Religion and social insurance: Evidence from the United States, 1970-2002

One of the major puzzles for political economy and a central question in this volume is why some governments adopt policies that intervene heavily to redistribute income from rich to poor and to provide social insurance against adverse events, while other governments do much less in either regard.1 Existing literature on the political economy of redistribution and the welfare state has identifi ed a number of plausible factors that can infl uence policy outcomes in this area. Th ese include, among others, prior levels of inequality, labor market structure, issue bundling and coalition politics, constitutional structures, and partisanship.2 Models produced by economists have also emphasized that countries with otherwise similar economic and political preconditions may nonetheless wind up with widely divergent welfare state outcomes due to learning or expectations mechanisms that generate multiple equilibria.3 In this chapter, we investigate empirically the argument that religious involvement and social spending can both serve to insure individuals against the eff ects of adverse life events.4 Specifi cally, we argue that social insurance and religious engagement are two alternative mechanisms that limit the costs of adverse life events. As a consequence, religious individuals on average will prefer lower levels of social-insurance provision than will those who are secular. Moreover, if policy outcomes refl ect variation in citizens' preferences, then we can also expect that countries with higher levels of religiosity will have lower levels of welfare spending.5 Th is hypothesis is consistent with important stylized facts, like the large diff erence in both social-insurance provision and religiosity between the United States and many European countries. Th is chapter discusses a wide array of evidence that goes well beyond this simple comparison in support of the hypothesis, drawing on cross-country evidence, variation in policymaking across the U.S. states, and individual-level data on religiosity and attitudes toward social spending. Our argument emphasizes diff erences between individuals who are religious, irrespective of their denomination, and individuals who are not religious, rather than emphasizing diff erences between individuals of different religious denominations. Th e latter approach has been more prominent in political economy, due to familiarity with Weber's arguments about Protestantism and capitalism, as well as to observations about the links between Christian Democracy and Catholic doctrine. We suggest that a distinct cleavage exists for the politics of social insurance in advanced industrial countries between the religious and the nonreligious, irrespective of denomination. Our predictions regarding religion and social insurance involve three core assumptions. First, adverse life events involving unemployment, workplace accidents, illness, or retirement income do not only generate monetary costs for individuals; they also generate psychic costs that can involve a loss of selfesteem, stress, or related phenomena. Second, we assume, consistent with a substantial theoretical and empirical literature in psychology, that religiosity provides some of the same psychic benefi ts as does being in good health, having a job, or having a suffi cient retirement income. Finally, as discussed in detail later in the chapter, we assume that individuals have a utility function such that monetary costs and psychic costs are not additively separable. More specifi cally, we assume, again building on recent empirical fi ndings, that the psychological benefi ts of religion are greater for those with lower incomes. If one made a more restricted assumption that individuals suff er both monetary and psychic costs from adverse events such as job loss, but their utility is additively separable across these two factors, then religion might lower the psychic costs of adverse life events, but it would not infl uence an individual's demand for state social-insurance provision. Our argument emphasizes how religious involvement can serve as an alternative to social insurance for individuals to buff er themselves against adverse events. In some cases religious participation will allow individuals to draw on communal material support in times of diffi culty. Although these strictly material benefi ts from religion may be important, we draw on theoretical and empirical work suggesting that religion can also limit the psychic costs of adverse life events. So, for example, if the psychic costs of unemployment involve a loss of self-esteem, then religion may help insulate individuals against this eff ect, because their self-esteem is linked heavily to their religious engagement. Likewise, if falling ill or suffering a shock to one's retirement income produces stress, then religion may also serve as a buff er against this type of psychic cost. We will discuss how recent theoretical work by psychologists has emphasized that religious individuals may appraise adverse events as being less threatening to their overall self-image, beliefs, or well-being than would be the case for nonreligious individuals.6 Th is theoretical work is supported by numerous recent empirical studies linking religiosity to higher levels of subjective well-being and lower incidence of depression.7 Although earlier studies by psychologists, and in particular Freud's classic contribution on religion, Th e Future of an Illusion (1927), take a more negative view of religiosity, they share a commonality with recent work in emphasizing how religion can function as a buff er against uncontrollable external forces. In choosing the title for his work, Freud was certainly aware of Marx's (1844) description of religion as providing an illusory form of happiness for the people. We should note that another classic psychological portrayal of religion, Th e Varieties of Religious Experience, by William James (1902) took a less negative view of religiosity than that expressed by Freud or Marx, but James too emphasized religion's role as a buff er against external forces. In our study we do not seek to establish whether one should view religion positively or negatively in an overall sense, and our principal theoretical propositions do not depend on which normative conclusion one draws about religion, provided one accepts it can function as a type of insurance. Under our assumptions, there should be a negative correlation between an individual's degree of religiosity and the extent to which he or she prefers government provision of social insurance. Further, to the extent that policy outcomes refl ect individual preferences, there should also be a negative correlation between aggregate levels of religiosity and social- insurance spending. We evaluate these predictions empirically using cross-country, cross-regional, and individual-level data, focusing our attention on presenting evidence from the United States between 1970 and 2002. Our cross-country evidence on religiosity is drawn from both the World Values and ISSP surveys and involves questions about both the importance of religion in individuals' lives and time devoted to religious activities. We fi nd that there is a signifi cant negative correlation between religiosity and levels of social spending in the advanced industrial countries. Despite the small sample size, these results remain robust when controlling for a number of other potential determinants of social-spending levels, including the proportion of the population over age sixty-fi ve, diff erences in the representation of religious denominations, and beliefs about the importance of eff ort versus exogenous factors in determining individual economic success. Although these results are strongly suggestive of a link between religiosity and welfare spending, they are subject to the usual limitations of cross-country evidence with a small number of cases. For example, to what extent are there unobserved and thus omitted characteristics of these countries that might cause a spurious correlation between religiosity and social-insurance spending? To begin to address this possibility, in this chapter we evaluate the link between religiosity and policy outcomes in the U.S. states. For some areas of social-insurance policymaking, such as workers' compensation, U.S. states make their own policy decisions or at least have control over signifi cant features of the programs that determine the extent of insurance.