Date of Award

January 2013

Document Type


Degree Name

Master of Science (MS)


Economics & Finance

First Advisor

Cullen Goenner


Prior research on the effect that candidate spending has on vote share has produced conflicting results, with some research finding that only spending by challenger candidates has an effect on the vote outcome and other research finding incumbent and challenger spending roughly equally effective. The paper explores two suspected causes of the discrepancy in prior results: measurement error and endogeneity in the spending variables. Using new spending data from the Federal Election Commission that can readily be categorized into different spending types, I find evidence of significant amounts of inconsequential spending that should be excluded from candidate spending totals. Additionally, I show that isolating spending on broadcast media reduces underestimation of the effect candidate spending has on vote outcomes, and employ a new instrumental variable incorporating differences in television advertising costs to find both incumbent and challenger spending have a positive impact on vote share, but to different degrees.