Below-replacement fertility: determinants and prospects in South Asia

Journal of Population Research, May, 2003 by Bruce K. Caldwell, John C. Caldwell

Although 10 countries and two of China's special administrative areas, totalling 1,528 million people or 44 per cent of Asia's total population, are now characterized by fertility rates below long-term replacement levels, no such countries are yet found in South Asia. This paper first examines the characteristics of 12 Asian administrations with very low fertility at various stages of their fertility declines and then compares the findings with the present situation in three South Asian countries, Sri Lanka, India and Bangladesh. This allows a prediction of when the South Asian countries will reach replacement fertility in accord with the trends in two key criteria, the percentage of girls in secondary school and the infant mortality rate. These conclusions are then buttressed for each country by the findings of anthropological demographic research programs in which the authors were involved. The predictions are that all three countries will attain a total fertility rate of 2.1 within the next 30 years and th at the UN 2000 Revision of the medium population projection is plausible in that regard. However, the authors part company with the UN projection in their assessment that the nature of these societies means that they will all subsequently fall to still lower fertility levels.

Keywords: Asia, Southern Asia, India, Bangladesh, Sri Lanka, fertility decline, below-replacement fertility, population forecasts, marriage postponement, education of women.

**********

The world is now experiencing unprecedented low fertility. By the end of the twentieth century, 44 per cent of its population lived in countries with below-replacement levels of fertility as measured by total fertility rates (TFRs). Whether the whole world will face declining population depends to a considerable extent upon what happens in the huge, poor countries of Asia.

The focus of this paper is South Asia, a key area in this regard in that it comprises one-quarter of the human race, living at a lower per capita income level than anywhere outside tropical Africa. Indeed, the two most populous countries examined, India and Bangladesh, have real per capita incomes equal to those of the United Kingdom in the early nineteenth and mid-sixteenth centuries respectively (Maddison 2001), long before that country began its fertility transition. Nevertheless, during the second half of the twentieth century the TFRs of these two countries halved from six or higher to a little over three, while rates in Sri Lanka, which is also addressed, declined from a little under six to two (United Nations 2001). The explanation even for India and Bangladesh was levels of schooling not achieved by Britain until the end of the nineteenth century, infant and child mortality rates not achieved until the 1920s, and the existence of national family programs strongly supported by governments and expressed public opinion. The question to be addressed is whether such advances can propel poor countries to below-replacement-level fertility.

The paper briefly examines the characteristics of those Asian countries which have already achieved below-long-term replacement fertility, and which constitute 42 per cent of the continent's population; it then compares this situation with that in three countries of South Asia which may reach this level within the next two or three decades. These three countries, Sri Lanka, India and Bangladesh, constitute another 32 per cent of Asia's population, and have been selected because the authors have undertaken anthropological demographic research there on family change and family motivation for limiting fertility. (1)

Table 1 examines those Asian countries (2) which have achieved below-replacement-level fertility and groups them according to historical experience. The groups are ordered by the dating of the onset of fertility decline. Judging by historical experience, important criteria for below-replacement fertility appear to be considerable levels of education, especially female education, and low child mortality. All these countries are characterized by over three-quarters of adult females being literate and, with the exception of Thailand, at least two-thirds of adolescent girls being in school. Except in China, infant mortality rates are around 20 or below. All exhibit adult female marriage with few women marrying under 18 years of age, and in industrial Asia, very late age at marriage. In Sri Lanka, and to a lesser extent elsewhere, postponed female marriage was a factor in the early stages of fertility decline (Alam and Cleland 1981). Everywhere, later female age at marriage has probably been a factor in allowing c ouples to practise contraception more easily. The lower level of contraceptive practice in Japan and ex-USSR Asia is explained by high levels of abortion.

Table 2 presents the situation in the three countries upon which we are focusing and adds comparative data from the below-replacement-fertility countries of Table 1. Sri Lanka falls everywhere within the range characterizing the below-replacement-level fertility countries except for the level of urbanization. The low level of urbanization is explained by the Sri Lanka settlement pattern where, to take the most striking example, continuous settlements along the southwest coast from which large numbers of workers commute to Colombo are classified as distinct residential areas, thus obscuring the fact that these are in effect quasi-metropolitan populations subject to city influences and lifestyles. Sri Lanka is probably about to fall below the replacement-fertility level, but there remains a question of why it has not already done so. The answer probably lies in a relatively comfortable lifestyle not seeming to necessitate urgent change, as well as the nature of the Sinhalese Buddhist society, with its emphasis on accepting one's life situation. The attainment of a TFR of 2.1 in Sri Lanka with only 44 per cent of couples practising modem contraception is explained by a long and largely successful tradition of using both withdrawal and periodic abstinence (Caldwell et al. 1987). In contrast, both India and Bangladesh fall outside the range of nearly every socio-economic index, and, unlike Sri Lanka with its long history of rising education levels, their recent increases in female education will take a very long time to shape the whole adult population. Nevertheless, it should be noted that India has reached the lower threshold of girls in secondary school although it will take years for it to do so in terms of adult female literacy and education. Bangladesh is currently attempting a major increase in its numbers of girls in secondary school and the degree of success so far suggests that the projected attainment of replacement fertility according to this criterion in Table 3 may prove too pessimistic.