29
What is more surprising is how quickly these shocks filter out, with the impulse
responses quickly heading towards zero by the second day. The effects of shocks to the
Bush series also close out quickly, though those effects are not statistically significant at
any point.
Figure 2 About Here
Why are weak rational electoral expectations found in the case of Bush versus
Gore? There are two main explanations. The first reason may be that the data used
actually starts too late. Bush and Gore were long expected to be the nominees of their
respective parties. The data used here starts on December 14
th
, 1999, while voter
expectations about whether or not Bush could beat Gore in the general election had been
forming for quite some time before that. The other candidates involved only began
entering into the national conscious in late 1999, so any match-ups dealing with those
candidates would be more appropriately measured using the starting point of the
Annenberg study. There is little that can be done about this problem, however.
The second explanation for why weak rational electoral expectations is found is
that because Bush and Gore had been expected to be the nominees for so long,
expectations about that race were relatively firm, and outside information would not be as
effective in changing these assessments as they might otherwise have been. As a result,
only some information would affect these assessments of Bush and Gore’s electability.
The finding of weak rational expectations, therefore, is the result of voters already having
firm ideas about the chances of Bush and Gore against each other.
Discussion
Convention