The influence of descriptive representation on voting patterns: a study of age and gender dynamics

Abstract

Polarised political landscapes have amplified identity-based narratives, emphasising traits like race, gender, and ethnicity. This article contributes to debates on representation by examining whether descriptive representation–in terms of age and gender–shapes electoral outcomes in municipal elections in Flanders, Belgium. Leveraging an original dataset of approximately 105,000 candidates from 300 municipalities across 2006, 2012, and 2018 elections, a fixed-effects regression model demonstrates that age proximity between candidates and voters significantly enhances electoral performance. We further uncover a persistent gender penalty, with female candidates systematically receiving fewer votes than their male counterparts. A two-stage DiD approach shows, however, that female mayors help mitigate this bias. Additionally, spatial panel regression models indicate that Flemish voters are influenced by neighbouring municipalities, suggesting elements of path dependence. These findings highlight that, even within ideologically driven contexts committed to gender equality, descriptive representation can influence voting behaviour and reduce discriminatory voting patterns.