New and old studies tend to regard civil wars as organized around one of two separate dimensions: ethnicity or ideology. The social landscape of a civil war is therefore assumed to be shaped by blood ties or imported ideologies, where aggrieved ethnic groups or superpowers’ ideological pawns face brutal governments in a war of all against all. Yet this approach is too simple to account for civil war environments that are characterized by cross-cutting ties and mostly instrumental adoption of identity by opposing actors. Fotini Christia brings a rare insight to the politics of civil wars in Alliance Formation in Civil Wars. She shows that the cleavages that seem to be the most salient dividing lines of a society may be the building blocks of warring organizations, but they are less than satisfactory explanations for the ways in which these groups relate to one another in multiparty civil wars.

Christia satisfactorily applies what has long been regarded as the core of neorealism, the balance-of-power theory, to explain inter- and intragroup dynamics in political violence. She shows that even in civil wars where we would expect alignments on ethnic and ideological lines, organizations seem to be making highly strategic alliance decisions. Nicely predicted by balance-of-power theory, warring groups build minimum winning coalitions. In such diverse settings as Afghanistan and Bosnia, there is a strong tendency to balance the power of the stronger coalition by joining with the weaker side independent of the identity of the group. This is a coalition that is usefully sized to win the war but also guarantees that its members will maximize their share of the war spoils without being exploited by a stronger party. Identities complement this picture in interesting ways. The leadership invokes perceptual frames solely to justify their alliance decisions to the common folk. Identities therefore emerge as strategic choices made by the leadership cadre from a range of options to serve their interests.

Christia combines historical and quantitative analysis in a multimethod framework. She adopts a research design that allows sufficient variation in the key independent variables, power and identity, for a comparative case study approach. The book tracks four civil wars in two countries, Afghanistan and Bosnia Herzegovina. The jihad (1978–89) in Afghanistan and the Bosnian civil war (1941–45) are ideological wars, whereas the intra-Mujahedeen war (1992–98) and the second Bosnian civil war (1992–95) are fought along identity lines. Despite such differences, similar outcomes are observed in all four wars. Not surprisingly, the book’s strongest chapters are those on the two Afghan civil wars (Chapters 3–5). the author’s extensive fieldwork in Afghanistan resulted in fine-grained data that are suited to testing alliance formation and group fractionalization on multiple levels of analysis. At the warring group level, alliances follow the relative distribution of power, whereas commander-level rivalries closely mirror the alliances at the center rather than the characteristics of commanders. While alliance formation is endogenous to the conflict, group fractionalization is not. When they face dramatic battlefield losses, groups fractionalize along identity lines that are easily identifiable before the war. Quantitative tests (Chapter 5) analyze the same framework with a cross-national data set of multiparty conflicts and show that the most significant separation in alliance formation is observed between hegemonic and balanced conflicts when controlling for politically salient cleavages.

Alliance Formation in Civil Wars goes beyond reviving balance of power and refurbishing it to fit intrastate contexts. In a book on balancing, the reader finds many helpful insights on identity and identity politics. Although coalition building between groups closely follows the relative distribution of power, a warring group splits along identity lines when its survival becomes at risk. Christia considers fractionalization as a response to the changes in the group’s environment. However, given that identity allows us to predict how the group will split, it still presents an underlying current that coordinates actions in subtle ways. Recruitment of individuals into insurgent organizations tends to closely follow sociological and biological ties. Intermediaries and local power brokers build coalitions around cousinage and blood ties. They rely on closed societies for recruitment and try to spread the word through these networks. The fact that entrepreneurs in civil wars take into account local realities and civil war dynamics simultaneously opens up a research avenue that may have remained underexploited in Christia’s framework. Identity may still be a relevant factor, but the mechanisms through which it plays a role in civil-war environments require more attention paid to the role of agency in mobilizing and reproducing them.

Midrange theories can be built from case studies through a meticulous process that requires back-and-forth between theory and evidence. When the analyst strives to test the theory in different contexts, there are challenges as well as rewards. While Christia’s results on the warring-group level are comparable across cases, one might wonder whether the commander/subgroup-level analysis provides a similar basis. In the Afghanistan case, data on commander characteristics collected from primary sources allow for a test of the theory on the “meso level.” In Bosnia, municipality, instead of commander, is the unit of analysis, and municipal data coded from prewar Yugoslavian maps only give an idea about the population sizes of ethnic groups before the war. Christia shows that conflict breaks out between ethnic groups in municipalities in accordance with the alliances at the center. Data-scarce environments certainly legitimize the use of proxies when necessary. Yet ethnic groups present a rather raw alternative to military commanders and may suggest few possible alliance combinations. Besides, in a theoretical setup where alliances are endogenous to the conflict, prewar measures may fail to account for groups’ responses to the changing dynamics of the war. At the onset of the conflict, when little is known about how warring groups will perform on the battlefield, alliances may be driven less by power considerations and more by existing ties.

What defines the arena of political competition between groups? Survival as the foremost goal coordinates warring groups’ alliance decisions and makes them especially sensitive to the changes in their environment. Group dynamics may take a back seat and become a salient factor only when external forces, such as battlefield outcomes, threaten survival. Christia’s book claims its spot in an emerging cottage industry of micro-level civil war studies initiated by the groundbreaking studies of Stathis Kalyvas (The Logic of Violence in Civil War, 2006) and Jeremy Weinstein (Inside Rebellion, 2007). (For further discussion, see Sidney Tarrow’s “Inside Insurgency: Politics and Violence in an Age of Civil War,” Perspectives on Politics [September 2007].: 587–600). With attention to the causal mechanisms at work, recent studies (e.g., Jason Lyall, “Are Co-Ethnics More Effective Counter-Insurgents?” American Political Science Review 104 [February 2010]: 1–20) brought context back into the study of civil wars in the international relations field by drawing from historical and sociological accounts. Alliance Formation in Civil Wars further shows that international relations theory is alive and can work along with other social science traditions to explain civil war processes.