Religion and the Ideal Number of Children in Developed Countries

Alicia Adsera, University of Illinois at Chicago

The relevance of religion to explain differences in fertility has long been recognized. However the debate and the empirical study over the role of religion on fertility behavior has been mainly limited to the US and to women's preferences. From this research, we learn that fertility behavior differs among women affiliated to any religious group and those without affiliation. Those differences increase when the level of church attendance is taken into account. The literature, nonetheless, has mostly focused in the gap between Catholics and non- Catholics, which seems to have significantly decreased in recent years (Westoff and Jones 1979, Mosher et. al. 1986). With few exceptions, little research has been done on how fertility varies across a wider set of religious affiliations (Mosher et al. 1984, Williams and Zimmer 1990, Lehrer 1996). To tackle these shortcomings of the previous literature, I study the role of religion and religious practice in fertility decisions in a multicountry dataset that encompasses answers of around 20,000 individuals of different ages in 13 OECD countries. These data are based on a survey conducted by the ISSP in 1994. This paper is the first multicountry study in the literature dealing with the role of religion in fertility decisions. I use variance in the ideal number of children across individuals, as reported by the respondent, to study the effect of religious affiliation and religious practice in preferences over family size after controlling for individual characteristics, such as age, sex or education. The analysis is conducted both using ordinary least squares and ordered probit with country dummies and the individual control characteristics. In the paper, however, only OLS results are reported for the easiness in interpreting the results. I distinguish among five main religious groups, namely, ecumenical Protestants, exclusivist Protestants, Catholic, those affiliated to other religions (mostly non-Christian), and those without an affiliation. The ISSP provides detailed coding of religious denominations, especially across Protestant denominations. I align Protestant churches along the ''exclusivist-ecumenical'' continuum defined by the strictness of their membership criteria. The study provides information on the regularity of church attendance. To control for religiosity, I construct a dummy for those individuals who attend church once a week and a second dummy that also includes those who attend at least 2 or 3 times a month to test the robustness of the results since weekly church attendance is not mandatory for some affiliations. Even though most of the previous studies only focus on the wife's fertility, I estimate separately equations on the ideal number of children for men and for women to analyze likely differences in the preferences among the genders. Moreover, during the last two decades fertility has sharply decreased in most developed countries and is now below the replacement level. The average total fertility rate in the OECD went from 2.9 in 1960 to 2.04 in 1975 and then plummeted to 1.6 in the late 1990s. Countries with traditionally higher rates, such as Spain or Italy, fertility rates are approaching the unity level. I estimate similar regressions across three age groups to study whether changes in preferences among generations, due to a fast secularization of the population, can explain those fertility trends. Williams and Zimmer (1990) point out to an increased polarization within the Catholic Church in the US that implies an augmented relevance of religious practice in fertility decisions. As a result, if the same were true for Europe, I expect to find that for younger generations practice is more important than before in determining their preferences with regard to children. Moreover, I anticipate the difference to be more significant within those churches with pro-natalist teaching such as Catholics and some conservative Protestant denominations. The lower level of participation in younger generations would then, partly, account for the decrease in total fertility. The main results are as follows. First, the ideal number of children is higher for conservative Protestants and Catholics than for ecumenical Protestants or individuals with no religious affiliation. Second, if both affiliation and the level of religiosity are included, religious practice is the most relevant in explaining differences in the ideal number of children. Further, religion interacts with gender to generate important differences between men and women. Religious affiliation alone has a larger impact for men than for women across all age groups whereas practice matters the most for women. Third, the degree of the different impact of both religious variables varies across age groups. While affiliation is relevant for women older than 50, religious affiliation not followed by religious practice becomes irrelevant as an explanatory factor among young women, especially those under 31 years. Among more pro-natalist groups, while the fact of being Catholic alone continues to matter for young men, it is practice that determines young women's choices within Catholics. Among affiliations, the sharpest decline in church participation has occurred among Catholics. While 56.5% of Catholic women over 50 attend weekly mass, only 25% of those under 31 do. The predicted ideal number of children for Catholic practicing women under 31 and for those over 50 is almost the same but given the lower weight of practicing Catholics within the young women, the average ideal number is lower among the young. Younger generations are more polarized depending on their acceptance of the birth control doctrine of the church. The fast secularization of young Catholic women and the increased role of young women in economic and contraceptive decisions within the couple may account for the new fertility trends, particularly in Southern Europe. Finally, while for men the number of ideal children clearly increases with age; for women, the relation of the ideal family size with age is nonlinear with a minimum for the 31-50 year old group. Civil status is relevant for women but not for men. Results imply that a divorced woman declares an ideal number of children around 5.5% smaller than a single woman and 2.7% smaller than a married woman. Most of the previous research has been directed to analyzing the Catholic and non-Catholic (mainly Protestant) divide. Here, by analyzing changes in the gap of Catholics and Protestants across different cohorts, I conclude that the recent lessening in the gap that has occurred is likely to be due to the steady decline of the difference between responses of Catholic and ecumenical women for both levels of religiosity. Among men the change in the gap has been minor. The reduction in the gap between Catholics and exclusivist Protestants has been smaller. Given the larger weight in society of Catholics and ecumenical Protestants than of exclusivist Protestants, the gap in the ideal number of children between Catholic and Protestant women has indeed decreased. (The paper has been emailed to the session chair).