High Fertility, Age Distribution, and Economic Development

II. Population Growth and Economic Development in the Third World

A. High Fertility, Age Distribution, and Economic Development

In this section of the course, we switch to a macroeconomic focus. In particular, we turn our attention to the Third World, and an examination of the relationships between population growth and economic development in developing nations.

The discussion begins with another, closer look at the concept of the demographic transition, and from there takes us to a consideration of the consequences of fertility behavior for age distribution and ultimately for prospects for economic development. The second part of the section will focus on population, land use, and the environment, with emphasis on two case studies from Rwanda and Zaire.

The demographic transition revisited: Determinants and consequences

At the beginning of the course we considered the concept of the demographic transition, referring to the transition from a demographic regime of high birth and death rates to one of low birth and death rates. During most of mankind's existence on earth, life was nasty, brutish, and short. In order for a given population to survive, high fertility was essential in order to offset high mortality (cf., micro incentives for high fertility).

On average (in the aggregate), fertility was slightly greater than mortality, and world population grew. However, the rate of growth was exceedingly slow and its pace very uneven --both natural and man-made disasters could wipe out in a very short span growth that had taken decades to accomplish.

Every industrialized country has undergone this demographic transition, and many developing countries are experiencing it as well. The stylized view of the transition sees the onset as being reflected initially in a decline of mortality, while fertility remains high. After a lag, this is followed by a stage during which both mortality and fertility decline, and finally they stabilize at low levels (low relative to historical experience).

Figure II.A.1 depicts fertility and mortality in major regions of the world within the context of this stylized view of the transition. (Although the figure shows an upward movement of the death rate rather than simply stabilization at the latter stage of the transition; this reflects a presumed age distribution effect to be discussed more below.)

The historical experience of individual countries or regions of countries did not always follow the stages of the stylized transition as just described. In addition, efforts to link the transition to a specific factor such as industrialization or literacy inevitably encountered contradicting examples (cf., France and the U.S. vs. Germany and England vis-à-vis industrialization).

However, explanations for the "typical pattern" of the transition emphasized both the differences in attitudes toward high mortality as compared to attitudes toward high fertility, and the different factors that influence mortality behavior and fertility behavior.

High mortality was inevitable for most of human history, but eventually this changed. Taking the long historical view, it is clear that improvements in economic organization facilitated trade and made humans less vulnerable to local crop failures. Improvements in the general standard of living and development of public sanitation systems helped bring mortality down, and by the 19th century advances in the treatment and prevention of infectious diseases further reduced death rates. The key aspect here that distinguishes the situation with respect to mortality as compared to that regarding fertility is that once high mortality was no longer inevitable, lower mortality was universally desired.

The desire for lower fertility, on the other hand, was much slower to develop. The Easterlin approach to fertility suggests that births represent a parental demand for surviving offspring, and hence mortality reduction will eventually bring about some reduction in fertility. However, it is likely that a recognition lag will be required before fertility responds to declining mortality. In addition, while mortality reduction will take place without any alteration in individuals' values and indeed (in some cases) without modification of individual behavior, the same cannot be said of fertility reduction.

To put the matter a bit differently: a declining death rate will emerge as a consequence of an improved standard of living and system of economic organization, and improvements in medical technology; however, in order to have a declining birth rate, while it helps to have declining mortality, it is also necessary that there be changing socioeconomic conditions and changing values conducive to low fertility.

Observers of the transition have often argued that considerable time is required before strongly pronatalist values change. However, declines in fertility in a number of developing countries in Asia and Latin America over the past 25-30 years suggests that the hysteresis of pronatalist values may not be as strong as implied by historical (19th century) experience in Europe and other industrialized countries.

A related issue of current interest here, particularly in sub-Saharan Africa, is the impact of economic decline on fertility behavior. Some scholars have raised the possibility of crisis-induced fertility transitions (cf., "nos enfants sont nos richesses" vs. "it costs a lot to send your kids to school"), and there is some evidence of at least short-term effects of adverse economic conditions contributing to reduced fertility, operating via delays in entry to marriage and in rates of first births.

In any case, there are two important consequences of the demographic transition that we need to consider. First is the emergence of sustained, historically unprecedented population growth.

Since the rate of natural increase of a population equals the difference between the birth rate and the death rate, the typical pattern of the transition as outlined above entails relatively rapid population growth from the onset of mortality decline until fertility falls to reach the level of mortality. Historically, the period during which this rapid growth took place often was several decades. Consequently, the resulting growth of population was, from an historical perspective, unprecedented and quite rapid.

A second important consequence of the demographic transition was a change in the age composition of populations. In particular, post-transition populations were distinctly older, on average, than pre-transition populations, with smaller proportions of young people and of dependents (young plus old).

The aging of populations did not come about because of the decline in mortality. As shown in Fig. II.A.2, mortality decline tends to be proportionately most significant at young ages, and by virtue of resulting in higher rates of survival of infants and young children mortality decline actually results in making populations somewhat younger rather than older (at least up to a mortality level corresponding to life expectancy at birth of about 60).

It is, instead, the decline in fertility that is responsible for the aging of populations (fertility decline in general has stronger effects than mortality decline). As we shall see below, declining fertility tends to reduce dependency burdens (defined as the number of dependents, young and old, relative to the population of adults of working age). At a macroeconomic level, this decline in dependency burdens, in turn, creates the possibility for increased investment aimed at enhancing economic growth, since a lower proportion of society's output will be required for current consumption needs.

The discussion of the demographic transition has, so far, emphasized commonalities as between the transition experienced in the West during the 19th century and the transition being experienced in much of the developing world at present. However, there is a key difference that is worth noting here: rapid population growth in developing countries today is quantitatively much larger than rapid population growth in 19th-century Europe.

More specifically, when 19th-century populations experiencing the demographic transition were growing most rapidly, their annual growth rates were on the order of 1-1 1/2 percent, implying doubling times of roughly 45-70 years. By contrast, in much of the Third World, growth rates of 3 percent per year and even higher (implying doubling times of 24 years or less) have been observed.

This difference in growth rates stems largely from the fact that modern public health techniques and treatment of infectious diseases, which developed slowly during the 19th century, can now be imported into low-income nations. This in turn effectively serves in large part to divorce mortality from the general standard of living.

Further, the much more rapid population growth in developing nations today alters the magnitude of urbanization and rural population growth, and essentially requires substantially more in the way of "demographic investments", defined as investments required to maintain per capita levels of consumption of goods and services.

Consequences of fertility decline for economic development: A simulation

I want to turn now to consideration of the potential impacts of declining fertility on a country's prospects for economic development. We will define economic development very narrowly, to mean simply the transformation of an economy from an agrarian base to an industrial base.

Typically, this industrialization has been accompanied by increases in the material standard of living of the population, as measured by the level of real GDP per capita. In addition to the demographic changes that are at least broadly associated with economic development, there also tend to be sharp increases in the educational attainments of the population as an economy develops, and corresponding reductions in illiteracy rates.

In examining high fertility, age distribution, and economic development, we seek to understand how high fertility and the resulting age distribution influence a traditional society's prospects for transformation from an agrarian to an industrial base, and for generating sustained improvements in the material standard of living of the population.

From a macro perspective, an economy's level of output may be regarded as dependent on the inputs of capital and labor, and on the state of technology (cf., Cobb-Douglas production function and the marginal product of labor). Over time, economic growth (growth in output) will occur in response to increases in any or all of these factors.

In order to achieve increases in the standard of living, however, it is necessary to have output growth that is greater than the rate of population growth. As a crude approximation, we can focus on growth in output per worker.

A key factor generating economic growth and the transformation to an industrial economy is the level of net investment that a society is able to maintain (net investment equals gross investment minus depreciation). Other things equal, greater investment should result in more rapid economic growth and a higher standard of living in the future.

Early work (in the late 1950s) on the relationship between population growth and economic development by Ansley Coale and Edgar Hoover was very influential. Coale and Hoover distinguished among three types of expenditures aimed at promoting devlopment: direct growth outlays, current welfare outlays, and investment welfare outlays.

Direct growth outlays consisted of expenditures seeking to enhance productive capacity and hence output immediately. This would include, for example, spending on investment in capital goods (plant and equipment) and on building of power plants.

Current welfare outlays represented expenditures that contributed to the well-being of the population at present while at the same time enhancing productive capacity in the future. This would include expenses for educational and health service supplies and for salaries of personnel in education and health care.

Investment welfare outlays consisted of investments in the well-being of the population. This would cover expenditures for construction of schools and health care facilities, as well as spending on housing.

These three types of outlays differ not only in their orientation, but also in the timing of their effects. Welfare outlays serve to improve the productivity of human resources, with effects spread out over a long period of time. By contrast, direct growth outlays constitute the expenditure likely to be most effective in the near term in generating rapid economic growth.

We noted last time that high fertility leads to relatively young populations, with high proportions of children (youth under age 15) and high dependency burdens (youth plus the elderly relative to the population of working age) as compared to low-fertility populations. This may be seen by comparing the age distributions in different parts of the world (see Table II.A.1 -- handout).

In sub-Saharan Africa, where fertility remains very high, fully 46 percent of the population is under age 15. Although only 3 percent of the population is age 65 or over, this means that the young and the old represent essentially half of the population.

At the other extreme is Western Europe, where those under age 15 are only 18 percent of the population, and not much more numerous than those age 65 and over. The sum of the young and the old, though, is only a third of the population. Clearly, there is a substantially higher proportion of the population that is dependent in Africa as compared to Western Europe (other countries are in between, as shown in Table II.A.1).

To explore the consequences of high fertility, Coale some years ago performed a simulation. He began by considering two populations that, at the outset, had the same high level of fertility and a moderate level of mortality. The key difference in his simulation was that one population (A) was assumed to continue with its high fertility, while the other population (B) was assumed to experience a (linear) 50 percent reduction in its total fertility rate over the course of a 25-year period, after which fertility would remain constant. Both populations were assumed to experience slow and steady improvements in life expectancy for a period of 30 years.

Coale's Figure 1 shows the consequences of these alternative assumptions for three key variables: the dependency burden, defined as the total number of persons divided by the number aged 15-64; the rate of growth of the population aged 15-64, which he considers to be the population of labor force age; and the total number of people aged 15-64 (the size of the population of labor force age).

As can be seen in the figure, reduced fertility begins immediately to affect the age composition of the population and hence the dependency burden. The sustained high fertility population experiences a slight increase in the dependency burden over time, as a consequence of the reduction in mortality, while the population in which fertility falls experiences a decline in the dependency burden, accelerating after year 10 and reaching a minimum level after about 30 years.

The rate of growth of the population aged 15-64 is identical for the two projections during the first 15 years, rising slightly due to the declining mortality. Once the impact of differential fertility is felt, the growth rate rises slowly for about 25-30 years to well in excess of three percent per year in projection A, while falling over a 50-year period (the medium run for the demographer) to a level of a little more than one percent per year in projection B.

Because of the fact that the impact of this growth on the size of the overall population aged 15-64 is only at the margin, it takes a long time for the differences between the two projections to be evident with respect to overall population. Hence, the difference is still quite small after 40 years, and it takes about 60 years before the high-fertility projection has a population aged 15-64 that is twice the size of that in the declining-fertility projection.

Having considered the demographic consequences of fertility reduction, we can now turn to an examination of the economic consequences. In the demographic short run, the declining-fertility population will experience a lower dependency burden while at the same time having a labor force (population of labor force age) of comparable size to that in the population with sustained high fertility.

Coale argued that, as a consequence, the population with reduced fertility would find itself more readily able to divert resources away from current consumption and towards investments. Further, both current and investment welfare outlays would presumably be less, since reduced fertility will have a strong impact on the aggregate demand for housing, education, and health services (cf., high health risk population). Hence, both the amount and the composition of investment are likely to be more favorable for promoting growth in projection B.

This dependency burden advantage is maintained in the medium to long run. Consider now the consequences of differential growth of the population aged 15-64, which emerges in the medium term. It was suggested earlier that improvement in the level of output per worker would be facilitated by investment that increases the amount of capital per worker. This is called capital deepening.

It should be clear that capital deepening will require more effort when the potential labor force is growing rapidly than when it is growing slowly. For any given level of investment, when labor force growth is rapid more of that investment will be required simply to maintain the existing amount of capital per worker (capital widening) than when growth is slow. What would go to maintaining the capital-labor ratio in the high-fertility scenario can be used to increase that ratio with low fertility. This can readily be seen in a quantitative illustration (to be discussed in class).

In any case, now in addition to the dependency burden advantage of the short term, we have an argument that the slow-growth projection should be more readily able to increase capital per worker and hence the standard of living as compared to the high-fertility projection. To put it differently, projection B should find it easier to invest than projection A, and even for comparable levels of investment B's growth in output per worker will be higher. Consideration of real-world problems of high unemployment and underemployment in developing countries, which are likely to be exacerbated by rapid population growth, further emphasizes the disadvantages of sustained high fertility.

In the long term, differences in the density of population relative to land and other resources will emerge. While it is clearly possible in some cases to maintain very high population densities, as in Hong Kong and Singapore, for example, it is also apparent from countries like Bangladesh, Haiti, and Rwanda (among others) that population pressure on the land and other resources can be a real problem.

Overall, then, Coale's analysis appears to provide a strong case for substantial economic benefits to reducing fertility and population growth. As noted earlier, his work was very influential in conditioning people's thinking about population growth and economic development during the 1960s and 1970s.

The prevailing view at that time, based largely on work by Coale and by Coale and Hoover, was that rapid population growth was likely to have serious adverse consequences on a country's prospects for economic development. This perspective led to a second meaning for the phrase "demographic investments": in this context, such investments were defined as expenditures seeking to lower the rate of increase of population and hence facilitate the transition to an industrialized economy (cf., Tunisian cost of children consultancy).

It turns out, however, that the real-world evidence of the relationship between population growth and economic development has not been nearly so compelling as implied by Coale's theoretical argument. In particular, cross-country correlations between population growth and economic growth covering the 1960s and 1970s typically failed to find a statistically significant negative relationship. This suggests that economic growth is driven by a lot more than simply population growth.

As Dasgupta notes, the absence of a strong negative correlation between population growth and economic growth led to what has been called a "revisionist interpretation" of the evidence. This is best exemplified in the 1986 National Research Council publication, Population growth and economic development: Policy questions. The NRC report was focused around nine specific questions:

1. Will slower population growth increase the growth rate of per capita income through increasing per capita availability of exhaustible resources?

2. Will slower population growth increase the growth rate of per capita income through increasing per capita availability of renewable resources?

3. Will slower population growth alleviate pollution and the degradation of the natural environment?

4. Will slower population growth lead to more capital per worker, thereby increasing per worker output and consumption?

5. Do lower population densities lead to lower per capita incomes via a reduced stimulus to technological innovation and reduced exploitation of economies of scale in production and infrastructure?

6. Will slower population growth increase per capita levels of schooling and health?

7. Will slower population growth decrease the degree of inequality in the distribution of income?

8. Will slower population growth facilitate the absorption of workers into the modern economic sector and alleviate problems of urban growth?

9. Can a couple's fertility behavior impose costs on society at large?

While the report's review of these questions did find some evidence of adverse effects of rapid population growth, for the most part these were deemed to be modest, and at least partially offset by benefits to rapid population growth (cf., question 5 and Boserup's induced technological change argument).

More recently, there is evidence for the 1980s of emergence of the sort of strong negative cross-country correlations between population growth and economic growth that Coale's approach would lead us to expect (Allen Kelley and Robert Schmidt, "Population and Income Change: Recent Evidence," World Bank Discussion Paper 249). Further, the authors conclude that over the entire period from 1960, both the direction and the size of the correlations vary by the level of economic development. The correlation is most likely to be negative in relatively poor countries, and positive in relatively wealthy countries.

In any case, Dasgupta clearly is not happy with the "revisionist interpretation" of the evidence, and he has specific concerns about environmental degradation and its relationship to population growth. This will be a focus of our next section. However, there are several other issues raised by Dasgupta that merit at least brief attention.

In his overview of birth control, Dasgupta emphasizes that "it is parental demand for children rather than an unmet need for contraceptives that in large measure explains reproductive behavior" (pp. 1889-90). (Note concept and measurement of unmet need.) This is an assessment of a debate concerning the relative importance of demand for children versus supply of contraceptives as factors critical to fertility decline. (Consider how the Easterlin approach would address this question.)

In the course of his discussion, Dasgupta cites a quote that out of context sounds peculiar: "contraceptives are the best contraceptive." To appreciate this quote, it is useful to know some background information that ties into what we've looked at above. Since the 1960s, reflecting the prevailing wisdom that rapid population growth was bad for developing countries, bilateral and multilateral development assistance have often sought to promote family planning programs as a means of contributing to prospects for improved economic growth (i.e., as "demographic investments" -- meaning #2).

In many cases, however, these efforts ran into resistance from those who felt that development assistance should be spent on promoting industrialization directly. In this view, borrowing from the notion of the demographic transition, it was argued that once industrialization emerged fertility decline would follow. The slogan of those with this perspective became "development is the best contraceptive."

In essence, this view is one in which development is seen as influencing fertility, whereas the Coale view is that high fertility can impede development. The possibility of these two aspects being jointly determined (reciprocal relationships) seems like the most realistic perspective.

Indeed, Dasgupta's view, as expressed at the outset of his article, is one in which "population growth, poverty, and the state of the local environmental resource base [are] interconnected... Our approach... regards none of the three factors to be the prior cause of the other two; rather, it sees each as influencing the others and in turn being influenced by them" (p. 1879).

A final issue of note here is the focus of question 9 of the NRC -- essentially, the question of what Dasgupta refers to as "reproductive externalities" or the "demographic free-rider problem." If indeed a couple's reproductive behavior can impose costs on society at large (over and above the costs to the parents themselves), this creates a divergence between the private costs of children and the social costs of children.

This is a classic example of a negative externality, and as with all negative externalities, its existence means that in an unregulated market there will be too much production (of children) by private decision makers (couples). This is referred to as a case of market failure, meaning that the unregulated market has failed to produce an optimal output.

The externality argument is evident from Figure II.A.3. As shown in the figure, optimal production from the social point of view (Qs*) is exceeded by optimal production from the private point of view (Qp*). To get private decision makers to produce at a socially optimal level, the solution is to internalize the externality. In this case, that means taxing parents so that they bear the full cost of their children. This solution yields higher utility than a quantity restriction (like a one-child or two-child policy with harsh penalties for failure to comply).

The analysis of the consequences of "reproductive externalities" is fairly straightforward. It is worth noting what the sources of such externalities might be. The NRC report identifies several, including externalities inherent in common-property resources (air and water, some forests and fisheries, and some agricultural land). If population growth contributes to long-run resource degradation, the costs fall on society as a whole, not just on those making childbearing decisions.

Provision of public goods, including transportation infrastructure and sanitation/water/electric infrastructure, is influenced by population growth, with rapid growth often resulting in congestion and excessive demands on existing infrastructure. Because these congestion costs must be borne by the society as a whole, rapid population growth may have net negative external effects on public welfare.

An additional source of negative externalities discussed in the NRC report is with respect to publicly-subsidized education. At least in the short run, such education disproportionately benefits families with children, and a couple's decision to have an extra child imposes costs on all taxpayers.

Finally, there are also intrafamily externalities and pecuniary externalities associated with high fertility. The arrival of later-born siblings can adversely affect the opportunities available to earlier siblings, by diminishing the resources available per child. Pecuniary externalities refers to the notion that families that are well-endowed with land or capital are advantaged by other families' high fertility, while those that rely primarily on labor income are disadvantaged.

Some of the externalities just discussed might not be considered as "true" externalities. Overall, however, the presence of negative externalities provides a rationale for government intervention aimed at reducing fertility. Subsidized family planning services that seek to assist couples in achieving and not exceeding their desired levels of fertility provide an attractive vehicle for efforts to lower fertility.