Election Forecasting in a Multiparty System

University essay from Göteborgs universitet/Institutionen för nationalekonomi med statistik

Abstract: This bachelor thesis in statistics covers the subject of election forecasting in a multipartysystem, using polling data, that is data collected to measure party support, and dynamiclinear models (DLMs) with Kalman filtering. In terms of decision-making the outcomeof an election can be thought of as an uncertainty. Forecasts of election results canreduce risks for decision-makers and thereby facilitate decision-making. To be able toforesee the outcome of an event can be of use for experts in several different fields, forinstance political strategists, nancial investors and policy makers. A DLM considers anobservable time series to be a linear function of a latent, unobservable series and randomdisturbance. In the case of election forecasting we can think of the observable series asbeing polling data, and the underlying series to be true measures of party support. Thepurpose of using the Kalman filter is then to retrieve the latent series representing trueparty support. Altogether three different models are explored in the thesis; a Gamma-Normal, a time-invariant and a multivariate time-invariant model. The main differencebetween the frameworks concerns the variance term in the distribution of the noise termsin the DLM. The models are applied to the Swedish election of 2018, using polling datafor the period stretching from September 2014 to September 2018. The polling data isthen disregarded for three di erent time periods; the last month, the last six months andthe last twelve months before the election. For those periods, we instead use simulated data which together with the polling data is the basis of our forecasts. We find that the Gamma-Normal model performs slightly better than the two other models, whenforecasting the election result one month ahead, while the multivariate time-invariantmodel is slightly better for the two other time frames. For the one year forecast this modelpredicts the election result with an average absolute prediction error of 1.28 percentagepoints for each party. Finally, the forecasting capability of the models are discussed andevaluated in the analysis section of this thesis.

  AT THIS PAGE YOU CAN DOWNLOAD THE WHOLE ESSAY. (follow the link to the next page)