A Checkpoint Effect? Evidence from a Natural Experiment on Travel Restrictions in the West Bank

Introduction

Does nonviolentrepression prompt subject groups to obey or rebel? By what mechanism does it do so? Scholarly debate over repression and dissent is highly developed, but it systematically overlooks nonviolent institutions; further, to date most research has been observational and inconclusive. In this article, we redress these shortcomings via a study of checkpoints— that is, nonviolent impediments to movement—in the West Bank. According to the United Nations Office for the Coordination of Humanitarian Affairs (OCHA), as of 2012 there were 540 “obstacles blocking Palestinian movement” in the West Bank, including 59 staffed checkpoints and 455 unstaffed impediments to movement, such as roadblocks, gates, barriers, and trenches (OCHA 2012b). We ask: Do checkpoints make Palestinians more likely to support diplomatic negotiation, or violence against Israel? As the epigraph makes clear, contending local interests argue, on the one hand, that checkpoints suppress violence and, on the other, that they perpetuate it. Which is it, and how can we tell?

We confront this problem by exploiting a natural experiment based on a policy intervention by the Tony Blair–led Quartet (i.e., the United States, the United Nations, the European Union, and Russia)4 in May/June 2009 toward the easement of checkpoints as a means of opening up economic corridors in the West Bank. The “Jenin First Initiative” was initiated as a pilot in the north, to be extended if deemed successful. This intervention presents the rare opportunity to draw causal inference on Palestinian political preferences, since while travel in the Jenin corridor was eased, restrictions outside of this area remained in place. To capture the effect of this initiative, we sampled populations (n = 599) before and after the intervention—some in villages near the Za’atara checkpoint, slated for easement (treatment), others near the Wadi Nar checkpoint, which would undergo no change (control). We then pursue difference-in-difference estimation to ascertain the effect of easement on Palestinian political attitudes. This design is experimental because the policy is as-if exogenous, or orthogonal, to Palestinian attitudes and because it occurred without corresponding changes to other institutions of repression. We are able to rule out rival explanations for the divergence in preferences outside of the treatment alone, thereby isolating a “checkpoint effect.”

This difference-in-difference design matches other natural experiments derived from administrative or jurisdictional boundaries (Asiwaju 1985; Card and Krueger 1994; Krasno and Green 2005; Miguel 2004; Miles and Rochefort 1991), using a village-based selection model akin to that employed by Posner (2004), with the noise of nonrandom assignment cleaned up via difference-indifference estimation (Ashenfelter and Card 1985; Smith and Todd 2005). In doing so, it provides us with unique insight into a centerpiece of the Israeli-Palestinian conflict, which remains grossly understudied due to issues of identification. In the West Bank, it is almost impossible to distinguish onefacet of controlfrom others; as a result, existing studies of checkpoints have been unsystematic and incomplete (Brown 2004; Bornstein 2006; Kotef and Amir 2011; Naaman 2006), with econometric reports going so far as to claim that “the losses of internal closures are difficult to measure” (Aranki 2004, 28) or that “quantifying the economic impact of current restrictions is difficult given the paucity of data” (World Bank 2008).

As with any natural experiment, it is incumbent upon us to prove that the treatment assignment can be considered as-if random, such that the expected values of treatment and control vary due to the intervention alone (Dunning 2008). We take extra care to address challenges to identification—a small price for such a unique research frame—drawing on extensive fieldwork to clarify our research design and support our claims.6 Additionally, we test the robustness of our findings against an independent panel (n = 1,200) conducted in four waves between 2007 and 2009 across a representative sample in the West Bank.

We find that Palestinians subject to checkpoint easement are significantly less likely to support violence against Israel, or the militant Islamist group Hamas over the secular-nationalist Fatah. This confirms a positive relationship between nonviolent repression and support for violent dissent, such that high levels of repression correspond to support for violence; as conditions of repression lessened, so did the subject population’s preference for dissent. We suggest one particular mechanism linking the experience of checkpoints and support for militancy: humiliation-. We find that Palestinians who feel humiliated by the experience of checkpoints are more likely to support violence against Israel than, for example, those who express fear or have suffered financial loss. Together, these findings have considerable practical significance, indicating that checkpoint easement may have a positive effect on future Israeli-Palestinian peace negotiations. To this end, they suggest a rethinking of Israeli security policy, as short-term concerns over Palestinian movement may be compromising Israel’s long-term interests—a Faustian bargain that would be devastating to perpetuate.

Read full article