Contents - Previous - Next

This is the old United Nations University website. Visit the new site at

Trends in urbanization

Levels of urbanization in South-East Asia are low by world standards, and even by developing-country standards (Table 3.8). Due to factors such as different definitions of urban areas, inter-country comparability is questionable, as indicated by the inclusion in the table of an alternative set of estimates for Thailand and Peninsular Malaysia which the author believes provides data somewhat more comparable with the other countries than the data employed by the United Nations (UN, 1989).

Since 1960, urbanization has been progressing steadily, but not dramatically. The growth rates of urban population, at 3-5 per cent per annum, however, have been high, because modest rates of rural-urban migration are added to still high rates of natural increase. The logistic curve is generally accepted as the standard path of urbanization and most South-East Asian countries are located near the inflexion point on the curve where accelerating urbanization can be expected, especially if rates of economic growth are rapid. Despite the slowing of population growth rates in most of these countries, this would mean a continuation of present rates of urban population expansion for some time, and the UN projections (Table 3.8) imply that a 4 per cent rate of urban growth (that is, a doubling of urban population in 17 years) will not be uncommon in the 1990s.

TABLE 3.8 South-East and East Asian Urbanization Trends and Projections, 1960-2005 Per Cent Urban

  1960 1970 1980 2000
World total 33.6 37.2 39.9 48.2
South-East Asia 17.6 20.2 24.0 35.5
Indonesia 14.6 17.1 22.2 36.5
Thailand 12.5 13.3 17.3 29.4
  (16.2) (20.8) (24.5)a (40.0)
Philippines 30.3 33.0 37.4 49.0
Malaysia 25.2 27.0 34.2 50.4
    (28.7) (37.4) (n.a.)
Myanmar 19.3 22.8 23.9 28.2
Cambodia 10.3 11.7 10.3 14.5
Vietnam 14.7 18.3 19.3 27.1
Laos 7.9 9.6 13.4 25.1
Papua New Guinea 2.7 9.8 13.1 20.2
East Asia 25.0 26.9 28.1 32.6
China 16.8 19.3 20.4 25.1
Japan 62.5 71.2 76.2 77.7
South Korea 27.7 40.7 56.9 80.5
North Korea 40.2 50.1 59.7 72.9

Urban Population Growth Rates (average annual percentage)

  1960-5 1970--5 1980-5 1990-5 2000-5
World total 2.97 2.65 2.42 2.55 2.52
South-East Asia 3.77 4.14 3.96 3.83 3.44
Indonesia 3.72 4.92 4.60 3.96 3.24
Thailand 3.53 5.59 4.66 4.02 3.70
Philippines 3.85 4.02 3.81 3.71 3.32
Malaysia 3.77 4.87 4.51 3.87 2.77
Myanmar 3.83 3.23 2.09 3.21 3.84
Cambodia 3.48 -2.09 3.54 4.19 4.30
Vietnam 4.18 2.86 3.27 4.16 4.38
Laos 3.14 5.52 5.58 5.45 4.77
Papua New Guinea 15.44 6.08 4.35 4.88 5.05

Sources: UN ( 1989); figures in brackets from Jones (1983: Table 2); Ashakul (1990).
a 1978 figure.
n.a. = Not available.

The relatively low levels of urbanization in South-East Asia, then, are the result of high rates of population growth in rural areas rather than the failure of cities to reach substantial size (Hackenberg, 1980). Indeed, three of the megacities of South-East Asia-Jakarta, Manila and Bangkok-will each have exceeded 10 million people by the year 2000 (Table 3.9). They will each have left Sydney and Melbourne, which were larger than any of them in 1940, far behind. Singapore will also be left far behind because of its lack of a rural hinterland and its now very low rate of natural increase.

TABLE 3.9 Growth of South-East Asian. South Korean and Australian Metropolises, 1940-2000

  Population (millions) Percentage Increase
Metropolis 1940 1960 1980 2000 1960-2000
Jakartaa 0 8 0.3b 6.5 12.0 300
Bandung 0.3 1.0b 1.8 3.4 240
Surabaya 0.5 1.0b 1.7 3.2 220
Hanoia 0.1 0.9 2.5 3.4 278
Ho Chi Minh Citya 0.5 2.3 3.4 5.0 117
Manilaa 0.9 2.4 5.9 11.5 379
Singapore 0.7 1.6 2.4 3 0 87
Bangkoka 0.6 2.3 4 7 10.3 348
Kuala Lumpara 0.2 0.4 1.3 3.8 850
Yangon (Rangoon)a 0.5 1.0 2.2 4.5 354
Seoul n.a. 2 4 8.3 13.0 342
Pusan n.a. 1.2 3.1 5.8 406
Melbourne 1.1 1.8 2.8 3.3 83
Sydney 1.3 2.1 3 2 4.1 95

a Conurbation
b 1961
n.a. = Not available

Proximate Causes of Urban Growth and Urbanization

Though South-East Asian planners, mirroring popular understanding, frequently attribute urban growth to rural-urban migration, there are in fact three components of urban population growth: natural increase of the urban population, rural-urban migration and the reclassification of areas previously defined as rural. These can be combined in different proportions to produce different rates of urban population growth. Using a l-year 'accounting' perspective, the contribution of each component can be measured fairly precisely, provided that appropriate data are available.

Natural increase provides a 'floor' for urban population growth rates, and rural-urban migration and reclassification supplement this growth. Where urban proportions of the population remain small, a constant, moderate rate of outmigration from rural areas can be decisive for the development of cities. Other things being equal, its contribution will decline as the urban proportion of the population grows. But other things are often not equal, and in countries whose urban population has reached 30 or 40 per cent of the total (as in much of South-East Asia), the role of inmigrants from rural areas may well increase for two reasons. First, cityward migration is encouraged by wide and persistent urban-rural income disparities and rapid economic growth in the urban core. Secondly, rates of natural increase of the urban population often decline sharply as urbanization proceeds, as they have, for example, in Thailand, Malaysia and Indonesia.

What has been the post-war South-East Asian experience in this regard? Even at only moderate levels of urbanization, the natural increase of city populations has played a major role in their growth. It was responsible for 61 per cent of urban growth in Peninsular Malaysia in 1957-70 and 65-70 per cent in Indonesia in the 1960s and early 1970s, compared with 60 per cent in a sample of developing countries in the 1960s (Ogawa, 1985: Table 3; UN/ESCAP, 1981: 72 4). However, a trend towards an increased contribution of rural-urban migration to the growth of cities was observed in Indonesia in the 1970s when natural increase was responsible for only 45 per cent of growth. Over the same period, natural increase accounted for only 50 per cent of urban growth in Peninsular Malaysia and 39 per cent in Thailand, although it was as high as 60 per cent in the Philippines (Ogawa, 1985: Table 3). In the longer-sometimes much longer-term, the share of migration in urban growth will inevitably fall as the impact of any given rate of rural outmigration on urban areas, which hold an increasingly large share of the total population, will lessen.

If the l-year 'accounting' approach is replaced with a longer time perspective, however, the relationship between the three factors in urban growth becomes conceptually quite complex (Jones, 1988: 139-40; Keyfitz and Philipov, 1981). It can be argued, on the one hand, that the role of migration is larger because much of the natural increase of city populations is attributable to migrants-most of them in the young reproductive ages-from a previous period (Rogers, 1982). On the other hand, the rural 'pool' from which migrants are being drawn is constantly expanding by natural increase. A corollary of this is that a decline in the rate of natural increase in rural areas, via a decline in the crude birth-rate, has a key contribution to make over the long term in the control of urban population growth. The role of reclassification is similarly clouded; there is clearly an interdependence between the three factors contributing to urban growth which is difficult to disentangle.

So far, the discussion has focused on urban growth, not urbanization (that is, the rise in the proportion of total population living in urban areas: for definitions, see Goldstein and Sly, 1975). In very broad accounting terms, natural increase contributes to urban growth, but rarely to urbanization because rates of natural increase in the urban areas are usually below those in rural areas. Using a short-term 'accounting' perspective, urbanization is normally attributable entirely to rural-urban migration and reclassification.

Underlying Causes of Urban Growth and Urbanization

What generates rural-urban migration and hence the tempo of urbanization? It is not intended here to review the massive literature on migration in the region (Hugo, 1984; Khoo, 1984; UN/ESCAP, 1980, 1981, 1982a, 1982b). The key point to make is that both studies dealing with individual or familylevel decision-making and studies at a more aggregative level with net migratory flows (for example, Arnold and Cochrane, 1980; Pernia et al., 1983; Titus, 1978) indicate that economic factors and employment opportunities are the main reasons for migration (Gugler, 1982: 185; Rogers and Williamson, 1982: 471). Theretore, the underlying explanation for urbanization has to do with changing employment opportunities as structural change takes place in the economy.

Over the course of development, agriculture's share of total employment will not necessarily decline (Booth and Sundrum, 1984). However, in most Asian countries over the 1960s-1980s, although there have been gains in per worker agricultural productivity, these gains have been surpassed by advances in productivity elsewhere in the economy; thus, agriculture's share of total employment has been steadily declining.

Since the predominant economic activity is agricultural in rural areas and non-agricultural in urban areas, there is obviously a close relationship between the shift in the balance of economic activity from agriculture to nonagriculture, and the shift in the residential location of the population from mainly rural to a higher proportion in urban areas. The link between structural economic change and urbanization, however, is not as strong as is often imagined (Jones, 1983, 1990; UN, 1980). Even in the most traditional rural areas, some people engage in trade, construction and other nonagricultural activities, and the proportion doing so tends to increase as an economy becomes more complex. Some urban dwellers also work in agriculture, though the proportion does decline as cities increase in size and traditional rural-urban linkages are broken.

Improvements in transportation break down the association between location of job and of residence by broadening the options for people who wish to participate in the urban economy. Many are able to commute from as far as 40 or 50 kilometres away from the city, and others can move regularly between urban and rural areas, in accordance with the demands of the job market and of family responsibilities, without the need to make a permanent change in residence. Advances in transportation, by easing rural access to external markets, also broaden the options both for the kinds of small-scale production and service activities in which farm-household members might engage, and for the location of large-scale industry and other forms of economic activity. Whereas in the past, there may have been no alternative, the development of road and rail networks has made it more feasible to locate these activities away from the cities (though the choice is often the pert-urban fringe areas which are 'rural' only by definition).

Other factors encouraging such a move include the increasing cost of urban land, the diseconomies of congestion in many urban industrial areas, and educational developments, which have given rise to a better-educated rural work-force than was previously available. A cross-national regression study of factors influencing rural, nonfarm employment (Blank and Parish, 1988) has found that density of transport systems did indeed have a significant effect on rural, non-farm employment, and that it was the only factor that did so, apart from the level of economic development.

The rural labour force undergoes major modifications in occupational structure during the course of economic development (UN, 1980: Table 29). In Japan, the proportion of the rural labour force in agriculture fell from 54 to 38 per cent in just 10 years (1960-70). In other countries, the trend may not be as sharp, but the direction is the same. Off-farm employment now represents a substantial share of farm-household incomes: 60 per cent in Japan, 50 per cent in Taiwan, 40 per cent in South Korea and even 50 per cent in the least industrialized Peninsular Malaysian state of Kelantan (Jones. 1983: 24-8: Shand, 1986).

Emerging Patterns of Rural-Urban Interaction

The rising non-agricultural share of rural employment, and wide range of mobility patterns characterizing rural-urban interactions in South-East Asia, must be seen in the context of changing structural linkages between urban and rural areas-indeed the blurring of the distinction between urban and rural areas-in terms of employment structure, morphology and access to facilities (Jones, 1983: 21-8). Improvements in transportation lead to ribbon development, expanded rural-urban fringe areas and what McGee (1991) refers to as 'zones of intense rural-urban interaction'. Beyond this, 'truly' rural areas have minibus services, television, primary schools and consumer goods, which greatly modify the formerly stark rural-urban contrast.

This suggests the possibility of alternative patterns of development which do not inevitably lead to such high levels of urbanization (and therefore requirements for urban infrastructure) as Western experience suggests is inevitable. Particularly appropriate in densely settled areas, such as Java, much of the Philippines and the Red River delta in North Vietnam, may be patterns pioneered in Japan and Taiwan which link urban and rural areas very closely

and increase off-farm employment opportunities of various kinds. These may involve a more diversified range of activities at the village level or in nearby rural areas, or it may involve the possibility of moving to cities on a daily 'commuting' basis or in a regular pattern of temporary migration. One of the main preconditions for such developments is a good public transportation network.


Urban primacy and megacity issues

Primate Cities

Much of the concern with urbanization in South-East Asia has really been with the growth of the 'primate city'-Bangkok, Manila, Jakarta and others. This concern is not misplaced; the size of megacities emerging in the region will 'take us, in planning terms, far beyond anything the world has yet seen and hence into realms of great uncertainty' (Jones, 1983: 3). Moreover, many metropolitan projections may underestimate the scale of the agglomerations that are developing. As Vining (1985) has stressed, to fully capture the population trend in the region surrounding a megacity, it is necessary to set the boundary of the 'core region' rather widely. When this is done, what appears to be a decline in the metropolis' share of the total urban population may merely reflect that growth has slowed in the city proper, but is accelerating on the periphery which, in a major metropolitan region with reasonable transport systems, can be far from the city centre.

Jakarta is a good example. Between the 1980 and 1990 Population Censuses, the population of the Jakarta Special Region grew by only 2.4 per cent per annum to reach 8.2 million, a much slower rate of growth than for the overall urban population of Indonesia. But the population of the regencies of Bogor, Tanggerang and Bekasi, which surround Jakarta and constitute the extended Jakarta planning region of Jabotabek, grew by 5.2 per cent per annum. More significantly, although somewhat exaggerated by changing definitions of 'urban' between the two censuses. the urban population of these three regencies grew by a remarkable 15.9 per cent per annum. Although the growth of Jakarta proper was relatively slow, the urban population of this 'core region' grew somewhat more rapidly than Indonesia's urban population as a whole (5.9 per cent compared with 5.4 per cent).

It is accepted among regional planners that there is no 'optimal' hierarchy of city sizes, despite the normative connotations often given to the 'rank size' rule. However, it is also generally accepted that efforts should be made to divert growth from the largest cities; growth-pole strategies, secondary-city strategies and 'agropolitan' developments have all been promoted as mechanisms to achieve this end (Jones, 1991). The attempt to block big-city growth by imposition of controls (a strategy adopted for a time in Indonesia in the early 1970s) distorts natural forces and entails large social costs. The adoption of such draconian policies also fails to recognize that diseconomies of metropolitan growth, such as congestion and inflated land values, if sufficiently serious, will affect private investment decisions (Kelley and Williamson, 1984). These market-based adjustments to 'excessive' metropolitan growth should be complemented by a strategy of eliminating biases in macro and sectoral policy that unnecessarily promote the growth of large cities.

Accepting that strategies will be adopted to foster the growth of smaller rather than larger cities, the potential impact of such policies in slowing the growth of the metropolis will depend very much on the city-size hierarchy in the country concerned. If the metropolis contains a large proportion of the total urban population, and there are no other fairly large cities capable of absorbing a big absolute increase in population, then the prospect of holding back the growth of the metropolis through promoting intermediate cities is, at least in the shots teen, rather dismal. South-East Asian countries differ markedly in the degree of concentration of their urban population (Table 3.10). Whereas Indonesia and Malaysia have well-developed urban hierarchies, this is less true of the Philippines and Vietnam and not true at all of Thailand, which is frequently cited as one of the world's outstanding examples of urban primacy.

Indonesia clearly has more options than Thailand for restraining the growth of the metropolis by fostering medium- and small-city development, because of the large number and substantial size of its intermediate cities. Moreover, in 1971-80, there was no strong trend towards urban primacy. The growth rate of cities over one million was not the highest; instead, cities of 200,000500,000 grew fastest. Not only this, but towns in every size category except 10,000 20,000 experienced growth rates exceeding 4 per cent per annum (Jones, 1988: Table 4). Implications of megacity growth in Asia are discussed by Brennan and Richardson (1989) and Jones (1988: 141-4). Although planners and politicians seem to be united in their distaste for the further growth of large cities, this is by no means the unanimous view in the academic debate. Why this divergence of attitudes?

TABLE 3.10 Indices of Concentration of Urban Population in the Largest City, in South-East Asian Countries and China. 1960, 1970 and 1980


Percentage of urban population Living in Largest City

4-city Primacy Indexa

1960 1970 1980
Thailand 41b 11.49 12.96 10.22c
Malaysia (Peninsular)d 29 0.73 0.88 1.39
Philippines 33 3.23 3.44 3.44
Indonesia 20 1.17 1.34 1.34
Vietnam 34 8.97e 4.60e 1.06
China 5 n.a. n.a. 0.43

Sources: Goldstein (1985: Table 7); Hugo et al. (1987: Table 3.15): Pernia et al. (1983: Tables 2.3 3.2; Thrifl and Forbes (1985: Tables 3-5).
a The ratio of the population of the largest city to the combined population of the next three larges cities in the same country.
b Revised estimates of the urban population (Robinson and Wongbuddha. 1980) have been used which increase the total urban population and reduce Bangkok's share.
c In computing this index for 1980. widened 'urhan planning area boundaries have been used for second. third and fourth largest cities, rather than municipal boundaries. The index would be even higher if municipal boundaries were used (UN/ESCAP, 1982b: 18-24).
d The figures for Jinjang and Petaling Jaya are added to the Kuala Lumpur figure in computing the 1960 and 1970 rates; whereas the metropolitan figure is used for 1980.
e The figures are for South Vietnam alone. The figures computed for all of Vietnam. comparable lo 1980, would be 2.07 for 1960 and 2.66 for 1970. These of course have little meaning as the country was divided.
n.a. = Not available.

Part of the bias against big cities is undoubtedly a reaction by the political elite, the bureaucracy and the upper-middle class in general to the problems they face as a result of the migration of poorer rural dwellers to the urban areas-problems including congestion, crime and potential political instability. Throughout South-East Asia, election results typically show that the strongest opposition to the government in power is found in the big cities where it is difficult to control. These cities are also the focus of riots and antigovernment demonstrations when they do occur. Through the Indonesian military's 'floating mass' strategies, political expression in rural areas can be held to a minimum. The heavy gerrymander in favour of rural electorates in Malaysia effectively neutralizes the electoral strength of the opposition parties in urban areas. Further growth in the size and relative share of big cities in the total population would weaken or place these strategies themselves in jeopardy.

The opposition of urban elites to metropolitan growth fuelled by migration of the rural poor is, of course, based on self-interest and should not be unquestioningly accepted as the basis of policy. Social justice requires that the fruits of metropolitan growth, even if that growth is accelerated by distortions in macroeconomic policy, be potentially available to poor rural dwellers. The more academically, if not politically, persuasive argument against rapid metropolitan growth is that 'siphoning off' some of the metropolitan growth into smaller cities and towns would lead to more effective linkages with rural areas and assist in the modernization and development of the countryside.

In the academic debate, widely divergent viewpoints are presented. Lipton (1977) and Todaro and Stilkind (1981), see an 'urban bias' in many social and economic policies, thus, leading to excessive rural-urban migration, often to the most favoured capital city. On the other hand, Vining (1985: 30), noting the statistical association between increasing primacy and faster economic growth, argues that 'the concentration of investment in the core region, which causes the ... [cityward] migrational flow, is the most efficient route to increased production. Investment can be diverted from the core only at the cost of retarding economic growth.' In similar vein, Mera (1978: 271), a strong advocate of regional laissez-faire, states: 'There is a fundamental conflict between high economic growth and decentralization of population. If a high rate of economic growth is to be achieved, further concentration of population into a few large metropolitan areas cannot be avoided.'

Implications of Megacity Growth

This view is supported by the finding that large cities are often more efficient and innovative than other urban centres. Several writers have sought to demonstrate that there is no 'optimum' city size beyond which further growth is undesirable, basing the argument on evidence that industrial productivity is highest in the largest cities, even when allowance is made for differences in capital per worker and size of enterprise (Alonso, 1968; Richardson, 1973). Data from China support the view that the returns on investment are higher in the metropolitan areas than in small towns or in rural areas (Kim, 1990). Based on such evidence. some economists have argued that government intervention in the distribution of economic activity is likely to waste scarce capital resources and thereby slow the rate of national economic growth. Thus, in the longer term, the country will be less able to redistribute income and solve the problem of poverty (Gilbert and Gugler, 1982: 176).

Even if it is true that national economic growth would be maximized by allowing the larger metropolises to grow to vast size, planners might nevertheless appropriately opt for slower national growth if rapid urban expansion meant seriously widening regional income disparities. It has been argued, for example, that Bangkok and the Central Plain region of Thailand is like an NIE (newly industrializing economy) set within an underdeveloped country. An argument against intervention to restrict the growth of urban agglomerations, however, is based on Williamson's (1965) finding that regional income disparities tend to first increase and then decrease as a more mature economic system evolves. In support of this finding, Richardson (1977) noted strong signs of what he called 'polarization reversal' in the urban systems of South Korea, Brazil and Colombia, though Cochrane and Vining (1988: 239-42) are more agnostic about recent trends in South Korea and Taiwan.

It is probably naive to expect 'polarization reversal' to happen automatically in most developing countries. Many countries may never reach the levels of income where regional income disparities tend to narrow; in such countries, these differences are greater than those characteristic of developed countries in the past. Convergence depends on effective government intervention, but many governments show little sign of interest or competence in remedying regional inequalities (Gilbert and Gugler, 1982: 177).

There may be reason, also, to doubt the evidence for the efficiency of large cities; for example, the diseconomies of metropolitan growth (such as dry-cleaning bills and cost of long journeys to work) tend to be counted in the regional income figures which are interpreted to demonstrate high productivity in the metropolitan area. 'Agglomeration economies' may derive from better urban infrastructure or higher-quality labour, but if equivalent infrastructure or labour were available in medium-sized centres, then the productivity of these centres might well rise. Additionally, high productivity among private firms in large cities may be more apparent than real because of indirect subsidization by the state. If the firms had to bear the full cost of the externalities they impose, they might find the large city less attractive and many would transfer the higher-productivity enterprises to intermediate cities, thereby reducing the apparent differential in industrial productivity between the smaller and the larger centres (Gilbert and Gugler, 1982: 177-8; Jones, 1990: 13). Moreover, many modern industries are located in large cities because of the advantages of better-quality infrastructure, import-substitution development strategies and a 'permitridden' environment requiring location near the sources of decision-making.

Given the theme of the book, the issue of environmental sustainability of megacity growth should also be stressed. Bangkok's problems of flooding and subsidence have been widely discussed, and Jakarta's serious problems of water supply, encroaching salinity and appropriate corridors for expansion are well covered in Douglass (1988). The economists reaction to such issues- that these diseconomies will in time be reflected in the locational decisions of individuals and firms and hence lead to slackened metropolitan growth-as an argument against government intervention to influence urbanization patterns holds only if the negative externalities are felt, and felt without much time lag, by those who create them. Otherwise, substantial social costs of excessive metropolitan growth will continue to be felt over a long period by those who are powerless to influence the trends.

Contents - Previous - Next