Few topics in the study of American politics attract more attention than divided government. Yet while we have a reasonably good handle on why divided government occurs, our understanding of the effects of divided government remains limited. For example, David Mayhew's (1991) Divided We Govern and the huge literature it spawned reveal a host of contradictory findings about divided government's impact on major policy production. Why is this? Perhaps because we usually treat divided governments as empirically indistinct when there are very good theoretical reasons to believe that divided governments differ—and that these differences directly affect politics. Reducing divided government to a mere dummy variable fails to account for huge variations in factors like congressional party margins, party cohesion, and party polarization.

In this valuable volume, Richard Conley takes significant strides toward developing a much richer perspective on divided government, particularly as it applies to the behavior and impact of the president. Looking at the post–World War II period, he argues that a president's legislative effectiveness depends on “presidential leverage in political time” (p. 7). Leverage runs along a strategic continuum from the positive (“coalition-building leverage”) to the negative (“veto lever-age”). The way divided (or unified) government interacts with the electoral, institutional, and environmental contexts deter-mines where a president resides on the continuum.

Conley goes on to argue that during the postwar period, these contexts aligned in various ways to create three secular eras: Bipartisan Conservative [1947–64], Liberal Activist [1965–78], and Postreform/Party-Unity [1979–98] (p. 36). Each era was characterized by its different influences on presidential effectiveness, especially under divided government. Factors such as declining presidential coattails, increases in the incumbency effect, increased centralization of power within Congress, generally decreasing deference to the president's agenda, and, of course, the dramatic rise in the internal cohesion and external polarization of the two parties served to put recent divided-government presidents on the defensive. So, for example, President Eisenhower effectively asserted positive leverage because he could build cross-partisan coalitions to pass legislation. In contrast, the first President Bush, unable to find enough Democratic votes to support his agenda, resorted to aggressive use of negative leverage: the veto.

Conley sets up his basic theory in the Introduction. Chapter 1 follows with a series of statistical analyses examining the impact of the contextual factors on presidential success in Congress. He combines this analysis with a separate examination of the way the three eras—interacting with divided government—affect success. His results largely conform to his arguments, but scholars familiar with the mammoth literature on presidential success will not find many surprises. His analysis of the president's contribution to the production of major legislation is more unique. Using Mayhew's familiar list, Conley identifies the subset of major legislation that originated with the president. He finds that a large proportion of important legislation that was passed during unified governments came from the president's agenda. Those important laws passed during divided governments are substantially less likely to come from the president's agenda, and this likelihood declined over time.

The rest of the book features careful and statistically informed case studies of various groups of presidents. A sequence of chapters examines the Truman, Eisenhower, Nixon, Reagan, Bush, and Clinton experiences under divided government. These are followed by two chapters on unified government Presidents Kennedy, Johnson, Carter, and Clinton.

A necessarily terse review such as this cannot do justice to what is a nuanced work. Conley deftly combines quantitative analysis with a keen and detailed sense of the strategic challenges different presidents faced. Nonetheless, early in the book he overstates some of his claims about era differences. As just one example, the claim that presidential coattails in opposition party districts “shrank rather considerably during the postwar period” (p. 17) is not supported by the data in Figure I-3.

Conley's use of three different eras is well grounded in the literature (though its association with Stephen Skowronek's well-known conception of “political time” is loose). I am uneasy when scholars endow time periods with meaning. At worst, an era is simply a summary of events that happened. This can quickly lead to tautology when we denote some period as “Activist,” for example, and then use an Activist Period variable to predict something like policy production. At best, and certainly the way Conley uses it, the denotation of an era is simply a convenient way to summarize some bundle of independent factors that happened to coincide. Thus, the Bipartisan Conservative Era is an easy way of describing a particular mix of electoral, partisan, institutional, and environmental factors that existed at a particular time. These underlying factors, and not the eras themselves, cause the variation in the phenomena we care about. In the context of this book, a measure that simply denotes an era necessarily masks important variation within the era, even after we distinguish between divided and unified government. Thus, for example, Conley finds that Truman and Eisenhower had very different divided-government experiences, even though both served during the Bipartisan Conservative Era. On a some-what related note, I am not sure that the interaction effects between the divided government variable and the three era variables would hold up if the underlying dummy variables, rather than just the interaction terms, were in the models. Arguably, they should be included.

Finally, it is unfortunate that this book fails to confront Keith Krehbiel's (1998) Pivotal Politics. Krehbiel argues that it is not divided government per se that is important. Rather, the underlying structure of preferences and institutions is key, and this structure may correlate with unified or divided government only weakly, if at all. Despite a very different theoretical approach, Conley's work shares some basic insights with Krehbiel's, but it differs dramatically in the way it conceptualizes divided government—and thus party—as crucially important, no matter the era. Conley passes up the opportunity to show why his approach is potentially superior.

These quibbles aside, Conley has written a fine book. The Presidency, Congress, and Divided Government should reinvigorate and redirect the literature on Congress and presidents.