RELIGION AND SOCIAL INSURANCE

It is well known that there are large differences between advanced industrial countries in terms of levels of income redistribution and social insurance provision. Far less certainty exists about the explanations for these differences.

Recent literature has not fully examined to what extent differences in degrees of religiosity might help account for both variation in individual attitudes toward social insurance provision, as well as for the sizeable cross-country differences we observe in actual levels of social spending.

Figure 1 presents a simple scatterplot of the level of social spending in percent of GDP for 22 OECD countries (average 1990–98), plotted against one measure of religiosity from the WorldValues Survey. Social spending here includes state provision of unemployment benefits, health spending, and retirement benefits.

The horizontal axis here corresponds to the average response in each country to the question “How important is God in your life?” with 1 corresponding to “not at all” and 10 corresponding to “very”.

As can be seen, there is a striking negative correlation between the degree of religiosity and levels of social spending. Figure 1 is visually striking, but it of course does not tell us why we observe this negative correlation.

In what follows we review existing arguments that may shed light on this question. We then outline our own argument.

Existing Arguments

Religious beliefs and activity might influence individual attitudes toward social insurance for a number of reasons. Explanations about religion and redistribution may emphasize differences in attitudes between individuals of different religious denominations, such as Catholic vs. Protestant.

Alternatively, explanations may emphasize differences in attitudes between people who are religiously active, irrespective of denomination, and those who do not hold religious beliefs.

Theories emphasizing the importance of denominational differences are well known in political economy, due in large part to Weber’s work on Protestantism and the development of capitalism. This type of theory may be relevant for understanding social insurance.

So, for example, in cases where a religious tradition suggests that worldly success is attributable to individual merit, one might find that believers prefer low levels of social insurance whereas this would not be the case for members of other religious traditions.

One might also suggest that the relative importance of different religious denominations helps determine the types of political parties that form in a country, and these parties themselves lead to different welfare state outcomes.

For European countries it has often been argued that the welfare policies adopted by Christian Democratic parties are influenced by Catholic doctrine and that they differ systematically from the types of welfare policies pursued by Social Democratic parties (Esping-Anderson 1990).

Our empirical results suggest that while there is a positive correlation between the percentage of Catholics in a country and levels of social spending, this correlation is generally not statistically significant. Nor is there a significant correlation apparent when considering individual-level data.

Beyond the issue of denominational differences, there are several possible channels through which religiosity, irrespective of denomination, might have an influence on the demand for social insurance.

One important possibility is if being religious prompts individuals to become more altruistic, advocating greater spending on the disadvantaged.

While religiosity may lead to greater altruism, and consequently greater advocacy of social insurance provision, for certain groups at certain times, the strong negative correlation between social spending and religiosity in Figure 1 strongly suggests that religiosity may also have other effects that work in the opposite direction, and our statistical tests further support this view.

Another possible influence of religion on social insurance is that it may lead individuals to draw particular inferences about how the economy functions.

So, for example, individuals who are religious may be more likely to believe that hard effort will be rewarded with a higher income and that exogenous factors like family background do not represent inherent obstacles to economic success.

As a consequence, they favor less provision of social insurance. Piketty (1995) argues that differing beliefs about the extent to which income is dependent on individual effort are an important determinant of individual attitudes with regard to income redistribution.

Subsequent empirical evidence has supported this claim, demonstrating that differing beliefs about the determinants of economic success are correlated with individual preferences with regard to income redistribution, and social insurance, and that these beliefs are also correlated with cross-country differences in levels of social spending.

While the model in Piketty (1995) emphasizes how past personal experience influences beliefs about the “importance of effort”, Benabou and Tirole (2006) have suggested that certain individuals may find it useful to maintain a belief that the world is just (in the sense that hard effort is well rewarded) even when faced with contrary evidence.

Maintaining a “belief in a just world” helps people with imperfect willpower to motivate themselves. Drawing on Benabou and Tirole’s contribution, one might argue that it is likely that individuals with religious beliefs (Protestant or other) have a particularly high propensity to believe the world is just.

This does not imply that economic beliefs held by religious individuals are necessarily inaccurate (it may be that non-religious individuals have more biased beliefs), but it does suggest a further mechanism through which religion might influence attitudes toward social insurance provision.

While there is strong empirical support for the proposition that beliefs about “effort” play an important role in determining redistributive preferences, the effect of religiosity on social spending does not appear to pass principally through this channel.

In our extensive empirical tests we find that our core empirical result that religious individuals prefer less state provision of social insurance is unaffected by the inclusion of this luck/effort variable. One further possible channel through which religion may affect social insurance provision is through an issue bundling effect.

If political competition involves both a standard left–right dimension where preferences are determined by income, as well as a second dimension, where preferences are determined by religious beliefs, then it may be the case that religious individuals who are favorable to social insurance support political candidates who are less favorable to social insurance but who nonetheless share similar views with regard to the second issue dimension.

The second issue dimension could involve questions like whether abortion should be legal, whether homosexual marriage should be legalized, or an issue involving church–state relations.

Roemer (2001) has presented a theoretical framework for considering such issue bundling effects, and in a previous contribution (Roemer, 1998) he provides empirical evidence that policy outcomes involving income redistribution may be affected by religion as a second issue dimension.

Arguments about issue bundling suggest that religiosity may be correlated with support for candidates who are unfavorable to social insurance, and hence religiosity may be an important determinant of cross-national variation in levels of social spending.

These arguments do not suggest, however, that religious individuals should be inherently less favorable to social insurance provision.

Since our empirical tests show a negative correlation of religiosity with both cross-country data on social spending and individual data on social spending preferences, this suggests that, while religion may influence social insurance through an issue bundling effect, this is not the only channel through which religion may affect social spending.

The relationship observed in Figure 1 between religiosity and social spending might also reflect a direct substitution effect.

According to this argument, individuals who are religious are no less demanding of insurance against adverse events than are other people, but because members of religious congregations receive material insurance benefits directly from their churches, they express less demand for social insurance provided by the state.

Several recent papers including Hungerman (2005), Chen and Lind (2005), and Dehejia et al. (2005) have used U.S. evidence to explore whether membership in religious congregations involves a form of monetary insurance where individuals give monetary contributions and members of congregations who suffer adverse events like unemployment receive tangible benefits from their church.

The most direct mechanism for insurance is through charitable spending by churches. Hungerman (2005) shows that for members of the U.S. Presbyterian church, charitable contributions have been negatively correlated with levels of state welfare spending.

Dehejia et al. (2005) show that the consumption effects of shocks to household income are attenuated for members of religious congregations, which is also supportive of the direct substitution argument.

While these papers support the idea that religious participation can provide an alternative form of monetary insurance to state-sponsored programs, it seems less likely that these findings also provide an explanation for the gaps in social spending observed in Figure 1 between highly religious and highly secular countries.

The simple reason for this is that, according to the data used by Hungerman (2005), Chen and Lind (2005), and Dehejia et al. (2005), even in a highly religious country like the United States, religious individuals on average give no more than 2% of their income to churches.

This means that in the aggregate the amount of religious giving in the United States pales in comparison with the differences in social spending levels observed in Figure 1 between the United States and more secular countries.

In sum, the direct substitution argument is plausible and supported by empirical studies, but it is insufficient to explain the cross-country differences in social spending that we observe within the OECD.

One final possibility regarding the link between religion and social insurance is that the relationship is spurious because both religiosity and social spending are endogenous to a country’s level of economic development.

The well-known secularization hypothesis suggests that as a country becomes richer its population grows less religious. It is also frequently observed that rich countries tend to have larger welfare states, on average, when compared with poor countries, though the theoretical reasons for this are not firmly established.

To the extent that both of these hypotheses are accurate, we would expect to observe a negative correlation between religion and social insurance provision, even if there was no causal relationship between the two variables.

In fact, within the sample of high-income OECD countries considered in Figure 1 the level of economic development (measured in terms of log GDP per capita) is only weakly correlated with either social spending or religiosity.

Our Argument and Its Assumptions

Though we believe that religiosity undoubtedly influences welfare state spending through several channels, we place particular emphasis on the possibility that religion and social spending are alternative mechanisms of insurance.

While government programs like unemployment insurance, health insurance, and pensions spending help cushion people against the effects of adverse life events, personal engagement in a religion may also dampen the extent to which people are affected by events like job loss, or ill health.

In some cases religious participation may bring communal material support. However, we argue below that, beyond any purely material benefits, religious engagement can also provide important psychic benefits for individuals who suffer an adverse event.

We focus on these psychic benefits in the formalization of our argument. The main cost of social insurance is that it needs to be financed by taxes, and some individuals inevitably wind up financing collective insurance without needing to draw heavily on its benefits.

One of the main costs of religious engagement involves the time it draws away from other activities that people may find pleasurable.

If one accepts that religion and welfare state programs have related, if not identical effects, and that they both have costs, then to the extent that individuals privately insure themselves via religion, they should logically prefer a lower level of insurance by the state.

Our proposition about the effect of religiosity on social insurance involves three core assumptions. First, events like job loss or major sickness do not only impose monetary costs on individuals, they also create psychic costs.

These psychic costs can involve damage to self-esteem, stress, or the loss of enjoyment from having a network of friends.

Empirical evidence supports this proposition.9 Our second core assumption is that religiosity provides some of the same psychic benefits as does being in a “good” state in terms of health, employment, or retirement income.

Religion may help insure individuals against adverse life events for several reasons. For one, people who are religious may derive psychic benefits from having a network of friends from their church, mosque, or synagogue, and such associations are likely to provide comfort during times of difficulty in the same way as would friends within the workplace.

But this type of psychic benefit should logically exist for any type of collective leisure activity. Religiosity may also have more profound psychic benefits that make it exceptional, if not unique, in influencing the way individuals “appraise” adverse events like job loss or ill health (Pargament 1997, Smith, McCullough, and Poll 2003, Park, Cohen, and Herb 1990).

So, for example, religious individuals may be more likely to judge that such events do not pose challenges to their self-esteem, their overall beliefs or life goals, and they may even see adverse events as a challenge offering opportunities for spiritual growth.

In an extensive study Pargament (1997) demonstrates how religiosity also influences the different “coping” mechanisms that individuals use to confront adverse life events. In making such arguments these authors draw on the theory of stress, appraisal, and coping developed by Lazarus and Folkman (1984).

Lazarus and Folkman define cognitive appraisal as “a process through which the person evaluates whether a particular encounter with the environment is relevant to his or her well-being and, if so, in what way” (Folkman and Lazarus 1986, p. 572).

This could involve a judgement whether an event poses potential harm to one’s self-esteem. They go on to suggest that “A range of personality characteristics including values, commitments, goals and beliefs about oneself and the world helps to define the stakes that the person identifies as having relevance to well-being in specific stressful situations.”

While Lazarus and Folkman did not themselves emphasize the importance of religiosity for appraisal, it is not difficult to understand why religious beliefs might have an important influence on this process.

These ideas of contemporary psychologists about religion as a buffer against external forces are also consistent with classic work in the field of psychology.

So, although Freud (1927(2001)) took a negative view of religion, he too emphasized its role in providing individuals with a mental buffer against external forces.

In addition to the strong theoretical arguments, there is clear empirical evidence to support the idea that religion has positive effects on the psychological state of individuals, and that it helps in responding to adverse life events.

A number of studies have demonstrated that individuals who describe themselves as being religious tend to have higher subjectively measured levels of life satisfaction. It is also interesting to note that a number of recent empirical studies have demonstrated that there is a lower incidence of depression in individuals who describe themselves as being religious.

In addition to the above empirical evidence, it is important to note that Clark and Lelkes (2004) have shown that individuals who are religious suffer from significantly lower estimated losses in subjective utility as a result of episodes like unemployment.

This supports our second assumption quite directly. Finally, at least one study has found that people who describe themselves as being religious tend to purchase significantly less life insurance than do non-religious people, which is also consistent with our assumptions (Burnett and Palmer, 1984).

Our third and final assumption is that individuals have utility functions where monetary consumption and “psychic benefits” are not additively separable.12 More specifically, we assume that the psychological benefits of religion are greater for those with lower incomes.

This is consistent with empirical studies indicating higher levels of religiosity generally and religious coping in response to adverse events specifically among the poor, elderly, minorities, and women (see e.g. Pargament 1997, p. 156).

The paper by Dehejia et al. (2005) provides further evidence in this regard from the U.S. National Survey of Families and Households. They find evidence that religious involvement may do more for low-income than high-income individuals to attenuate the negative effects on subjective well-being of adverse events like unemployment.

This is consistent with our non-additive separability assumption, as is their conclusion that the psychic insurance effect of religion is more clearly observed for African Americans in the United States than for whites.

Location: Indonesia