After news broke that Russia allegedly paid militants in Afghanistan to kill US soldiers, many commentators were quick to condemn the Trump administration’s lack of response. Since then, President Trump has denied he was briefed on this intelligence, but he recently admitted that he did not use the opportunity of a phone call with Russian President Vladimir Putin to rebuke Russia’s use of bounties.

Russia has long supported non-state armed groups, or “proxies,” to accomplish various military tasks, but their use in Afghanistan would mark the first time Russia has employed proxies against the United States since the end of the Cold War, despite engaging in other activities to disrupt American strategy in war zones. Mercenaries or volunteer groups made up of Russian citizens have operated in wars in which US forces are also present, from Kosovo to Syria, and even in some cases clashed with those forces—although the degree to which Russia’s top brass directed the activity of those forces remains ambiguous. The new evidence that Russia provided direct payments to local armed actors in Afghanistan to target American forces, especially when considered alongside Russian assertiveness elsewhere, adds to growing fears that Russia has been shifting its strategy to more directly confront American forces.

Many experts are deliberating the best way to respond to this real threat to US forces, but scholars know far more about the risks of possible responses to a state sponsoring a proxy group than which options might work to make the state sponsor to stop. It’s a Goldilocks problem: A lack of response leaves open the opportunity for the sponsor to continue its disruptive behavior, but a response that is too severe could spread instability or escalate an indirect war into a direct one.

Any administration hoping to compel a state to stop providing assistance to a proxy can choose from a menu of at least seven options. The options are not mutually exclusive, but each presents its own challenges.

Option 1: Do Nothing

Inaction is often overlooked as a strategic option. Although for some foreign policy decisions inaction may be the wisest decision, targeted states presented with strong and public evidence of a state’s sponsorship do not often let it go unanswered. However, since news outlets first reported Russian-backed threats against US forces, all indications are that the Trump administration prefers not to act. The president and administration officials, including from the military and intelligence community, have cast doubt on the veracity of reports that the Russian government was behind these payments, perhaps in a deliberate effort to give the White House flexibility in its response (including abstaining from any). This deniability may be out of a tried-and-true playbook between adversaries caught in proxy war: Sometimes both the state sponsor and the target choose to keep the state sponsor’s involvement secret in order to absolve the target from the burden of retaliating, which risks mutually costly escalation. Sure, no response avoids the risk of escalation with a nuclear-armed rival, but overreaction is not the only alternative to inaction. And in this case, inaction could cost American lives by allowing Russia to continue to harm US forces and doing nothing to stunt Russian opportunism in the future.

Option 2: Deny the Proxy a Target

Research shows that one of the main factors that drive states to support a proxy is interstate rivalry. States often jump at opportunities to sponsor a proxy to bleed a longtime adversary when that adversary is already supporting a warring party on an opposing side or has intervened with its own troops in a foreign civil war. Viewed through this lens, the American military presence in Afghanistan was a standing invitation for Russia to collaborate with the Taliban. In 2017, Putin implied that Russia could retaliate to the US provision of weapons to Ukraine by sowing instability in another conflict. Continuing with plans to withdraw US troops from Afghanistan would remove targets from harm’s way and decrease Russia’s incentive to make these bounty deals—although it would do little to discourage future Russian opportunism.

Some observers may see this option as unacceptably allowing Taliban abuses to go unchecked and capitulating to Russian threats, perhaps reducing American credibility for future bargaining with Russia. However, analysts disagree about what Russia wants in Afghanistan: The Pentagon assesses that Russia wants to expedite a withdrawal, but some experts argue that Russia prefers to keep the United States bogged down in Afghanistan.

Ultimately, if a goal is to eliminate Russian-backed Taliban threats to US forces, then removing the ability of Russia to target American forces in Afghanistan through withdrawing is one way of doing so, despite how emotionally unsatisfactory that option might be. Skeptics may object that Russia has other US targets to attack in other parts of the world, but Russia can more easily hide and deny its disruption of American activity in war zones like Afghanistan, where American enemies are plentiful.

Option 3: Diplomatically Rebuke the State Sponsor

States facing proxy threats often directly confront the sponsor of those threats, although it is unclear to what degree admonitions alone convince sponsors to relinquish a proxy strategy that they see as a low-cost way to bleed and distract a rival. In the early days of Russia’s annexation of Crimea in 2014, President Obama called for Russia to stop supporting Ukrainian separatists or else face continued diplomatic isolation; when Russia continued its disruptive activity in Ukraine, the Obama administration eventually sought other means to counter it. Though open questions remain about their efficacy, the advantage of diplomatic rebukes is that they cost little—which raises more questions about why the Trump administration has refused to use this basic option. A menu of options prepared for the president in response to the intelligence on the bounties reportedly included “diplomatic censure,” according to Trump’s national security adviser, Robert C O’Brien. Although Secretary of State Mike Pompeo has claimed that he confronted the Russian government about its involvement in Afghanistan, the president’s refusal to confront the Russians at the highest levels undercuts the credibility of any diplomatic signals from other parts of his administration.

Option 4: Punish the Proxy

The aggrieved state could target the proxy to convince it that continuing to act on behalf of its state sponsor costs more than it is worth. The Trump administration has opted for this option in the past. In December, in response to attacks by Kataib Hezbollah (KH), an Iranian-backed militia in Iraq, on coalition bases, the Trump administration retaliated by bombing some of KH’s compounds in the country. KH continued its campaign of attacks, which the United States responded to with more violence—but the attacks have persisted. In the Afghanistan case, however, punishing the Taliban for Russian-backed targeting of American soldiers not only risks being ineffectual but also could risk undermining peace negotiations, increasing Taliban violence and spoiling progress toward a withdrawal. Some observers even see a direct trade-off between holding the Taliban accountable and leaving the country, and recommend punishing the Taliban for accepting and acting on these bounty deals by stopping negotiations and withdrawal plans. For decision-makers considering this option, the key question is whether it is more important for the United States to stop Russian assistance to the Taliban or to maximize the chances of an agreement that enables US forces to withdraw from Afghanistan.

Option 5: Sanction the State Sponsor

Sanctions are increasingly the tool of choice in US foreign policy, especially to hold other states accountable for violations of international law. Sanctions impose costs on the target to enhance leverage but are seen as less escalatory than military options. In the face of evidence that Russia may have provided payments to Taliban fighters to attack American troops, some administration officials reportedly considered sanctioning Russia, and former National Security Adviser John Bolton and Congress have also called for the use of this option to punish Russia and convince it to change its behavior. However, as summarized by Elizabeth Rosenberg and Jordan Tama, research shows that sanctions “rarely produce capitulation” and are instead better designed for deterring future action.

Even if sanctions theoretically could work, the number of existing sanctions against Russia for various other reasons—from its involvement in Ukraine to meddling in American elections—could crowd out this particular request. As with all coercion by punishment, how much more pain is enough to get Russia to stop paying proxies, which it denies anyway?

Sanctioning Russia for this behavior presents an added complication that could spoil its effectiveness: Publicly punishing Russia until it capitulates means Russia has no face-saving way out. To relieve sanctions, Russia would have to implicitly admit to offering these bounties—hurting perceptions of Russian resolve in future crises.

Whether economic sanctions would succeed at coercing Russia to stop its disruptive sponsorship of the Taliban in Afghanistan is an open question. Even if it does not change Russia’s behavior in this particular instance, however, many commentators calling for sanctions may consider the very act of punishing Russia in this way to be worthwhile, arguing that doing so could deter Russia or other states from doing the same thing in the future. However, how much another state might care about how the United States acted in the past is unclear, given mixed findings about whether states consider their rivals’ past reputations in assessing credibility during crises.

Option 6: Pressure the State Sponsor in an Unrelated Arena

To pressure foreign actors, states often use their leverage over issues other than the one at stake. In this way, the United States could punish Russia for its sins not by addressing its proxy sponsorship head-on but, rather, by punishing Russia on some other front over which the United States has an advantage. Because the issue of Russian bounties to the Taliban is more important to the United States than to Russia, research suggests the key is to flip this dynamic and find another arena that matters more to Russia than to the United States. For example, the United States could threaten to increase military aid to Ukraine unless Russia stops its assistance to the Taliban. Other research cautions that any effort to link issues to coerce an enemy should be wary of publicly embarrassing Russia, which would make it hard for Russia to capitulate and keep its reputation intact. To avoid painting Russia into a corner, the United States could privately signal to Russia a threat or reprieve in another issue area in exchange for decreased Russian opportunism in Afghanistan.

Option 7: Escalate to Deescalate

The most drastic option in response to a proxy threat is to directly attack the state sponsor that has elected to use these indirect means. Some people—including Trump and Bolton—have said they consider any Taliban-instigated attacks incentivized by Russian bounties to be “direct attacks on Americans,” even though states have historically operated through proxies in part to escape the consequences of direct enemy engagement. Even international law, in debates over when a state that supports an armed group can and cannot be held legally responsible for the latter’s actions, makes a distinction between states and the militias they support. Any attempt to retaliate against Russia directly for attacks on US forces will be complicated by the fact that the American intelligence community disagrees about the strength of evidence of Russia’s complicity—a feature, not a bug, of secretive state-proxy relationships.

In the case of Iran-backed proxies, the Trump administration preferred to retaliate against the state sponsor directly. The administration first laid the groundwork to hold Iran directly accountable for the actions of its proxies through designating the Islamic Revolutionary Guard Corps (IRGC) a terrorist group in April 2019. Then, in January of this year, after weeks of attacks from Iran-backed militants against US forces in Iraq, Trump approved a strike against IRGC-Quds Force Gen. Qassem Soleimani, then-manager of Iran’s vast proxy network. The move set off a tense standoff with Iran and a series of retaliatory Iranian missile strikes against US targets in Iraq. Although the crisis did not escalate further, the tensions set off by Trump’s decision to escalate to deescalate have failed to stem continued attacks on US forces and Iraqis—the original reason for targeting Soleimani. Exactly how the United States would “escalate to deescalate” in response to Russian-backed proxy attacks in Afghanistan, however, is unclear, given that Russian forces are not operating in Afghanistan. Retaliating against Russian forces operating in other conflicts—such as Syria or Ukraine—risks continuing a spiral of tit for tat across the globe; Russia has already shown itself willing to retaliate to perceived American provocations in one arena by responding in another.

The Merit of Discretion

Any administration facing public anger over proxy violence would have to balance the simultaneous goals of keeping American troops safe and avoiding continued instability or war with Russia. It is unclear, however, whether the retributive responses that would most placate an angry public would be effective at countering Russia’s provocation—and prevent further escalation (and future threats to Americans and American forces) in the process. The options with the greatest likelihood of success—denying a target through withdrawing from Afghanistan or using bargaining chips on other issues of interest to Russia—may happen behind the scenes in back channels between Washington and the Kremlin, and their success may depend on their receiving deliberately little fanfare.

Sara Plana is a doctoral candidate in political science at the Massachusetts Institute of Technology and a predoctoral fellow at the International Security Program at Harvard Kennedy School’s Belfer Center for Science and International Affairs. Her research focuses on proxy warfare, armed-group and military organizations, and control over the use of force. Prior to MIT, Sara worked for the US Department of Defense.