Election Forecasting Using Macroeconomic and Social Indicators via Machine Learning Lucas Guan1 and Ganesh Mani # 1 Palo Alto High School, Palo Alto, CA, USA # Advisor ABSTRACT A comparative analysis of machine learning models is executed for the forecasting of incumbent party losses during federal elections in democratic countries. A proprietary dataset that encompasses a wide array of poten- tial economic and social factors affecting election outcomes is compiled, and the most significant factors are identified and evaluated. A myriad of the most popular machine learning models for supervised learning are applied to the dataset, utilizing them as classifiers to predict whether the incumbent party stays in power during federal elections for eleven of the world’s most populous and democratic countries: the United States, Canada, the United Kingdom, the Netherlands, Austria, Norway, Sweden, Denmark, Australia, India, and New Zealand. The results show that the most significant factors for election outcomes are inflation growth rate, unemployment growth rate, and voter turnout growth rate. Multilayer perceptron produces the most accurate classifications. Additionally, Gaussian models such as Gaussian process classifier and Gaussian naive Bayes have the poorest classification accuracy. Introduction Forecasting federal elections has been attempted in long-standing democracies across the world. Although a wide array of research exists on the use of macroeconomic indicators to forecast federal election outcomes, significantly less leverages machine learning. In fact, with recent advances in artificial intelligence and super- vised learning, scholarly attention has turned away from economic factors to sentiment analysis using social media platforms. However, the state of a national economy appears to be an important one, as many incumbent parties have lost national elections during times of economic crisis, such as the Canadian Progressive Conserva- tives in 1993, the Republican Party in the United States in 2008, and the Greek New Democracy Party in 2015. This research applies ten of the most common supervised learning models to predicting democratic election outcomes using macroeconomic and social indicators. To ensure geographical diversity and comparability in the countries studied, 11 of the world’s most long-standing, developed, and populous democracies are chosen based on the Democracy Index: the United States, Canada, the United Kingdom, the Netherlands, Austria, Nor- way, Sweden, Denmark, Australia, India, and New Zealand. Related Works Early election predictions focus primarily on the effect of economic conditions on election outcomes. Specifi- cally, the academic debate revolves around conflicting findings on whether economic forces and perceptions play a key role at the ballot box. Lewis-Beck and Stegmaier (2000) find that past research centered on a single country has consistently shown that economic forces exert “a heavy and variegated” impact on democratic elections worldwide. By contrast, Paldam (1991) concludes that economic results “are either insignificant or Volume 11 Issue 3 (2022) ISSN: 2167-1907 www.JSR.org 1