This study suggests that performance voting is characterised by extensive individual heterogeneity. Most economic voting studies to date treat voters as rather homogeneous in their reactions to economic performance of incumbents. Yet, a large and well‐established line of research from the American context demonstrates the conditional impact of political sophistication and salience on voters' political attitudes and behaviour. Building on this work, this article explores individual‐level variation in performance voting due to political sophistication and salience. Utilising cross‐national data from the Comparative Study of Electoral Systems (CSES) including 25 democracies, performance voting is examined across an array of policy areas including the economy, social welfare, immigration and national security, and it is shown that political sophistication and salience are key moderators of performance voting. The findings suggest that holding governments to account for past performance is mainly the prerogative of the highly sophisticated and thus may be more laborious than previously assumed. At the same time, the results indicate that the sophistication gap in performance voting narrows when voters attach a higher degree of salience to a policy area. As long as voters care enough about government activities in a particular policy area, incumbents can expect credit or blame for policy outcomes. This should provide at least some impetus for responsive policy making.

Although the theory of retrospective voting receives wide support in the literature on voting behaviour, less agreement exists on voters’ time horizon when assessing the government's performance – that is, whether voters are myopic. Previous studies on voter myopia tend to focus on aggregate‐level measures of the economy, or use an experimental approach. Using panel data, this article offers the first investigation into voter myopia that uses individual‐level evaluations of government performance in a representative survey at several points during the electoral cycle. The study focuses on The Netherlands, but it also provide tests of the generalisability and robustness of the findings, and a replication in the American context. The results indicate that voter satisfaction early in the government's term adds to explaining incumbent voting. Thus, rather than the myopic voter, evidence is found of the abiding voter – steady at her or his post, evaluating government performance over a long length of time.

Recent literature has shown that the long established link between economic performance and electoral outcomes is conditioned by a country's institutions and government, what is often termed ‘clarity of responsibility’. In this article two distinct dimensions of the clarity of the political context are identified: institutional and government clarity. The first captures the formal dispersion of government power, both horizontally and vertically. The second captures the cohesion of the incumbent government. Analysing survey data from 27 European countries, it is shown that voters' ability to hold governments to account, for both the economy and management of public services, is primarily influenced by the extent to which there is an identifiable and cohesive incumbent, whereas formal institutional rules have no direct impact on performance voting.

Electoral accountability is typically identified with retrospective economic voting even though it is widely recognized that explaining electors’ assignment of responsibility also implies considering issues other than the economy. Recently, scholars have also stressed the role of election quality in contributing to democratic legitimacy of elected authorities. In this perspective, electoral integrity as a valence issue would influence voters’ behaviour, structuring attitudes about accountability in substantial ways. This effect would also be moderated by individual- and country-level factors. I test these assumptions in 23 countries worldwide using a multilevel analysis of data from the sixth wave of the World Values Survey. Results indicate that the strength of the link between perceptions of electoral integrity and vote for the incumbent seems to be affected by individual characteristics such as partisanship, while it is also moderated by specific contextual characteristics such as government clarity of responsibility and pluralism of the media.