ten - What if we waited a little longer? The dependent variable problem within the comparative analysis of the welfare state revisited

Summary

Introduction

Generally regarded as the main trigger for the ‘welfare modelling business’, Esping-Andersen's (1990) The three worlds of welfare capitalism incidentally also heralded the so-called dependent variable problem within the comparative analysis of the welfare state (henceforth, dependent variable problem). In what has become one of the most cited critiques, Esping-Andersen (1990, pp 19-21) stated that expenditure-based summary measures of welfare state effort ‘are epiphenomenal to the theoretical substance of welfare states’, and indicate neither a ‘commitment to social citizenship and solidarity [nor] full employment.’ Furthermore, Esping-Andersen (1990, pp 19-21) argued that ‘if our aim is to test causal theories that involve actors […] it is difficult to imagine that anyone struggled for spending per se’. Indeed, in my earlier study on the dependent variable problem covering 21 Organisation for Economic Co-operation and Development (OECD) countries and a time period between 1980 and 2001, I found that widely used expenditure-based summary measures of welfare state generosity produce inconsistent findings at the country level, and are hampered by conceptual limitations, missing values and imprecise operational definitions (Kühner, 2007). As for analyses of welfare state change, I showed that considerable asymmetries in the findings produced by different expenditure-based summary measures, but also increasingly between quantitatively informed research and qualitative comparative historical case analyses, persist.

At the same time, recent progress in the broad availability of social rights-based summary measures (see Scruggs and Allan, 2006, 2008; Scruggs et al, 2013) has enabled a seemingly preferable perspective on welfare state activity across time and space, but largely failed to bring about a decisive end to long-fought debates about best and second-best measures in comparative welfare state analysis. Instead, it is now commonplace to argue that scholars should take more seriously the multidimensional character of the welfare state and welfare state change, and perform analyses of different dependent variables whenever feasible. By doing so, the notion of broad welfare state retrenchment across high-income countries is overstated, but the same is true for claims of welfare resilience: while some countries have shown signs of sizeable cutbacks, some have remained stable and others have expanded welfare provision, noticeably since the 1980s. Importantly, my earlier study concluded by arguing that (1) Esping-Andersen's (1990) regime typology does not provide a sufficient explanation of these different reform processes and (2) the trajectories of disaggregated welfare state components are relatively independent.