It is estimated that in February alone, the Nigerian group Boko Haram was responsible for the deaths of more than 300 people. In that month, the group attacked at least three villages, Konduga, Izghe and Doron Baga, killing around 130 people. In Buni Yadi, they were also responsible for setting ablaze a boarding school, killing nearly 60 students. At the same time as he declared the killings “heinous, brutal and mindless” and the culprits “deranged terrorists and fanatics who have clearly lost all human morality and descended to bestiality,” Nigerian President Goodluck Jonathan was busy imploring members of Boko Haram to cash in on the government’s amnesty program and begin negotiations with the state. This appears to be part of a broader trend.

The popular adage that governments “do not negotiate with terrorists” appears to be untrue, at least in civil war. In a new study published in the American Journal of Political Science, I find that governments embroiled in domestic conflicts in Africa between 1989 and 2010 are more likely to hold negotiations with rebel groups when they engage in more acts of terrorism. Rebels are also likely to gain more concessions from their governments when they execute more terror attacks.

These findings are important because they weigh in on a long-standing debate regarding the effectiveness of terrorism. Specifically, recent studies find that violent organizations are less likely to see significant policy achievements (concessions) when using terrorism to accomplish their goals. This conclusion, however, begs the question of why groups continue to utilize terrorism when it appears to systematically lead to fewer advances for groups that engage in such behavior.

One plausible explanation is that scholars are focusing on the wrong question. Instead of asking whether terrorism is effective, we should be concentrating on when and for what purpose is terrorism effective, especially since the empirical record shows that terrorism has both hurt and helped the causes of violent organizations that have employed the tactic. Very little extant research, however, helps us understand this variation.

One of the limitations of current research is that it relies on concessions as the sole measure by which the success of acts of terrorism can be evaluated. That is, studies consider groups successful when they are able to extract a great deal of concessions from their targets, and unsuccessful when they are not.

Examining whether policy concessions are a direct result of a group’s use of terrorism, however, limits our ability to understand the impact of such attacks as terrorism can be used to further multiple strategies and achieve multiple goals, including but not limited to concessions. Researchers (for example, Andrew Kydd and Barbara Walter) argue that groups have also used terrorism to disrupt peace deals (spoiling) and to divert support from other organizations (outbidding). In a civil war, groups can also use terrorism strategically to gain control over the civilian population and territory. While related, both can separately boost rebel’s bargaining positions and help them achieve better deals in negotiations.

The conundrum is that while terrorism can offer rebels bargaining power, which may ultimately culminate in concessions, groups need to first make it to the negotiating table. Most governments, however, are reluctant to engage in talks with terrorists (a label often given to any group engaging in violent opposition with the state), suggesting that attaining negotiations is not a trivial task. Being “terrorists” makes it harder to gain seats at the table, although using terrorism may also make it easier.

Why would terrorism increase the odds of rebels participating in talks? In short, it is because it hurts. Recurrent acts of terrorism undermine the state’s credibility and control by underscoring that the government is either unwilling or unable to protect civilians from violence. Terrorism can engender support for rebels by incentivizing civilians to look to them for protection. It can also make governments with inefficient counterterrorism policies look especially bad. Nigeria, for example, has undermined its own credibility by using a heavy-handed approach when dealing with Boko Haram. While Nigeria’s actions have invited national and international condemnation, they have also boosted Boko Haram’s recruitment and control in Borno State.

When both terrorism and counterterrorism inflict massive costs on civilians, the population is left with a choice of two bad options. Given that preexisting grievances are likely to have facilitated rebels’ mobilization efforts in the first place, governments are doubly likely to lose in such a contest. For these reasons, governments should be expected to pursue negotiated settlements to stop the pain caused by terrorism and to strike deals while their bargaining positions still permit favorable deals. New research by Reed Wood and Jacob Kathman similarly demonstrate that victimizing civilians may help groups achieve desirable outcomes in civil wars.

While a possible interpretation of these findings would suggest that terrorism helps groups “win” in war, it is important to note that the effects of terrorism are largely contingent upon a government’s response to it. If governments avoid feeding into terrorists’ strategies and instead engage in sound counterterrorism practices, they may be able to stymie the growth of terrorists (or at least not aid it). By offering protection and resisting the urge to use violence against civilians when responding to terrorists, governments should be able to undermine terrorists’ ability to undermine them.

Another way governments might deal with the problem is by finding effective ways of ending civil wars once they start. If governments resolve their conflicts, terrorists lose their power to hurt as war provides the context in which terrorism can work. As much as terrorism is a problem, a government refusing to seek peaceful settlements is also a problem. By dealing with the latter, governments can also address the former and resist “rewarding bad behavior.” Providing concessions to civilians or even more moderate armed groups in society, for example, may do enough to limit some of the aforementioned effects of terrorists’ strategies.

Jakana Thomas is an assistant professor of political science at Michigan State University. This piece is based on the article “Rewarding Bad Behavior: How Governments Respond to Terrorism in Civil War” published by the American Journal of Political Science. The data collection for the article was funded in part by the Centre for the Study of Civil War at the Peace Research Institute Oslo.