Indian Politics & Policy Volume 3, Number 1, Spring 2020 | Page 25

Economic Evaluations and the Incumbent Vote in India’s Parliamentary Elections (2014, 2019)
Figure 3: Logistic Regression [Model 3: Dependent Variable – Vote for BJP in 2019]
Figure 2A indicates that the coefficients
for all variables retain expected
signs and significance. Two key results
from Model 2 are (a) the marginal effect
of a positive evaluation of household
economic conditions in 2019 is not
statistically different from that in 2014
and (b) the marginal effect of a negative
evaluation is statistically greater in 2019
relative to 2014. For instance, the probability
that a positive evaluation increases
the likelihood of an incumbent vote
is about 4 percent in 2014 and increased
to about 6 percent in 2019 (as seen in
Figure 2B). The difference between the
two time periods is not statistically different.
The effect of a negative evaluation
on the likelihood of an incumbent
doubled from 7 percent in 2014
to about 14 percent in 2019. A negative
evaluation had a significantly larger effect
in 2019. The effect of partisanship is
strong, although it drops marginally to
around 60 percent, as does the effect of
Year to 17 percent. 19
The final model (Model 3) focuses
exclusively on the BJP vote in 2019.
21