Religious Faith and Economic Growth: What Matters Most-Belief or Belonging?

Theoretical Approaches

Regarding the conceptual or theoretical approaches to the connection between religion and political economy, there are two causal directions that analysts tend to think about. This work has appeared especially in the literature on the sociology of religion.

One important line of research posits that religion (or measures of religiosity) are dependent upon developments in the economic and political aspects of contemporary life. This research suggests that events in an economy--levels and standard of living or governmental market interference--influence such things as church attendance or religious beliefs.

The second theoretical approach looks at the connection between religion and economic and social life from the other direction. Religion is thought of as being the independent variable, influencing something about outcome on the economic, political, and social side. For example, Max Weber's famous theory about capitalism is in that line: Religiosity influences economic performance, and perhaps political institutions.

Secularization Hypothesis

Going back to the first type of model, in which religion is seen as being dependent on social and economic factors, there are two important sociological theories about how religion responds to these factors. One approach is called the "Secularization Hypothesis." It's a part of what is often called "Modernization Theory," which looks at how the economies of developing countries develop institutional capabilities to alleviate poverty and rationalize markets.

The Modernization theory posits that as an economy develops and gets richer, certain societal institutions and features change in a regular way. The Secularization hypothesis applies this theory to religiosity: As economies develop and get richer, people supposedly become less religious. "Less religious" is measured either by participation in organized religion (e.g., church attendance) or by certain indicators of religious belief.

The Secularization theory (or "Secularization Hypothesis") is widely held among analysts. The United States is frequently mentioned as an exception to this idea: The United States is very rich and it is also highly religious. Therefore, the U.S. is not thought to fit the Secularization view.

Religion Market Model

The second important approach in the sociology of religion literature is often called the "Religion Market Model." It speaks about the way government interacts with religion and influences the extent of participation in religion--or even the extent of religious beliefs. Thus, sometimes the government regulates the market, possibly promoting a monopoly religion or making it difficult for other religions to flourish.

Under this theory, the government might make it difficult for people to practice their religion by going to services. On the other hand, it might subsidize religious activity. Yet, in one way or another, the government will be influencing the amount of formal religious activity. One example of this influence is the establishment of an official religion in a country.

There are, however, different views in the literature about whether such governmental action will promote religiosity or detract from it. The argument that it might detract from religiosity is as follows: If you have an established religion, you tend to have a monopoly, and monopolies tend to function inefficiently. For example, it is thought that a monopoly church--such as the Catholic Church in predominantly Catholic countries or the Anglican Church in England--doesn't perform efficiently and is not attractive to practitioners. In response, people are thought to participate less. This argument even appears in Adam Smith's Wealth of Nations, so it's quite an old idea.

On the other hand, established religion tends to go along with government funding of religious activities. The government might be paying for church buildings or church personnel. In this case, one might predict that subsidizing something might result in greater religious participation.

In general, this Religion Market Model argues that the way state and church interact is quite central. Extreme examples are communist regimes, including the Soviet Union and China. But many Eastern European countries also tried very hard to eradicate organized religion and apparently had some degree of success. That's another way in which the government might influence religious activity.

A Combination of Approaches

Our empirical analysis of the determinants of religion combines these two kinds of approaches. It combines the Secularization view's emphasis on the demand for religion--and how religion is influenced by economic development--with the Religion Market Model's focus on the supply of religion--particularly as it is influenced by government policies in this area. Our results combine these two theories rather than trying to determine which is the most valid.

Regarding how religion influences economic performance, I would stress religion's influence on economic growth--as in Max Weber's work on the "Protestant Ethic." This view underlies an important part of the empirical work we're doing. This theory posits that religion affects the economy by influencing certain individual traits. These traits, in turn, may make people more or less economically productive.

Weber particularly stressed the Protestant work ethic. In general, if religiosity influences the willingness to work and to be productive, that might be an important factor.

There might also be influences working through honesty, ethics, and other kinds of values. Additional effects are thought to operate through thrift, whether people are open to strangers or not, and other factors. Those would be the kinds of individual characteristics that might influence the economy and in turn be influenced by religious beliefs and values.

There are other aspects of religion that might be negative for economic performance. For example, organized religion may prohibit certain kinds of economic transactions or be able to persuade the government to prohibit them--for example, credit markets, insurance markets, or regulations related to the incorporation of businesses.

Most of our study is empirical. Thus, I want to say a little bit about the kind of information that we're using to carry out our analysis.

First, we're trying to quantify some of these ideas. There are two main kinds of individual-level dimensions with respect to religiosity. One is about participation in formal religion--which might have to do with things like attendance at formal religious services or at personal prayer--and the other is about religious beliefs of certain kinds. When thinking about how religion might influence economic outcomes, we think about these outcomes as working through certain beliefs that influence traits like work ethic, honesty, and other characteristics.

Sampling and Sources

In order to get a broad cross-country sample of the extent of religiosity, the things we are measuring come from six important international surveys of values and other activities. These were carried out from the early 1980s through 1999. Three of these surveys are waves of the so-called World Value Survey: 1981, 1990, and 1995.

The World Value Survey now covers around 50 countries and surveys 1,000 to 2,000 individuals in each country to get an idea of their values in various respects. We use data about attendance at formal religious services in addition to a number of specific religious beliefs.

The data we are looking at concern beliefs--related, for example, to an afterlife or whether people believe in heaven/hell. There are some more general questions, such as belief in God, and there are also questions that are more robust across religions: for example, whether or not you consider yourself to be a religious person.

There exist these three waves from the World Value Survey, and there are two waves from the International Social Survey program in the 1990s. Those data are mostly for richer countries, and only include a few less-developed countries.

There is something of a trade-off between these two sources because the World Value Survey tries to be more internationally inclusive. As a consequence of the goal of inclusiveness, the quality of the data is probably not as good: It is more difficult to do the surveys in a scientific manner in some of these poorer countries. We also have information from Gallup--from their so-called Millennium Survey that refers mostly to 1999.

I might also note there's a current wave of the World Survey that will soon be available which applies mostly to 2001. For this survey, there are affiliates in various countries who actually do the surveys and collect the data. To provide incentives for people to participate in that project, they get a three-year head start in using all the data, and only after a three-year lag do they release the data for public use. That release will occur in the spring of 2004.

Thus, a few months from now, the data for the 2001 wave will be available. That is particularly interesting because the new wave contains many more countries than before. It includes about 70 countries, and it has more representation than previous surveys in Muslim countries and some countries that practice Eastern religions, such as Buddhism.

We also use information on religious adherence across countries. That information lets us know about people who profess to be Catholic, Protestant, Muslim, Buddhist, Hindu, and so on. The Barrett World Christian Encyclopedia, which has two editions, is the source of these data. Even though it has the word "Christian" in the title, it covers all world religions.

Variable Constructs

I mentioned before that part of the theoretical framework for this project is from the Religion Market Model, which focuses on the nature of the state's regulation of religion and how that affects the religion that people can practice. We have a number of variables that we have constructed to try to quantify that influence. I will mention them briefly.

• I have a measure about whether or not a country has an official state religion (an established religion) that typically applies around 1970. There are some controversial cases, but I have countries classified as either having an established religion around 1970 or not. For example, most Muslim countries, except for Turkey, are classified as having an established religion. Some Catholic countries have established religions and some do not. The Scandinavian countries all have the Lutheran religion as the established religion around 1970. Thailand is an example of the Buddhist established state religion. Of course, many countries do not have an established religion.
• We also have a measure of the way that government regulates the market for religion, particularly in terms of entry. This measure focuses on whether or not the government has to approve leaders of the church in the country or actually makes the appointments.
• Another influence is about Communist countries: Communism has an important negative effect on religiosity. We have a lot of examples of Communist countries in the sample, especially Eastern Europe. It is also interesting to look at important regime changes in the 1990s. Almost all of those countries moved away from Communism. What was the nature of the recovery of religiosity, or did it recover? That is part of what we'll be examining.

The Nature of the Data

Let me tell you something about the data that we're using. These are some of the variables I just discussed. (See Table 1.) This first variable is this indicator for whether or not a country had an established state religion around 1970. Most of the cases are straightforward. A few are not and are subject to a lot of argument.

For example, some people think that Italy and Spain don't have an established religion around 1970. There was clearly a lot of liberalization away from the previous regime by then. We're classifying those countries as having an established Catholic religion in 1970.

This other variable is about whether or not the government is regulating the market for religion in the way I described before. And the final column is another variable that's important in this religion market approach: It is the extent of diversity in a country with respect to the religious denominations among the adherents to some religion.

For example, if you take any of the Scandinavian countries, a substantial proportion of the people say they have no religion. However, if you look at people who say they have some religion, it turns out that most of the people in those countries are Lutheran. There is very little religious pluralism there because most of the people who adhere to some religion are of the same type. In Turkey, for example, almost everybody is Muslim, and that's true of most of the Muslim countries, except for Malaysia: A very large fraction of the population who adhere to a religion is Muslim.

There are many Catholic countries that are very heavily Catholic, like Ireland, Spain, and many South American countries. Countries like that have a pluralism indicator in our table that's close to zero. Italy, for example, is almost a zero, which means that everybody there who professes some religion claims to be Catholic.

In contrast, there are other countries where there is a great deal of diversity among the religions in the population. The United States is often mentioned as an example of that. The U.S. contains many religions: Catholic, Protestant, and an array of others.

Germany is an interesting case because it is about 50/50 Protestant and Catholic. That ends up being a fair degree of pluralism in the way that we mentioned. But you might think it is different to have two religions that everybody claims, rather than a greater number, so there are different ways we might measure this. We think that pluralism makes the religion market more efficient and, therefore, ultimately makes people more religious.

That's just a little sketch of the data. Let me tell you some of the results that we find. We're looking across a large set of countries, at various different points in time, but it's in the 1980s and 1990s that we have the data on measures of religiosity. We particularly examined what might determine things like church attendance or attendance at an analogous formal service, as well as various religious beliefs.

Findings

One thing we find is that if you look at economic development as a single kind of indicator--like per capita gross domestic product or real income per person--you definitely find a strong pattern whereby more economic development and higher per capita income are associated with lower religiosity. That's true whether you're talking about church attendance or various individual religious beliefs that we examined. To that extent, I think that there is some validity in the secularization view. The overall evidence that we have from the "World Group of Countries" is that richer places are less religious.

We also make an argument in our current paper that causation runs from economic development to religiosity. That is, getting richer means you practice less religion, rather than the reverse. You might think there is also a reverse effect, from religiosity to economic development. But we think that the effect that we've isolated here is really from higher per capita income to lower religiosity; that we have some evidence about the causal arrow in this particular case being that way.

To that extent, we have results that look like the secularization view, but the effect is not all that important. Quantitatively, it doesn't really explain very much of the differences in religiosity across the world. The U.S. is definitely an outlier in that relationship. The U.S. looks more religious than you would expect, as measured either by beliefs or by church attendance.

Let me tell you about some other results that we have here. We look at the consequences of having an established state religion. On net, we find that that is actually positive, both for church attendance and for religious beliefs. To some extent, that goes against what Adam Smith said. Smith stressed that established religion would promote monopoly, poor service, and decreased service attendance. He particularly inferred that from looking at the Anglican Church in England.

We find, however, that the net relationship is actually positive, and we think that is basically because state religion tends to be accompanied by the state subsidizing religious activity in various ways. I think an economist, particularly Adam Smith, would generally accept the idea that something that is subsidized will tend to occur more often than something that isn't. It's the same as saying that something that is taxed will tend to occur less often.

However, consistent with Smith, we find that government regulation has an important negative effect on religious participation and beliefs. That's more in line with the Smith results.

Also consistent with this approach and the religion market model is that more of a pluralist structure--a greater diversity of religions represented in the country-- is associated with more religiosity, more participation, and firmer beliefs. That also goes along with the idea that competition and greater choice will produce a better outcome, or a more vigorous one at least.

The Effects of Communism

We find that the communist regimes had an important negative influence on religiosity. Of course, this pattern applies in the context of how people answer questions on surveys, and I'm assuming that the responses are accurate.

In that sense, the communist governments were successful in promoting communism as an alternative religion and suppressing the normal kind of religion. Related to that effect, there seems to have been a recovery in most countries of religiosity in the 1990s after the communist governments were abandoned--roughly speaking about half the way back from where they might have been otherwise after 10 years.

I might mention Poland as a big outlier in that relationship. Poland is the one communist country that was extremely religious during the communist period, before 1990 in particular. Poland and Ireland are the most religious countries in Europe by far. After the communist regime was dropped in Poland, religiosity actually declined. This was unlike the usual pattern, in which religion tended to recover after the communist governments were eliminated.

Let me tell you something about those results before I move on. This is an example of what you can explain or not explain with this empirical approach that I've been sketching here.

Using data from 1990, from the World Value Survey, I've looked at two measures. One is a measure of formal participation--that is, the monthly rate of church attendance--which is the fraction of the total population of a country that says that they attend formal religious services at least on a monthly basis.

The second measure is an example of a kind of religious belief that we can measure from the cross-country surveys. This one is about belief in heaven. The number here is the fraction in the population that say yes to the query as to whether or not you believe in heaven.

Using numbers for the countries that have the data, the model shows the estimated participation or belief, given things like the level of economic development and whether you have a state religion, the variables that I just went through.

One point is that the United States really is an outlier in this relation, sometimes referred to as "A Great Exception." For example, the fraction of the population that says they attend church at least monthly is 58 percent, which is quite high, particularly for a rich country. The model says it should be 36 percent, so that's the extent to which the model does not explain the United States very well. In the broader world data, higher income goes along with lower religiosity. It's just that the United States is not so well-explained by this analysis.

The same is true with respect to belief. The United States has an extraordinarily high belief rate in this particular concept, "Heaven," but also in other ones: 87 percent say they believe in heaven, which is actually quite remarkable if you compare with other places. The model says it should have been 57 percent, so that's another gauge of the extent to which the U.S. is an outlier.

Actually, Poland is the biggest outlier: 85 percent of the population attending church at least monthly in 1990, and the model doesn't explain that at all. People think the Catholic Church played a special role in Poland, and that's why people went to church, but that's a very special kind of explanation.

We have tried to put a finer breakdown into what we mean by economic development. Instead of looking at per capita income, you can look at variables that typically move in a regular way in the process of development. For example, as countries get richer, they tend to have more education; the population is more urban; health is better (including higher life expectancy); and fertility rates tend to go down. Hence, the fraction of the population that is represented by children tends to be lower, and because people are healthier, a greater fraction of the population is old.

I mention those variables because many theories about what determines religiosity focus on those variables: things like education, urbanization, life expectancy, age, etc., and whether or not you have children present. If you look at the U.S. with these finer details about economic development, the fit for the United States improves.

One of the problems with this analysis is that, if you go to this finer detail of economic development, it becomes less clear whether development is affecting religiosity--which I'm convinced is true in terms of economic development overall--or religiosity is affecting the various dimensions of development. Fertility and health, for example, or education are things that often are thought to be influenced by religiosity. So the model fits better if you have this finer breakdown, but it also causes other problems in terms of making firm conclusions about causation.

Religiosity and Economic Performance

Let me now turn to the other kind of analysis, which is "How does religiosity affect the performance of the economy or other aspects such as the political structure and democracy?" We've looked at measures of economic growth across countries based on the change in the per capita gross domestic product. This is part of an earlier research project that I've been working on for 10 or 12 years, and I think I know a lot of things that may influence economic growth and why some countries grow faster than others.

I've tried here to extend that analysis, to examine whether there is a separate influence starting from where I was before, from culture or specifically from religiosity, as an influence on growth. I'm holding fixed a lot of other things that I think I've been working on, things that influence economic growth, and I'm trying to see whether there is a separate influence related specifically to religion in terms of effecting economic performance. That's what I've tried to investigate.

In a way, it's testing the thesis that Max Weber put forward about a hundred years ago. We entered religiosity as a potential influence on growth in two dimensions. One is about participation in formal religion, measured by rates of church attendance, and a second one is about religious beliefs. It might be belief in an afterlife or in hell or heaven or something like that.

Conceptually, the way we're thinking about this is that if religion matters for economic performance, it's going to work by affecting beliefs that affect certain character traits that I referred to before: work ethic, honesty, etc. I'm thinking about religious beliefs as being the important output from the religion sector with respect to a potential influence on economic performance. It might be, for example, that formal religion or other things influence beliefs, and those might work in a positive direction to encourage productivity and growth.

If I hold fixed beliefs--and, therefore, character traits such as work ethic and honesty--then more church attendance signifies more resources being used. That probably means that there should be a negative connection with growth. That is, if I'm already holding fixed the thing that is being produced by the religion sector--beliefs--then if I have more input, more time, and more resources, I expect that to be negative on growth.

This is not the only way you could imagine this whole process. You might think that the formal religion sector produces other things, like social capital, networks, or some kinds of social organizations. You might think those are particularly important for economic activity--that is, a channel that does not work through the beliefs. You might think that the religion sector is important for other reasons, for example, influences on education. But my conjecture is that the main effect--and the thing that's special about religion--is the way it affects beliefs, values and certain character traits. That is why I'm focusing on that and that is going to affect my interpretation of these findings.

Let me tell you the kinds of results we get. Conceptually, I'm looking at the determinants of economic growth, across countries and over time, measured by growth of per capita gross domestic product. There are a number of things that I think influence growth, based on a lot of previous research by others and myself. I'm holding those constant, and I'm looking at the separate contribution to growth from religiosity measured by the two kinds of variables I discussed before: one is about formal participation--here it's measured by monthly church attendance--and the other one is some kind of religious belief, which I think matters for character traits. In this particular case, a belief in hell turns out to have the most impact in terms of economic performance. I don't know whether to be happy or sad about that.

The kinds of results that we get are summarized by Figures 1 and 2. Think of this positive relationship as the impact on economic growth, if belief were to go up while holding fixed the amount of formal participation in religion--as measured by church attendance. If you look across countries, some countries are very religious and some are not very religious, measured either by church attendance or beliefs. Therefore, the biggest variation in the data has the two measures of religiosity moving together.

That is not what the figure represents. It represents the effect of firmer beliefs--in this case belief in hell--for a given level of participation. For example, there is a recent book by Grace Davie discussing religion in Britain. She argues that this is a case of people who have high levels of belief in various things but don't attend church. Therefore, formal participation, in terms of church attendance or other metrics, is low, but beliefs are relatively high.

That is the kind of thing we are picking up: a situation in which belief is high for a given level of formal participation, measured by church attendance. That is where we find this positive relationship contributing to economic growth.

Figures 3 and 4 apply "for a given level of belief," which I think of as the principal output of the religion sector: it shows the effect on when you have more participation--meaning more church attendance--which we find to be negative. My interpretation of this negative effect is not that church attendance is negative overall for economic performance. Rather, if the church attendance is doing something--if the formal sector is doing something--then the greater church attendance and participation should produce firmer beliefs. Therefore, you would not actually be holding this other thing constant. This is supposed to be the effect of resources being used, given what the religion sector is producing in terms of beliefs.

If the religion sector had also been important in terms of social organizations and social capital, you might have expected that more participation would have a positive effect, even given beliefs. But we don't find that. We find that it has this negative kind of relationship.

You could also ask: What is the overall consequence for economic growth of a country being more or less religious? I mentioned that the typical pattern is those countries that are high on attendance are also high on beliefs, and vice versa. If you look at countries just being more religious overall or less, the relation with economic growth is weak. There's not much relationship overall between whether countries are just more religious or less religious and whether they grow fast or slow.

With respect to beliefs, however, there's a little bit of evidence that the impact of hell (which you can think about as kind of a negative incentive to avoid bad things instead of a positive incentive to do good things) is more potent for economic performance than some other beliefs. I wouldn't push that very far. I don't think you can really discriminate that well between such a belief and other factors. We do find in general that beliefs in an afterlife, heaven and hell, all work in a similar way; and they all look similarly positive. Whereas, if you look at a more general question such as "Belief in God" or whether people regard themselves as religious, that actually has no relationship to economic performance.

Most people feel like they have to answer yes to the question, "Do you believe in God?" So the average belief in God across the whole sample is close to 90 percent. However, it doesn't seem to mean anything in terms of behavior and economic results, whereas some of these other beliefs are much less frequent and seem to have much more significance in terms of the relationship between actual behavior and economic results. It's not the case that all religious beliefs work in the same way in our cross-country findings.

We also make an argument in our paper about why the direction of effect flows from religion to economic growth rather than the reverse. I mentioned before that there is a two-way causation. We think that these effects represent the effects of either church attendance or belief on growth rather than the reverse, and we make some arguments in our research about why we come to that conclusion--about why this is not just a correlation and why we think there is actually some causation from religiosity to economic results.

Conclusion

Let me mention in conclusion some additional kinds of research work that we're currently looking at. In addition to looking at the effect of religion on economic performance, we are trying to look at the effect of religion on various social and political indicators. For example, we are examining whether religion has some influence on the preservation or occurrence of democracy and civil liberties. We are also looking at whether religion has an important country-wide effect on variables like fertility, health, education, etc.

We are also looking at the question about why some countries have state religions and some do not. Many state religions date back hundreds of years. As an example, Sweden instituted the Lutheran Church as a state religion in the 1500s, along with the Reformation, and stuck with it for more than 400 years--even though the country became very non-religious. Sweden eliminated the Lutheran Church as the official religion only in the year 2000. Many countries with state religions have them because something happened 400 years ago--for example, Henry the VIII decided he didn't want the Catholic Church anymore. Yet there are a lot of examples of countries dropping state religions. For example, if you look at the years between 1900 and 2000, many countries that once had state religions have abandoned them.

Even after the long period of communist governments, the state religions of the past often re-emerged. That is particularly true in some of the Asian republics that were formerly in the Soviet Union.

We are looking at people who are non-religious and at people classified in the data as those with no religious affiliation and those people who say they are atheists. I guess that to be an atheist is to take an active position: Being anti-religion and being non-religion is perhaps something different, but we're trying to investigate that, along with this other kind of analysis. We are also doing some work to determine religious conversion within a country. Some countries have a fair amount of conversion--such as the United States. Some countries have very little, and we're thinking about the determinants of conversion.

JOSHUA MITCHELL:
It seems to me that the question before us--the larger question-- is: Why religion, now? The storybook tale of relationship between politics and religion, especially between the period of 1989 to 2001, could have been summed up by Francis Fukuyama in The End of History: The idea was that, somehow, all of these deep longings and beliefs would go by the wayside. Yet this storybook tale is something that pre-dates Fukuyama, obviously.

I call this storybook tale the "Fable of Liberalism," and by that, I mean a deep tale that we tell ourselves in the West. It goes all the way back to Montesquieu and Voltaire and Constant and to Tocqueville and Adam Smith and Immanuel Kant, and it can be summed up in Voltaire's Letters on England, written roughly in 1750.

What he's trying to do, among other things, is to move us away from archaic religious beliefs toward a tamer, calmer, more rational social order. He says, "I go to the English stock market. There I see Jew, Christian and Moslem and nobody cares about each other's beliefs as long as their money is good." Voltaire is saying that we need to move in this direction, move away from religion entirely, because it only gives us intractable oppositions, and that perhaps economic and political liberalism are the way out.

We also have, in more contemporary language and largely since World War II, a whole generation of scholars who developed "Modernization Theory." There, too, you have the view that religion is going away. So I think this research by Professor Barro may beg the question: Why religion now? Why are we so concerned about it?

There's a larger debate about the relationship between religion and economics that was alluded to in the presentation. First you have Max Weber's work in 1906 in a very subtle book that everybody should read, The Protestant Ethic and the Spirit of Capitalism. Weber's larger question was: Why capitalism now? Why in the West does it occur?

Weber is concerned with the universalizing force of instrumental rationality, and his question is a large one for us today: Does something that emerges in the West, for peculiar reasons, have to do with the West, or is it universal?

I will note that, in the empirical analysis here, you have Protestant as one category. Weber, however, was very clear to parse out all of the different Protestant categories: Calvinists, Lutherans, Ana-Baptists, and so forth. Anyone who has listened to A Prairie Home Companion will know that Lutherans like those in Lake Woebegone are not the kind of hard-driving economic actors that you find in Calvinist New England.

This larger theory that Weber has of capitalism and the advance of instrumental rationality, of course, ties in with the secularization theory. And then, in 1776 you've got Adam Smith's Wealth of Nations, where one finds the claim that religions thrive when they're subject to market forces. Here religion is understood largely as a good, as a product. You can see the influence of Smith on Professors Barro and McCleary.

However, the authors don't entirely rely on Smithian competition. They move in the direction of another very, very important insight, namely that there are what could be called deep character traits, which help or hinder a person's development in certain ways, both politically and economically.

This insight turns us toward an author that was not mentioned by Professor Barro: Alexis de Tocqueville, who is, I think, a hero of a number of us. Tocqueville understood this idea very well, that economic institutions and all the others--family, local government, religion, even the legal system--are important for shaping character of a certain sort.

Moreover, Professor Barro has a deeply Tocquevillian conclusion that these two domains of life--mainly politics and religion--should be kept separate if you're going to have a thriving commerce and, Tocqueville would say, a thriving democracy as well. Tocqueville's reasoning here is important, and it's a question that I think looms large for us in the 21st century.

One argument for the separation of these two domains, which Professor Barro does not cite, but which Tocqueville does, is predicated on that notion that religion is not a preference. Religion cannot be understood in the modality of everyday life of choosing and having preferences in the market model of human life.

Tocqueville understood that the reason why religion obtains at all in human experience is because it speaks at those moments when everyday life is suspended: moments of deep suffering, of loss, of death. His argument is that religion is eternal to the human heart, and therefore, insofar as we can keep these two modalities of experience separate, we can have both strong religion and strong economic growth as well. His critique of the unification of these things is addressed to Europe, where he thought that the political order had overlapped with the religious order, to the detriment of both.

So the two conclusions that we find in this paper--that religion forms traits conducive to commerce; and that of the separation of religion and politics is also necessary--are quite consistent with the conclusions of that great thinker, Tocqueville.

The question I would pose, then, is this: Is it possible, really, to comprehend religion in terms of economic analysis? My intimation is that there are a number of us in the audience who would raise some variant of that question. To be sure, there are some things which can be touched upon when you do this, but it's not clear to me ultimately that we get to the thing itself by treating religion in econometric ways.

The exercise of understanding social reality is a messy sort of thing, Tocqueville would say, because democratic souls--this is my usage, but Tocqueville would agree--want the kind of univocal measure of things. And I think the great temptation in the democratic age is that we want to have a single way of comprehending all sorts of disparate phenomena. Tocqueville's argument is that religion is a different phenomenon, that it can't be comprehended in terms of preferences, and, therefore, we do injustice to it if we try to proceed in that way.