Academic journal article Contemporary Economic Policy

Determinants of the Saving Rate: An International Comparison

Academic journal article Contemporary Economic Policy

Determinants of the Saving Rate: An International Comparison

Article excerpt


The saving rate is a key performance indicator for development policy. Despite the crucial role that saving plays in the development process, economists have not reached conclusive answers about the role of various economic and demographic variables. The mainstream literature on saving behavior also has not considered explicitly the effects of life expectancy (or mortality decline). The analysis here (i) uses a life-cycle framework to incorporate empirically the effect of a life expectancy increase on the saving ratio, (ii) uses recent data and replicates earlier results to provide additional evidence on the existing debate, and (iii) incorporates investment in human capital to broaden the definition of saving.

The life-cycle hypothesis postulates that individuals spread their consumption evenly over their life spans by saving during the earning span and dissaving during retirement (Ando and Modigliani, 1963). In that model, the saving ratio is independent of the level of per capita income but is a function of the population growth rate and of the income growth rate resulting from productivity growth. As Modigliani (1970) shows, the saving ratio is a decreasing function of the ratio of retired to working population (R/W). This age-structure variable represents the proportion of nonproductive dissavers who constitute the dependency burden. Coale and Hoover (1958) demonstrate that a high fertility rate's unfavorable effect on the age structure and on the dependency ratio adversely affects a country's ability to save and invest. Leff (1969) explicitly analyzes the effect of the dependency burden (both in terms of very young and very aged population) by using international cross-section data. According to Leff, dependency ratios exert a significant negative impact on saving. Ram (1982) and Kelley (1988) replicate Leff's model using alternative samples, time periods, and specifications and find no evidence or very weak evidence of dependency burden's negative impact on the saving rate. Kelley extends the definition of saving to include physical as well as human capital and for this broad measure of saving finds that the young dependency burden is statistically significant but quantitatively unimportant. Leff (1980) takes a new look at the issue in the light of the controversy generated by his first article and again finds support for his original hypothesis. Schumaker and Clark (1992) examine the stability of dependency estimates over the time period 1975--1985 and find that for the full sample, the structure of the model is not stable between 1975 and 1980. A Chow test for regional variations reveals that the structure of the saving model is significantly different for Africa and Latin America and for Africa and Asia and is not stable for Africa and Latin America from 1975 to 1985.

Section II describes the model, data, and samples and provides a priori expectations for the relation between life expectancy and other explanatory variables and the saving ratio. Section III presents results for the basic model and for the extended model that includes human capital. Section IV offers conclusions.


A. The Model

According to the Ando-Brumberg-Modigliani and Houthakkar (1960), the aggregate saving ratio is a function of population growth--more specifically, of age-structure resulting from population growth--and of productivity growth. Most empirical works consider only young and old dependency ratios and level and growth of per capita income (Leff, 1969; Ram, 1982; Kelley, 1988). The analysis here adds life expectancy as one more explanatory variable. Thus, the aggregate saving function is:

[MATHEMATICAL EXPRESSION OMITTED] where s/y is the ratio of gross national saving to the gross national product (percentage) averaged over 1981--1988; [D.sub.1] is the percentage of total population 14 years old or younger and represents young dependency burden; [D. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed


An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.