The study focused on ethno-nationalist terrorist groups operating in the Middle East and North Africa since 1980, building on past efforts of both the Global Terrorism Database project and the Minorities at Risk/Organizational Behavior data set. For this project, (1) data was collected and coded on the range of counterterrorism policies enacted by governments in Israel, Lebanon, Egypt, Turkey, and Algeria during the period 1980 through 2004; and (2) multivariate statistical analyses were conducted, integrating these new data as well as existing data on terrorist organizations and their behavior, terrorist activity in general and political conditions within a state to determine the effect and effectiveness of counterterrorism policies under a range of conditions.
We find that, in general, Israel's use of conciliatory tactics has reduced Palestinian terrorist attacks, although these effects seem to be cumulative over time. Purely repressive actions have led to increases in terrorism. Our study is the first to show empirically that conciliation toward Palestinians yields lower levels of subsequent Palestinian violence. The implications of these findings are that when facing a protracted terrorist threat, states should consider raising the expected utility of abstaining from terrorism through rewards. This is especially true when the rewards are distributed indiscriminately - that is, when they are distributed to the terrorists' constituency (not just individual perpetrators of terror attacks). The results also suggest that conciliatory actions must be sustained if they are to effectively reduce terrorist violence. A few conciliatory efforts will unlikely show an effect within the first week; however, an ongoing consistent campaign of conciliation will likely lead to a drop in terrorism by the following month. While this is good news, it can easily be dismantled by Israeli repression, as the backlash can be swift, especially if the repression was violent and affected innocent Palestinians. The full set of findings reinforces Braithwaite's (2005) speculation that terrorist organizations benefit from repressive actions, and likely strategically elicit repressive responses that would sabotage any good-will that might have been developing between Israel and the Palestinians. If the Palestinian people begin to trust Israel, the terrorist organizations will lose their base of support. Thus, we expect that Palestinian terrorists rely on Israel's hawkish reactions to terror in order to preserve their longevity. These findings may vary over time and space. Follow-up analyses of the effects of conciliatory and repressive counterterrorism policies on terrorist violence in Turkey, Egypt, Algeria, and Lebanon are forthcoming.
The data were collected by an automated text collection and coding process and cleaned by research assistants. The analysis described above was conducted on monthly and weekly time series data using Generalized Additive Models for a non-parametric assessment of the relationships, and Negative Binomial Regression for a parametric test. We controlled for four lag periods of Palestinian terrorist attacks to reduce the chances of simultaneity bias. The use of both non-parametric and parametric models helps assess the robustness of the findings. A similar methodology is being used to analyze the findings for Turkey; however, because Turkey and the other countries have fewer government actions, it is difficult to identify a unit of analysis that has enough variation without losing the needed statistical power. The non-parametric findings for Turkey are similar to those for Israel, but the parametric tests are null, likely due to low statistical power.