Population People

Posted April 1999

Ageing of Rural Populations in South-East and East Asia, Part 1

by Ronald Skeldon
Consultant, FAO Population Programme Service
and Adjunct Professor
Institute for Population and Social Research, Mahidol University
Thailand

This paper demonstrates the linkages between rural ageing fuelled by fertility decline, rural-urban migration and agricultural development as envisaged by the Population Programme Service of FAO (SDWP). The perceptive comments of Jacques du Guerny, David Iaquinta, Guillaume Lanly and Alain Marcoux on an earlier draft of this paper are gratefully acknowledged. The paper covers the countries of East and South East Asia excluding the Democratic People's Republic of Korea, Mongolia and Taiwan Province of China. Hong Kong China and Singapore are only briefly considered as their rural populations are negligible.

Nineteen-ninety-nine is the United Nations International Year of Older Persons. As part of FAO's contribution to this year it was thought that it would be of interest to focus on the ageing of rural populations in a region where this process was rapidly advancing. Furthermore, the ageing is not only of interest per se, but is important in relation to the changes in agriculture in the countries covered.

Background

The ageing of human populations has emerged as one of the most significant and universal demographic processes as we move into the twenty-first century. By the year 2000, almost one in ten of the world's population will be aged 60 years or over but, considering the more developed countries only, that proportion jumps to virtually one in every five persons [1]. The implications of having such large proportions of populations which have surpassed the traditional age of retirement are not yet fully understood. Although the proportions of older people are clearly highest in the more developed parts of the world, and hence in those areas most financially capable of supporting them, in terms of absolute numbers there are more elderly people in the developing world simply because that is where the bulk of the world's population is to be found. By the year 2000, it is estimated that there will be 229.5 million people 60 years of age and over in the more developed regions of the world compared with 373.3 million in the less developed regions. The proportions which fall into this category for East and South-East Asia in the year 2000 are estimated to be 11.0 and 7.2 per cent respectively, up from 7.3 and 5.3 per cent in 1960. These relatively small increases over 40 years in percentage terms mask substantial increases in absolute numbers, with an increase from 60 million in East Asia in 1960 to 163.8 million in 2000, and from 12 to 37 million in South-East Asia.

The principal mechanism bringing about this increase in the proportion of older people in the world has been the sustained decline in fertility over the last 30 to 40 years. The fact that each cohort of women begins to have fewer children does not initially give rise to fewer children as a whole because, given prior rapid rates of population growth, there are more women entering the reproductive age groups. Put another way, more women in the reproductive age groups can give rise to greater numbers of total births, even when each one of them is having fewer children than their mothers. However, when sustained over the medium term, declining fertility must give rise to fewer births, decreasing the size of the younger cohorts and bringing about the increasing proportion of older age groups. When combined with declining mortality at older ages giving increased survivorship to older populations, the result is a marked ageing of human populations.

Countries in East and South-East Asia have witnessed some of the most dramatic declines in fertility among all countries over this period and these are shown in table 1. By the late 1990s, not one area in East Asia (excluding Mongolia) had a total fertility rate (a period measure defined as the number of children the average woman would be expected to have had by the end of her reproductive life) above the replacement level of 2.1. Fertility rates in South-East Asia were significantly higher, with the poorest countries of the former Indo-China, Lao PDR and Cambodia, showing the highest fertility, and the larger countries of South-East Asia more moderate levels.

Table 1. Patterns of fertility decline: total fertility rates, selected countries, 1950-1995
1950-55 1955-60 1960-65 1965-70 1970-75 1975-80 1980-85 1985-90 1990-95

China

6.11 5.48 5.61 5.94 4.76 3.26 2.50 2.41 1.95

Hong Kong

4.44 4.71 5.31 4.02 2.89 2.32 1.80 1.31 1.21

Japan

2.75 2.08 2.01 2.00 2.07 1.81 1.76 1.66 1.50

Republic of Korea

5.18 6.07 5.40 4.52 4.11 2.80 2.40 1.73 1.73

Brunei Darussalam

7.00 7.00 6.72 5.94 5.40 4.40 3.80 3.40 3.07

Indonesia

5.49 5.67 5.42 5.57 5.10 4.68 4.06 3.31 2.90

Malaysia

6.83 6.94 6.72 5.94 5.15 4.16 4.24 4.00 3.62

Philippines

7.29 7.09 6.61 6.04 5.50 4.96 4.74 4.30 3.93

Singapore

6.40 5.99 4.93 3.46 2.62 1.87 1.69 1.71 1.73

Thailand

6.62 6.42 6.42 6.14 5.01 4.27 2.96 2.57 2.10

Cambodia

6.29 6.29 6.29 6.22 5.53 4.10 5.06 5.25 5.25

Lao PDR

6.15 6.15 6.15 6.15 6.15 6.69 6.69 6.69 6.69

Viet Nam

6.05 6.05 6.05 5.94 5.85 5.59 4.69 4.22 3.87
Source: United Nations 1995 World Population Prospects: The 1994 Revision, New York, Department for Economic and Social Information and Policy Analysis, Population Division, ST/ESA/SER.A/145.

The difference in proportion of elderly populations between East and South-East Asia observed earlier is largely explained by these patterns of fertility, and inter-country differences in the proportion of elderly population can be largely seen in terms of the date of the onset of fertility decline (compare tables 1 and 2). The pattern for Japan is unique among the countries under consideration. It last saw a total fertility rate of over 5 children per woman in the mid-1920s and had essentially reached below-replacement-fertility by the late 1950s, a time when fertility was generally peaking in the other economies in the region.

With the exception of the poorest countries of Indo-China, fertility decline was under way throughout the region from the first half of the 1960s, being most marked in Singapore, followed about five years later by Hong Kong and the Republic of Korea, and five years after that by China and Thailand. This highly generalized pattern obscures a host of intra-country variations. The reasons accounting for the onset and magnitude of the fertility declines are too complex to summarize here, and have been examined at length elsewhere. Suffice it to say here that government policy was a not insignificant factor among the matrix of explanatory variables.

Throughout the region under consideration, the proportion of the population that was aged 60 years and older in 1960 was less than 10 per cent of the total population, and typically between 4 and 6 per cent (table 2). By the mid-1970s, only in Japan had that proportion risen to over 10 per cent of the total population. For several countries in the region that proportion had even declined somewhat over the 15 years from 1960, reflecting the impact of the high fertility of the time. The 15-year period from 1975 saw an increase in the proportion of the elderly populations of all countries in the region, with that increase being most marked in those areas where fertility decline had either been earliest or most rapid and pronounced. Only in Japan, however, had the numbers of elderly approached one in five of the population, and for the majority of the countries in the region the numbers were still fewer than one in ten of the respective populations in 1990. The ageing of the populations of the East and South-East Asian region is thus recent and, with the exception of Japan, had yet to make a truly major impact on the age structure of the populations, until 1990 at least. This overall trend, as we will see, hides significant differences between rural and urban sectors.



Table 2. United Nations estimates of the proportion of persons 60 years of age and older, selected countries, 1960-1990
1960 1975 1990

China

7.2 6.9 8.6

Hong Kong

5.9 8.8 12.6

Japan

8.9 11.7 17.4

Republic of Korea

5.3 5.8 7.7

Brunei Darussalam

6.1 5.6 4.3

Indonesia

5.2 5.3 6.3

Malaysia

5.3 5.6 5.8

Philippines

4.9 4.3 5.1

Singapore

3.7 6.7 8.4

Thailand

4.5 4.7 6.9

Cambodia

4.5 4.6 4.7

Lao PDR

4.1 4.6 4.9

Viet Nam

6.9 6.4 7.2

Source: United Nations 1997 The Sex and Age Distributions of the World Populations: The 1996 Revision, New York, Department of Economic and Social Affairs, Population Division, ST/ESA/SER.A/162.

The basic patterns of fertility decline, of changing demographic structure and of the ageing of the populations, as described above, are well developed in the existing literature. A substantial literature, too, already exists on the ageing of Asian populations. See, for example, United Nations (1997; 1996a, b, c, d, e; 1994; 1993; 1991); Knodel and Debavalya (1992; 1997); Phillips (1992); Chen and Jones (1989). Most of these articles focus primarily, and very correctly, on the demographics of the process or on the availability of family support systems or private or public welfare provision for the elderly. While recognizing that significant intra-country variation by region and by urban and rural sector exists, the literature has relatively little to say about sub-national differences in the pattern of ageing or on the consequences of that ageing for particular subregions or sectors. An exception is the demographic description of ageing on the rural sector of developing countries by Marcoux (1994). This paper will extend this study but focus on the differential patterns and consequences of ageing for the rural sector in countries of East and South-East Asia, and particularly upon the role of migration in modifying the effects of changes in fertility and mortality.

Two scenarios

At the risk of overgeneralization, the interpretations of the impact of migration on rural areas can be broadly divided into two types. The first would see the migration as essential for development and its impact as positive. Among these approaches are the neo-classical, two-sector models of urbanization, migration and development which saw excess population removed from a labour-surplus rural economy of a developing country through migration to a labour-deficit urban sector. Once a critical threshold had been reached, and the marginal productivity of labour in the rural sector ceased to be zero, real wages in that sector would rise, eventually to reach those of the urban sector. Migration from the rural sector to the urban sector would then slow to reach a new equilibrium. This idealized model seldom, if ever, described the real experience of developing countries and has been the subject of modification and refinement. However, this need not detain us here. The critical point is that, in this scenario, migration is generally seen to be positive for development and, presumably, the ageing rural populations would eventually benefit from the movement of people from rural to urban sectors.

The above simple model ignores the selective nature of migration, that it is the youngest and the brightest who tend to leave the villages for the towns, albeit with spatial and temporal variation in the specific patterns of selectivity. Thus, a second very different interpretation builds in a demographic component. In this scenario, migration not only exacerbates the ageing effects of fertility decline by removing the youthful and most energetic cohorts, but it also removes those with initiative who are most likely to improve the condition of the rural and increasingly elderly poor. In this scenario, migration is negative for development and is likely to lead to the erosion of rural production and a deterioration in the welfare of the elderly village populations. The exodus of the youthful cohorts ultimately reduces the capacity of the village to reproduce itself, leading to a stage of decline and depopulation as the remaining elderly populations die off.

These two scenarios are clearly ideal types that try to capture the essence of what is occurring while ignoring the detail. As will become apparent in subsequent discussion, there is an element of truth in both viewpoints depending upon the development potential of the area concerned and upon the stage of a hypothesized development sequence. That sequence is related to the phases of the demographic transition as reflected in the patterns of fertility decline. However, although all the societies are experiencing a transition towards ageing populations, that transition is variable in both its pattern and its impact. Populations with different proportions of elderly will be examined before a more general assessment of an ageing transition model is made in the conclusion.

Before the relative impact of ageing on rural societies at the country level can be attempted, a major difficulty which underlies the analysis of the patterns of rural and urban ageing needs to be examined. This difficulty is essentially methodological and goes some way to explaining why the issue of differential rural ageing and the impact of migration have been poorly developed in the literature. The methodological difficulties relate to problems of data availability and to definition.

Methodological issues

Although the majority of countries in the East and South-East Asian region publish data on population by age and sex by sector, there are several important qualifications that have to be borne in mind. The first relates to the well-known issue of the lack of a common definition of rural (or urban) from country to country. Hence, direct comparisons from country to country of proportion of rural population can be deceptive. Changing definitions of urban and rural over time for a single country can also complicate any analysis of sectoral change and China is a particular case in point in Asia (see Chan 1994; Zhang and Zhao 1998). Nevertheless, despite the definitional difficulties, distinct rural and urban sectors exist in each country. It is not at all clear, however, that distinct rural and urban populations exist. In every census and large-scale survey, respondents are allocated to a rural or urban sector of residence, depending upon rules of enumeration, giving the impression that it is in that sector that they live and work. In reality, there is tremendous interchange of population between the two sectors, not just of long-term rural-to-urban migration, or the converse, but of a multitude of more short-term movements of human circulation. These migrations are paralleled by flows of goods and money in the form of remittances.

Thus, substantial numbers of those living in the urban sector may return to the villages at weekly, monthly or more long-term intervals and regularly send remittances in cash or kind back to their families in the rural sector. The short-term circulation of people is generally not captured in the majority of censuses or large-scale surveys. Hence, simply comparing the population structure of a rural sector, as derived from such data sources, may give a very misleading impression of the very real dependency relationships that exist across sectors. The problem will be more acute in those countries adopting de jure systems of census and survey registration, where the population is enumerated where it "usually lives" and significant numbers are recorded in their places of registration rather than where they normally live. This problem is particularly acute in China but is also present in other countries in South-East Asia that follow a de jure census enumeration strategy such as Indonesia, the Philippines and Thailand. A more refined definition of migration adopted in the National Migration Survey of Thailand, for example, virtually tripled the proportion of the migrant population of Bangkok from 8 to 22 per cent when the survey results were compared with those of the 1990 census of population (Chamratrithirong et al. 1995: 18-19). Using these different definitions of migrants brought out a difference between the wet and dry season populations of Bangkok, a city of some 6.6 million in 1994, in the order of 10 per cent.

One could perhaps argue that where the population is enumerated at place of registration, rather than where it is actually living at "census moment", may give a better impression of commitment to place of enumeration and thus of dependency relationships. That is, the de jure system, although omitting a greater proportion of short-term mobility than the de facto system of census enumeration, may in fact give a better impression of the "real" structure of the population based in the rural sector. Unfortunately, such a hypothesis must remain "not proven" at this stage as we do not have the data to test the idea adequately. In the following discussions, the available data on the rural population structure will be used but readers must be aware of the weaknesses inherent in the application of such data. Deficiencies in the data, however, must not be used as an excuse for lack of analysis and, despite their weaknesses, the existing data certainly suffice to allow the broad trends of the impact of migration on rural populations to be drawn out.

Rural-urban differences in population structure [2]

In all the East and South-East Asian countries for which data exist, the populations of rural areas are older than those of urban areas (table 3). It is in Japan and the Republic of Korea, those two most developed countries, with the highest proportions of elderly population, that the greatest differences between the proportions of populations 60 years of age and over in urban and rural sectors are to be found. The Republic of Korea is perhaps an extreme case, with almost one in five of the rural population classified as elderly, compared with only about one in every 14 in the urban population. In the Asian countries under consideration, fertility, while declining overall, as we have seen, is generally higher in the rural than the urban sectors (table 4). Although the differences in several countries are hardly pronounced, they are enough to suggest that the variations in the sectoral patterns of ageing should, if fertility decline were the only factor and other factors were equal, be in the opposite direction to that observed. Other factors are not equal, however, and migration from rural to urban areas is clearly critical in transferring youthful cohorts from the villages to the cities to generate the observed patterns. Urban areas are made youthful by accretion and rural areas age as residual populations through migration. That this statement requires some qualification will become clearer in the discussion of South-East Asia below.

Table 3. Total population, proportion urban, and proportion of elderly in urban and rural populations, various countries, latest available years
Total population Proportion urban (%) Urban population:
proportion 60 years and older (%)
Rural population:
proportion 60 years and older (%)

Japan (1995)

125,570,246 78.1 19.5 25.4

Republic of Korea (1995)

44,553,710 78.5 6.9 17.9

Brunei Darussalam (1991)

260,482 66.6 3.5 5.3

China
(1990)

1,130,510,638 26.2 8.1 8.7

Indonesia (1995)

194,754,808 35.9 5.6 7.4

Malaysia (1991)

17,498,091 50.6 5.3 6.7

Philippines (1990)

60,559,116 48.6 5.0 5.5

Thailand (1990)

54,532,300 18.7 6.3 7.7

Cambodia (1996)

10,702,329 14.4 5.0 5.4

Lao PDR (1995)

4,574,848 17.1 5.7

Viet Nam (1992)

69,175,080 19.5 7.4 7.7
Source: United Nations 1998 1996 Demographic Yearbook, New York, Department of Economic and Social Affairs, ST/ESA/STAT/SER.R/27, Table 8. Incomplete data for Lao PDR from "A demographic perspective on women in development in Cambodia, Lao People’s Democratic Republic, Myanmar and Viet Nam", Asian Population Studies Series No. 148, New York, Economic and Social Commission for Asia and the Pacific, ST/ESCAP/1869, table 39.

Table 4. Patterns of urban and rural fertility, various countries, mid-1970s
Total fertility rate: rural Total fertility rate: other urban Total fertility rate: major urban

Indonesia

4.9 4.3 4.6

Malaysia

5.0 4.5 3.5

Philippines

6.0 4.0 3.5

Republic of Korea

5.0 4.2 3.3

Thailand

5.0 3.6 2.5
Note: The three categories, rural, urban and major urban, come from definitions adopted in the World Fertility Survey. Urban areas were identified through individual national definitions, while major urban centres included cities exceeding 1 million population, all national capitals regardless of population, and the largest city in any country where no city exceeding 1 million inhabitants existed. Source: United Nations 1987 Fertility Behaviour in the Context of Development New York, Department of International Economic and Social Affairs, Population Studies No. 100, ST/ESA/SER.A/100, p.193.

It is also interesting to observe that in several countries where there is little observed difference between urban and rural patterns of ageing, such as China and Viet Nam, the censuses there too severely underestimate the volume of population mobility. In China, census enumeration through the system of household registration tends to omit vast numbers of the "floating" population, perhaps as many as 100 million, who have left their places of registration to look for work in the largest towns. As most of these "floaters" are young adults, the impact of migration upon the age structure of the rural populations of these countries is underestimated in the figures. However, as implied above, the question about the commitment and continued attachment to origin areas by the more short-term migrants becomes a critical issue in the assessment of the impact of migration on rural populations and areas. As we will see below, the volume of remittances sent by the young adults in the towns and cities of Asia to the villages may be an important factor in the continued viability of the rural economies and of the survivorship of the increasingly elderly populations.

Migration can also influence the age structure of rural populations in another way, not simply in terms of the exodus of the young adult cohorts. Return movements of those who have spent several years away, either overseas or in the urban sector, will contribute to the replenishment of older age cohorts and the aggravation of potential problems of supporting older populations in the villages by local sources of labour. Such potential problems, however, must be balanced against any wealth, knowledge or other resources, such as pensions, that they may bring back to the village communities. Return migrants, although older, also have learned skills that may encourage them to invest in small businesses or in land development.

The effects of the changing age structures are clearly reflected in the dependency ratios, or the ratio of the potentially economically active population 15-59 years old to those who are assumed to be dependent, or those 0-15 and 60 years of age and older. The dependency ratio is primarily a demographic index and it is well recognized that, in the rural areas of developing countries in particular, many of those 60 years of age and older (together with many younger members of those populations) are truly "economically active". This proviso and the accepted close interrelationship between urban and rural sectors apart (through which the urban-based economically active can support the rural-based elderly), the older age and youth dependency ratios for urban and rural sectors give some idea of where the burden of ageing is falling in the countries of the region (table 5). Looking at the rural areas, it is clear that Japan and the Republic of Korea are in a class by themselves, with a very high proportion of elderly relative to the numbers of economically active in that sector. There are considerably larger numbers of rural old relative to rural youth in Japan while proportions in the Republic of Korea are virtually equal. In all the other countries, the pattern is still one of the large proportion of youthful dependants, and ageing as an economic burden has not yet emerged as a significant issue. As might be expected, overall dependency in the urban areas is considerably lower than in the rural sector except in the cases of Brunei Darussalam and Viet Nam. Excluding Japan, there is also much less variation in the pattern of old age dependency. Thus, ageing as a potential burden is primarily to be found in the rural sectors of the most developed economies of the region.

Table 5. Old age and youth dependency ratios, by urban and rural sector, latest years
  Rural Urban
Old age dependency ratio Youth dependency ratio Old age dependency ratio Youth dependency ratio

Japan (1995)

44.1 29.1 29.4 24.1

Republic of Korea (1995)

29.0 32.6 10.0 34.2

Brunei Darussalam (1991)

8.5 57.8 5.7 55.3

China (1990)

14.1 48.0 11.7 32.1

Indonesia (1995)

13.0 62.2 9.1 48.9

Malaysia (1991)

12.1 74.7 8.7 55.1

Philippines (1990)

10.5 80.7 8.7 63.2

Thailand (1990)

11.3 48.4 9.0 32.5

Cambodia (1996)

10.7 88.8 9.3 74.2

Lao PDR (1995)

11.8 nd 9.7 nd

Viet Nam (1992)

14.5 75.0 14.0 74.4
Note: Old age dependency ratio is the ratio of the population of 60 years and older to the working population 15-59 years old. Similarly, youth dependency is the ratio of the population younger than 15 years old to the working population 15-49 years old. Source: United Nations 1998 1996 Demographic Yearbook, New York, Department of Economic and Social Affairs, ST/ESA/STAT/SER.R/27, table 8; United Nations 1998 "A demographic perspective on women in development in Cambodia, Lao People’s Democratic Republic, Myanmar and Viet Nam", Asian Population Studies Series No.148, New York, Economic and Social Commission for Asia and the Pacific, ST/ESCAP/1869, tables 7 and 8; and Lao Census 1995: Country Report, Ventiane, 1997, National Statistical Centre, pp.47-48.

The ageing of the rural sector also has a gender dimension. That women live longer than men is virtually a universal characteristic of human populations, even if there are certain exceptional areas where this is not the case. All the elderly populations in the countries under consideration, excluding Brunei Darussalam, follow this generalization and have a greater proportion of women, with that trend being most emphasized in the populations of highest longevity and greatest ageing, Japan and the Republic of Korea. There are slightly fewer elderly men per elderly woman in the rural sector of Japan when compared with the urban sector, while that situation is reversed in the Republic of Korea. The clearest difference, as far as the gender of the elderly populations is concerned, is found in Malaysia, the Philippines and Thailand. There, the elderly in the rural sector are much more balanced in terms of sex ratio than are the populations in the urban sector which are much more female-dominant. This difference may be due to a greater incidence of return of male rural-to-urban migrants to the rural sector, although why this pattern should not also be evident for Indonesia, where the phenomenon of male retirement migration back to the villages has been identified as being important (Hugo 1992: 216), is not clear from the data presented here (table 6).

Table 6. Sex ratio of populations 60 years of age and older by sector, latest years
Urban Rural

Japan (1995)

77 74

Republic of Korea (1995)

64 69

Brunei Darussalam (1991)

101 122

China (1990)

95 89

Indonesia (1995)

85 87

Malaysia (1991)

83 95

Philippines (1990)

82 95

Thailand (1990)

79 96

Cambodia (1996)

248 75

Lao PDR (1995)

90 92

Viet Nam (1992)

70 77
Note: The very high masculinity for urban areas in Cambodia is a result of the recent tragic events in that country which saw the virtual emptying of the cities during the Pol Pot years. Men have been among the first to enter the urban sector with a return to more stable conditions.

Source: United Nations 1998 1996 Demographic Yearbook, New York, Department of Economic and Social Affairs, ST/ESA/STAT/SER.R/27, table 8; United Nations 1998 "A demographic perspective on women in development in Cambodia, Lao People’s Democratic Republic, Myanmar and Viet Nam", Asian Population Studies Series No.148, New York, Economic and Social Commission for Asia and the Pacific, ST/ESCAP/1869, tables 7 and 8; and Lao Census 1995: Country Report, Ventiane, 1997, National Statistical Centre, pp. 47-48.

An alternative scenario would involve differential migration of the elderly women themselves. In Indonesia the data suggest a lower propensity for older women to migrate to towns than in Malaysia, the Philippines or Thailand. The migration of the elderly is always going to affect a very small minority of total movers but some attention may need to be given to the relative provision of traditional support services for elderly women in South-East Asian countries where, upon the death of their spouse, elderly women may have to seek support from urban-based family in some of these countries. More positively, grandmothers can also provide babysitting and other domestic services for their sons and daughters in town that would allow the adult female members to pursue urban employment. This situation might apply particularly where fertility decline has been marked and there are few children to look after younger siblings. Perhaps significantly, fertility in the major urban areas of Malaysia, the Philippines and Thailand is lower than in Indonesia. Hence, the differential migration of elderly women may indeed be a factor in explaining the observed differences and their migration and social situation warrant further study.

Clearly, the impact of ageing in the rural sector is not going to be evenly spread, either within the region or within any particular rural sector in any single country. We will now consider the impact of ageing on the rural populations of the East and South-East Asian region, looking first at the countries where ageing is most pronounced and then moving on to those areas where the impacts are as yet more potential than real. Particular attention will be given to the situation in the Republic of Korea and in Thailand.

Back to Top FAO Homepage