Responsive Politicians? The Thermostatic Model Applied to Politicians

Research on the policy preferences of voters has shown that the expressed spending preferences of the average voters tend to respond to spending relevant information, i.e. the actual level of public spending (Wlezien 1996, Soroka and Wlezien 2011). But what about the responsiveness of those who make the spending and policy decisions, the politicians: do their spending preferences respond to relevant information? Do center-right (-left) politicians always prefer less (more) spending on public welfare, regardless of the current spending on welfare? Or do the policy preferences of politicians actually respond to policy relevant information, including information about the current levels of public spending? To advance our understanding of these questions about the responsiveness of politicians’ policy preferences we utilize in this paper a unique dataset based on several representative surveys of the policy preferences of Danish local politicians. Our modeling approach is inspired by the work on public opinion responsiveness by Wlezien and Soroka. Particularly, for seven different issues we investigate how politicians adjust their relative spending preferences to the actual level of public spending. As argued by Soroka and Wlezien (2011) if you are asked whether more or less money should be spend on a given issue, then the actual level of spending on that issue should be the perhaps most relevant informational cue. More particularly, we apply this basic idea to the local level of government in Denmark, where we have surveys of spending preferences of local politicians covering more than 90 municipalities, four years, seven issues, and a total of 4,479 responses. Utilizing this unique data set on politicians’ spending preferences furthermore allows us to link the spending preferences with very valid and precise data on local government spending per issue.