How do Government Employees Vote? A Comparatives Study of Fifteen Post-Industrial Economies

Abstract

The bureau-voting model (BVM) predicts government employees’ fiscal preferences to outweigh private sector employees’ preferences, thereby resulting in larger public budgets. Empirical evidence on bureau-voting remains fragmented and inconsistent. Instead of assuming an unconditional relationship, we argue that government employees’ incentive and ability to exploit bureau-voting depends on contextual variation. We decompose the BVM into four hypotheses, testing their validity with respect to variation across time, types of government employees and national administrative systems. Applying binary logistic regression on cross-national public opinion data we find that, although fiscal preferences and voting behavior of government employees differ from private employees, the substantive effect is rather small and is conditioned by earnings and the national administrative system. In none of our 15 countries the causal chain predicted by the BVM could be confirmed. Merging individual-level estimates with aggregate spending, we conclude that the BVM fails to explain the size of governments.

Keywords: fiscal preferences, party preferences, public employment, public/private cleavage, voter turnout

JEL Classification: D72, D73