The third problem with security assistance is that it risks further destabilizing already unstable situations and actually countering U.S. interests. As in Syria, we may train soldiers who end up fighting for the other side or provide equipment that eventually falls into enemy hands. Our assistance may also create haves and have-nots within a local force, exacerbating political or sectarian divisions. That plagued our efforts to rebuild Iraqi army and police units from the start, and resulted in the creation of well-trained and well-equipped forces that moonlighted as sectarian partisans in Iraq’s civil war.

Defense officials frequently talk up the value of having foreign military officers attend U.S. military schools. And it may seem helpful when an American general is able to call a foreign general during a crisis based on their shared school experience. Yet when we help to strengthen uniformed leaders and not civilian ones, such as politicians and police chiefs, we make foreign militaries more likely to prevail and seize power in future political skirmishes. Research by political scientists Jonathan D. Caverly and Jesse Dillon Savage suggests that American military training “can nearly double the probability of a military-backed coup attempt in the recipient country,” as seen recently in Mali and Burkina Faso.

The flip side of every argument for assistance ought to be a dispassionate assessment of how the aid might be wasted or lost — or worse, how it might ultimately hurt U.S. interests. Such an assessment must take both a short and a long view, to capture risks like those we see now in Iraq and Syria, as well as the decades-long blowback we’ve experienced in Afghanistan and Pakistan after our efforts to arm anti-Soviet rebels there in the 1980s. We should also assess whether our biggest programs, such as the one that provided more than $1 billion in security assistance to Pakistan last year, help or hurt U.S. interests over the long term.