How Money Drives US Congressional Elections

Abstract

This paper analyzes whether money influences election outcomes. Using a new and more comprehensive dataset built from government sources, the paper begins by showing that the relations between money and major party votes in all elections for the U.S. Senate and House of Representatives from 1980 to 2014 are well approximated by straight lines. It then considers possible challenges to this “linear model” of money and elections on statistical grounds, resting on possible endogeneity arising from reciprocal causation between, for example, popularity and votes. Extending the analysis of latent instrumental variables pioneered by Peter Ebbes and recently analyzed by Irene Hueter, the paper tackles this much discussed problem by developing a spatial Bayesian latent instrumental variable model. Taking a leaf from discussions of event analysis in economics and finance, the paper also examines the light thrown on the model’s usefulness by studying changes in the gambling odds on a Republican takeover of the House in 1994. Both approaches suggest that reciprocal causation may happen to some degree, but that money’s independent influence on elections remains powerful. A concluding section of the paper considers the alleged “centerist” leanings of American large corporations by comparison with members of the Forbes 400 and evidence that the effect of money in House elections has dropped slightly over time, though it remains extremely strong.