The Equality Positions of Parties and Voters: Do Parties Respond to Voters?

• In Event: New Directions in the Study of Inequality and Democratic Representation

Abstract

It is crucial for a functioning democracy that the political system is responsive to the preferences of voters. Previous research generally confirms that Western democracies are responsive. At the same time, studies show that the preferences of the rich and better educated are more closely aligned with the preferences and policy outputs of political elites. This indicates a shortcoming because, in a functioning democracy, all citizens should be treated alike.

We evaluate the causal relationship between public opinion on redistribution and the equality positions of parties voiced in their election manifestos. We make two contributions. First, our analysis explicitly focuses on economic equality and relies on new data on the equality positions of parties from 12 countries, 1970-2020. These were obtained through crowd-coding of 850,000 party statements. This offers novel insights because previous large-scale quantitative analyses had to make do with party position measures that muddled the crucial distinction between economic equality, equality of opportunity, and equal rights.

Second, and in contrast to previous research, we employ causal identification strategies to analyze the inter-relationship between party and citizen positions. Previous studies from the responsiveness literature offer valuable insights, but it remains unclear whether the match (and mismatch) between public opinion on the one side and elite preferences and policies on the other side results from a causal connection (party responsiveness) or mere correlation (representation). Furthermore, it is equally likely that citizens’ preferences are responsive to the preferences and policies of parties (voter responsiveness) as voters tend to align their preferences with political elites as an informational shortcut.

We harmonize individual-level survey datasets from multiple sources that measure the demand for redistribution and electoral preferences of citizens, which we link to the equality positions of parties. To assess the causal influence of citizen preferences on parties, we instrument the redistribution preferences of citizens with their socio-economic profile (income, education, age, etc.). We confirm the prevalence of this causal relationship relying on the assumption that the socio-economic structure affects party preferences only via the policy preferences of voters. Parties are responsive to their voters, but not the public in general. However, we cannot confirm that citizens with more education and income enjoy better representation, which contrasts with previous research.

To assess the causal impact of party positions on voters, we rely on difference-in-difference (DID) and regression continuity (RD) designs where we check whether citizens adapt their redistribution preferences to (their) parties after a an unusually pro-equality manifesto is released during an election campaign. The RD design assesses whether citizens adapt their preferences exactly after a manifesto was released and the DID design whether voters attached to a party increase their demand for redistribution relative to other voters. Both analyses cannot confirm a causal relationship. Our analyses therefore suggest that citizens affect the equality positions of parties, but not the other way around.

Authors

• Leo Ahrens, University of Konstanz

• Alexander Horn, University of Konstanz