War Chests and Information
Beyond anecdote, Sorauf states that war chests may have dual purposes for incumbents:
“[Incumbents] raise money early to discourage would-be challengers both by displaying their
financial prowess and by setting a financial hurdle for any challenger. Even in short run
terms, incumbents raise large sums as a form of catastrophe insurance against the sudden
emergence of a strong and well-financed challenger...” (1988, 160). Furthermore, Sorauf
continues, “The incumbent may also simply be saving for a future campaign for the present
office” (1988, 161).
1
Previous theoretical models of war chests have considered only the deterrence or the
savings motivation. Epstein and Zemsky (1995) develop a formal model of campaign fund-
raising, and show that although fund-raising can deter strong challengers in some situations,
it will be difficult to observe this empirically.
2
In this case, war chests are an instrument
(exogenous to the model) used to test the comparative statics of their model. Thus, war
chests per se do not deter challengers, but they may help the analyst determine whether pre-
emptive fund-raising deters challengers. The potential problem with this characterization is
that the very fact that an incumbent has a war chest may have something to do with incum-
bent strength and challenger quality of the previous election–things Epstein and Zemsky
want to explain. War chests are a conscious decision by incumbents not to spend money,
and thus should not be treated as an exogenous variable in any comparative statics test.
In the model I present, war chests are an endogenous result that occurs under specifically
1
Sorauf also notes that war chests can be used to build a nest egg for retirement (possible for those elected to
the U.S. House before 1980 who retired before 1992), to run for higher office, or for unfavorable circumstances
after reapportionment.
2
It would be difficult to observe deterrence empirically because their model has both pooling and separating
equilibria under the same conditions.
2