Michael Horowitz (Univ. of Pennsylvania), Allan Stam (Univ. of Virginia), and Cali Ellis (Univ. of Michigan) have each written extensively on military power and war. In Why Leaders Fight, they adopt a distinctive approach to their subject: "Our statistical model shows that the probability that a country gets into a military conflict, roughly doubles when the leader has prior rebel experience" (137). The book is full of statistical charts, including one ranking over forty of the most conflict-prone leaders between 1875 and 2004 (80–81).[1] The preface informs readers that

Chapters 1 and 2 present an overview of how leaders matter and outline how the authors assess the probability of their going to war. To make their points, they refer to examples like Nikita Khrushchev and John Kennedy.[2] Chapter 3 concerns the variables of time and place as they affect leaders' behavior, while 4 and 5 evaluate the influence of military and rebellion backgrounds, age when leaders come to power, education, childhood and family experiences, and the role of gender.

Finally, chapter 6 provides context for judging whether leaders acted wisely or foolishly in choosing war. It also indicates the value of the authors' work for public policy making and future research. Much of this will appeal more to historians, political scientists, and other scholars fond of statistical approaches than to general readers.

The authors conclude that, regardless of structural and institutional constraints, "states' leaders are the most important determinant of state action" (127). Type of government, too, is critical: for instance, "revolutionary regimes are more likely to start military conflicts" (42), as are leaders in autocratic countries.

The authors also find that "Leaders who have seen combat are more cautious about military engagements than those who served but did not see combat" (22), citing John Kennedy and John Kerry; but they admit that this is just a general tendency and not, as the career of Adolf Hitler demonstrates, an iron-clad determinant. They observe that, contrary to conventional wisdom, older leaders are "substantially more prone to militarized behavior" (45), particularly those over age seventy. Not surprisingly, the success or failure of previous military conflicts also affects leaders' inclination to risk war. Though educational and family experiences may all influence a leader's propensity for violence (the book provides many examples), the authors discern only one pattern flowing from these factors: living through a war as a child makes a leader more likely to initiate one in later life.

Despite the perception that women might be less hawkish leaders than men, the authors' data reveals no significant difference between the sexes in this regard. They make educated guesses about the reason for this, speculating that

This subject is especially germane at present, as former Secretary of State Hillary Clinton conducts her presidential campaign.

Chapter 5 is devoted to five female leaders: the Philippines' Corazon Aquino, Pakistan's Benazir Bhutto, Canada's Kim Campbell, Israel's Golda Meir, and Great Britain's Margaret Thatcher. The authors caution here that their conflict-prone rankings and other tables and figures should be used with care. The case of Margaret Thatcher is a good example. Though the authors' Leader Risk index places her in the bottom half of leaders likely to initiate conflict, her character and background "inclined her toward militaristic behavior" (71). Moreover, her response to aggression (not measured by the index), such as Argentina's attack on the Falkland Islands in 1982, was vigorously belligerent.

P.V. Narasimha Rao's background made him, by the authors' own indicators, likely to start conflicts with foreign powers, but he never did: "In this case, the characteristics of the system and domestic institutions overwhelmed any potentially aggressive preferences of the man" (111). The authors note that Rao dealt very aggressively with domestic rebelliousness, but that they have not applied their methods of predicting risk acceptance to leaders' reactions to internal threats.

The authors argue that many factors in Woodrow Wilson's background inclined him avoid armed conflict, but that the course of events and his ideology of liberal internationalism left him "no choice" but to ask Congress to declare war in 1917 (102). They grant that "his most important contribution was his ideology" (104), but their statistics do not attempt to measure that important influence on leaders.

Though they admit the limits of their statistical methodology, the authors forcefully claim that

They lament that the "general trend among historians has been to focus less on political leaders" (25).

This is far too sweeping a generalization, as the myriad historical and biographical studies of, to cite just two figures, Adolf Hitler and Joseph Stalin amply attest. Nor is the authors' LEAD dataset, though a useful creation, as valuable a tool as they maintain. Human beings and their institutions are simply too complex for any statistical approach to allow us to "see the complete elephant."

The esteemed British political thinker Isaiah Berlin once wrote that it is most essential in making political judgments to clearly grasp "what the result is likely to be in a concrete situation of the interplay of human beings and impersonal forces…. What matters is to understand a particular situation in its full uniqueness, the particular men and events and dangers, the particular hopes and fears which are actively at work in a particular place at a particular time."[3] Such a subtle and existential perspective is also necessary to fathom "why leaders fight." No dataset can be more than a welcome ancillary in the struggle to comprehend their motives.