Do Several Winning Coalitions Exist in a State for Senators of the Same Party? Evidence from an Event Study.

• Author: McGarrity, Joseph P.
• Position: Statistical Data Included

Joseph P. McGarrity [*]

Armand Picou [+]

We argue that U.S. senators from the same state and in the same party form different winning coalitions. We also develop a theory that stipulates that parties encourage these senators to form very distinct constituency coalitions. Parties use committee assignments as a carrot to give these senators an incentive to represent different groups. In our empirical analysis, we find that there are fewer overlapping committee assignments among senators in the same state when they are from the same party. We also consider the case of John Heinz and Arlen Specter, both Pennsylvania Republicans. When John Heinz died in a plane crash, Political Action Committees (PACs) that bought influence from Specter but not Heinz now had the possibility that the new senator replacing Heinz would include them in his resource constituency. The resulting competition by the new Pennsylvania senator and Specter for campaign resources would lower the price these firms had to pay for representation, thus improving their expected future earn ings.

1. Introduction

   When two senators hail from the same state, they may potentially serve the same voters and the same interest groups. If the Denzau and Munger (1986) model is correct, interest groups make campaign contributions to senators to get them to deviate from their voters' preferences and voters punish these deviations with less support at the polls. In this framework, two senators representing the same voters face identical opportunity costs in lost votes from accepting money from any particular interest group. Because interest groups buy influence from the senators who offer the cheapest service and these two senators from the same state have identical reservation prices for going against their voters' wishes, they are equally attractive to any single interest group. An interest group will want to either buy representation from both senators or from neither senator. Further, both senators will want to represent the same interest groups because these groups provide each senator with campaign money at the lowest cost in foregone votes.

   If the two senators offer their services to the same voters and interest groups, they are forced to compete for electoral resources. As in any market for a homogeneous good served by more than one producer with identical costs of production, competition among senators allows the buyer to pay a lower price (campaign money and votes) and to receive more output (Senate representation). These two incumbent senators are vulnerable to challengers who include a different set of voters and interest groups in their winning coalitions. These challengers can charge the new voting blocks a monopoly price (number of votes) for their services. While the challenger will face competition from politicians running in other senate races, these outsiders represent different constituents, giving him/her a cost advantage over some of these politicians toward some interest groups. This politician will be spared direct competition against a legislator with identical opportunity costs for accepting campaign money.

   When two senators from a state are in the same party, there is a heightened risk they will try to represent the same voters and thus the same interest groups. They may try to serve the same voters simply because some groups of voters in a state tend to be republicans and some groups tend to be democrats. Senators from the same party would gain if they could agree to differentiate their electoral and resource constituencies as much as possible to avoid competing with one another. However, this agreement is not enforceable and the cartel by itself will collapse. Parties, while unable to enforce the collusive agreement, can provide incentives for senators from the same state in their party to choose to represent some different voters and thus some different interest groups. The party, by allocating different committee seats to members from the same state, influences each senator's choice of which voters and interest groups to represent. A seat on a committee lowers the cost to a senator for serving voters and in terest groups concerned with the policy under the committee's jurisdiction, leaving extra electoral resources for a senator who chooses to represent this group.

   In this article, we develop a theory that stipulates parties use committee assignments to lessen the competition for electoral resources among senators from the same state. By giving these senators different committee assignments, the party provides incentives for senators in their party to serve different groups of voters and different interest groups. Dividing the market increases each senator's electoral resources, providing an edge in the legislator's reelection efforts. In our empirical analysis, we find evidence consistent with our theory. For instance, in the 101st and 106th Congresses, there are fewer overlapping committee assignments among senators in the same state when they are from the same party; this result suggests parties try to limit competition among members from the same state. We also consider the case of John Heinz and Arlen Specter, both Pennsylvania Republicans who had different committee assignments, different electoral constituencies, and different resource constituencies. We employ a n event study and find that corporations with PACs that contributed to Heinz but not Specter experienced an abnormally large drop in their stock prices after Heinz's unexpected death in a plane crash on April 4, 1991. We also found that companies that gave money to Specter but not Heinz experienced an abnormal gain in their stock prices after the crash. These results suggest Heinz and Specter offered interest groups a different price to provide representation. Each senator's choice of which voters to represent as well as his committee assignments impacted the PACs to which he or she could cheaply sell influence. PACs who were able to buy cheap influence from Heinz had their access to influence put at risk by Heinz's death. Contrastingly, PACs that could cheaply buy political representation from Specter now had the possibility that the new senator who replaced Heinz would put together an electoral constituency much like Specter's. These PACs may soon have another senator who would compete with Specter for inte rest group money, driving down the price Specter can charge for his representation and giving these interest groups the benefit of senatorial competition.

2. There Can Be More Than One Winning Coalition of Voters

   Black (1948) found that, under certain conditions, majority rule yields stable electoral outcomes. His analysis assumed voters have single peaked preferences and one issue is being decided on at a time. He found that, in a pairwise election, the candidate who adopts the optimal policy position of the median voter is the condorcet winner in a majority-rule election; that is, another candidate picking a different policy position cannot beat the first candidate.

   However, when we relax either of Black's two assumptions, the stable outcomes disappear and no unique electoral equilibrium emerges. When voters have double (or more) peaked preferences and outcomes are determined by majority rule, more than one election result may take place. For example, during the Vietnam War, citizens may have had preferences that gave them the most utility at two different extreme policy decisions. Citizens may have wanted us to either stay out of the conflict or go into the region with immense firepower and take care of the job with overwhelming superiority. [1] A candidate can win an election by picking either platform. Second, when voters consider more than one issue at a time, a candidate can create a platform to attract a majority of the voters. However, several winning platforms exist and outcomes will vary over time. [2]

   Nevertheless, whether or not majority-rule outcomes produce stable platforms for the winners of Senate elections is primarily an empirical question. Many studies have documented the wide spread occurrence of split party representation in the Senate. For example, Brunell and Grofman (1998) report that 28% of the states had split delegations between 1788 and 1996. The range for this data was 4% and 60%, which occurred in 1904 and 1830, respectively.

   Several reasons have been given for divided representation. Brunell and Grofman (1998) report evidence that split representation in a state's Senate delegation is often caused by staggered elections extending the period of time it takes for a realignment to translate into changes in representation. Fiorina (1992) claims a moderate electorate may find it rational to send a split delegation to Washington. Cox and Katz (1996) and Gelman and King (1990) claim this phenomenon may be a result of the increased importance of incumbent performance and the relative decline in the importance of party ties.

   Only two studies have attempted to empirically determine whether different winning platforms exist for members of the same party. [3] However, these articles do not measure the composition of a legislator's winning coalition directly. The first article used an index of roll call votes as a proxy for constituent interests, in effect assuming politicians vote the interests of their voters. This article by Goff and Grier (1993) finds the differences in the absolute value of Americans for Democratic Action (ADA) scores between senators from the same state (they use ADA scores to measure different constituent interests) can in part be explained by party as well as various measures of the heterogeneity of the voters in a state. Their results suggest a more diverse electorate allows politicians more choice of which groups to represent. The second article used a lower level of aggregation than is typical. In this article, Peltzman (1984) used county-level data instead of statewide data to control for constituent...