How voters evaluate public services: the case of the National Health Service in England

Abstract

A key condition for the public to be able to hold democratic governments accountable is that voters must be able to evaluate government performance. This thesis improves our understanding of the operation of this first step in the retrospective accountability process. It asks whether voter perceptions are responsive to real differences in performance, and which factors condition this responsiveness. Drawing on social-psychological and valence literatures, I examine how partisanship inhibits responsiveness of perceptions and whether the types of information that voters receive about performance is sufficient to change perceptions.

My empirical evidence comes from analysis of voters' evaluations of the National Health Service in England between 1983 and 2019 as recorded in cross-sectional and panel survey data, combined with official performance statistics. This is a good case study because the issue of healthcare is highly salient, and clear metrics exist to measure the service's performance over time and between different areas of the country.

I find that although voter evaluations of NHS performance do depend on partisanship, this relationship is not so strong as to crowd out responsiveness to changes in real performance. Voters evaluate the NHS more positively when national and local performance is better, and that recent users of services are more satisfied with health services. However, the publication of plans to downgrade hospital services in some parts of England did not result in changes in opinions among people living nearby. This thesis, therefore, argues that voter perceptions are indeed responsive to public service performance, but only under certain conditions - primarily when there is a consistent flow of credible information for voters to process.