Publication by Virginia Deane Abernethy

INTRODUCTION

The fertility rate is declining almost universally and explanations vary (Robey 1993; Eberstadt 1994; Obermeyer 1995; Bongaarts and Watkins 1996; Potts 1997; Abernethy 1979, 1993). Weaknesses in simplistic socioeconomic interpretations of fertility trends have given prominence to the default position, that provision of family planning services is the uniquely correct and effective way to reduce fertility rates worldwide (Robey 1993; Potts 1998). Family planning technology and services do indeed offer a humane way to limit family size – possibly the best ever devised by human societies – but it does not follow that modern contraception is causing the fertility decline.

Individual and family decision-makers do not use contraception, even if it is convenient and free, unless they intend to space or limit childbearing. But what determines intention? This paper contends that socioeconomic approaches, which have become increasingly sophisticated, provide a good understanding of rising or falling fertility rates, and should be tested for their predictive power.

Pursuit of causal relationships within the socioeconomic paradigm requires one to specifically address certain simplistic but durable claims made by earlier models. These are, especially, that modernization, low infant mortality, education (especially for women), and a high standard of living directly cause low fertility (Notestein 1945; United Nations 1988; Caldwell et al. 1992; Sinding 1992; Kennedy 1993; Pritchett 1994; Hodgson and Watkins 1997; Gelbard 1997). These propositions are understandably popular because they would combine a better quality of life for many people with a long-sought demographic result. Nevertheless, they have proved inadequate. Some demographers began to note, relatively early, that fertility rates in the third world did not decline as had been projected. Indeed, decades of jump-starting modernization and economic development failed to induce lower fertility (Davis 1963; Teitelbaum 1975; Kanter & Teitelbaum 1977; Coale & Watkins 1986).

Such disappointments do not, however, justify dismissal of socioeconomic accounts of fertility. Much recent research is nuanced, traces multiple pathways to fertility decline, and provides valuable insights into individual societies. The scholars engaged in this work bring perspectives from demography, anthropology, biology, economics, epidemiology, geography, political science and history, at a minimum. Many studies address causality and, although no single case study constitutes proof of a causal pathway, the accumulation of numerous congruent examples encourages consideration of a strategy that might be called qualitative meta-analysis. A collaborative effort would probably be needed to accomplish a work of that magnitude.

Demographers and political scientists have long seen individual motivation as a key to explaining fertility rates (Malthus 1798, 1803, 1830; Davis 1963; Demeny 1968, 1988). The central importance of micro-level analysis is confirmed by the finding that the number of children a couple wants explains 85-90 percent of actual family size (Pritchett 1994). A cultural perspective is implicit in this paradigm because, where a critical mass of couples shift their family size targets, the culture itself evolves adaptively to accommodate the new choices (Abernethy 1979). In very stable societies, moreover, cultural beliefs, values, and customs may substitute for individual decision-making (Nag 1963).

Cultures often contain internal contradictions. For example, even stable cultures may foster large family-size values at the same time as they prescribe behaviors - such as long post-partum sex taboos – that reduce a woman’s exposure to pregnancy. These contradictions may be functional, however, because child-spacing is often the best protector of women’s health, and it may maximize successful reproduction, including child survival rates, in the long run.

The common theme of much recent literature is that, although large families are often preferred, difficulty in providing for children at the desired level of quality damps fertility. Parents take into account not only the emotional and economic benefits of children, but also the cost of raising them including the opportunity cost of parents’ time – in Gary Becker’s terms, the "price of children" (Becker 1992). It follows that a couple’s family size target approximates the number of children that, in their estimation, can be adequately raised, taking into account aspirations and expectations. The causal connection envisioned could be described by the concept of "fertility opportunity."

The idea that couples (or individuals) seize upon apparent opportunities for childbearing builds upon or is consistent with microeconomic decision-making models (Becker 1960, 1992; Davis 1963; Demeny 1988; Easterlin 1962, 1971, 1975; Stokes 1995). It also assumes the correctness of Patricia Draper’s observation (1989, p.147) that "the critical factor affecting reproductive decision making is the individual’s perception of resource quantity, rather than absolute resource quantity…. Cultural understandings about the availability and location of resources appear in many cases to be a more powerful predictor than are the ‘actual’ resources themselves." Perception and motivation are, thus, key variables.

A great variety of circumstances can create a perception that the resources needed for raising children are scarce. People appear to learn to value a particular standard of living, including leisure time and educational opportunity, so perceptions about the adequacy of resources for raising children reflect aspirations. They are not only subjective, but also culturally specific. Perceived resources, optimism or trepidation about the future, and estimates of the cost of raising a child all influence desired family size. The decision to allow or delay childbearing appears to result from interplay between the natural predisposition to reproduce and socioeconomic factors, with the implicit goal of providing offspring with the best start available given the alternatives.

The frequency and wide distribution of behavioral patterns that limit child-bearing (Nag 1968) suggest that restraint (Princen 1997) is a widespread human adaptation. An evolutionarily-selected predisposition to exercise restraint under appropriate circumstances seems most likely in species where parents invest heavily in the rearing of offspring, that is, where there may be a trade-off between quantity and quality of offspring. A frequently-seen individual reproductive strategy is to delay the first birth, space births, and limit total family size.

Laboratory and field studies show that when a limiting factor in the environment is removed, offspring numbers increase. In biology, the unexpected lifting of an environmental constraint is called "ecological release," and the expected result is prolific reproduction. Under stressful conditions, by contrast, prospective parents in many species appear to hoard resources by delaying or restricting reproductive effort. The description of how socioeconomic factors and motivational pathways influence human fertility is congruent with theoretical principles of natural selection.

The perspective that couples seize on apparently better reproductive opportunities but exercise restraint under ordinary circumstances suggests that desired family size is shrinking because perceived scarcity in most regions and sectors of industrialized, industrializing, and third world countries is making couples cautious about undertaking marriage and childbearing. Even where gross national product is rising rapidly, the distribution of wealth may be very skewed, or modernization may bring new aspirations, so that the majority of the people have a renewed sense of limits. Whether on the edge of subsistence, in relatively affluent societies, or in very affluent societies, more and more couples believe that large families are not affordable.

Understanding sometimes requires one to "step into another’s shoes." How does it feel, for example, to be middle class with some professional training in a society where duplicate numbers of oneself are legion but well-capitalized jobs are relatively few? Except by answering this and similar questions, how can one understand perception, incentives and motivation? Resolution of such research problems often depends upon naturalistic observation, so continuing data collection and evaluation of alternate explanations is essential.

Emphasis upon in-depth observation does not preclude the design of testable hypotheses. Often, however, these tests will be opportunistic, depending upon the occurrence of something dramatic or unique. For example, following the summer of 1997 currency crashes, wipe-outs of savings, rising unemployment, and even food shortages (Indonesia) in countries that had come to be known as the Asian tigers (Japan, South Korea, Thailand, Malaysia, the Philippines, Singapore, and Indonesia), I predicted declines in their fertility rates (Abernethy 1998). Taiwan and China escaped much of the economic turmoil so were not included in the prediction. If expectations are born out, it will be the more remarkable (in my opinion) because most of these countries already enjoyed near-replacement level fertility. Of this set, only Malaysia and the Philippines have much room to shift down.

Certain cases of declining fertility appear to support simplistic accounts of the socioeconomic model. Most cases, however, present significant difficulties and yield only to the sophisticated treatment that characterizes much recent research and theory. This theoretical work bears on policy because, seeing population stabilization as essential to long term sustainability (Atiyah & Press 1992; McNicoll 1995), the United States government and numerous non-governmental organizations remain active in the population arena. Provision of family planning assistance allows couples to avoid repeated childbearing humanely, if that is their wish, but the basis for wanting larger or smaller families continues to be debated and is key.

Without cultural or individual inhibition of childbearing, the natural human fertility rate might hover in the vicinity of 8 or 9 births per woman. This was the rate observed in a 1940s survey of Hutterite women in a prosperous northwestern U.S. farming community that encouraged large family size (Heer 1968). The fertility rate also rose to nearly 9 children per woman in an Amazonian tribe, the Shipibo, shortly after encounters with non-indigenous Peruvians began to alter their traditional culture (Hern 1991b). Fertility usually remains below the natural rate. But why?

EARLY SOCIOECONOMIC MODELS

Early and still popular socioeconomic models called for effecting a worldwide fertility decline through achievement of low infant mortality, widespread literacy, equality for women, urbanization, and general prosperity (Caldwell et al 1992; Kennedy 1993; United Nations 1988; Sadik 1990; Sinding 1992; Gelbard 1997). Most advocates of the model anticipated a modernization process that was benign.

WEAKNESSES IN EARLY, POPULAR MODELS

The rationale is persuasive but supporting evidence is slim. Simplistic socioeconomic models are heavily reliant on the 1930s observation that developed, western countries had much lower fertility than any other major region of the world (Davis 1937; Notestein 1945). Regrettably, little theoretical significance was attached to the severe economic depression that, during the 1930s, engulfed every western nation including the United States and which was probably the source of reluctance to bear children (Easterlin 1962).

European historical demography showing when and under what circumstances western countries adopted their low fertility pattern was largely unavailable to early theorists. It is these historical studies that now help to qualify beliefs in the simplistic causal relationships claimed for modernization and economic development (Coale and Watkins 1986; Demeny 1968, 1988; Knodel and van de Walle 1979).

Allegedly causal factors could not have caused the French demographic transition because they had still to emerge when the shift to low fertility and nearly universal marriage was completed. France, the first western country to adopt a modern family structure, had acquired its new fertility regime by 1850 while still characterized by an agricultural economy, widespread poverty, and high infant mortality. Indeed, the transition had begun while child abandonment was so great a problem that Napoleon ordered turnstiles installed at orphanage entrances so babies would not be left in the cold. This end-of-the-eighteenth century reform increased the number of babies given up, and mortality within the first two years of life reached 80 percent in some institutions (Langer 1972). Decline in the national infant mortality rate began after 1900, 50 to 70 years after a low fertility rate was well established (Rollet-Echalier 1990).

Nor could decline in infant mortality have caused the German fertility decline. Fertility fell between 1870 and 1930 in most of Germany's 71 administrative units, but no systematic relationship with the decline in the infant mortality rate is apparent. In about two-thirds of cases, the fertility rate declined before or at about the same time as the infant mortality rate (Knodel 1974).

In contemporary third world countries as well, states epidemiologist Maurice King (1990), infant mortality often declines after the fertility decline, and therefore is not a possible cause. Standing the development-first rationale on its head, ample data suggest that people having fewer children attach more value to each one and, all else equal, have more resources per child to invest in their healthcare and education.

Education, particularly for girls and women, is another economic development variable receiving much attention. The 1994 Cairo conference on population and environment espoused the goal of raising the educational as well as legal status of women, in part because of the assumed effect on fertility. Indeed, the negative correlation between women's education and the fertility rate is substantial. Nevertheless, correlation does not establish causality. Martine’s analysis of the fertility decline in Brazil notes the simultaneity of fertility decline and increasing education, but rejects a simple causal conclusion (1996, p.63). The Population Reference Bureau takes a similarly, and newly, cautious position, stating that women's education and employment result in lower fertility "only when they bring women more power" do (Riley 1997, p. 14).

Comparisons among modernizing countries in South-East Asia also indicate that educational level, itself, is not a magic bullet for reducing fertility. Specifically, about 70 percent of Philippine children attend secondary school (the third highest after Taiwan and South Korea) and the 1997 fertility rate was 4.1 children per woman. In Thailand only about 40 percent of children attend secondary school (South-East Asia's Learning 1997) but the Thai 1997 fertility rate was 1.9, below replacement level and down .2 of a child in one year (World Population Data Sheet. Population Reference Bureau, various dates). Similarly, decades of education programs provided to the Kikuyu in Kenya did not block a fertility increase to about 8 children per woman after elimination of polygyny (Reining and Tinker 1975; Reining, personal communication 1997). But in the late 1980s, after cuts in international aid forced retrenchment in education subsidies, the fertility rate declined (Martha Campbell, Packard Foundation, personal communication, 1997).

A more serious flaw in simplistic models is the apparent presumption that only modern societies can have low fertility rates. To the contrary, many fertility-limiting mechanisms operated in non-industrialized societies. Mechanical means, social arrangements, and cultural practices which kept fertility well below the natural rate in pre-modern (which is to say, illiterate and rural) societies are amply documented (Richards and Reining 1954; Reining and Tinker 1975; Nag 1968; Scheffel 1988; Hern 1991b).

The three following examples suggest the range of possibilities for fertility control in pre-modern societies. In each case, restrictive behaviors were embedded in the culture. In stable environments, generations of experience select for patterns that lead to avoiding demographic disaster. (And note indications that balance may be disrupted by exogenous and therefore misunderstood influences) (Abernethy 1979, 1993).

Oral tradition in the Pacific Island of Yap recalls a large population beset by intense famine and death approximately fifty to one hundred years after first contact with western navigators. By 1930, the fertility rate appeared to be at replacement level or lower. Japanese public health records reveal concern about depopulation. Venereal disease appears to have been a contributing factor but, in addition, the culture penalized childbearing and put obstacles in the way of sexual intercourse. Low fertility was the result of magical beliefs in the danger of sex and of mixing gender and age categories in ritual, including the preparation and eating of everyday meals. The probability of childbearing was reduced by a 7-year post-partum sex taboo, shame attaching to pre-marital pregnancy, mechanical interference with conception including coitus interruptus, abortion, rules which limited marital sexual activity, and arcane rules for food collection and preparation that motivated women to remain single or, if married, to prefer small family size (Hunt 1954; Nag 1968). The clustering of behaviors and beliefs that limit childbearing is not unexpected in societies that have been "shaped" by the experience of devastating overpopulation, although the Yapese case is extreme.

A second case, Emilio Moran’s (1991) description of the ecologically impoverished "blackwater" people of the Brazilian Amazon, also suggests that limited resources produce a culture that essentially guarantees low average lifetime fertility. Unlike most undisturbed Amazon tribes, the blackwater people enjoyed few good housing sites or productive hunting or fishing locations. Delayed marriage and low fertility rates resulted from a hierarchically organized society that used inheritance rules to restrict access to scarce resources. Marriage was prohibited until a man acquired resources sufficient to support a family, and no provision was made for support of children born out of wedlock.

Epidemiologist Warren Hern (1991a, 1991b) shows that modernization may raise the fertility rate if it leads to abandoning parts of the traditional culture, as occurred among the Shipibo in the Peruvian Amazon. Before contact, polygynous marriage was standard and resulted in 4.5-month longer intervals between births and several fewer births over a woman’s lifetime than with monogamy. Shipibo villages are now typically monogamous and the fertility rate has risen to about 8 births per woman. Women suffer severely impaired health as a consequence of unremitting childbearing, and the newly-established population doubling time is about 17 years. Population may or may not have been in equilibrium with resources in the traditional culture, but today’s situation appears unstable.

The theories of natural selection and sociobiology postulate that behavior, as well as physical characteristics, is subject to natural selection. It follows that behavior which enhances survival and successful reproduction is positively selected.

Lots of births do not necessarily maximize successful reproduction. In species such as human, where parental care and investment contribute to optimal development and future success of young, successful reproduction is maximized by many factors including careful timing of births. Quantity and quality of resources available to offspring often depend upon selecting an opportune moment for reproduction. Badly timed reproductive effort is likely to be wasted. Thus, first births may be delayed to allow time for accumulation of resources (natural and human capital). Subsequent births may be spaced in order to maximize the resources available to each offspring. The total number may be limited, to the same effect. To the extent that parental investment affects the chances that offspring thrive, parents in many species appear to trade off quantity against the advantages that can be bestowed upon a fewer number (Trivers 1972; Wilson 1975).

Such responses are not easily detected among humans because occasions for hypothesis testing are serendipitous. Such occasion fortunately presented itself with the observation that relatively isolated Finnish communities on the archipelago of Aland and Aboland had the highest rate of fraternal twinning of any reported in the world. In contrast to identical twinning, which can be genetically-determined or not, fraternal twinning (dyzogotic) is always genetically-determined. This means that differential rates of twinning in different environments, which would include fraternal twinning, reflect natural selection. To test for the selective advantage of twinning or not, Lumma et al. (1998) compared child survival and lifetime maternal reproductive success with singletons and twins within and between archipelago communities and the Finnish mainland, using church parish records from the pre-industrial period. On the archipelago, twins had as high a survival rate to age 15 as singletons, and women giving birth to twins had as great a lifetime reproductive success as those giving birth to singletons. Matched communities on the mainland had equivalent survival rates for singletons but lower child survival and lifetime maternal reproductive success rates with twins. The authors conclude that frequent famines on the mainland had selected against twinning, whereas the archipelago’s steady and ample food supply from fishing allowed twinning to be an equivalent reproductive strategy even after taking into account increased maternal mortality with twins.

The foregoing describes selection at the individual level. Group selection is relatively rare because it must satisfy special conditions (Wilson 1975). Nevertheless, natural selection operating against whole groups might be inferred from the remnants of Easter Island (Diamond 1995) and the Anasazi of the U.S. southwest (Kohler 1992). These societies probably satisfy the conditions for group selection: a high degree of kin-relatedness and genetic isolation. Both once flourished but then declined. Invaders certainly did not scatter the inhabitants of Easter Island, whose population declined from about 7000 in 1550 to approximately 100 individuals three centuries later. Nor, probably, were the Anasazi losers to an invading group. Climate change cannot be ruled out, but the primary source of failure was probably endogenous. Cultures and social arrangements capable of balancing population size, consumption, and resources over the long term did not develop, so the Easter Island and the Anasazi civilizations died out. A few descendants of the once-large Easter Island population remain, and individual Anasazi may have dispersed.

Most findings lack crystalline demonstration of the operation of natural selection - that is, rewards and penalties for particular strategies. Nevertheless, the congruence of observed traits with expectations generated by theory is valuable supporting evidence. Behavior that is predicted by the theory of natural selection includes, under very specific circumstances, the practice of infanticide.

Infanticide by parents becomes more likely when a child already has an unusually low probability of survival and/or future reproductive success, and parents are likely to have other, better reproductive opportunities (Hausfater and Hrdy 1984). For example, unmarried mothers (but not others) among the South American Ayoreo reportedly practice infanticide. Ayoreo culture and social arrangements make it unlikely that an illegitimate child would be well cared for. Moreover, never-married Ayoreo women are young women who can usually look forward to future reproductive opportunities. They would be inclined, therefore, to wait for the likeliest best use for the time and energy expended on parental investments (Bugos and McCarthy 1984). Infanticide is also practiced in certain Brazilian tribes when infants are closely spaced (Hern 1991a), again a course predicted by sociobiology because such infants have a relatively low probability of survival. A string of infanticides in Hungary, which came to light in 1997, is blamed on the upsurge in "social, economic and psychological havoc caused by Hungary’s transition to a free-market economy." Many people appear to have been pushed "to the edge" by an unemployment rate of 10 percent and an inflation rate of approximately 25 percent, which amounted to a 12 percent decline in real wages (Agovino 1997, p.5F). Knowing neither the age, marital status, nor number of surviving children of women (?) who committed infanticide, it cannot be said if this was a reproductively "smart" decision.

Nancy Scheper-Hughes (1992) describes less overt behavior: the ready resignation of mothers to the deaths of young children in poverty-stricken areas of Alto, in northeastern Brazil. Infants who appear passive are said to not have the will to live and are, in fact, neglected so that the "diagnosis" becomes a self-fulfilling prophecy. Similar insouciance, even the abetting of extraordinarily high infant mortality, characterized Labrador Inuit women after a century of disruption of their culture and impoverishment by Moravian missionaries (Scheffel 1988). Where the probability of any particular infant surviving is low, sociobiology predicts that parents will concentrate their resources on those who exhibit traits most likely to reward parental effort.

Increases in fecundity and fertility when conditions are good and most infants seem likely to survive are the opposite side of this coin. Increases in the maturation, growth and survival rates, and/or population growth in response to more resources per individual among non-human species, while perhaps tangential to this paper, are consistent with the socioeconomic account of reproduction. Under experimental conditions, simultaneously lifting the constraints of both limited-food and predation increases the population significantly more than when either condition is manipulated separately (Rosenberg et al. 1993; Stenseth 1995; Valone and Brown 1995). Among primates, orangutans appear to adjust offspring numbers to the availability of resources (MacKinnon 1974).

Ronald Lee (1980) reports comparable findings from research on pre-industrial England. Analyzing the relation of wages to the rate of natural increase, Lee concludes "that a doubling of the real wage would increase the population growth rate by about 1.25% per year, ceteris paribus"(p. 540). As far back as the sixteenth century, "short-run variations in the real wage" explained 20 to 30 percent of the short-run variance in nuptiality and marital fertility rates (p. 539). He adds that evidence on the relationship between wealth and marriage and fertility rates is mixed, although on balance it appears "that wealthier couples may not only have married earlier, but also have had higher fertility within marriage" (Lee 1980, p.537). Diane Macunovich finds, similarly, that when more-educated American women earn high incomes, they, too, have higher fertility than average (Zachary 1998). Recent reports from China also acknowledge "many known cases, especially in southern China, where rich businessmen deliberately have second, third and fourth children by different women, as a way of displaying wealth and position through traditional means." In some regions, moreover, it has become routine for officials to encourage prosperous families to have two or three children, because fines collected from violations of the "one-child-policy" belong "to the local government and may sometimes be spent by the same officials who levy the fines" (Faison, 1997, p.1, 6). It goes without saying that both sides to such transactions must be willing.

Human sensitivity to cues about the state of the natural and socioeconomic environment appears to be the mediating factor for many of these relationships. Sensitivity, or attentiveness to early, marginal signs, is illustrated by Timothy Dyson’s analysis of five major famines, spanning over a century, in the Indian sub-continent. Dyson shows that small price increases for staple foods led to behavioral changes which caused the fertility rate to fall significantly by the time the full force of famine materialized. The mechanisms included delayed marriage (possibly because dowries were more difficult to accumulate) and the departure of married men to seek work in less affected areas (Dyson 1991a, 1991b). Strategies for coping with scarcity often entail the separation of spouses, but economic cues would be unlikely to influence human reproduction if people were not attuned to the balance between their family size and resources.

Although reduced fertility is an effective response to limited resources, it often is not the preferred strategy. Humans are able to affect their environment, and one of humanity's best answers to perceived constraints in the local environment is invention of new technology, even though this sometimes entails increased work (Boserup 1965). Another solution is migration to richer territory. But where these responses to perceived scarcity fail, the motivation to reduce fertility appears to strengthen (Davis 1963; Friedlander 1983; Brittain 1990, 1991; Low 1996).

AMBIGUITY

Cases are often ambiguous. The source of ambiguity is often the different meanings put upon the concept of scarcity. The "sense of scarcity" is subjective, by definition. But simplistic models view scarcity as incidental and unfortunate, whereas some recent socioeconomic accounts of fertility portray scarcity as a virtually existential part of the human experience.

Although the sense of limited resources is virtually ubiquitous, it can be dispelled temporarily under certain circumstances. Technological advances, migration toward opportunity, or unaccustomed wealth from any source, often engender the perception that resources will expand almost limitlessly. The key word is "unaccustomed," because after habituation to a particular level of affluence, fewer resources or even a slower pace of acquiring wealth may be perceived as privation. "Scarcity," indeed, describes the shortfall between wanting and having. Thus, scarcity is experienced within a wide range of economic circumstances (Abernethy 1979).

Scarcity approaches an absolute among large families in poverty-stricken regions where per capita income or access to land is insufficient to provide life's necessities. In addition, perceived scarcity is associated with the low and declining real wages that usually accompany rapid growth in the labor force. Stark income inequality is often a function of rapid growth in the supply of labor, so that it outstrips the creation of new, good jobs (Lee 1980, 1987; Borjas 1996). Therefore, rapid growth of the population, which soon flows into the labor market, makes countries vulnerable to a process that polarizes rich and poor, capitalist and worker. Portes (1997) sees inequality as a necessary corollary of modernization (although labor market dynamics suggest that it can occur at any stage of economic development so long as the population is growing rapidly or jobs are contracting). Martine (1996) suggests that "proletarianization" is an inevitable part of modernization and conducive to lower fertility, whether or not it is associated with real pauperization.

In very affluent societies with broadly distributed wealth many people experience scarcity in terms of relative deprivation. Resources may be insufficient to satisfy wants because of aspirations derived from one's parents' lifestyle (Easterlin 1987). Any standard associated with a subjectively accepted reference group, as well as new consumption opportunities displayed through advertising, can induce the sense of scarcity. Keeping up with the Joneses (or the idealized TV family) and consumerism make lots of people feel poor.

In the United States, consumerism puts the goal of acquiring things into competition with the value of extra children (Macunovich & Easterlin 1990). But consumerism, which environmentalists lambaste in the United States, has made prophets of those who say that economic development causes fertility decline. The sense of not having enough induced by consumerism accompanies declining fertility in the growing middle class of South and Central America, North Africa, and Asia, that is, in some countries where economic growth rates have set records (Portes, 1997; Wysocki, 1997). Martine concludes that, "Increasing aspirations can be as significant as objective deprivation in the formation of pressure to reduce family size" (1996, p.65). The consumer culture makes people prefer goods to children.

Acquisition of more goods may be, in part, a means of allaying anxiety that recent gains can be as quickly lost. Johannson (1987) suggests that "status anxiety," fear among Europe's wealthiest, landed families that wealth would be dissipated if heirs were too numerous, accounts for depressed fertility rates visible in the upper classes by 1600. (In this connection, however, recall Ronald Lee’s conclusion that evidence on the relationship between wealth and fertility rates is mixed.)

A few cases illustrate the ambiguity of data, where the causes of fertility decline are difficult to isolate and sometimes revolve around local understanding of scarcity. The Indian province of Kerala, Asian countries that modernized rapidly after World War II, much of South America, and certain Yoruba villages of Nigeria depict the difficulties of interpretation, and these are far from being the only examples of a mix of factors appearing to contribute to demographic transition. Even if the immediate incentives for limiting family size are feelings of scarcity and relative deprivation, perhaps it should be counted as a success for simplistic socioeconomic models that these feelings are induced by the large wealth inequalities and consumerism that have often accompanied modernization in the third world.

The Caldwell et al. studies that trace development and fertility in selected Yoruba villages in Nigeria are notable for their depth and breadth of data. During the early to mid-1980s, oil wealth and international assistance rolled into Nigeria, healthcare and education programs became widely available, the infant mortality rate declined, and many more women had access to education. These developments were expected to reduce fertility rates.

As the 1980s rolled on, however, early marriage, closely spaced births, and disinterest in contraception showed no sign of abating. While disappointing, the lag also seemed understandable and, on revisiting their study population in 1990-91, Caldwell et al. (1992) found significant declines in the fertility rate which were readily interpreted as a response to modernization. Later marriage and a new level of receptivity to modern contraception – used pre-maritally to delay age of marriage, for birth spacing, and to terminate childbearing– were viewed as a belated response to the earlier interventions.

A different interpretation is possible. The view that individuals adjust childbearing and family size to perceived opportunity would have predicted the early to mid-1980s delay in the fertility decline on grounds that the average person was enjoying unprecedented advantages and optimism was pervasive. For this hypothesis it also is significant that declining fertility followed the late 1980s economic downturn, diminishing international aid, reduced subsidies for education and healthcare, more difficulty in finding remunerative employment, and widespread disappointment.

Caldwell et al. address the problem: "The most difficult issue is the role of the present economic crisis in affecting fertility change. The predominance of childbearing and marriage deferral and of birth spacing in the demand for contraception in Ado-Ekiti suggests that it is education and the rise of the modern economy which underlie the rising demand. Nevertheless, most contraceptors say that an important factor in their practice is the hard economic times" (pp. 236-237). Specifically, "Most of the respondents believed that child mortality had risen over the previous five years of economic difficulty" (p.226), and "Two thirds of all respondents claimed that the major force behind marriage postponement and the use of contraception to achieve it was the present hard economic conditions" (p.229).

Respondents’ self-reported reasons for limiting fertility, including the rising risk of child mortality, are clearly difficult to integrate into simplistic socioeconomic explanations of declining fertility. Putting the best possible light on such accounts, one might conclude that modernization underlies the fertility decline because the immediate cause of the decline is fear of losing previous gains. Fertility is depressed by shrinking employment opportunity, more expensive schooling, and more difficulty in raising children to adulthood, but these are perceived as difficulties only because modernization created new tastes, new expectations, and new standards for child survival.

Kerala, India, presents a still more ambiguous case. The facts of low infant mortality, good healthcare, a broadly-educated populace, significant equality enjoyed by women, and all in conjunction with low fertility, are widely reported. Proponents of modernization say that Kerala is proof that socioeconomic development "works."

On the other hand, the historical record is not well reported so it is difficult to pinpoint other possible causes of decline. It also is true that average family land holdings are the smallest, and Kerala is the most densely populated state in India. Moreover, "Unemployment and the lack of industrial development have been well known sore points" for decades (Wallich 1995). Comparing themselves with a relevant reference group – other parts of India, for example – Keralans may feel poor.

It may be significant that Kerala shares with Bangladesh, Rwanda, and Malawi, countries not noted for economic development, an acute hunger for land. The receptivity to modern contraception and fertility decline in Bangladesh are well documented. Kalipeni (1996) and May (1995) state that land hunger and the inability to colonize new territory explain recent fertility declines in Malawi and Rwanda, respectively. Griffith Feeney (1994) suggests that land hunger accounts for the negligible resistance, even in rural areas, when China’s 1970s "later-longer-fewer policy" was first promulgated. This campaign led to fertility that dipped briefly below 3 children per woman and was, of course, a precursor to the 1979 one-child per family policy.

One may ask, is land hunger in an agricultural economy sufficient to account for fertility decline? Does the disconnect between resources available– including land - and resources needed to support an additional child equally well explain the low fertility rate in Kerala?

Unlike Kerala where the modernization variables are primarily social, that is, improved education, healthcare, and women’s status, other Asian countries including Japan, Taiwan, South Korea, Singapore, Thailand, and Indonesia have had rapid economic growth. Their fertility declines, as well, are frequently cited as proofs of the causal influence of socioeconomic development. Of all these, Japan’s fertility history is best documented and it, as shown by Kingsley Davis (1963), makes an ambiguous case, at best, for simplistic accounts of the decline.

Japan led post-World War II Asia in both industrialization and fertility decline, but the relevant history begins earlier. In the aftermath of Japan’s opening to westernization in the late nineteenth century, their ecological niche expanded rapidly through industrialization, agricultural intensification and, lastly, emigration. Kingsley Davis (1963) suggests that a growing national product as well as outlets for the rapidly expanding population encouraged high fertility. The total fertility rate (TFR) was approximately 5.l in 1925 (Kennedy 1993). By 1930, births were exceeding deaths by nearly 1 million annually.

Population growth apparently overshot domestic resources and became a factor fueling not only emigration but also Japan’s military ambitions (Kunii 1994). Initiation of the Greater Asian Co-Prosperity Sphere was designed to secure rubber and oil, and it facilitated colonization of Korea, Indonesia and other neighbors.

Japan’s industrial and military strategies did not suffice to alleviate growing hardships among ordinary people, forcing the average family to bear underemployment, rice famines, and other hardships not only during, but also prior to World War II. Chojiro Kunii (1994) recalls how these trials became an incentive to avoid large family size. By 1940, fewer than half of Japanese women age 20 to 24 had married, a shift to later age at marriage, says Davis, that was "prompted by personal rather than national goals" (1963, p.350). Feeney (1994) concurs: "It is apparent that the Japanese fertility decline began during the 1930s and was about two-thirds complete by 1950" (1994, p.1519). Two factors, adds Feeney, suggest that parents’ decisions to limit family size were a rational response in terms of the labor needed to work family farms: "excess children" had become problematic because of better child survival and structural changes in the economy. The latter included the necessity of better educating children if they were to find good off-farm employment.

After World War II, Japan had to reabsorb several million expatriates from former colonies into its now-shattered economy (Statistics 1986). By 1955 only one third of women age 20-24 had married and abortion became common. Japan’s gross reproduction rate (daughters per women) declined from 2.7 in 1920 to 0.99 in 1959 (Davis 1963, p.349).

Thus, the cycle of Japanese fertility rates is nearly the opposite of that predicted by a simplistic account of socioeconomic development and fertility. Early twentieth-century industrialization and improvements in the education and healthcare systems are associated with rising or continuing high fertility. The fertility decline was barely evident until shortly before World War II, so it did not come about in proximity with the leading edge development factors. It did coincide with increasing hardship, perceived resource scarcity, and the felt need for territorial expansion. Further, the decline preceded by decades the trend toward equality for women, which is relatively recent.

The perspective that higher fertility is a response to broadening opportunity does account, however, for high fertility in the early part of the twentieth century – as Japanese industry, the military, and opportunity expanded – and for declining and continuing low fertility as population growth outran resources and good, new jobs. Middle class Japanese worked hard under conditions of a large labor force relative to the number of jobs both before and after World War II. They paid high taxes, saved much of their earnings because the government offered minimal old-age assistance, and acquired advantages through an extremely competitive educational system where one succeeded only with hard work and concentrated parental investment. Moreover, even in the heyday of Japan’s Asian tiger status, reports James Fallows, "Very little of Japan’s new wealth spreads out to ordinary families. The country may be three times as rich as it was, but its families still live in little houses – and must work harder than they did before to remain competitive…" (Fallows 1995, p.25). In 1998 life has become harder for ordinary people and one would not be surprised to see further fertility declines. The effective economic hardship experienced by the average person may be expected to determine the severity of downward pressure on fertility rates.

Other countries, formerly designated "Asian tigers," present a more ambiguous pattern. The fertility decline appears to have accompanied urbanization and industrialization in most of the region and some countries, such as Taiwan and Indonesia, have had active family planning programs in place since the 1960s and 70s. In some countries, including Taiwan, Malaysia, and Singapore, declining infant mortality rates and better education favor the interpretation that the social and healthcare variables caused the fertility decline. But they apparently are not necessary conditions because fertility rates also declined in countries such as Thailand and Indonesia, where relatively little social progress has been made.

A sense of scarcity and difficulty in making a living may be the more general sentiment in the region. In Thailand, indeed, 200,000 to 2 million women are said to engage in voluntary or involuntary prostitution, which means, at a minimum, 1 in 40 of all women aged 13-29 engaged in prostitution (Poor 1991) and – until recent popularization of condoms - at high risk of sexually transmitted diseases. In addition, wealth is very unevenly distributed and some people are probably living worse that formerly (Moffet 1994). Except for Taiwan and Japan, which began to limit family size four or five decades ago, the former Asian tigers’ economies are swamped with labor, so wages are low and the ordinary "middle class" worker enjoys relatively little of the new national income.

The average person is spending little on consumption, whereas the wealthy (at least until the 1997 economic collapse) spent but also saved. Most Southeast Asian countries had an extraordinarily high savings rate. Before Thailand’s recent economic meltdown, up to 40 percent of the gross domestic product had been reinvested (Krugman 1994). The high level of saving and reinvestment is possible because the concentration of wealth leaves some few people and companies with ample discretionary income. But it occurs at the cost of a low standard of living for the majority of the population.

In Thailand, traditional sources of wealth had begun signaling scarcity even as the economy grew: "...population growth has shrunk per capita holdings of farmland in rural Thailand [and there is awareness of] diminishing land for succeeding generations to inherit" (Moffett 1994, p. 148). Parts of Thailand most resemble Bangladesh which, with very high population density and poverty rates, a primarily agricultural economy, and lagging progress in both social reforms and industrialization, has been notably successful in introducing family planning.

The fertility decline in many Latin American countries is similarly open to alternate interpretations. A compendium on Latin America repeatedly addresses the theme of falling living standards as a partial cause, at least, of regional fertility declines; other factors such as aggressive family planning programs and economic development are variably present (Guzman et al., eds. 1996).

As in the cases of Malawi (see below), Kerala, and Bangladesh, Brazil's recent 50 percent decline in fertility is concentrated in a region of high effective population density, the region where Scheper-Hughes (1992) reports high infant mortality. Writes George Martine (1996), "Brazil's poorest socioeconomic region, the Northeast, has undergone the fastest fertility reduction over the last 20 years: there, the TFR has fallen from around 7 in 1970 to 3.7 in 1990" (p. 49). Addressing common cause and effect hypotheses, he reports that women's education and labor force participation are negatively correlated with declining fertility, but no causal effect has been proved. Martine suggests that the mechanism underlying the fertility decline is an aspect of modernization, concluding that is a secular decline continuing through good and bad economic conditions. The development process, he suggests, creates new needs and wants, and sets up large sectors of the population for disappointment.

In sum, certain of the data can support simple attributions of causality to a few socioeconomic variables, but sophisticated models that encompass aspirations seem more satisfactory. The widespread 1997-1999 economic crisis is creating circumstances that favor hypothesis testing. The fertility opportunity hypothesis predicts that the fertility decline will accelerate where modern aspirations are dashed, underemployment deepens, and savings are wiped out by currency devaluation and its sequelae (Abernethy 1998). Additional tests might focus on major oil-producing countries, where fertility rates remained high through a long period of widely-distributed prosperity and broadening access to education and healthcare, but have recently tracked oil prices downward. If cash-short governments are further forced to withdraw subsidies for social goods that the people have come to expect then, all else equal, the fertility rate should fall rapidly.

In 1966, when the first census was taken, Malawi’s population was 4 million. By 1995 it was 10 million. Ezekiel Kalipeni (1996) su