Canadian federal election expenses modeling
LE3 .A278 2019
Bachelor of Science
Mathematics & Statistics
Money and election results are related. Candidates who spend more money win more elections. This thesis explores the relationship between election spending and percentage of the vote won, at the level of individual candidates. A regression model is estimated using data from the 2011 and 2015 Canadian federal elections to predict percentage of the vote won for federal ridings. In addition to spending, variables such as political party, incumbency, rural ridings, perceived leader quality and region are used as predictors. After adjusting for these variables, there is a strong and nonlinear relationship between spending and percentage of the vote won. Model predictions are used to determine the probability that a candidate will win a riding. Using these predictions as inputs to a linear programming problem, different spending strategies are identified to find the best allocation of funds for major political parties. The analysis uses the model to determine optimal allocation of fundraising dollars in various scenarios for the 2015 federal election, including how the Conservative party could possibly have allocated funds differently to win more seats, preventing a Liberal majority in 2015.
The author grants permission to the University Librarian at Acadia University to reproduce, loan or distribute copies of my thesis in microform, paper or electronic formats on a non-profit basis. The author retains the copyright of the thesis.