Get access

Equilibrium Party Government


  • This article has benefited enormously from discussions with John Aldrich, Christopher Berry, Ethan Bueno de Mesquita, Barry Burden, Ernesto Dal Bo, Sean Gailmard, Will Howell, Maggie Penn, Keith Poole, David Rohde, Eric Schickler, Jasjeet Sekhon, Jim Snyder, Ken Shepsle, Craig Volden, and Alan Wiseman, as well as the comments of seminar audiences at University of California, Berkeley, University of Chicago, Duke University, and Harvard. Thanks are also due to Keith Poole and Howard Rosenthal for making their data available. As always, the author retains sole ownership of all faults, omissions, and errors contained herein.

John W. Patty is assistant professor of government, Harvard University, 1737 Cambridge St., Cambridge, MA 02138 (


In this article, I present an equilibrium model of party government within a two-party legislature. The theory is predicated upon members of the majority party having potentially conflicting individual and collective interests. In response to this potential conflict, the members of the majority party endogenously choose a degree of control to grant to their leadership. The equilibrium level of party strength is decreasing in the size of the majority party and increasing in the strength of opposition among members of the minority party. The theory implies that the average performance of W-Nominate estimates of majority party members' ideal points will be a decreasing function of the size of the majority party while the performance of these estimates for members of the minority party will not be affected by the size of the majority party. Using data from the U.S. House and Senate between 1866 and 2004, the theory's predictions are largely consistent with roll-call voting in both chambers.