans with patience and containment as the least-bad policy while waiting for this awful war to play itself out.

If that’s the case, the most important contribution Americans can make in Syria and Iraq might well not be on the battlefield at all. A smart containment strategy should include serious efforts to assist regional powers in coping with the humanitarian fallout of Syrian and Iraqi violence, to limit the risk that neighboring states suffer those countries’ fate, and to encourage long-run political settlements where possible. Providing support to Iraq, Jordan, and Turkey to help them provide reasonable conditions for refugees so their skills do not atrophy and their children are educated must be part of any successful policy. So too must be active diplomatic engagement with the Iraqi government to ensure it provides reasonable governance to its Sunni population, something it did not do for three years prior to the Islamic State’s military successes in Iraq in 2014. (Indeed, recent polling shows that while the vast majority of Sunnis across Iraq oppose the Islamic State, those in ISIS-occupied Mosul do not want to be “liberated” by the Iraqi Army, likely because they fear and distrust state security forces as well as Kurdish and Shiite paramilitaries.)

And humanitarian assistance might also help reduce the terrorism threat in the West. The best intelligence system for detecting plots like the ones that have so recently traumatized Brussels, Istanbul, and Paris is the active cooperation of the populations within which the terrorists seek to hide. The greater the misery among dispossessed, poorly housed, poorly fed populations, the greater the anger that can fuel terrorism and the less prone such populations and expatriate communities in the West will be to assist Western counterterrorism efforts. Humanitarian aid is not just an appropriate response to suffering—given the limits on what available military options can accomplish, aid may be one of our most important counter-terrorism tools, too.