Printer Friendly

Economic development and family size: a comment.

Rios (1991) provides a useful overview of Latin American fertility trends and addresses the possibility that increasing incomes will lead to fertility decreases as happened in Europe. His estimates suggest a conclusion which places increased income at center stage regarding population: "Future economic development in Latin America will cause income to rise, improve the educational attainment of women, and further reduce the proportion of the population living in rural areas; then, as in Europe, family size will drop. Further decline in mortality, not accompanied by economic development, will encourage family size to increase." The equation supporting this conclusion (Rios's Equation 2) is replicated as Equation 1 in Table 1. The data set includes 19 Latin American countries and 15 European countries.

These findings are consistent with Rios's stated conclusion. The negative coefficient on GNP is also consistent, however, with a very different hypothesis: that high fertility rates lead to increased population and, thence, to decreased income. That is, the causation may be opposite to that suggested by the author. As it turns out, however, we need not dwell on attempts to explain what a statistically significant relationship between these two variables might mean; the data offer little evidence that any such relationship exists.

Rios includes both European and Latin American countries, so estimation imbeds the maintained assumption that the linear relationship is the same in both sets of countries. Given the very different levels for both per capita GNP and high-school attendance, and the length of time for which it has persisted, different behavior might be expected. (Further, even if the underlying behavioral relationships are the same in both Latin American and European countries but are non-linear, then imposing linearity can badly bias estimators.) Indeed, adapting this equation slightly to allow different behavior for European and Latin American countries changes the picture significantly. (A Chow test strongly rejects the null hypothesis that the same model applies to Europe and Latin America; see Table 1.) The more flexible model provides little evidence that income per se affects fertility. Table 1 reports the estimated relationship for Latin American and European countries. The estimates result from an equation of the form:

FR = ||Alpha~.sub.0~ + ||Alpha~.sub.1~GNP + ||Alpha~.sub.2~HS + ||Beta~.sub.0~E + ||Beta~.sub.1~E*GNP + ||Beta~.sub.2~E*HS + |Epsilon~.

Here E is a binary variable, equal to 1 for European countries; E*GNP and E*HS are interactive terms with GNP and HS indicating per capita GNP and the percent of young women who attend high school, respectively; and |Epsilon~ is a random error term. These estimates suggest that the apparent relationship between GNP and FR in equation 1 is spurious, resulting from forcing the same parameters on data from different populations. GNP does not have a significant impact in either subset.
TABLE 1
Total Fertility Regressed on Development Indicators
 Equation 2(**)
Independent Equation 1(*) Latin
Variables America Europe
Constant 5.933 6.522 2.238
 (0.421) (0.438) (0.900)
Per capita GNP -0.095 -0.314 0.046
 (0.045) (0.260) (0.049)
% women in H.S. -0.038 -0.041 -0.0005
 (0.009) (0.011) (0.012)
R-squared 0.749 0.856
* Equation 1 is Rios's Equation 2, Table 7.
** Equation 2 allows different intercept and slope values for
European countries. The F-score on the test that coefficients
are the same in Europe and Latin America, 6.887, allows
rejection of the null hypothesis.


In contrast, there is still evidence that high-school attendance by young women does reduce Latin American birth rates. This is consistent with an opportunity-cost explanation of decisions regarding childbearing. The education variable has no appreciable impact in European countries. This result is not surprising, since high school attendance is so nearly universal in Europe that the marginal impact of increased attendance is negligible.

Together, these results confirm part of Rios's conclusion, that factors which increase the proportion of young women who are educated cause reduced fertility rates. They cast doubt on the proposition that increasing income leads inexorably to slower population growth. Rather, to the extent that reducing population growth is a policy goal, the estimates suggest implementing policies favoring the education of young women.

Reference

Rios, R., "Economic Development and Family Size," American Economist, Vol. 35, Number 2, 1991.

J. Wilson Mixon, Jr. is a Professor of Economics, Berry College.

ECONOMIC DEVELOPMENT AND FAMILY SIZE: REPLY

In his comment of my article "Economic Development and Family Size" Mr. Mixon raises several conceptual issues and makes one statistical point. He sees my article as contrasting the effect that rising income has on fertility with the effect that declining mortality has on fertility. My intention was different, I meant to contrast the effect that economic development has on fertility with the effect that declining mortality has on fertility. Economic development is a broad process of change which includes rising income, urbanization, additional schooling, and less attachment to traditional values.

GNP changes are a convenient proxy for the modernization and development, which I believe are essential for fertility to decline. Them are other variables that meter development with equal or greater precision than income, for example schooling. A year of schooling has roughly the same meaning from one Latin American country to the next, while differences in per capita income are distorted by unequal income distribution. To check Mr. Mixon's claim that GNP does not have an impact on fertility among the Latin American countries I divided these countries in three groups: those with a less equal distribution of income, those with an average degree of inequality, and those with more equal distribution of income. Two dummy variables were introduced to differentiate across groups.

Controlling for inequality makes the income variable more precise, and in this formulation it shows a negative coefficient that is significant. As expected countries with a more unequal income distribution have higher fertility, while countries with a more equal distribution have lower fertility. The schooling variable carries a negative coefficient that is not significant. But I will not argue based on these results that Latin America must strive to raise incomes and neglect education. Development will change them both and will also bring down fertility.
Regression Results -- Dependent Variable: Total Fertility Rate
(19 Observations, only Latin American countries are included)
 Standard
Variable Coefficient Error
Constant 5.437 0.577
Schooling -0.003 0.017
GNP -0.735 0.325
More Equal -0.970 0.460
Less Equal 1.118 0.512
Adjusted R-squared = 0.73


There remains the point as to whether the coefficient of the income variable is negative because development lowers fertility or because high fertility hinders development. Without denying that the latter makes a lot of sense it is the former that is correct. Prior to the turn of the century Latin American countries were very similar in their basic demographic rates, yet quite different in income levels. The income disparities have historical roots: countries like Argentina, Uruguay, and to a certain extent Chile had been thinly populated at the time of the conquest and were poor in precious metals. In these the Spaniards eliminated rather than subjugated the natives; after independence these countries attracted large numbers of European immigrants and moved quickly into modern ways of production and organization. In the Andean countries and Mexico, sites of the Inca and Aztec empires, the colonial system was entrenched, and has proven hard to displace. And there are those who are rich by accident, like lucky oil rich Venezuela. Comparatively few are like Haiti, made poor by its high fertility.

Robert J. Rios Assistant Professor of Economics, Lehman College, CUNY
COPYRIGHT 1993 Omicron Delta Epsilon
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1993 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Title Annotation:includes author's reply; response to Roberto J. Rios, American Economist, vol. 35, no. 2, 1991
Author:Mixon, J. Wilson, Jr.
Publication:American Economist
Date:Mar 22, 1993
Words:1270
Previous Article:Allen's arc vs. assumed isoelasticity, pedagogical efficacy vs. artificial accuracy: a comment.
Next Article:Industrial organization: survey of reading materials in universities in Canada and the United States.
Topics:


Related Articles
Testing the quadratic relationship between economic development and the fertility rate: a comment.
A citation-based analysis of economists and economics programs.
Research in brief.
Essays on Economics and Economists.
Getting noticed in economics: the determinants of academic citations.
Hoover Institution Fellows On Israeli-Palestinian Peace Talks, China Trade Bill, and Education Reform.
Productivity. (Program Report).

Terms of use | Copyright © 2017 Farlex, Inc. | Feedback | For webmasters