the existing system of social protection. The implication of this argument would seem to be that
welfare spending should be more politically contested in LMEs than in CMEs or, in other words,
that the interests of the core constituencies of Left-leaning and Right-leaning parties should
diverge more sharply in LMEs than in CMEs. More decisively, the VofC approach leads us to
expect that partisan effects have become more salient in LMEs over time, as international
competition has intensified and capital mobility has increased. In CMEs, by contrast, the VofC
approach seems to predict little or no change in partisan effects, whatever the size of these effects
may have been in the “pre-globalization era.”
We test these predictions by estimating the effects of government partisanship on welfare
spending in LMEs and SMEs separately, again using moving-windows analysis (with fixed-
length windows of 15 years) to capture over-time changes in partisan effects. On account of the
exceptional features of the Japanese case, as regards public welfare provision, we have chosen not
group Japan with the European CMEs and use the term “SMEs” to signal this departure from the
standard VofC classification of countries (cf. Rueda and Pontusson 2000, Pontusson 2003). For
our present purposes, SMEs can simply be defined as “CMEs without Japan.”
2
Using the
Cusack-Franzese measure of government partisanship, robust estimation with data for the period
1962-95 yields enduring and positive effects of Left government on social spending growth in
SMEs, but not in LMEs. However, with this and other measures of government partisanship,
which enable us to extend the analysis to 1998, we find significant effects of changes in
government partisanship in LMEs as well as SMEs. In general, newly formed governments
influence the rate of social spending growth across varieties of capitalism, but these effects
appear to be more transitory in LMEs. As for the question of change over time, our analysis
indicates that in SMEs more leftist governments were less prone to promote social spending
growth than more rightist governments in the 1960s and 1970s, but this changed completely in
the 1980s. Since the early 1980s, more leftist governments have been consistently more “pro-
welfare” in SMEs. In LMEs, by contrast, we observe a rather steady decline of partisan effects
2
Convention