John Myles

Political Theories of the Welfare State

We review the main theoretical conclusions from a quarter century of comparative studies of welfare states in the affluent democracies. We contrast early debates over the relative importance of industrialization, economic growth, and social classes for explaining welfare state differences with contemporary claims about the role of globalization, postindustrialism, and gender relations in shaping their futures. We evaluate the claims against recent empirical evidence with the aim of highlighting both important lessons from the past and promising directions for future analysis.

STASIS AMIDST CHANGE: CANADIAN PENSION REFORM IN AN AGE OF RETRENCHMENT

Faced with aging populations and especially heightened fiscal constraints, large scale pension reforms were implemented in many affluent democracies during the 1990s. Canadian reforms, by contrast, were quite modest and old age security benefits emerged largely unscathed. Drawing on the comparative experience of other OECD nations, we highlight four characteristics of the Canadian pension system and the policy environment to account for this relative stability:(1) the comparatively modest scale of Canadian public sector pension expenditures; (2) relatively greater reliance on general revenue as opposed to payroll taxes to finance these expenditures; (3) the availability of other expenditure targets, notably health care, post-secondary education and social assistance, that could be cut with less political backlash; and (4) a pension design that allocates the public sector share disproportionately to the bottom end of the income distribution, precluding the emergence of the oppositional politics that fueled public debate elsewhere.

Welfare States and Public Opinion

has acquired an ontological status as a social fact that is well beyond the political experience of earlier generations. And a growing body of evidence on policy responsiveness to public opinion has given new credibility to claims that mass democracies work much as advertised: the people get what they want. As Paul Burstein (1998) has demonstrated, however, sociologists have not shown much interest in the role of mass preferences in shaping policy outcomes in general and national differences in welfare state policies in particular. Among 49 articles on policy outcomes published in leading sociology journals between 1980 and 1997, only 10 articles discuss the role of public opinion and only 3 actually incorporate some data on public opinion (Burstein 1998). In this issue of ASR, Brooks and Manza (2006; henceforth B&M) provide an important corrective to this neglect. First, they establish an intellectual link to the growing body of political science literature on this topic (for reviews, see Burstein 1998 and Manza and Cook 2002). Second, they challenge comparative welfare state scholars to attend to an issue that has been mostly dormant since they turned against the “national values” tradition associated with Lipset (1963) and others. As B&M highlight, popular preferences (and how they are shaped) have remained part of the interpretive apparatus used to explain national differences in welfare state outcomes but have been rarely studied directly.1 The central claim of B&M is the discovery of a robust and large effect of public opinion on cross-national differences in welfare state outputs despite fundamental cross-national differences in political institutions. Their contribution goes beyond the identification of the correlation between public opinion and welfare state outputs. The fact that Swedes are more likely than Americans to favor government intervention to reduce economic inequalities is well known from earlier analyses of International Social Survey Programme (ISSP) data. Rather, their important and original contribution arises from their effort to assess this correlation within an explicitly causal framework. To my knowledge, their study is the first attempt to embed the public opinion thesis within a multivariate model that does the following: (a) tests for the presence of (short-term) endogeneity (i.e., does public opinion simply reflect current levels of social spending?); and (b) controls for a set of other likely suspects that account for welfare state variation. Their conclusion is that the correlation is not only robust and large but also causal, an index of the “policy responsiveness” of political elites to mass opinion. So will publication of these results put the public opinion thesis on the research agenda of comparative political sociologists once again? Will, as B&M urge, welfare state research begin to incorporate mass policy preferences into welfare state theory? I suspect there will be resistance. If the reviewers’ responses to B&M are a guide, the surface debate will focus on technical and statistical issues. Whether and under what conditions one can make a case for “causality” with correlations of this sort is a highly contested issue (Shalev, forthcoming). But the main hesitation in taking up their challenge may be theoretical. Burstein (1998) puts his finger on the issue. “Democratic theory,” as he calls the reigning paradigm of the 1950s and 1960s, produced considerable skepticism in the generation of scholars who developed the field of comparative welfare state Comment on Brooks and Manza, ASR, June 2006

What Justice Requires: Pension Reform in Ageing Societies

designed under conditions of high fertility and sustained labour-force growth might require some serious tweaking once those conditions have disappeared. Prudence suggests we should do something about the social-policy legacy we leave to our children now that we are having so few of them. And that legacy will inevitably be normative as well as economic. The physical, human and environmental capital we pass on to our heirs will be embedded in a social infrastructure – a set of rules and institutions – that bear the imprint of our (competing) vision(s) of the ‘good society’. There is no choice. Since ‘overlapping cohorts’, to use Schokkaert and Van Parijs’s (henceforth S&P) phrase, are part of the human condition, we could not start them off from a normative (i.e. institutional) tabula rasa even if we wanted to. So S&P are correct to insist that ‘no serious discussion of possible pension reforms can dispense with some conception of what justice requires’. S&P’s discussion of the normative foundations for pension reform was initially presented as a ‘response’ to Chapter 5 (Myles, 2002) of Why We Need a New Welfare State (EspingAndersen et al., 2002, hereafter AGHM), written for the Belgian presidency of the European Union. As anyone who compares the original texts will conclude, disagreement between us is minimal. Nevertheless, there are differences. Since much of the attention has focused on our ‘rediscovery’ of a paper written by Richard Musgrave (1986), that is where I begin. The intergenerational contract and the Musgrave rule