Forecasting Canadian Federal Elections

In recent decades, the scientific forecasting of election outcomes has made great strides in a number of advanced industrial democracies. One country that has not received much attention to date is Canada. In this article, we present a vote function model to forecast Canadian federal elections. We explain our model's theoretical underpinnings and assess its statistical properties and forecasting capabilities against all federal elections held between 1953 and 2008. We then explore potential ways for improving the model's accuracy. We conclude by discussing the forecast of different hypothetical scenarios for an upcoming federal election.