Nathalie Giger et al have conducted a series of studies that suggest that that the empirical evidence in support of a central assumption in the ‘New Politics’ theory is weak at best: Voters do not seem to punish government parties that have implemented welfare retrenchment that harshly. These studies have many strong points, but I will argue that they also have serious methodological problems.
Most importantly, I argue that voters cannot be expected to punish governments for welfare retrenchment that they are not aware of. Many of the reforms Giger et al study are by and large what I would call untransparent (cf. Lindbom 2007). When estimating voter punishment, we should focus on transparent retrenchment. The independent variable in my study is closures of emergency wards in Swedish health care. The closures are an extremely transparent type of retrenchment.
Studying variation within Sweden holds a number of contextual variables constant. Furthermore, since Swedish general elections on three levels take place on the same day, there are also considerable possibilities to isolate the electoral effects of unpopular reform proposals made on the intermediate level responsible for health care.
