Posted January 1999
These complications explain the hesitant way in which the link between population growth and poverty has been summarized in the recent research literature. The dominant tone is caution and uncertainty in the face of a complex relationship. There is little in this literature to support a strong poverty-based rationale for reducing fertility. The most that the 1992 UNFPA Consultative Meeting of Economists could conclude on the matter was that "The causal links between population growth and absolute deprivation are not well understood.... Research has not established a strong causal link running from high fertility to poverty" (United Nations Population Fund 1993:50-51). World Bank economist Lant Pritchett (1997:51) is brusquer: solid evidence that population growth is "a cause or even an exacerbating condition of poverty" cannot be given "because there is none".
Nevertheless, while acknowledging the complexity, it is likely that most social scientists would accept that there is a connection between population growth and poverty. In particular, achieving lower fertility, at least in some circumstances (probably including those of most of the world's poorest countries), is likely to help alleviate poverty. The reverse relationship-of poverty alleviation on fertility reduction-is also widely accepted. Hence, for instance, the Programme of Action of the 1994 International Conference on Population and Development in its section entitled "Population, sustained economic growth and poverty" (3.10-3.22) asserts that efforts to slow population growth and to reduce poverty are mutually reinforcing.
We therefore have a case in which common sense views about a particular consequence of demographic change rest on an inconclusive body of research. This is unfortunately a familiar situation. It underlies the so-called revisionist arguments that hold population growth to be a comparatively marginal factor in the development process. The 1986 report on population and economic growth of the U.S. National Research Council was summarized by its co-chair as "support [ing] the conclusion that population growth has little or no negative effect on economic development" (Johnson 1994: 524).
An intriguing development in the 1990s has been the second thoughts about the effects of population growth that have emerged from a close look at the strong economic performance of East Asia in the 1970s and 1980s. The early fertility decline and consequent fall in youth dependency in most of these countries is now seen by some economists, echoing the classic Coale-Hoover result, as contributing to their surge in investment (see Higgins and Williamson 1997). Significantly, East Asian poverty rates also fell sharply over this period. However, the case is still circumstantial: there are various competing accounts of the "East Asian miracle"-indeed, since the "meltdown" of some of these same economies in 1997, pushing large numbers of people back into poverty, many would deny that there even was a miracle.
The prima facie empirical case for the unimportance of population to economic change has come from cross-country analysis. Scatter plots of countries on axes representing population growth rates versus per capita GNP or more refined indexes of income poverty are famously unpersuasive. Figure 1 is a typical example: high-population-growth countries are nearly all poor, but the association is anything but tight. (It is certainly negative, but that is hardly surprising given the overall demographic contrast between rich and poor countries.) The same opaque relationships can be reproduced using numerous pairs of demographic and economic variables (see McNicoll 1984: 211).
Both mortality and fertility contribute to the confused picture (migration does too, in the case of some outliers). There are steep income effects on mortality at low incomes-see Preston (1980)-and few exceptions to low fertility at high incomes, but wide variations in each over a broad middle income range. Poverty measures with substantial demographic content, like the Human Poverty Index, are by definition linked to mortality. Considerable noise is added to this and other similar bivariate relationships by the variation among countries in political and economic systems and social institutions-which together affect both economic performance and capacities to influence demographic outcomes.
Multivariate analysis does little to clarify the matter. Theoretical support can be found for various causal paths between population and poverty, in both directions. Moreover, the underlying models must assume cross-country regularities in development patterns that are belied by the actual diversity among countries. Unsurprisingly, no sharp picture emerges. The careful aggregative modeling by Rodgers (1984) concludes on the overall slightness of the adverse population growth effect. In a later piece he writes: "there is little empirical evidence to support the proposition that population growth is a major constraint on the alleviation of poverty... Population growth is both a cause and a consequence of poverty, and unraveling the pattern of causation may be an almost hopeless task" (Rodgers 1989: 9). Ahlburg (1996) finds no association of population growth with changes in poverty. Virtually the only exception in this literature is a study by Michael Lipton (1997), which reports a "large" positive effect of fertility on subsequent poverty in a 28-country analysis. Timmer (1994: 262) comments on population-poverty research that "because of path-dependency and important dynamic feedback effects, cross-sectional evidence can offer only limited insights."
Is aggregative analysis then of no help in illuminating the population-poverty relationship? That would be an overstatement, on two grounds. First, an explanatory calculus that finds little to be explained may simply be unsuited to the analytical task. Second, a murky cross-sectional picture does not prove the absence of a relationship. We may never obtain simple conclusions on population and poverty, but it does not follow that the relationship is unimportant. Complexity does not imply inconsequence.
If cross-country studies tell us little, we are thrown back on single country analysis or on cautious and selective use of country comparisons. The difficulties of dealing with counterfactuals-investigating what would have happened had conditions been different in specified ways-then have to be faced, but that task is familiar to historians and differs mainly in degree from the one routinely encountered in doing sensitivity analysis.
The second and third routes take account of possible changes in the environmental and social setting. Many population-related environmental changes have effects on poverty. We can write:
D Population --> D Environment --> D Poverty
(where D signifies "change in"). A somewhat distinct triplet would link population and poverty through effects on social organization, though this one as yet has been far less researched than the other:
D Population --> D Social organization --> D Poverty.
Poverty, in turn, would often have consequences for both environment and the social order, and environmental change and social (in)stability are themselves sometimes connected. These additional relationships rapidly complicate the picture. Moreover, the D notation leaves ambiguous the matter of contingency: it does not specify what else is being held constant. Thus, to identify any relationship working through environmental change involves making assumptions about social organization-at a minimum, a ceteris paribus stipulation. Analogous assumptions on the environment are required in tracing causal paths through social arrangements.
Poverty links to environmental change and to social instability are discussed separately below.
There is of course no simple connection between rural poverty and population growth. There are often conflicting Malthusian and Boserupian interpretations of past trends, and possibilities for "tragedies of the commons". Moreover, it is not just population numbers that potentially matter: poverty outcomes may be influenced, for example, by disproportionate out-migration from the most economically active age groups. Consider the range of responses to population growth in a subsistence agrarian economy. They might be categorized as follows:
The list is not exhaustive but covers the principal possibilities. It mixes "effective" responses, in which agriculture proves a dynamic growth sector and a potential source of investment for the rest of the economy, and "ineffective" responses that yield poverty, social disorder, and environmental degradation. Commonly, responses would encompass elements of several of these categories, the balance varying over different social groups. The actual outcome in a particular case depends on resource endowments, inherited institutional and cultural practices, the course of political development, competency in public administration and development planning, and some measure of chance. There would be usually be winners and losers but a rough consensus could probably be reached on whether rural poverty had risen or fallen. Where there is a net income gain, it is at least theoretically possible for the losers to be compensated so that no one is worse off, or for the same to happen through trickle-down effects.
Consider some instances on the ineffective side. Population growth in 19th century Ireland played an underlying role in creating that region's extreme vulnerability to crop failure, with its outcome at mid-century of starvation and mass emigration. Nepal and the Philippines in recent decades offer pictures of population pressures leading to ecologically unsound spread of upland cultivation, producing severe environmental degradation and out-migration-to the Terai and to Mindanao, respectively (see Eckholm 1976, Seddon 1989, and Ooi 1988). In the lowland tropics, even where erosion is not a problem, the soil quality of much newly-deforested land falls off sharply after a few seasons, offering little prospect for subsequent intensification-indeed, often ending in abandonment. The case of Brazil is most familiar, with rapid population growth in Amazonia (the migrants coming from elsewhere in the rural sector) producing these outcomes. Bangladesh has witnessed a long process of impoverishment through plot subdivision and extension of cultivation on to hazardous delta islands, as rural population densities have risen to extreme levels. Sub-Saharan Africa presents notable instances of the breakdown of rural agricultural systems and subsequent impoverishment, but for the most part these do not belong in this category: the environmental outcomes have been byproducts of destruction of social infrastructure, leading to endemic insecurity, population displacement, and heightened mortality.
It is important to recognize that environmental destruction also occurs for reasons that have little or nothing to do with population. For example, ecological degradation in Nepal and elsewhere in the Himalayas can be explained in some part without invoking human agency. The deep lowland soils of the northern subcontinent come from eons of heavy but "natural" erosion, leading some alpine ecologists to dismiss "ecocatastrophe" accounts such as Eckholm's (see Ives and Messerli 1989). The apparently irreversible environmental degradation of the agricultural region surrounding the Aral Sea-and the shrinkage of the sea itself-while it has clearly led to widespread rural impoverishment, is almost entirely attributable to drastically ill-conceived planning decisions (Kasperson et al. 1995). Again, natural increase of population had little to do with the Brazilian experience of falling incomes of small farmers as landholdings were consolidated and operations commercialized in the 1970s. Both the frontier migration and the expulsion of labour out of the agricultural sector altogether were consequences of this severe contraction in labour demand (see Merrick and Berquo 1983: ch.4; Wood and Perz 1996).
Deterioration of the natural systems on which agriculture depends, such as watersheds and aquifers, can be potent causes of increased poverty in expanding rural populations. Even if these factors are comparatively stable, however, behavioural responses to rising densities may impede economic progress. Unquestionably the rural populations of, say, Bangladesh or India would be much better off today if they had not more than doubled their numbers over the last few decades. There must be structural reasons why the society allowed itself to expand rather than following the path of macro-level rationality. The exploration of such reasons, however, calls for research at a less aggregative level.
A discussion of poverty and the environment should also take account of the urban sector-where half of the world's population live and nearly all of current population growth in the world is taking place. If industrial employment generation fails to keep up with growth in the labour supply, the urban working age population backs up in low-productivity activities and all-but-open unemployment-that is, in the usual understanding of the term, in the informal sector. This sector is seen as a kind of labour sump, "a refuge for those unable to find jobs in the modern sector" (Portes 1997: 249).
There is an alternative, far less pessimistic, view of the urban informal sector. Sachs (1990: 99), writing of Brazil, notes the intricate organizational structures that develop and the social capital that can be mobilized. The result is "a maze of interconnected labour, service, and goods markets, ranging from organized business to organized crime, a non-market household sector, as well as an incipient non-market social sector based on mutual help." Indeed, the informal sector can be seen less as a labour sump than as a dynamic source of microscale entrepreneurship for the economy. This view first came to prominence in the work of Hernando De Soto (1989), to whom informal economic activity appears as a dynamic response to the institutional rigidities of over-regulated or statalist economies. Unregulated enterprise, far from being the survival behaviour of a near-destitute labour reservoir, becomes the entrepreneurial feedstock of development. The poverty issue does not wholly vanish but is relegated to the status of a comparatively fringe concern. In Portes and Schauffler's (1994: 55) paraphrase of De Soto, "the informal sector is not part of the problem of underdevelopment, but part of its solution."
In any particular case, these contrasting pictures must somehow be reconciled with the single empirical reality. Part of that reality is a physical environment that for many cities is dominated by shantytowns, bidonvilles, and favelas. Many cities in the developing countries have been growing at double the rate of natural increase of population: rates of 5 percent per year are common, a doubling time of 14 years. Under those conditions the quality of the built environment is bound to decline; public health is likely to suffer along with it. Using DHS data, Brockerhoff and Brennan (1998) show the adverse effects of rapid population growth on infant mortality in cities, a link found to be especially pronounced in Latin America.
Plausibly, the two pictures are only superficially inconsistent. Portes and Schauffler (1994) propose a compromise. Their interpretation calls for a typology of informal activities, recognizing both the dynamism of many microenterprises and the precarious lives of many people in the sector. Some amount of dynamism and off-the-books organization in economic activity and elaborate networks of mutual assistance can coexist with a broad degree of material poverty. How much is hard to gauge: by their nature, estimates of the size of the informal economy, and thus of the incomes of its participants, are error-prone. (Mortality rates, or more general measures of "human poverty", escape this problem for population groups.) The static aggregate situation, however, would be of less interest than the detailed directions of change: whether poverty was a diminishing phenomenon along the life cycle of most individuals or whether it permanently identified a substantial subgroup.
Recent writings making a population-growth/conflict case include those of Homer-Dixon (1991, 1994), Goldstone (1991, 1997), and, in more popular vein, R. Kaplan (1996). Numerous country studies have been undertaken in recent years investigating the likely political consequences of upcoming scarcities in renewable resources. Population growth is one of the causes of scarcity, out-migration is often one of the effects. (See, for example, the study of water scarcity in China by Economy 1997.) Goldstone investigates the tensions generated by population expansion under inflexible or maladapting institutions in a series of historical and contemporary contexts.
Environmental scarcity, in Homer-Dixon's usage, can result from change in access to resources or a change in distributive systems as well as from resource depletion or degradation, so that demographic causes would rarely stand alone. In turn, scarcity raises the financial and political demands on governments and reduces their revenues. "A widening gap between state capacity and demands on the state, along with the misguided economic interventions such a gap often provokes, aggravates popular and elite grievances, increases rivalry between elite factions, and erodes the state's legitimacy" (Homer-Dixon 1994: 25).
Drawn-out civil conflict and resulting decay of state legitimacy and capacity would nearly always be accompanied by drastic economic deterioration. The African examples of Somalia, Liberia, and Sierra Leone are cases in point. (The converse argument, that impoverishment will generate civil unrest, finds relatively little support.) However, conflict does not necessarily lead to impoverishment. For example, the economic impact of a violent change of regime would usually be severe but may be fairly transitory and affect mainly the more capitalized sectors.
Economic demographers have spent much time modeling the demographic calculations of individuals and families. But understanding the family economy and its demographic responses to changed circumstances and opportunities does not in itself shed much light on population-poverty relationships. The prices and other incentives that determine the costs and benefits of demographic behaviour are set by economic forces, social organization, and cultural patterns that an individual family neither controls nor much influences. We thus need to set the family within its material and social environment, and explore the dynamics of this somewhat larger system. The aim is to identify the kinds of organizational patterns that tend to connect population growth and poverty.
As development proceeds, the economic disadvantage of high fertility to parents should become ever more apparent. Notably, parents come to recognize the increasing need to invest in education-in the "quality" of children rather than the quantity. Yet here also the family-level evidence is weak. Lloyd (1994) finds only a small negative effect on children of large family size-though the average disguises substantial selective disadvantage to some siblings, particularly females. Kelley (1994: 57) concludes from an exhaustive review of quantitative studies that "evidence on the impact of family size on educational outcomes is mixed, showing no convincingly consistent and strong impact, one way or the other."
Notwithstanding this ambiguity, there are particular circumstances in which the costs of a family's high fertility clearly fall on the children. This is the case where high fertility is associated with an early start to childbearing and close birth spacing, or with "unwanted" births. Poor families are likely to be especially susceptible in these respects, through initially lower education, less access to health care, and often less effective birth control. Children's health as well as their education may suffer: for example, parental malnutrition can lead to low birthweight children and greater risk of stunted development. The likely outcome is to reproduce poverty in the next generation.
Oberai (1993: ch. 6) cites various studies that find high fertility in poor urban families-two or more sons-to be in part a strategy aimed at combating poverty by enhancing economic security, leading to a vicious circle: "low incomes ensure poor education, nutrition and health, which in turn lead to low productivity and low incomes" (p. 191). Efforts to achieve low fertility as a path to economic mobility-as argued, for example, by those who perceive "poverty-induced" fertility declines (see A.M. Basu 1986)-are consistent with this story.
At least as important as formal education in the intergenerational transmission of poverty are the intangible qualities of child rearing. Low human capital may be perpetuated within a family or social group through poor parenting, economic insecurity, and cultural impoverishment. Thomas Sowell (1983: 255) has stressed the effect of disadvantage here, in contrast to family surroundings "where the values and patterns of life were a human capital that made economic success more readily attainable." As fertility comes to acquire an inverse gradient by income, the proportion of children raised in conditions of disadvantage rises substantially.
Hardin's aim, like many subsequent writers on the case, is to illuminate processes of ecological degradation in rural poor country situations. The common property resources in question are usually assumed to be open-access forests (as sources of fuelwood and leaf fodder) and local aquifers (for well water). For the dry regions of India, Jodha (1985, 1990) lists community pastures, community forests, waste lands, common dumping and threshing grounds, watershed drainage, village ponds, and rivers. With overuse, the productivity of these resources declines, with immediate and disproportionate impact on the poor. The population connection is evident. From an early age children can undertake such essential family tasks as gathering fuel, carrying water, or caring for livestock. But the population growth thereby sustained further worsens the environment's capacity to supply fuel, water, and pasture-thus maintaining or even raising the value of children to parents. Dasgupta's (1993) treatise, An Inquiry into Well-Being and Destitution, discusses the situation in detail, and formalizes the underlying theory.
An urban parallel can be drawn to the Hardin case, identifying a similar kind of prisoner's dilemma. The urban analogue to common property resources in rural communities is the condition of physical security and comparative social stability that provides the setting for the informal labour market and allows predictability in patterns of social relations. These background conditions in poor urban areas are as potentially subject to "degradation" as the rural environment.
The circumstances in which population growth does actually lead to impoverishment through these kind of mechanisms are those where cultural or institutional factors impede collective or unilateral action at the local level to limit degradation. Privatization of the resource (alienation or enclosure) or cooperative or state management of it are among the options that can halt the decline, an outcome produced by proper alignment of the various economic and political interests in play. Unfortunately, one route for preventing overexploitation is through simple exclusion of some category of claimants. Poverty for all, Hardin-style, may then be avoided but not for the excluded group. Clearly critical for a poverty-alleviating outcome is a remedy that also protects against adverse distributional effects. Finally, for longer-run viability, the incentives that are set up should be such as to maintain or increase per capita returns-by raising production or limiting further population growth.
Dasgupta (1993: 351) describes how enmeshed custom can create a social equilibrium that gives rise to significant externalities:
In many societies, there are practices encouraging high fertility rates that no household desires unilaterally to break.... [S]o long as all others follow the practice and aim at large family sizes, no household on its own will wish to deviate from the practice; however, if all other households were to restrict their fertility rates, each would desire to restrict its own fertility rate as well. Thus, there can be multiple social equilibria, each sustained by its own bootstraps, so to speak, and a society can get stuck in one which, while it may have had a collective rationale in the past, does not have one any more.
This picture, novel as it may be to modern economists, is close to the perception of classical theorists like Malthus and J. Stuart Mill who saw education as the input that could shake a society out of a poverty equilibrium. It recognizes that values and preference sets are not fixed as assumed in most welfare theory but are determined in the course of socialization and acculturation and are modified by experience and opportunity, all processes that are not independent of societal design.
A different kind of poverty equilibrium is vividly portrayed in the analysis of social development by Robert D. Putnam (1993). Putnam describes an n-person prisoner's dilemma over whether or not individuals and families should put trust in civic institutions. Expectations of economic progress take for granted a social infrastructure based on some measure of trust-for instance, belief in security of contract and in a degree of legal and administrative impartiality. This amounts to a cooperative solution to the dilemma under which both the economy and social capital can grow. But there is also a stable equilibrium of shared distrust, with minimal social capital beyond the family and kin-group and a self-reinforcing pattern of exploitative behaviour that produces a meager economic performance.
A social equilibrium would seem to bear some resemblance to the problematic idea of a class culture, and a low-level social equilibrium to that of a "culture of poverty." These concepts have fallen into disuse in research, though they clearly persist in popular understanding of, for example, the urban underclass. Traits making up a culture of poverty included authoritarianism, present-time orientation, and fatalism (O. Lewis 1961), with the all-too-easy implication that they formed a coherent belief system resistant to change. Critics of the concept preferred to see the behaviour of low-income people as a pragmatic response to deprivation and stress-as "responsive to the conditions of life rather than to cultural guidelines" (Rodman 1968: 335). Policy is thereby given something tangible to work with. "Traps" and similar equilibrium concepts are always contingent on a modeling structure that, even if initially valid, trends in the real world may be rendering obsolete.
Members of one community can worsen conditions in another through environmental spillovers-for instance, through poor agricultural practices whose local effects can be avoided but that are felt elsewhere through, say, pollution or flooding. They can also do so by exertion of political power or direct force to gain an economic advantage, for example through alienation of what previously had been common-property resources. In a weaker sense, labour migration also represents a transfer of costs of population growth, in this case on to participants in the recipient area's labour market. (Employers in the recipient area may of course stand to benefit from greater labour market competition, complicating the cost calculation.) Rural-urban migration can sometimes be viewed in this way, as in the Brazilian case noted earlier. "Communities" here would usually refer to territorially defined entities like villages or regions; quite analogous arguments, although for a possibly different range of transfers, can apply to groups defined by economic status, such as class or caste groups. Indigenous minorities are often the involuntary recipients of offloaded costs associated with population growth elsewhere. In some circumstances, women may be too.
Unilateral transfers with implications for poverty outcomes can also be intertemporal. The next generation in a society "inherits" from the present generation its physical and human capital, natural resources, and environmental amenity. Under rapid population growth, population numbers would often double in a generation, imposing a major dilution effect on this inheritance. It is the case of intra-family transfers writ large. (There are many conceptual and even ethical problems in delineating this case. Dasgupta [1993: ch.13] gives a brief treatment of them.)
The common feature of unilateral transfers is the inability of those on to whom the costs are imposed to organize effectively to protect their interests. This may reflect the structural weakness of their position in the larger society, geographical remoteness, or, in the intertemporal case, inherent obstacles to intergenerational contracting.