Out or sample prediction errors
Because the model in my AERpaper uses insample errors, it may
underestimate the uncertainty in the upcoming election results. To see
how important this is, I estimated the outofsample errors. The average
absolute prediction error in the state level vote share increases from 3
percent insample to 3.7 percent outofsample. Most of this is due to
larger outofsample national swings.
This has large impacts on the estimated probability of winning. For
example, in the estimates made October 1,
using variance estimates based on
the outof sample errors, Obama’s chance of winning drops by around 10
percent compared to when using insample predictions. This is because
large national prediction errors are more likely when measured out of
sample. Obama's average
voteshare and number of electoral votes remain at the same levels.
While using the outofsample errors does make a difference for the
probability of winning, it does little to the optimal campaign
strategies. Graph 3 to the right shows the equilibrium visit shares for
based on variances using the distribution of the outofsample errors.
For
the exact description of how I proceeded, see
here.
