When Do Voters Stop Caring? Estimating the Shape of Voter Utility Function

TL;DR

This paper investigates the true shape of voter utility functions, finding that reverse S-shaped functions like the Gaussian better predict voting and abstention behavior than traditional concave models.

Contribution

It provides empirical evidence favoring Gaussian-shaped utility functions over concave loss functions in modeling voter behavior.

Findings

Reverse S-shaped utility functions fit observed data better.

Concave loss functions fail to predict actual voting and abstention.

Empirical analysis uses 2020 U.S. election CVR data.

Abstract

In this paper, we address a longstanding puzzle over the functional form that better approximates voter's political utility. Though it has become the norm in the literature to represent the voters' political utility with concave loss functions, for decades scholars have underscored this assumption's potential shortcomings. Yet there exists little to no evidence to support one functional form assumption over another. We fill this gap by first identifying electoral settings where the different functional forms generate divergent predictions about voter behavior. Then, we assess which functional form better matches observed voter and abstention behavior using Cast Vote Record (CVR) data that captures the anonymized ballots of millions of voters in the 2020 U.S. general election. Our findings indicate that concave loss functions fail to predict voting and abstention behavior and it is the reverse S-shaped loss functions, such as the Gaussian function, that better match the observed voter behavior.