Weber, Work Ethic And Well-Being

Following Max Weber’s seminal work, much recent work has turned to religious values to explain socio-economic developments. We present a test of Weber’s original thesis that addresses fundamental limitations of previous research. A novel method that builds on happiness research is used to measure a religious work ethic in terms of the psychic costs of unemployment. The resulting ‘experienced preferences’ provide strong support for Weber’s original thesis: for both Protestants and Protestant countries, not having a job has substantially larger negative happiness effects than for other religious denominations. This provides a Weber-type channel relating religion to socio-economic outcomes. More than a century after its publication, Max Weber’s The Protestant Ethic and the Spirit of Capitalism (1930 [1904-5]) continues to inspire social scientists in many disciplines. A large stream of work in the social sciences has built on the idea that religious values can explain social and economic developments. Over the past decades, this research has received an important impetus by the publication of large scale, cross-cultural values surveys (e.g. Hofstede, 1980,2001; Inglehart, 1990). The availability of data about value differences between countries has made it possible to investigate the relations between values and socioeconomic outcomes empirically. Paradoxically, much of this research has failed to find support for the thesis of Weber that originally inspired the literature. Any convincing relation between Protestantism on the one hand and work ethic or economic development on the other has yet failed to materialize in the data (e.g. Lehmann and Roth, 1993; Iannoccone, 1998; Delacroix and Nielsen, 2001). In fact, many researchers have reported evidence that Protestants overall value work less than people from other religions do (Norris and Inglehart, 2004; Weil, 2008). Furthermore, many authors have found negative associations between practices and values retrieved from values surveys, implying that a strong work ethic does not translate into higher employment rates or longer working hours (Bertrand and Mullainathan, 2001; House et al., 2004; Javidan et al., 2006). At the same time, the practice of using values surveys in order to measure cultural or religious values has come under fire from several directions. Methodological critiques have been forthcoming ever since values surveys became public (see Hofstede, 2001, p. 73 for an 1 In so far as people have found support for the idea of a link between Protestantism and economic prosperity, it has been argued that the more plausible route runs via literacy levels rather than the prevalence of a ‘capitalist spirit’, as Weber wants it (Becker and Wößmann, 2007). overview). A particular damaging line of criticism has been the revelation that scores in values surveys turn out to be extremely dependent on societal conditions (e.g. Clarke et al., 1999; Davis et al. 1999; Duch and Taylor, 1993). Recently, it has been argued that this problem has its roots in the failure of values surveys to distinguish between deep-rooted value traits and marginal preferences. Values surveys have been shown to elicit not the importance attached to objectives such as work, but only the importance attached to a little more or less work on top of the currently enjoyed quantity (Maseland and van Hoorn, 2008). In this paper, we argue that these conceptual and methodological issues are responsible for the failure to find empirical support for the Weber thesis, and we develop an alternative method to overcome these problems. We show that the counterintuitive results of earlier studies are intelligible when values surveys are interpreted in terms of marginal preferences. Moving away from values surveys in favour of an approach focusing on differences in the effects of situational factors on happiness, we find support for the thesis that Protestants value work more. We conclude that, when values are measured properly, the Weber thesis is confirmed by the data. The structure of our argument is as follows. In the first section we discuss the Weber thesis in more detail, looking at contemporary interpretations and efforts to find empirical support for the thesis. Section 2 scrutinizes the values surveys approach to measuring values and introduces an alternative method, which one might dub ‘experienced preferences’. We apply this new method to the problem of the relation between Protestantism and work ethic. Our empirical strategy and the results of the analysis are discussed in Sections 3 and 4. We round off with a discussion and conclusion, in which we elaborate the methodological and theoretical implications of our findings for future research into values and the economy. 1. RELIGION AND WORK ETHIC: WEBER’S THESIS The Protestant Ethic and the Spirit of Capitalism goes into history as one of the most frequently cited books in social sciences. Weber’s argument about the relation between Protestantism and capitalism has spawned an extensive and diverse literature dealing with the effect of religious values on economic performance. Some of this literature has retained Weber’s original focus on Protestantism. Other contributions have reworked the argument, applying it to other religions such as Catholicism (Tawney, 1926) or, more recently, Confucianism (Harrison, 1992; Harrison and Huntington, 2000; Kahn 1979). Still others have quantitatively assessed the role of non-religiously specified sets of values (e.g. Granato et al., 1996). More recently, general associations between religion and economic outcomes have been analysed empirically (Barro and McCleary, 2003; Guiso et al., 2006). Throughout this literature, Weber is commonly referred to as the person starting the debate about the link between religious ethics and economic growth. Before we delve into empirical tests of the Weber thesis, it pays to briefly go back to the original argument of the Protestant Ethic. As Giddens (2001, p. xx) and many others have emphasized, Weber’s famous essay can be approached on many different levels. The Protestant Ethic establishes a historical relation between the emergence of capitalism as a dominant economic system in Western Europe and North America and the Protestant reformation centuries earlier. More in particular, Weber draws attention to the peculiar ascetic ethical system propagated in Protestantism. Here originated the idea of a ‘calling’, a perception of one’s work and other economic activities as God-given duties. The emphasis on worldly activity as a means to prove one’s faith eventually evolved, through a process of rationalization, into what Weber calls the ‘spirit of capitalism’; the idea that working for the purpose of profit is a moral good in itself. As Weber writes: ‘... one’s duty in a calling, is what is most characteristic of the social ethic of capitalistic culture, and is in a sense the fundamental basis of it. It is an obligation which the individual is supposed to feel and does feel towards the content of his professional activity, no matter in what it consists, in particular no matter whether it appears on the surface as a utilization of his personal powers. Or only of his material possessions (as capital).’ (Weber, 1930, p. 19). In contrast to the common interpretation of the Protestant Ethic (e.g. Becker and Wößmann, 2007; Granato et al., 1996; Weil, 2008), there is little in the original text to suggest that Weber saw a causal relation between being Protestant and enjoying economic prosperity. Rather than that, he was trying to explain the initial origin of modern industrial capitalism in Northwest European and North American societies, relating it to values historically retraceable to a specific religious ethics. Weber argues that modern capitalism’s outstanding feature—compared to previous capitalist practices—is a set of values that is religious in origin, but has been rationalized and secularized over time. What is more, a link with economic growth and prosperity is largely absent in the work. Nevertheless, the ‘Common Interpretation’ (Delacroix and Nielsen, 2001) of the Protestant Ethic, seeing in Protestantism a cause of economic progress, has taken a life of its own. 2 Weber (1930, p. 27) writes: ‘the cultural consequences of the Reformation were to a great extent, perhaps in the particular aspects with which we are dealing predominantly, unforeseen and even unwished-for results of the labors of the reformers. They were often far removed from or even in contradiction to all that they themselves thought to attain.’ 3 The awareness of such a link may have been Weber’s motivation to write The Protestant Ethic, however. In addition, although, the present paper by focussing on Weber’s original thesis differs from recent quantitative work in economics, it is also strongly motivated by the possible link between religion and economic performance. A problem with this common interpretation is that it transforms Weber’s argument into two controversial theses rather than one, rendering testing difficult. First, it argues that Protestantism results in a strong work ethic, and second, it claims that a strong work ethic is a main determinant of economic growth and prosperity. Elsewhere, it has been shown that the latter relation is not that straightforward. Authors have reported evidence that people in poor countries attach more importance to work than people in developed countries do (e.g. Weil, 2008). Furthermore, it has been pointed out that inhabitants of some of the most rapidly growing economies of the past century (for example Japan, South Korea, Malaysia) have historically been viewed as lacking in work ethic (Alatas, 1977; Chang, 2007; Landes, 1998). Finally, caution is due as arguments about any relation between a work ethic and economic prosperity have often been used to legitimise rather than explain inequality (Alatas 1977; Said 1978). For this reason, we adopt a more limited focus here, dealing with the relation between Protestantism and work ethic, without going into the effects on economic performance. Empirical investigations into this relation have been conducted amongst others by Norris and Inglehart (2004). In that study, the idea is that if Weber’s thesis is correct, Protestantism should ‘have left an enduring legacy in values that still remains visible today’ (p. 162). To test whether this is indeed the case, they first construct a multidimensional measure of work ethic using a selection of items from the World Values Survey. Specifically, their index combines questions asking about, amongst others, which aspects people find important in a job (e.g. ‘an opportunity to use initiative, ‘good hours,’ and ‘good pay’), and the extent to which people agree with the statement that ‘work is a duty towards society,’ and that ‘people who don’t work turn lazy’ (p. 163). To their surprise, the analysis shows that people living in Protestant societies have a weaker work ethic than many individuals from other religious cultures. Norris and Inglehart (2004) conclude that the Weber thesis is to be dismissed. However, since this study relies entirely on values survey data for its results, there are doubts about the validity of this conclusion. 2. MEASURING THE WORK ETHIC 2.1 Why Values Surveys Do Not Work Equally controversial as the Weber thesis is the use of values surveys to study differences in values between societies. The assumption of a link between what respondents say in surveys and their deep-rooted values has long been questioned (e.g. Clarke et al., 1999; Davis et al., 1999). On top of this methodological criticism, it has been argued recently that values surveys are likely to be conceptually misguided (Maseland and van Hoorn, 2008). Values surveys tend to mistake marginal preferences (the importance attached to somewhat more or less satiation of any objective) for values (the importance attached to an objective in general). While we can expect a positive relation to exist between values and practices, marginal preferences are likely to decrease with rising satiation of an objective. Due to the principle of diminishing marginal utility, the importance attached to work falls with the amount of work performed. Elsewhere, it has been shown that a negative relation indeed exists between values and practices, suggesting that values surveys elicit marginal preferences rather than values (see, for example, Bertrand and Mullainathan, 2001; House et al., 2004; Javidan et al., 2006). This puts the results of Norris and Inglehart (2004) in a different light. If the World Values Survey is interpreted as eliciting marginal preferences rather than values, Norris and Inglehart’s (2004) findings are actually in line with Weber’s original insight. Were Protestants to value work higher, they would express this preference by working more, which causes their marginal preference for work to fall. A lower score on values surveys items about work is thus not at odds with the argument ascribed to Weber. It merely indicates that values surveys are not appropriate for testing the Weber thesis. We need a different strategy. 2.2 Well-Being And Unemployment For an alternative approach to measure a work ethic, we turn towards the literature on subjective well-being. Subjective well-being, commonly abbreviated as SWB, refers to ‘a broad category of phenomena that includes people’s emotional responses, domain satisfactions, and global judgments of life satisfaction’ (Diener et al., 1999, p. 277). The SWB construct is often used synonymously with happiness, though happiness is generally associated somewhat more with hedonic experience and the affective part of SWB. There is a great deal of evidence showing the reliability and validity of indicators of SWB, which often involves simply asking people how happy or satisfied with life they are (see, for example, Diener et al., 1999 and Frey and Stutzer, 2002 and references therein). What we are interested in here is heterogeneity in the structure of SWB. Various situational factors have been shown to have an impact on SWB, and unemployment is one of them (Diener et al., 1999 and Frey and Stutzer, 2002). Clark and Oswald (1994) analyse data from the British Household Panel and find that unemployed people have much lower levels of mental well-being than those in work. Similar results have been reported by Helliwell (2003) using data from the World Values Survey. What is more, it is clear that the effect runs from unemployment to SWB. Using longitudinal data, Clark et al. (2008) demonstrate that individuals who lose their job find their level of SWB decreasing substantially upon becoming unemployed, while they do not have lower SWB to begin with. Unemployment not only affects the well-being of the people losing their jobs, but also has an indirect impact on the population as a whole. This effect can be linked to notions of solidarity, fears of getting unemployed, reduced opportunities to change jobs, reduced chances of obtaining promotions or salary increases, and rising crime rates (e.g. Di Tella and MacCulloch, 2008). Although the negative effect of unemployment on SWB is a persistent result in the literature, the size of this effect has been shown to differ between groups of people. For instance, there is evidence that being unemployed is easier for people living in a region with high unemployment or for younger people (Clark and Oswald, 1994; Winkellmann and Winkellmann, 1998). Clark (2003) explains these results on the basis of reference groups and social norms, arguing that the more common unemployment is among your peers the weaker the stigma the unemployed suffer. An alternative interpretation of these findings would be that the causality runs the other way around: groups of people for whom the psychic costs of unemployment are lower may be making lesser efforts to find or keep jobs. In this interpretation, lower psychic costs of unemployment are indicative of a weaker work ethic. 2.3 The Protestant Work Ethic And The Psychic Costs Of Unemployment If we define a work ethic as the importance attached to having a job, differences between groups in size of the effects of unemployment on well-being can be interpreted as differences in work ethic. People attaching a lot of importance to work are hurt more by losing their job than people who think work is unimportant in life. This is what differences in the effect size of unemployment on SWB indicate. Religion has been shown to have an impact on these effect sizes. Apart from direct effects of religiosity on well-being (Diener et al., 1999; Helliwell, 2003), there is evidence that the impact of economic factors on well-being differs between religious and non-religious groups. For instance, religious beliefs seem to shield against part of the negative well-being 4 For theoretical work on the relation between unemployment and psychological well-being see, for example, Hayes and Nutman (1981) and Darity and Goldsmith (1996). effects of stressors like unemployment (Clark and Lelkes, 2005). Also, Lelkes (2006) reports that the effect of economic variables including income on happiness is smaller among the religious than among the non-religious. Apparently, religious people value work and income less than non-believers do. The question is whether we can find such differences between Protestants and nonProtestants as well. Such a finding would allow us to reach a verdict on the Weber thesis. This results in the following hypothesis. Hypothesis 1: Holding everything else constant, reported happiness ratings of Protestants are more influenced by being unemployed than those of people from other religions. Hypothesis 1 is useful for testing one of the more common interpretations of the Weber thesis—i.e. that those being Protestants now are likely to have a stronger work ethic than people currently holding different religious beliefs. However, we have noted that Weber’s original argument did not so much focus on Protestantism in the present but on a Protestant ethic as a historical factor, having evolved into a rational, secular ‘spirit of capitalism’. An hypothesis closer to the original argument is therefore: Hypothesis 2: Holding everything else constant, reported happiness ratings of people from historically Protestant regions are more influenced by being unemployed, than those of people from other regions. These hypotheses state that there is systematic heterogeneity in the way in which unemployment is transformed into (un)happiness. The focus hereby is on the direct psychic costs of unemployment and how this negative happiness effect is moderated by religious denomination, in particular by Protestantism. To make sure that we measure the pure SWB effect of unemployment, in the empirical analyses we correct for indirect effects of being unemployed, which might run through other factors such as income. The next section discusses our statistical method and data in more detail.