Free beta access. 50 free AI tokens each week. Optional top-ups available. See what is included
Opening Word document…

Horowitz is right

Weeks Professor of Political Science and H. Douglas Weaver Chair in Diplomacy and International Relations 17

(Jessica L. P. Weeks, In Why Leaders Fight, Michael Horowitz, Allan Stam, and Cali Ellis approach a fascinating question: To what extent do leaders’ personal experiences and backgrounds influence their decisions about military conflict?, International Politics Reviews volume 5, pages 39–41)

Why Leaders Fight is a monumental contribution to this endeavor. The book introduces the LEAD dataset, a major data collection effort that records the characteristics and life experiences of leaders around the world during 1875–2004. Using these data, the authors study how leaders’ characteristics affect their likelihood of initiating an international military dispute. The book gives scholars and students not only ideas about how the attributes of leaders could matter in international relations, but provides a dataset that scholars can use to make their new own discoveries. The applications of the dataset are virtually unlimited and will allow future scholars to study how leaders’ attributes affect any number of international and domestic policies and outcomes.

Aside from these core theoretical contributions and impressive dataset, the book has many additional strengths. It anticipates and addresses a number of potential objections to the authors’ arguments that leaders matter, including whether leaders are simply conduits for policy rather than affecting policy independently. The authors take various steps to account for the fact that countries could be selecting leaders with certain attributes, which could affect the correlation between leader attributes and the initiation of conflict. The authors are also admirably clear that leaders do not determine outcomes, but are among many factors that contribute to decisions about conflict, sometimes interacting with other variables such as domestic institutions. I also find both the leader risk scores and the analysis of systemic vs. leader-based factors to be interesting and insightful, and its conclusions to be useful for policy makers. Finally, the book contains helpful and interesting narrative examples of how the backgrounds of specific leaders from around the world might have influenced their foreign policy decisions.

That said, like any book, Why Leaders Fight has some limitations. I will focus on three here, all of which open up rich new areas for future scholarship.

The first is that the book leans rather heavily toward the empirical rather than the theoretical. While this is understandable given the wealth of data the authors have to explore, their theoretical expectations about specific leader-centric variables are often somewhat equivocal. For one, it is not always evident whether the authors have a specific logical reason to expect a certain direct relationship between a leader’s characteristics and conflict initiation; the theoretical discussion is often agnostic, presenting theoretical conjectures as possibilities to be evaluated empirically rather than on logical grounds. In some cases, the authors barely discuss an anticipated theoretical relationship at all, even when the results (shown in detail in the helpful appendix) are suggestive of an empirical relationship. For example, why is having a “creative” career associated with greater conflict initiation? Why do those with experience as an interpreter initiate less conflict? Why do older leaders initiate more conflict in relatively democratic regimes, but not in autocratic ones? These are theoretical questions that deserve more attention, although the sheer number of variables admittedly makes that difficult.

A second, related issue has to do with the “why” – the causal mechanisms linking specific leader attributes to conflict initiation. Returning to the example of leader age, what evidence do we have for why age matters? The authors suggest that the answer has to do with the leader’s time horizons, but is that mechanism consistent with their empirical finding that age does not matter in autocratic regimes, and only in relatively more democratic regimes? (Figure 4.4). It seems that time horizons should be less important in democracies, where leaders face relatively short time horizons regardless of their age, due to term limits. The authors offer ideas about the mechanisms linking many leader attributes to conflict initiation, and their historical illustrations often contain clues, but future work must assess why certain leader attributes are correlated with conflict initiation.

A third concern is about what causal conclusions we can draw from the evidence in the book. The authors are careful to refute the potential objection that the results are due to selection bias, but an issue that they do not address (at least in the book or the appendix) is about specific modeling choices, particularly the decision to include so many variables in the statistical analyses of specific leader attributes. One potential problem in this vein is that many of the predictor variables in the analyses may be causally related to each other (Angrist and Pischke 2009). Rather than designing specific models to test the effects of specific attributes, the analysis often errs on the side of inclusion at the risk of bias. For example, the authors report that the findings do not support a relationship between education and conflict initiation, but the models used to reach that conclusion include measures of occupation. Education is clearly related to many of the occupations included in the analysis, such as law and medicine. The causal effects of age are also difficult to discern. Regime type probably predicts leader age (if not monotonically) because in some kinds of autocracies, very young leaders can come to power. In democracies, in contrast, there tend to be strict lower age limits. Age probably “causes” military service, combat experience, witnessing a national war, divorce, and number of total spouses. The military career and military service variables are also probably causally related, and the military service and national war participation variables may have a causal relationship as well.2 Finally, regime type appears to be highly predictive of the leader’s gender, with extremely few nondemocratic leaders having been women. Including all of these variables in the same model makes it difficult to parse out the causal effect of each one.

In sum, Why Leaders Fight is an important contribution to our understanding of the causes of military conflict. The authors bring to bear impressive amounts of data to explore the relationship between individual leader attributes and military conflict, while staying attentive to the possibility that systemic and domestic variables could matter (or condition) the importance of individuals as well. These ideas and data provide countless avenues for future scholarship on when, why, and how leaders matter for a host of questions about international and domestic politics.3

Sign in to recut, update, or save this card.