In this paper we tested the hypothesis that the performance of the economy is a determinant factor in the voter's preferences. The results obtained enable us to corroborate this notion, since the model has a high predictive value, and the behavior of the economy, both in terms of level of activity as well as in its changes, comes up consistently as a determinant factor of voting behavior. These results are also consistent with those obtained through Chilean polls in which it has been ascertained that economic events, particularly in the case of those who favored the anti-Pinochet option NO, were of utmost importance in the voters' decisions. The results also confirm the idea that, in Chile, women tend to give economic issues a lesser weight in their political preferences.
An implied additional insight is that voters are relatively sophisticated when it comes to interpreting the available information. The most significant economic variables are that of average unemployment throughout the last three years prior to the election, while the variable associated to changes in unemployment over the last year is of lesser consequence. It follows that voters have a
good memory
and willingly write off economic trends which could be deemed as pre-electionary in nature. This entails that, even though economic variables are a key factor in voter's preferences and electoral results, there is no possibility for governments to manipulate these variables and benefit from opportunistic behaviors.