I am grateful to Philip Roeder for his generous reading of Nation Building and the important avenues for future research that he sees emerging from it. In this short response, I would like to address the five theoretical points he raises.

First, he deplores the somewhat underdeveloped treatment of the various mechanisms underlying the theoretical model. I agree. The book aimed for the general picture, rather than the particularities of how the micro-mechanisms operate.

Second and third, Roeder wonders what holds the different parts of the theory together (second) and whether modernization theorists could read the book as support for their views (third). Political modernization is indeed the key concept that holds the various parts of the model together, as I elaborate in the theory chapter, where the book claims the mantel of early (Weberian) modernization theory and gives it a relational twist by focusing on the nature of political alliance structures. All key factors and variables that generate inclusionary ethnopolitical ruling coalitions are indeed political modernization processes: the rise of centralized states in the first step and the development of civil society organizations, state capacity to provide public goods, and linguistic homogenization in the second step. Inclusionary ruling coalitions, I argue, then encourages national identification of the population at large. This revised political modernization theory of nation-building contrasts with other approaches that emphasize political institutions (such as democracy), international influences (such as globalization), the nature of ethnic cleavage structures (as in the Lipset tradition), premodern national sentiment (as Adam Smith has it), and so on.

Fourth, Philip Roeder wonders if the causal arrow could go in the other direction: that a strong sense of national belonging will make inclusionary ruling coalitions, the building of civil society networks, state capacity to deliver public goods, linguistic homogeneity, and even the emergence of centralized states before the nineteenth century more feasible. I try to address this important question in the book in various ways. In the case of Switzerland (which Roeder also references in other parts of his review), I show that Swiss national identity, at least at the mass level, was rather late to develop, crucially long after civil society networks (the core intermediate factor in the Swiss case) had emerged. In the cross-national, statistical analysis, I use long time lags of about 100 years and instrumental variable regressions to show that the causal arrow points in the direction my theory foresees.

Fifth, Philip Roeder wonders which of the three crucial intervening elements (civil society development, state capacity to provide public goods, and linguistic homogeneity) are “necessary, sufficient, or simply conditioning factors” for nation-building. As I discuss in the introduction, the case studies show that none of them is necessary, but they are individually sufficient. Switzerland is linguistically heterogenous, but early civil society development produced an inclusionary ruling coalition. Botswana lacks civil society development, but its record of post-independence public goods provision led to nation-building. China was not very efficient at public goods provision and lacked civil society networks below the elite level, but scriptural homogeneity allowed the emergence of alliances across linguistic divides, which in turn produced inclusionary ruling coalitions. On a side note, this is what makes Han China and Russia an appropriate pair of comparison: although “ethnically” homogeneous, the Han population is linguistically as diverse as that of the Russian Empire. Non of the Han language groups ever developed an ethnonationalist program, whereas Russia’s linguistic minorities did so and broke away from the common state both during the Russian Revolution and again in 1989—precisely the kind of comparative puzzles that Nation Building seeks to resolve.