29
In Model 2.3 we can observe how the clarity of responsibility does not have the
effect predicted by PW. The interactive variables between the clarity context and the
economic variables lack statistical significance, as expected. Although the Non-Myopic
Growth coefficient loses significance, high unemployment and inflation still lower the
probability of reelection of the incumbent party. Finally, in model 2.4 one can observe
that even without controlling for the institutional context, at least with respect to inflation
and unemployment, one can observe economic voting. However, growth does lose
statistical significance, which may mean that at least for some type of economic
performance, it does have an impact. Nonetheless, the models in table 2 show that
contrary to PW, bad economic performance will be punished independently of the level
of clarity that voters are facing. This implies that nothing will save the incumbent party
from punishment if it makes a bad economic performance.
Finally, from the models I ran with the extended data set that includes presidential
and parliamentary elections in 41 democracies (Table 3) one can conclude the following:
the clarity context does not have the impact on economic voting predicted by PW. In fact,
the positive signs of Clarity and the interactions of Clarity*M Inflation and Clarity*Non
M. Growth (although they lack statistical significance), would suggest that a high clarity
context, benefits, rather than harms the incumbent party. Incumbents will have a higher
probability of reelection in high clarity contexts than in low clarity contexts, however,
nothing will save them from economic punishment. This confirms the previous results of
model 2.3. Another interesting result from models 3.1 and 3.2 is what the coefficient of
the ideology interactive Right Wing*M.Inflation suggest. Contrary to what PW found,
and in line with Kiewiet, one can observe that left wing governments have a lower
Convention