Congressional organization and politics seems to change roughly every generation. The literature has identified 12 eras of congressional organization—outlined in Table 1 (see Galloway 1976, Hinckley 1988)—that can be classified by their degree of centralization of power. In some eras, strong party leaders (such as Joe Cannon or Newt Gingrich) control legislative organization and policy outputs. In other eras, control is decentralized to committee chairs, subcommittee chairs, nonpartisan coalitions, and so on.

We present two polar models of agenda power. In what follows we contrast these two models with each other and test their predictions head-to-head. Using outcomes from the 45th to 99th Congresses (1877 to 1986), we find that we can reject the floor agenda model. In contrast, there is substantial and credible evidence supporting the cartel agenda model. In particular, we show that negative agenda control is indeed a largely invariant advantage of majority status; and that positive agenda control is variable, changing with the internal homogeneity of the majority party. Variations in these two aspects of agenda control, we suggest, explain some of the basic tensions and historical fluctuations in congressional organization.