Additional incumbent spending really can harm (at least some) incumbents: An analysis of vote share maximization

The literature on the effects of campaign expenditures on electoral outcomes implicitly suggests that incumbent spending cannot have a negative marginal impact on the incumbent's vote share. Indeed, that literature has spent a great deal of effort finding positive and significant effects of incumbent spending. This paper shows that there are circumstances under which theory predicts zero and even negative impacts of incumbent spending. Estimating equations derived from the theory provide strong support for the base model, though only weak support for the extensions which predict nonpositive marginal products for incumbents.