Tim Hellwig makes some thoughtful comments on our book and raises wider issues about the study of electoral behavior in the future. The valence politics research program we have developed over time has implications for electoral studies in all democracies, not just in Britain. In this brief response, we would like to draw attention to two issues that are particularly important. The first is that in theoretical and empirical research on elections and in comparative politics, the spatial model generally dominates the picture in comparison with the valence model. This is despite the fact that the evidence from Britain, the United States, and elsewhere shows that the valence model has considerably more explanatory power than its spatial rival when it comes to explaining electoral choice. Spatial issues are, of course, important in explaining why people vote, but they play a secondary role in comparison with the big valence issues of security, economic performance, and public service delivery, where there is a widespread consensus about the goals to be achieved. The fact that electoral politics is primarily about delivery and competence and only secondarily about disagreements and promises needs to be recognized if we are to make progress.

There is an “end of ideology” argument to which Tim Hellwig alludes to that suggests that ideological divisions may have dominated electoral politics in the past, but now this may have given way to valence politics. In fact, the evidence from past election studies in Britain, which go back to the early 1960s, suggests that valence politics has always been more important than the ideology/class politics analysis originally introduced by David Butler and Donald Stokes in Political Choice in Britain (1969). Indeed the evidence from their own panel surveys conducted in the 1960s supports that conclusion, despite the fact that their analysis did not recognize it. There is a parallel here with current debates about the polarization of electoral politics in America. Divisions in U.S. electoral politics over cultural issues and domestic and foreign policy concerns should not detract attention from the fact that the delivery of outcomes, particularly in relation to the big valence issue of the economy, is the key driver of American voting behavior.

The second, wider lesson is that electoral choice can no longer be studied with low-response-rate, face-to-face probability surveys that take months to collect and cost a great deal of money. Electoral choice is dynamic, and therefore sociological variables such as class, race, and gender play a relatively minor role in explaining behavior in comparison with “choice” variables, such as partisanship, economic evaluations, and judgments about the effectiveness of political leaders. What is needed is high-frequency survey data such as the Continuous Monitoring Survey of the British Election Study, preferably with a panel element to identify effects. There are clear methodological problems associated with modeling such dynamics, notably the ubiquitous problem of endogeneity, but this makes it all the more necessary to abandon the slow and highly restricted face-to-face surveys. Methodological tools should reflect the theoretical needs, and electoral analysis now requires data and modeling strategies for understanding these dynamics.

So we thank Tim Hellwig for his thought-provoking remarks on our work and agree wholeheartedly with his desire to bring individual actors into the task of understanding the big issues of globalization and policymaking in contemporary democracies.