The question of the role institutions play in party aggregation has tended to be dominated by discussions of the role and effects of different electoral systems. In a refreshing change from this preoccupation with electoral systems, Pradeep Chhibber and Ken Kollman turn our attention to the impact of federalism, or, more appropriately, the degree of centralization or devolution of power in governmental systems, on the fragmentation of party systems. Their focus is dynamic: They are interested in the formation of national party systems and in how party systems change over time with respect to their extent of nationalization and in relation to the migration of political authority to the center or from the center to states, provinces, or regions. Highly nationalized party systems are ones in which parties receive similar vote shares across different levels of vote aggregation: district, regional, and national. Conversely, in weakly nationalized party systems, parties are differentially competitive across levels.

The authors have chosen to examine their questions about national party systems using four countries that share a similar electoral system—the single member plurality system—for elections to their lower houses, but that have had different histories with respect to centralization and “provincialization.” Their data for measuring the nationalization of party systems are district returns for lower-house elections from the outset of democratic elections in Canada, India, and the United States and from 1885 for Britain. Analyses of the centralization/decentralization trends in governing across the four countries are based on a (necessarily) eclectic mix of data on government activities and secondary sources on federal/provincial/regional government relations and activities. The focus of their study is mainly directed toward looking at differences within each country over time, although they also refer to some differences across the countries, especially why American politics is dominated by two national parties while this is not true of the other three countries in their study.

The analysis begins by looking at Maurice Duverger's two-party prediction for single member plurality systems and examining whether it holds at the district level. Although the discussion of Duverger's law and the formal theoretical models that have followed from it, as well as the findings both with respect to the extent of and limits to two-party competition at the local level, are interesting, they are less than central to the primary concern of the book, which is the degree of similarity in party support across levels, and not the number of competitive parties.

In Chapter 3, the authors elaborate their theory of party aggregation, which is based on the incentives that exist for voters and candidates to make links across districts. They argue that when national governments centralize authority and create policies that routinely affect peoples' lives, because they want to influence the direction of these policies, candidates and voters have incentives to create and support nationwide linkages in the form of national parties. On the other hand, with decentralization, where state or provincial governments enact important policies, it becomes “more likely that there will be an increase in voting for parties with regional, state or provincial labels” (p. 80). With this focus on the authority of state or provincial governments in accounting for voting support for regional parties, what is missing is any discussion of the possibility that voters and candidates may link up in regional parties because of regionally based interests that are affected by national government policies.

Chapter 4 describes how the principle that national parties are formed to influence policymaking at the national level worked historically in each of the four countries, tracing the movement among legislators from notables to party members. Chapter 5 then details shifts in government centralization in each country, categorizing periods as centralizing, decentralizing, or ambiguous. Chapters 6 and 7 lay out the authors' evidence for their central thesis on the relationship between centralization/decentralization and party system change. Their critical indicator of party nationalization is one they call party aggregation. They use Laakso and Taagepera's (Markku Laakso and Rei Taageper, “‘Effective’ Number of Parties: A Measure with Application to West Europe,” Comparative Political Studies 12 [1979] 3–27) measure for the effective number of parties, and define party aggregation as the difference between the effective number of parties at the national level and the average of the effective number of parties at the district level. While the simplicity of the measure is its virtue, it can overestimate the degree of nationalization in a system because it takes into account only the effective number of parties contesting at each level. As a result, the authors describe the Canadian party system at the end of the last century as “curiously nationalized” (p. 189) because their measure suggested it was nationalized, even though the party system had a “lack of cross-provincial voting” (p. 190) and, by virtually all other counts, was highly regionalized. Still, with this and other measures, they are able to illustrate a general relationship between centralization and nationalized party systems, especially in the earlier years covered in each of their countries.

What is more difficult to establish is the direction of causation in this relationship. Does the degree of centralization or provincialization cause party system change, or does the regionalization/nationalization of party systems affect government centralization? Chhibber and Kollman acknowledge that sometimes threats from the rise of regional parties do contribute to devolution, but they contend that the chronologies in the countries they have studied clearly indicate that growing centralization contributes to a more national party system. Moreover, they argue that even with the more difficult case of provincialization, their chronology shows that regional/provincial parties emerge after regional governments have gained power. Yet how quickly this occurs seems quite variable in their chronologies, and why it occurs seems undertheorized. Although the authors do refer to “more opportunities for regional, state or provincial parties to thrive” (p. 21), and to regional parties finding “the political space to form and survive” (p. 101) in decentralized systems, a more extended discussion of how decentralization works to support regional parties would be welcome.

With its focus on state power and party aggregation, The Formation of National Party Systems addresses an issue that has not, until now, received sufficient attention. It presents a number of useful theoretical discussions about party aggregation and provides a wealth of data on the topic. It documents, for example, that the United States did not have a two-party system until 1930, and it illustrates that even in a country such as India, which is very diverse within districts, Duverger's law still tends to hold at the district level. Although some country specialists may quibble with particular interpretations of the timing of changes in their variables, the authors' general thesis is largely born out. Overall, this book makes an important contribution to the literature on party systems and system change.