Cette page existe uniquement en anglais
Home > Activities > Committees >Population and Poverty > Exploratory Mission Report
By Nora Lustig (co-ordinator), Nancy Birdsall and Monica Das Gupta
December 4, 1998 (2)
The purpose of this report is to highlight the state of current knowledge and research gaps in the area of population and poverty. It is our view that IUSSP is well placed to make a contribution in advancing the knowledge in this area.
The following are some of our ideas on how the IUSSP could help achieve this end:
The Demographic Setting
Declining fertility, now underway to one degree or another in all regions of the world, is one of the most important demographic phenomena in the world today. This decline will result in substantially changed age structures and distribution, with gradually reduced proportions of the population under age 15 and enlarged proportions over age 65. As countries move through the demographic transition, they first face a period of increasing child-dependency ratios, then of decreasing child and old age dependency ratios as a larger proportion of the population moves through the working ages, and finally of increasing old-age dependency ratios.
The effect of fertility decline in the second intermediate stage is a one-time "demographic bonus" or "window of opportunity" --a period during which the ratio of the working age population to the dependent population is unusually high. After a country has passed through this period, it returns to a stable dependency ratio at new lower levels of both fertility and mortality.
The timing and pace of fertility decline affect both the size and the duration of the "window of opportunity." The faster the decline, the quicker a country will reach the window, but the shorter the period the window will remain open. Depending on the determinants of growth, this "window of opportunity" can materialise in transitory or permanently higher growth rates of output per worker (more on this point below). Everything else equal, the level of output per worker should rise during the demographic transition (at least during the phase of lower dependency ratios).
In essence, all of the developing world has, is or will be passing through the lower dependency ratios phase in the latter 20th and early 21st century. The decline in dependency ratios occurred earliest in East Asia, followed thereafter by Latin America and will happen latest in Africa (3). The Middle East and South Asia are at intermediate points between Latin America and Africa. The challenge for policy makers in the developing world is to create the conditions for the "window of opportunity" to materialise in lower poverty rates. Subsequently, the challenge will be to devise mechanisms to cope with an ageing population as large portions of the working age population in developing countries are not part of formal social insurance arrangements and informal mechanisms of family support get eroded with modernisation.
Population Changes and Poverty: the Economic Growth Effect
The demographic transition is characterised by sharp falls in (infant) mortality followed by falls in fertility at a later date. These changes result in: i. an initial increase and subsequently a fall in the rate of growth of population; ii. an initial increase in the ratio of children to working age adults followed subsequently by a fall; iii. a significant increase in the share of older adults to working age adults during the last stage of the demographic transition. These aspects of the transition affect poverty through three mechanisms: i. the effect of population changes on the rate of growth of output (income or consumption) per person (the economic growth effect); ii. by changing the distribution of income (the distribution effect); and, iii. by altering the well-being or capabilities of the poor at a given level of income (the conversion effect). In this section, we will discuss the connection between population changes and economic growth.
It is an established fact that one of the most powerful ways to reduce poverty is economic growth. What is the relationship between population changes and economic growth? In the standard neo-classical growth model a decrease in the population growth rate yields a higher equilibrium output per worker, a transitory increase in the growth rate of output per worker, and no change in the rate of growth of output per worker once equilibrium is established. The same results occur with a higher saving rate. Incorporating the life-cycle saving model in the neo-classical growth model introduces yet another connection between population changes and economic growth. If life-cycle saving is dominated by pension motives, slower population growth leads to an older population and lower aggregate savings. In contrast, if slower population growth reduces the burden of supporting children, it may lead to increased saving.
The prediction from standard growth theory is that population and economic growth should be inversely related during the transition, but changes in population growth should not affect the long-run growth rate of output per worker. What does the empirical evidence show? up until this decade researchers failed to find any significant relationship between population growth and output growth (4). The interpretation was not that population did not have an effect on economic growth, but that the effects were complex and tended to offset one another. Even in the cases where a negative relationship was found, it varied considerably by country and over time, and the impact of population was small by comparison with other factors affecting output growth. Hence, the traditionalist view that population growth was bad for economic growth was not strongly supported by empirical evidence.
Recent research, however, arrives at a different conclusion (5). After assessing, replicating and extending a number of studies, Kelley and Schmidt conclude that there appears to be a sizeable inverse relationship between population growth and economic growth over the period 1960-1995 (6). Specifically, declining population growth, fertility and mortality have had a sizeable positive impact on economic growth. Changing age distributions and rising population density and size also had a positive impact, although the contribution of size is small. Of all the demographic variables, improvements in life expectancy constitute the largest single impact on changes in output growth. Furthermore, the authors also find some evidence that the impact of demographic change is not solely transitional but that it also affects long-run economic growth. This result would imply that mechanisms other than those captured by the standard neo-classical growth model are at work in the real world. One such mechanism could be the relationship between population growth and resource degradation.
These results are good news for the many countries that are now passing through the demographic transition. If true, they predict a boost to economic growth and an opportunity to accelerate the pace of poverty reduction. However, as emphasised by the authors, these results should be taken as preliminary. The Kelley and Schmidt study is careful in incorporating the dynamics of demographic change in their analysis (7). The impact on economic growth of a new birth varies over a lifetime: usually negative during the child rearing years, then positive during the labour force years, and possibly negative during retirement. Modelling of demography must account for the patterns of birth and death rate changes over time. Some of the early "no correlation" findings can possibly be due to the fact that no care was taken in decomposing population change in its various factors. But, the question still remains whether the results obtained by Kelley and Schmidt (and others) are robust with respect to different statistical specifications and modelling variants. This is an area where further research is welcome.
The analysis based on aggregate models is useful to identify empirical regularities and to disentangle "on the surface" causality. However, in order to understand what factors lie behind the results obtained in aggregate modelling exercises, the behaviour of families and individuals, producers and consumers and policymakers must be understood as well. For example, what is the relationship between demographic change and savings? Do families behave according to the life-cycle model implicit in many of the aggregate models? In countries where this question has been analysed, there appears to be some evidence that changes in age structure produce significant changes in the savings rate (8). What is the impact of demographic changes on the form and pace of technical change? Does high population growth stimulate innovation or cause natural resource degradation and produce the "tragedy of commons?" Some authors argue that available evidence points to the conclusion that demographic change may have a negative impact on natural resource availability, but that the effect is successfully mediated by collective action (9). However, much more research is needed in this area.
It is clear from existing studies that fertility decline is not only a possible cause of more rapid economic growth, but is an outcome of factors associated with economic growth, including increased education, better functioning markets and so forth. Recent analyses take into account this possible reverse causality in a better way, which if not handled correctly, can lead to overestimating the effect of lower fertility on growth. Improved econometric methods make the recent results more credible than before, yet the possibility of reverse causality cannot be fully eliminated so the risk of some overstatement remains. On the other hand, the very likelihood of reverse causality is linked to our intuition that there is probably a mutually reinforcing effect over time between fertility decline and economic performance as fertility decline augments growth, which in turn contributes to further fertility decline. As such, it is possible that the long-run effect of an initial fertility decline and the initial boost to growth it may provide is understated. Analysis at the micro level should shed more light on the quantitative importance of the various determinants of fertility and mortality, and whether these determinants are exogenous or endogenous.
One of the reasons why many of these questions remain unanswered for a very large number of countries, particularly those in the developing world, is the lack of panel data. Hence, the importance of promoting the generation of longitudinal surveys should be an essential part of the research agenda.
There are some additional issues that need to be considered when assessing the implications of the "window of opportunity" for specific countries. First, what is the role of the initial conditions: that is, are poorer countries better situated to benefit? Second, the positive effect of the demographic changes associated with the demographic transition probably depends strongly on the quality of economic policy that accompanies the transition. The East Asian countries were able to exploit the opportunity presented by the "demographic bonus" because of a combination of reinforcing policies that ensured healthy returns to physical and human capital and to higher levels of participation in the labour force. In contrast, rapid demographic change in Latin America, including rapid fertility decline in the last two decades, has not been so clearly associated with improved economic performance. Fertility decline and other demographic changes may encourage economic growth but are far from sufficient to guarantee growth. A sound policy regime is essential. What policies have to be in place in order to exploit the "window of opportunity" provided by the demographic transition? Are those policies being implemented in the developing world? Are they part of the policy recommendations put forward by the international community? What are the most effective ways to entice countries to adopt the right policies?
Furthermore, has it been shown that governments divert resources from productive investment in response to population pressures? Will lower levels of population growth free resources to invest in physical and human capital of better quality more permanently? Or, will the higher "savings" obtained during the low-dependency ratio phase be used up subsequently by an ageing population? These questions are particularly relevant to assess the impact of demographic changes on poverty.
Population Changes and Poverty: the Distribution and Conversion Effects
Is high fertility a cause of poverty? In order to answer this question, economists have looked at how fertility changes might affect poverty through the growth effect, the distribution effect and the conversion effect. We have discussed the growth effect in the previous section. We now turn to the other two.
One established fact is that the poor enjoy falling infant mortality and rising prospects for female education later than the non-poor in the same country, and thus rationally delay their fertility transition. Therefore, poverty reduction from the fertility transition is also lagged. Also, the effects of the demographic transition on reductions in poverty are different at different stages of the transition: harmful to poverty reduction in the early stages as population growth accelerates due primarily to mortality decline, and helpful in the later stages as fertility declines and population growth slows. As fertility decline spreads to poor households, the benefits of fertility decline for reducing poverty increase even more.
It follows that during the early stages of the demographic transition, income differentials between poor and non-poor households may in fact become worse. But as the transition extends to all groups in the society and fertility declines in poor households, the poverty-reducing and inequality-reducing effects increase. As the dependency ratio and the cost of childbearing declines, more income is available for consumption and savings, particularly where women enter the labour force and contribute to increased family incomes.
The Distribution Effect
The impact of the demographic transition on the distribution of income occurs through the acquisition effect and the dependency effect of changes in fertility. The dependency effect measures the contribution of the higher overall fertility of the poor on their consumption per capita. That is, it refers to the worsening of the income distribution if the extra births are concentrated in the poorer households. The effect also will be present if the fertility transition for the poor takes place at a later date, which is an established fact. If the dependency effect is zero, then higher fertility can worsen the distribution of incomeand, thus, raise povertyonly by worsening the distribution of income or consumption per non-dependant. All the impact in such a case would be a pure acquisition effect.
Eastwood and Lipton use data based on eighteen Poverty Assessments of the World Bank that have separate dependency ratios for poor and non-poor households, and find that there was only a slight indication that the dependency effect was responsible for the distribution-channel linkage of higher fertility and greater poverty (10). However, the authors note that a problem with the data is that one cannot distinguish between children and adults over sixty-five in the dependants group. Because the proportion of older dependants is lower for developing countries, and there is a higher proportion of them in non-poor households, the measured impact is probably underestimating the impact of the dependency effect. This is an area where new research based on better information on the age distribution of household members would be very useful.
The acquisition effect measures the impact on the distribution of income resulting from the reduction in the relative ability or willingness of poorer households to acquire a given level of total household consumption as a result of higher fertility. It operates through four channels: marginal child costs; labour supply response within families; factor returns; and, savings (11). The marginal child cost measures the cost associated with an extra child in the household resulting from the extra burden of child care (assuming total income does not change) for example. If this is higher for poorer households that for the non-poor, this will result in a higher acquisition effect. If higher fertility causes more adult members to work or work harder, this may lower the acquisition effect.
The impact on factor returns is that labour in particular works its way through the labour and food markets. During the first stage of the fertility transition, the impact on the supply of unskilled labour is indeterminate given that the two previously described effects work in opposite directions; the dominant effect will depend on the relative size of each effect individually. However, as the extra children become adults and enter the work force, labour supply will rise and there will be a downward pressure on real wages. This effect may be exacerbated since a larger population will shift the demand for food, resulting in higher prices. Since the poor spend much larger shares of their income on food, their command over consumption suffers disproportionately.
Finally, higher fertility may affect the consumption of poor households disproportionately because they cannot absorb the costs of extra children by drawing on savings as the more asset-rich households can.
The impact of the demographic transition on the distribution of income operating through changes in factor returns is best explored at the macro-level while the other three mechanisms, as well as the dependency and conversion effects, require household-level data.
A review of the existing empirical literature indicates that there is little evidence based on economy-wide data that population growth or high fertility have a statistically significant impact on the incidence of poverty. For example, Ahlburg concludes: "Little direct evidence on the impact of population growth exists." (12) McNicoll finds little support in the extensive research literature for an association between birth rates and poverty (13). The united Nations Population Fund concluded that: "[r]esearch has not established a strong causal link running from high fertility to poverty." (14) Pritchett argues that there is no solid evidence showing population growth to be "a cause or even an exacerbating condition of poverty." (15)
A recent important exception to these assessments is the study by Eastwood and Lipton (16). Based on analysis of economic and demographic data for forty-five developing countries, the authors find that high fertility increases absolute levels of poverty both by retarding economic growth (thus slowing growth-induced poverty reduction) and by skewing the distribution of income against the poor. They estimate that had the average country reduced its birth rate by five per thousand throughout the 1980s (as did in fact many countries), the average country poverty incidence of 18.9 percent in the mid-1980s would have been reduced to 12.6 percent between 1990 and 1995. The statistical work suggests that about half the estimated decline in poverty over the period in the countries studied can be attributed to increases in economic growth and half to changes in the distribution of income in favour of the poor. The results reported by Lipton also show that the poorer the country and the higher its initial level of fertility, the greater the effect of declining fertility on a decline in absolute poverty.
Given that it reaches conclusions opposite to most of the other work, the Eastwood and Lipton analysis needs to be replicated in various forms to test its robustness. More work at the micro level needs to be done in order to fully understand the mechanismsfor example, how changes in wage rates affect labour force and fertility decisions of the poor--through which the observed macro outcomes occur.
Moreover, even if new empirical evidence indicates that fertility and poverty may be inversely related, this does not immediately warrant policy intervention. Policy intervention is justified if higher fertility in poor households raises the poverty levels of other households through the "Malthusian connection" in labour and product markets. Second, even in the absence of externalities, policy intervention is justified if the opportunity set is restricted: for example, if access to information on reproductive health is lacking or if there are market failures in the financial sector.
There is little debate that poverty and large family size go hand in hand. There are hundreds of empirical studies confirming that in todays developing countries, larger households have higher poverty incidence; and, that among poor households, those that have more children systematically invest less in childrens education and health, and we also see worse health outcomes for mothers associated with pregnancy. But economists have long cautioned that the association does not in itself indicate causality. High fertility in poor families may reflect parents sensible decisions to trade off current consumption for greater future family income when children begin work, or for greater old age security. Or it may simply reflect parents decisions to enjoy children rather than other forms of consumption. The fact that large families tend to have lower incomes should not be construed as meaning that they either are, or that they regard themselves as being, objectively "worse off."
On the other hand, studies and analysis over the last decade raise several countervailing arguments, increasingly shifting the burden of proof to those who argue that high fertility is the result of choice made by poor households and should be assumed to reflect optimal levels of welfare for the family. Those who argue that at least some fertility among the poor may not be optimal for all members of the family base their arguments on the following: In the face of poorly developed capital and other markets, poor households cannot borrow against future earnings of better-educated children, and literally cannot afford to choose few children, even recognising that their fewer children might face better prospects. (17)
Second, given the poor state of markets for information, poor households may lack information on the changing probability of infant mortality, on increasing returns to schooling, and on improved financial markets as a mechanism for old-age security. That is, poor households may lack information on a variety of changing conditions that would lead them to choose fewer children. They also may lack the capacity to adjust their intergenerational allocations during times of rapid societal change. (18)
Third, men may dominate in the choice of number of children, while not internalising the costs a kind of negative externality within the family that is assumed away in traditional unitary household utility functions. Cultural and institutional factors may lead to differing interests among household members and unequal capacity to participate in household decisions, particularly for women. In many low-income settings, there may be gender imbalances in the capacity to decide whom and when to marry; who in the household gets access to health care and education; when and what kind of contraception to use; and the power to negotiate safe sex when the risk of sexually transmitted diseases and HIV/AIDS infection is high. There is increasing evidence that the autonomy women have in making reproductive decisions varies considerably in different cultural contexts.
Finally, there is ample evidence of at least some unwanted pregnancies based on more than two decades of surveys, and also of increasing high incidences of abortion, including among the poor -- even where abortion is illegal and dangerous.
In addition to the fact that the opportunity set for the poor may be restricted, there are other reasons why one might be concerned about the effect of poverty on large families even if no spill-overs or externalities are present. In particular, one may be concerned for the negative impact that higher fertility may have on education, health, and the nutrition of household members, particularly children. According to Lipton and Eastwood, "[A]lmost certainly, high fertility reduced conversion efficiency for the poor and near-poor, i.e., sib crowding effects outweigh economies of scale in consumption for them." (19)
The causal link from high fertility to worse and less education for children in larger households appears in numerous studies (20). In particular, it has been found that high fertility and large households particularly damage the educational prospects for girls (21). Likewise, it has been found that large households provide less care per child, less access to health care, and show more gender discrimination in food distribution (22). There is much greater risk of under-nutrition in large households. Large households are associated with bad outcomes not only because of female (mother) exhaustion or sib crowding but because larger households tend to be more authoritarian, less altruistic and more hierarchical by age and gender (23).
Nevertheless, these results are not found everywhere. For example, based on a review of thirty-six econometric studies, Kelley concludes that the link between educational outcomes (enrolments and attainments) and household size is not pervasive. There are a good number of countries displaying positive, negative and statistically insignificant coefficients. Whenever the relationship is statistically significant, however, a negative correlation is more likely. Nevertheless, the impact is low.
Even if the evidence is not fully conclusive, the mere fact that high societal fertility and in particular high household fertility may exacerbate poverty, especially for children, combined with evidence that some high fertility is unwanted, provide justification for PUBLIC financing of non-coercive programs that would help reduce fertility.
Summing up: A Research Agenda for the Future
The following is a list of broad questions that the research community should address in order to increase the scientific knowledge on the relationship between population and poverty. These are some of the main questions extrapolated from the results and debates underlying current work on the subject.
Population Change and Economic Growth: Is the new evidence that there is an inverse relationship between population growthits various components--and the growth rate of output per capita robust? How do initial conditions affect this relationship? What demographic variables are truly exogenous? Are developing countries taking advantage of the window of opportunity provided by the demographic transition?
Population Change and Income Distribution: Is the new evidence of an inverse relationship between population growth and the incidence (and intensity) of poverty robust? Does it override previous findings that revealed there was no statistically significant relationship? Through what mechanisms does the relationship operate? Is the effect the result of the spill-over of higher fertility on factorlabour in particularreturns? Which labour supply effect at the household level dominates?
Information, Markets and Choice: Do poor households have access to adequate information on reproductive health and demographic trends? Are there failures in the credit markets that limit the poor households ability to smooth consumption over their life cycle? Are women in the households given equal rights in decisions regarding marriage, fertility, education and labour force participation? What factors come in between access to adequate information, credit markets and equal rights within the household? Can the obstacles be changed by policy intervention? What policies are best from the cost-effectiveness, efficiency and social point of view?
Family Size and Outcomes: What is the relationship between family size and the educational attainments, health and nutrition of their members when longitudinal data is used? What is the effect of family size on gender relations and on the situation of women? Do the effects differ under different regimes of gender treatment?
Migration and Poverty: What is the impact of international migration on poverty both in source and recipient countries? Does the option to migrate affect fertility decisions? For how long are the pressures to migrate to richer countries likely to continue at high levels? What policies can be put in place to manage those pressures?
Natural Resource use: What are the effects of rapid population growth on natural resource use and sustainability, particularly in conditions of poverty?
Collective Action, Policies and Institutions: under what conditions are social groups more likely to accommodate rapid population change with improved collective action and improved policies and institutions?