The following forecast is a revision of the…

The following forecast is a revision of the Belanger and Godbout (2010) model. My revised model improves their work by lowering the forecast error of 4.2 to 2.6 and does a better job of picking up shifts in Canadian Federal Elections. Figure 1 illustrates the original model and Figure 2 is my revised model. Note the revisions do a better job of capturing the dynamics of Canadian elections when compared to the original model.

The revised utilizes the following variables:

Percent Vote for the Incumbent Party in Parliament=60.1-1.8*Unemployment rate – 2.4*Months in power for the Incumbent Party+ 3.35*Incumbent popularity and Leadership Factor ± 2.6 (error)

Final Forecasts for the Conservative Party on May 2, 2011:

Predictions
Estimate Lower Upper
Seats 119.3 108.0 130.6
Vote 35.3 32.1 38.1
Percent Seats 38.7 35.1 42.4

Forecast errors
vote: 4.4
percent seats: 4.3

Market: Conservatives Win the Most Seats
Probability Estimate Lower Upper
Most Seats 89% 65% 98%

Avg probability of Conservatives win most seats
84%