The Spending Game: Money, Votes, and Incumbency in Congressional Elections

This paper takes a game-theoretic approach to the analysis of the spending-votes relationship in Congressional elections to reinvestigate the surprisingly weak effects of incumbent spending measured in previous studies. Rather than focusing narrowly on the impact of spending on electoral outcomes, we attempt to take account of the reciprocal effect of (anticipated) closeness on spending using several statistical approaches. We also offer improvements in the specification and measurement of the vote equation, by using a better measure of district party strength adjusted for year-effects, and by including a variable that measures the heat of the campaign in terms of total spending by the incumbents and challengers. The latter measure partially corrects for the simultaneously determined (and highly positively correlated) levels of incumbent and challenger spending. A more rigorous multiequation simultaneous equations model, identified by uncorrelated errors, provides even more leverage for sorting out the effects of incumbent and challenger spending on votes. That analysis indicates (in a complete turnaround from findings reported elsewhere) that incumbent spending effects are highly significant and of a magnitude that is, if anything, greater than challenger spending effects. The paper concludes by using a game theoretic model to estimate the effect of anticipated closeness on spending and to estimate differences in campaign financing costs between incumbents and challengers