These groups, however, may gradually lose touch with their constituents because they do not rely on them to raise money or supply recruits who are not in it for the salary. Groups like Al Saiqa, a Palestinian group strongly backed by Syria, faded out in part because they could not replenish their ranks after Israeli strikes decimated the existing cadre.

These risks mean that it is often the most-desperate and least-popular groups that seek outside support—a proxy war version of adverse selection. It is not surprising then that the Palestine Islamic Jihad is willing to take the opprobrium associated with ties to Iran because it is broke and lacks the networks to acquire significant weapons on its own. Hamas, in contrast, also works with Iran, but as a far stronger group it was able to cut those ties for several years when Iran’s support for the Syrian regime made it widely hated among Sunnis as war spread in 2012. In general, Hamas has proven it is not an Iranian puppet despite taking Tehran’s money and weapons.

Even groups that retain their credibility often regret taking on foreign support. At the very least, many sponsors impose limits, fearing at least some retaliation or damage to their reputations. Iran and Pakistan, both of which rely heavily on proxies, have limited the arms they provide. Iran, for example, has not transferred chemical weapons to Hezbollah despite their exceptionally close relationship. Tehran also recognizes that if one of its known proxies attacks the U.S. homeland, America might retaliate by going after Iran. Not surprisingly, Iranian proxies have focused on American forces in war zones like Iraq, not the U.S. homeland, in order to limit the chances of unwanted escalation.

As various Kurdish groups can attest, foreign supporters are fickle. The United States turned its back on Iraqi Kurds, whom it had backed in the early 1970s against Saddam Hussein’s regime, after the Shah of Iran cut a deal with Saddam in 1975. Washington is now reducing support for the YPG in Syria and may end it altogether, due to pressure from Turkey. The YPG successfully fought the Islamic State on behalf of the United States and, having succeeded (at least for now), they are no longer necessary while Turkey remains an enduring interest. As the Kurdish experience suggests, states are usually more concerned with the policies and goodwill of other governments, and will sacrifice proxies as necessary to advance their ends.

States are particularly likely to abandon proxies (or back rival ones) when their goals differ. Pakistan provided support for the Jammu and Kashmir Liberation Front (JKLF) in the 1980s when it rebelled against India. However, the jklf wanted an independent Kashmir, not union with Pakistan, so Islamabad supported its rivals that favored Pakistan’s agenda. The JKLF eventually collapsed, caught between Indian security forces and pro-Pakistani militants. Support is often an alliance of convenience, not a close relationship.

Just as states should think twice about supporting proxies, so too should proxies think twice about relying on foreign powers. At best they can hope for resources and assistance, but they should never count on it lasting over time.

IF THE United States is going to engage in proxy warfare—and it will—then Americans need to recognize these limits and problems. At the same time, the United States should not overestimate Iran, Russia or other adversaries and the proxies they support. They too will face willful, abusive and incompetent proxies, and their abilities to achieve their goals are likely to suffer as a result.

One step the United States should take is to constantly highlight its adversaries’ use of proxies to reduce the rewards of deniability. Hezbollah may be helping the Assad government in Syria, but it is doing so at the direction of Iran. Rebels in the Donbass may be fighting the Ukrainian government, but without Russian support they would have collapsed. Their fighting, and also the human rights abuses these groups commit, should be laid at the feet of their sponsors.

Similarly, if the United States and its allies choose to fight back, they should not feel compelled to limit their escalation to the narrow war zone chosen by the proxy’s backer. If Iranian-backed forces attack Israel from Syrian or Lebanon, Israel should have the right to target Iran itself. States usually will decide such escalation is not in their interests, but when they do fight back their decision to target the sponsor should be seen as legitimate.

Even as it publicly ties proxies to their sponsors, the United States should recognize the common divisions and try to widen and exploit them. Some proxies can be bought off, and others can be diverted from the wishes of their would-be masters. The Houthis, for example, are hardly loyal Iranian puppets, and the right mix of threats and bribes might drive the two apart.

When backing proxies, the United States should recognize from the start the risk of being sucked into a conflict. Proxy war may seem a cheap, indeed almost costless, way of waging war, but it can be the thin end of the wedge. Often, the political necessity of talking up the cause and the proxy’s virtues translates into an obligation when the war goes south—how, after all, can this noble cause and these brave freedom fighters be abandoned in their hour of need? President Obama tried to avoid intervening in Syria and provided limited support to Syrian opposition forces as a way to limit U.S. involvement. Nevertheless, the U.S. role increased to air strikes and then the deployment of ground forces as the proxies, alone, did not achieve American objectives. At times, the United States may feel compelled to rescue an embattled proxy, as it did with the Kurds when Saddam Hussein’s forces menaced them in 1996.

In addition, the United States can increase its own capacity to use proxies. Already, the wars in Iraq and Afghanistan have given special operators unparalleled leadership roles in the U.S. military. Given the prevalence of proxy war in the post–Cold War era, its nature should be elevated in professional military education and training. U.S. intelligence can gather more information on proxies, both foes and supposed friends, to learn their true goals and capacities. If they are working with America, it is vital to determine how much they cooperate in reality; if they are foes, then it is necessary to identify how to increase the differences they have with their paymasters.

Often proxy war promises to hit the political sweet spot between doing too little and too high a cost. In reality, however, it is an imperfect form of warfare. Better policies can improve our track record, but it is bound to disappoint many of its proponents, and at times put a country on the road to an unwanted conflict.

Daniel Byman is a professor and vice dean at Georgetown University’s Edmund A. Walsh School of Foreign Service and a senior fellow at the Brookings Institution’s Center for Middle East Policy.