Indian Politics & Policy Volume 3, Number 1, Spring 2020 | Page 23

Economic Evaluations and the Incumbent Vote in India’s Parliamentary Elections (2014, 2019)
independent variables on the probability
of an incumbent vote. Note that
the negative values of marginal effects
(in Figure 1B) should be interpreted
as representing a negative effect on the
probability of an incumbent vote.
I find that economic evaluations
have statistically significant effects in
expected directions on the incumbent
vote, controlling for incumbent
party attachment and the passage of
time. Voters who evaluate their household
economic conditions as having
improved over the past five years are
more likely to vote for the incumbent
relative to those who see conditions as
unchanged. Note that the coefficient
for Econ.Impr falls to the right of the
vertical zero line and its confidence
interval does not intersect the vertical
line. Conversely, voters who evaluate
their household economic conditions
as having declined over the past five
years are less likely to vote for the incumbent
relative to those who see conditions
as unchanged. In this case, the
coefficient for Econ.Wors falls to the left
of the vertical zero line and its confidence
interval does not include zero. I
also find that attachment to the incumbent
party has a statistically significant
and positive effect on the vote for an
incumbent. This result too is expected
and confirms the effect of partisanship
on vote choice. Finally, the results
also show that the incumbent vote increased
in 2019 relative to 2014. That
is, voters were more likely to choose
the incumbent in 2019 compared to
2014. The positive effect of time on the
incumbent vote might be capturing,
for instance, an (unobserved) improvement
in the ability of a government to
leverage incumbency advantage in mobilizing
votes (or a similar incumbency
related effect on vote choice). It may
also suggest an increase in the ability of
the BJP to mobilize votes in 2019 (a BJP
specific effect). And without additional
time periods, we cannot say more
about this technical change.
The substantive (or marginal) effects
of positive and negative economic
evaluations are presented in Figure 1B
as probabilities. A positive evaluation
increases the likelihood of a vote for the
incumbent by 5 percent and a negative
evaluation lowers this likelihood by 11
percent. A negative evaluation results
in a larger decline in the likelihood of
voting for an incumbent compared to a
positive one. Attachment to the incumbent
party or coalition has the largest
effect on the incumbent vote. A voter
expressing attachment to a party is 62
percent more likely to choose the incumbent
party. Similarly, a voter is 21
percent more likely to choose the incumbent
in 2019 relative to 2014.
Is the effect of the economic vote
stable or is there a difference across the
two time periods? I test for effects of
economic evaluations conditional on
time by including two multiplicative
interaction terms to Model 1. These
terms are generated by multiplying the
economic evaluations variables with
the time dummy variable, i.e., Econ.
Impr*Year and Econ.Wors*Year. The
results from Model 2 are presented in
Figure 2A and associated marginal effects
are presented in Figure 2B.
19