This is a helpful addition to the literature on congressional parties. After dispensing with the research problems inherent in studies that have answered “no” to the question of “Do parties matter?” Steven Smith sets out to redefine the research agenda for congressional scholars. To begin, researchers must more rigorously delineate when parties matter and how parties matter.

This is not an easy task. It is easy for legislative scholars to find anecdotes to support their claims about congressional parties. For example, it is impossible to read Robert Caro's Master of the Senate (2002) and argue that parties and their leaders have not affected the decisions of individual legislators. Caro's description/recounting of Lyndon Johnson twisting arms to the point of costing senators reelection are strong examples of senators not just following their ideological beliefs or maximizing their chances of reelection. But over the history of Congress, there are anecdotes to suit every theory.

The search for systematic empirical proof of congressional party influence is of course more challenging. This may be generally true, but Smith explains how finding statistical evidence of party influence is particularly difficult. There have been thousands of roll call votes in congressional history with dozens or hundreds of members participating in each roll call. Even if party influence were occurring in its simplest form and out in the open—perhaps if C-SPAN's cameras could capture leaders exerting party pressure through some version of the “Johnson Treatment”—we would still find the number of cases where legislators went against their particular interests overwhelmed by the cases where they were simply left to make their own decisions. And party influence can be wielded well in advance of any roll call being taken; indeed, parties may have their greatest influence in preventing a roll call from occurring or structuring how the vote will occur.

Smith is careful to sort through the ways in which we should look for evidence of party effects and the areas where we might find them. This is complicated by variation across issues, rules, eras, and chambers. In fact, his call for paying more attention to the Senate in studies of congressional parties is one of the key ways he sets the scholarly agenda. He explains how both direct and indirect forms of party pressure can be exerted and that we should expect the need for such exertions to depend upon majority status and upon the relative sizes of the party delegations.

One interesting question addressed by the book is whether the primary purpose of congressional parties is to maximize policy outcomes, electoral success, or both. Smith makes a case for “both,” citing classic studies of party leadership elections. This differs from the conditional party government (CPG) approach where policy change (or, more precisely, maximization) is the paramount goal. Maximizing party seat share is helpful toward this goal, and some versions of CPG stress Richard Fenno's view that parties serve legislators who themselves have multiple goals. More recent discussions of cartel theory begin with the view that electoral goals are paramount but that policy maximization is central to that goal, and so the cartel model focuses on policy outcomes. The more recent theory of strategic party government (SPG) posits that parties seek to maximize seat share and that winning votes and changing status quo policies is an instrumental goal—parties do these things to bolster their reputation but not per se for the satisfaction of party members.

This leaves us with the subtle question: Do parties enact laws so that they can gain seats, or do they gain seats so that they can maximize their policy goals? We frequently observe behavior that is directed at both policy change and winning elections, but perhaps for the sake of parsimony we can reasonably treat one goal as paramount and the other as instrumental. The empirical question is whether a significant amount of legislative behavior can only be explained by a double-goaled approach: parties forcing through policy changes that are adverse to their electoral interests and parties acting in ways that promote their electoral interests while betraying their policy preferences. In Party Influence in Congress, Smith does not carry this empirical exercise very far, but he does lay out a feasible alternative to the prevailing theories on congressional behavior, and in doing so, the biggest contributions here are the structure he gives to the search for party influence and the challenges he makes to congressional researchers to engage the agenda he sets out.