At first blush, the authors of these two books on territorial politics appear to be living in and describing two very different worlds. Motivated by the desire to understand why so many countries in recent years have decentralized and what this means for economic development, Jean-Paul Faguet and Caroline Pöschl have brought together pairs of policymakers and scholars to analyze the causes and consequences of decentralization. Ken Kollman, in contrast, focuses on what he sees as a dominant trend in the opposite direction, namely, the nearly inevitable tendency of a wide variety of federated institutions to become more and more centralized over time. It would hardly seem possible that the world today is simultaneously becoming more decentralized and more centralized. While the perspectives adopted by these two books can thus be difficult to reconcile, one crucial commonality is that both reflect a healthy dose of skepticism about national government officials who claim that they have decentralized. “True” decentralization after centralization is quite rare, according to Kollman, and “cynical” as opposed to “sincere” decentralization can actually facilitate continued rule by the center, in the words of Faguet and Pöschl.

One of the major contributions of The Perils of Centralization is that it connects the literature on territorial politics (i.e., decentralization, centralization, federalism, and multilevel governance) with the literature on presidentialism and separation-of-powers systems more generally. Kollman’s core claim in this regard is that we can understand changes in the vertical distribution of power between levels of government only if we also pay attention to horizontal relations between actors situated at the same level in the institutional hierarchy. The theoretical argument that Kollman develops is a simple yet powerful one. It begins when autonomous subunits in federated systems assent to the vertical centralization of authority to representative institutions, where they maintain a voice over policy decisions. Over time, however, what he calls representative centralization is replaced by executive centralization, as representative bodies cede authority to executives who develop capacities for consolidating authority that eventually “lock in” centralization. Partisanship is the lubricant that facilitates this mostly inexorable—but not deterministic—process of locked-in executive centralization. Partisanship develops “when representatives of subunits begin to act less as supporters of subunits and more as independent agents in a game for control of the central unit” (p. 29). In many ways, then, Kollman’s theoretical framework resembles a marriage between path dependence and principle agent perspectives, though he does not explicitly use vocabulary from either tradition, and leaves it to the reader to ponder how these insights from rational and historical institutionalism might be integrated.

The most noteworthy and thought provoking—indeed remarkable—aspect of this book, however, is not its theory but, rather, its research design, particularly Kollman’s decision to examine how centralization takes place in four very disparate cases of federated institutions, three of which have experienced “lock in.” Two of his cases are more obvious and familiar: 1) the U.S. government as it experienced representative centralization over the course of the nineteenth century and then executive centralization under the two Roosevelt administrations, and 2) the European Union, which has on occasion come close to embracing executive centralization but has so far stepped back from the precipice—even though Europe-level partisan affiliations have begun to deepen. More unexpected are Kollman’s other two cases: the Catholic Church and General Motors. Most intriguing is his account of the power struggles between popes and bishops that took place over the course of many hundreds of years of church history, marked by synods, councils, schisms and anti-popes. These struggles ultimately “locked in” papal dominance, thanks to Vatican I and II in the 1870s and 1960s, after which “Rome can revoke any autonomy the bishops have” (p. 102). The General Motors story is more truncated but no less fascinating; though founded in 1908 on the very principle of divisional autonomy for the various auto companies that came together to form GM; already by the early 1930s, central managers had wrested decision-making authority away from the component “subunit” auto divisions.

Precisely because it is conceived as the lubricant that facilitates the various centralizing processes that Kollman describes, partisanship is the concept in this book that bears a lot of weight and deserves the most critical reflection. It is also where a more sustained engagement with William Riker would have been useful, given Riker’s influential argument that partisanship often shapes how federalism operates. Reflecting the diversity of his four cases (two of which are nongovernmental organizations), Kollman describes partisanship not as loyalty to a political party but, rather, as the more general “self-identification with a group seeking common policy goals” (p. 29). Defining partisanship as support for common policy goals would seem to imply ideological or ideational affinity, and yet what partisanship really seems to mean for Kollman is the emergence of actors in subunits (i.e., U.S. states, GM divisions, local churches, and European countries) who deem that their interests would be furthered by centralization.

Why some subunit actors come to this conclusion remains an open question. One idea is that they do so when their copartisans capture the central unit, but this would seem to assume, rather than explain, the existence of partisanship. The focus on partisanship as the allegedly critical factor also raises questions about other causal factors like economic crisis and fiscal resources, which play important roles in the empirical cases but are relegated to less important roles in the theory. Could economic crisis or fiscal revenues under the center’s control substitute for partisanship in the locking in of centralization? Finally, the development of partisanship among existing subunits is difficult to detect in the chapter on the church, and appears not to have occurred at all in the case of GM, where central managers were able to take autonomy away from the auto divisions, not because copartisans emerged in these subunits who wanted to empower them but because the managers unilaterally moved to add new “corporate subunits that produced industrialized parts used by all the auto divisions” (p. 119).

Why politicians decide to reassign authority between levels of government is also a major question posed by Faguet and Pöschl, though they are interested in changes that decentralize rather than centralize authority. The volume that they edited is divided into three main parts organized around three main questions. First is a set of chapters that focuses on the decision to decentralize, including an extended interview with the man responsible for one of the most ambitious and radical episodes of decentralization anywhere in the world: Bolivian President Gonzalo Sánchez de Losada. Part II asks about the design of reforms once reformers have decided to decentralize, and extracts useful lessons about how to assign taxes, transfers, and expenditures to different levels of government that are based on the experience of more than 10 mostly developing countries. The third part of the book poses the logically succeeding question: Once it has been adopted and designed, what are the effects of decentralization on the provision of governmental services, including health, education, and infrastructure?

The edited volume makes a number of distinctive contributions toward its overarching goal of pushing the literature on decentralization back toward “large, nuanced, complex questions that rank among policymakers’ first concerns, but are methodologically difficult to answer” (p. 22). One such contribution can be seen in Thomas Bossert’s fine chapter on decentralized health systems. Whereas the literature has emphasized the importance of variation in the level of decentralization across policy areas (fiscal vs. administrative, health vs. education), Bossert shows that we should be thinking about variation within policy areas and across the specific functions within those areas. More specifically, based on health outcomes in Guatemala and Ghana, he argues, functions like inventory and logistics should be centralized, while planning and budgeting are better decentralized—all within the health sphere. Also innovative are the two chapters that examine attempts to encourage competition between municipalities. One chapter on the Philippines argues that when neighboring municipalities are required by law to publish standardized, comparable information about their spending decisions, “yardstick competition” can improve performance and lessen elite capture. Another chapter, sounding a more cautionary note, finds that Area-Based Competitions in China have generated unintended and negative consequences, as when competing cities demolished older houses in the name of “city beautification.”

Is Decentralization Good for Development? provides a number of persuasive answers to the question posed in the title, but it succeeds less well as an attempt to connect the decision to decentralize with substantive outcomes. First, the distinction between “cynical” and “sincere” decentralization may not be a useful one. By “cynical,” the editors refer to politicians who decentralize not because they are genuinely interested in devolving authority but due to a series of short-term political calculations. By “sincere,” they refer to policymakers, including those who participated in the edited volume, who were motivated by normative commitments to democracy and development, and whose sincerity generates better outcomes. In practice, it can be difficult to determine the cynical versus sincere nature of decentralization, and Sánchez de Losada is a case in point. Although he is offered as an example of a sincerely motivated president who decentralized in order to rein in separatist regional politicians, it is also possible to interpret his support for decentralization, as Kathleen O’Neill has done, in more cynical terms and as an attempt to benefit his own political party disproportionately.

Second, given the complexity, if not impossibility, of determining whether politicians’ motives are sincere, the political science literature on decentralization has probably been correct to assume cynicism in the form of self-interested, career-oriented politicians. More generally, it is not at all clear, as the editors claim, that the effect of decentralization on development is “very much determined by the motives” (p. 7). In fact, one of the most interesting aspects of decentralization is the Pandora’s box it tends to open; cynical politicians seeking limited change often trigger sequential dynamics that end up devolving significant power to local officials, as Tulia Faletti has demonstrated. At the same time, politicians at the center setting out to sincerely and fundamentally transform their country’s politics often manage to change very little. As Perils of Centralization demonstrates, politicians’ initial motivations set in motion dynamics that they have a hard time controlling.