Price Theory, Vote Scores, and Campaign Expenditures: Signaling or Ideology?

Abstract

The empirical literature on legislator behavior attributes significant explanatory power to interest group vote scores. The literature has also dichotomized the interpretation of these scores as either ideological shirking or market signaling. Most studies simply regress a residualized vote score on a left-hand roll call vote. This paper innovates a new way of discriminating between signaling and ideology on the basis of residualized vote scores. The structural model considers a congressional seat as a unique, non-divisible good, and candidates' campaign expenditures as the unit price bid for the good. Introducing exogenous control variables, and executing 2SLS, the reduced form model yields a convenient test: if a more extreme voting record is negatively related to the bid, this indicates a substitution relationship and supports the market signaling hypothesis. This paper finds support of this kind in the 1994 House elections.

JEL Classification: D72