As any professor who teaches about political violence can attest, it is often challenging for students to flip the script regarding their preconceived notions of the causes and nature of civil conflict. Because fighting produces appalling outcomes perpetrated by nefarious actors, many people are reasonably reluctant to embrace explanatory models with political economy foundations. Yet accepting that violence results from rational actors seeking to maximum their expected utility, even though this creates harm for society, generates many an “aha!” moment in the classroom.

In Voter Backlash and Elite Misperception: The Logic of Violence in Electoral Competition, Steven Rosenzweig seeks to flip the script once more. In a rare but revealing preface, he admits to his own conceptual blinders that took for granted the supposed benefits of election violence, as he set out to conduct field research in Kenya: “As the data came in, it became increasingly clear that my initial thinking was not at all in line with reality” (xv). He discovered that violence rarely helped candidates win and even produced voter backlash. Why, then, do candidates routinely use violence?

The book recounts Rosenzweig’s journey to a set of important answers that balance the assumption that office-seeking candidates and voters are indeed rational with explaining why politicians nonetheless perpetuate violence even when it hurts their electoral chances. Establishing scope conditions, he convincingly shows that election violence is a common feature in all manner of emerging democracies with competitive elections. He leaves aside settings in which elections face more severe authoritarian repression and those where nonstate actors, such as insurgents, spark conflict. This allows him to study the candidates and parties who commit violent acts before, during, and after an election, as well as the populations subject to this violence.

The book weaves a rich tapestry of theory, rigorous methodology and data collection, and policy insights to measure and explain the relationship between election violence and voting outcomes, including why violence is ultimately a costly strategy for politicians and why they so often fail to realize this. Previous explanations simply assume that elites must instigate violence because it works in their favor, but in a first empirical chapter Rosenzweig shows this has not been the case in Kenya, a country that forms the book’s empirical focus. Using constituency-level violence and electoral data, he shows that although the ruling Kenya African National Union (KANU) party used violence in transitional elections of the 1990s, this violence did not improve KANU’s electoral chances (contrary to popular scholarly and public opinion). Furthermore, he exploits information leaked to WikiLeaks regarding roughly two dozen politicians named in investigations done after international mediation helped end postelection violence in 2007–8. Although their identities were generally known among the chattering classes of Kenya’s elite, media circles, and voters, most were never successfully prosecuted. But by measuring how these suspects fared as candidates in 2013 compared to 2007, Rosenzweig concludes that “candidates with a history of violence appear to perform worse at the polls than candidates without it” (85).

The book then explores specific channels through which violence might affect outcomes, particularly through its effects on turnout and vote choice. Rosenzweig notes two prevailing sets of hypotheses. The first views violence as a coercive means to influence voting behavior through force; a second examines the persuasive quality of violence, suggesting that voters desire candidates with qualities reflective of violent tendencies or those who signal group rivalries through divisive ethnic rhetoric. Using administrative data and a set of clever experimental manipulations and vignettes embedded in surveys conducted with Kenyan voters, he discounts coercion in favor of persuasion, but shows that neither work to improve election outcomes for menacing candidates. Instead, he reveals evidence of a “large and broad-based voter backlash against violence, including among the coethnic voters that politicians rely on to win elections” (114). And although ethnic rhetoric may play a role in increasing the likelihood of an actual outbreak of violence, “it is not a useful strategy for mobilizing coethnic support” (ibid.).

If violence neither works to mobilize or persuade, why do candidates believe it does? Rosenzweig’s answer rests in politicians’ cognitive biases. Using experimental treatments with actual Kenyan politicians, he shows that they underestimate how much voters reject violence and ethnicized appeals. Even when given information about the potential for voter backlash, politicians’ perceptual lenses still prevent them from internalizing the costs of employing violence.

The last empirical chapter offers one explanation for the stickiness of these misperceptions. It broadens the scope of investigation beyond Kenya by using time-series cross-national data and short case studies in Indonesia, Pakistan, Ghana, Nigeria, and Brazil. Rosenzweig shows that violent founding elections will increase the likelihood of future election violence because of persistent misperceptions on the part of politicians about its efficacy.

Rosenzweig’s contribution to theory is novel and important by allowing for rational calculations on the part of politicians and voters at the same time as explaining why violence ultimately does not form a winning strategy. If politicians’ cognitive biases remain, we should expect election violence to continue, and Rosenzweig calls for more work to understand the sources and resilience of these biases.

His empirical strategy is rich, detailed, and multifaceted, demonstrating the best of contemporary comparative politics that combines deep case knowledge, microlevel data collection, and careful attention to inference with clearly delimited scope conditions and attention to broader implications. With its careful attention to detailed technical matters while achieving readability and clarity, Voter Backlash and Elite Misperception is also a masterclass for PhD students wanting to learn how to successfully turn dissertations into first books.

I highly recommend this book to students and scholars of contemporary democratic politics and wish to outline two wider contributions of interest. First, many of Rosenzweig’s results demonstrate that ethnicity is not the sole or even an especially relevant factor that explains electoral behavior in diverse societies. He is careful to demonstrate this empirically because he admits it will be surprising to many observers of Kenya. But whether looking at the tendency to choose nonviolent non-coethnics over violent coethnics, or the rejection of violent coethnic appeals, he joins a growing literature in African politics that underscores the limits of ethnic theories.

Second, the mismatches between what politicians believe voters want and what they actually want are striking. In his concluding chapter, Rosenzweig draws the importance of his study to understanding the types of policy interventions to overcome this gulf. Rather than relying on normative pleas, he suggests “appealing to, rather than competing against, political elites’ electoral incentives … correcting those misperceptions can effectively reduce the incidence of violence associated with electoral competition” (161–62). But what then causes politicians to update their priors? Although his experimental results suggest that entrenched elites’ cognitive biases may be too powerful to change, his policy prescriptions posit the importance of providing resources to new entrants to the political scene who abjure violence.

The mismatch no doubt has important implications for democratic consolidation. It first suggests that politicians’ blinders are a more powerful source of undermining democratic institutions than voters’ behavior. Given growing concerns over backsliding that focus on the public’s attraction to chauvinistic or authoritarian messaging, this points back to the leaders pushing that messaging as the problem. It also highlights the role that voters play in holding politicians accountable. Previously violent candidates may succeed in evading prosecution for alleged crimes, as in Kenya, but they are unlikely to succeed in evading a voting base willing to punish them at the ballot box.