This paper develops a test of the hypothesis that voting is an investm
ent by voters in desired electoral outcomes (the 'investment theory'),
ag opposed to the alternative hypothesis that voting is a consumption
good (i.e., the act of voting itself gives utility to the voter, akin
to the 'warm glow' theory of provision of public goods). It is shown
that the investment theory implies that voter abstention should be a f
unction of indifference between the opposing candidates, but not of 'a
lienation' from all candidates. In other words, the difference in util
ity the voter expects to receive from the two candidates, and not the
absolute level of utility from his favored candidate, matters in the i
nvestment theory. Micro survey data from the 1976 U.S. Presidential el
ection support the hypothesis that, on the contrary, the absolute leve
l of utility had a significant effect on abstention, while the differe
nce in utilities had no effect. Thus the evidence supports the 'consum
ption' theory and tends to reject the investment theory. An implicatio
n of this result is that the act of voting may not be as paradoxical a
s is commonly believed, because the free-rider problem is irrelevant i
f voting is a private, consumption good.