points of select House committee members “tend to move counter to those of the President” and
that preference outliers are more numerous on House committees under conditions of divided
government (also see Kedar 2003 for a parallel argument with respect to voters in multi-party
parliamentary systems). Indeed, our model implies that when House members negotiate amongst
themselves, they may gain by providing extreme factions with positions on powerful House
committees, especially if such positions make the members more likely to serve as conferees.
Conclusion
One way to think about our research is that it answers the question, “If a tree falls on the
President of the United States, does the Speaker of the House get hurt?” While established
theories of legislative organization in the House are based on the premise that the well being of
House members and the President are independent, a constitutional theory of legislation reveals
otherwise. In it, a change in the President can alter the bills that foresighted conferees send to the
Oval Office. Such alterations, in turn, can affect the expected benefits of holding powerful
positions in the House, which can change the bargains that House factions are willing to strike
with one another when organizing the chamber. Such changes can, in turn, affect how much the
Speaker and his faction must offer to other factions in order to pursue their own policy objectives.
The health of the President and the well being of the Speaker, therefore, are not independent. A
change in the Senate can have analogous effects.
16
While our model is one of few attempts to consider House-Senate bargaining as a non-cooperative game,
it joins a growing list of such models. Ansolabehere, Snyder, and Ting (2003), for example, examine the
extent to which malapportionment in the Senate yields and unequal distribution of public expenditures. The
model represents an important advance. They find (2003:2) that “if both chambers operate under majority
rule and proposals originate in the lower house, then the expected share of public expenditures is the same
in all districts--the expected division of expenditures is not skewed toward the areas that are over-
represented in the Senate.” Since the models are set up to address different questions, they differ in
important ways. For example, their model offers far more detail on how House-Senate decisions affect
constituents in the House’s perfectly apportioned districts and the Senate’s malapportioned districts. Three
other differences make our design more suitable for questions of how changes in the Senate and Presidency
affect organizational decisions in the House. First, the House-Senate negotiation in our model can be
affected by Presidential preferences while their model features no president. Second, actors in their model
play a divide-the-dollar game. In our model, the total value of coalescence (i.e., the value of the pie that the
coalition factions ultimately divide) depends on the spatial relationship between the House factions’, Senate
and President’s ideal points -- where the latter two factors affect the pie’s value through their impact on the
conference bill. Third, in their model the Senate must consider the House’s bill under a closed rule. In our
model, the conference bill is the result of an efficient bargain between House and Senate.
27
Convention