8.

Rural Migration and Urbanization in Developing Countries:

About half of the world's population lives in cities. But by 2025, nearly two-thirds will live
in urban areas. All of the fastest-growing cities are found in the developing world. By 2015, seven
cities will have grown to more than 20 million inhabitants. All but one of these will be found in the
developing world. These are some of the findings of the UN, presented to the Habitat II conference
in Istanbul in 1996. (For a full discussion of the challenges of urbanization, see Todaro and Smith,
chapter 8.)

Urban population growth in the developing world is far more rapid than population growth
generally; about half the urban growth is accounted for by migrants from the rural areas. Cities in
the developing world are growing far more rapidly than those in developed countries. What drives
migration? What effects does this have? The cases of India and Botswana are instructive in showing
the value of the development theory of migration, as well as some areas where it needs to be
extended.

Relative rates of urbanization are not higher in developing countries today than they were
in the period of rapid urbanization of today's developed countries. But in today's developing
countries this migration is different for at least two important reasons. First, is much higher in terms
of absolute numbers of migrants, and second, in recent years it has been taking place, in many
cases, against a background of stagnant living standards rather than one of growth. Problems
caused by urbanization like stress on infrastructure and other urban services are likely greater.

Shantytowns and similar makeshift settlements represent over one-third of developing

About half of the urban labor force works in the informal sector of low-skilled, low-
productivity, often self-employed jobs in petty sales and services. Still, this sector may generate up
to a third of urban income, and features a low capital-intensity, low-cost training, waste recycling,
and creation of surplus as well as employment creation.

Any economic or social policy that affects rural and urban incomes will influence migration;
this in turn will affect sectoral and geographic economic activity, income distribution, and even
population growth. Before the Todaro and Harris-Todaro migration models were introduced,

74
migration was widely viewed as irrational or driven by noneconomic motivations, sometimes termed
the "bright city lights." Noneconomic factors influence migration decisions, but economic factors are
now understood to be primary.

In the economic version of the bright city lights theory, people rationally migrated on the
basis of costs and benefits. In this approach, it was assumed that if migrants appeared to be worse
off, this was because other benefits were being overlooked, with the effect of making the migrants
feel better off (or raising their overall utility).

The Todaro migration models postulate that observed migration is individually rational and
that migrants respond to urban-rural differences in expected rather than actual earnings. Urban
modern sector earnings are much higher than rural earnings, which may in turn be even higher than
urban traditional sector earnings. Migration occurs until average, or expected rather than actual
incomes, are equal across regions, generating equilibrium unemployment or underemployment in
the urban traditional sector.

The extension of the model to consider equilibrium and effects of actions like increases in
wages and probability of employment in the urban areas, undertaken by Harris and Todaro, shows
that under some conditions, notably elastic supply of labor, creation of employment opportunities
in cities can actually lead to an increase in unemployment by attracting more migrants than there are
new jobs.

Despite being individually rational, extensive rural-urban migration produces social costs
including lowered output and strains on limited infrastructure. In addition to its costs to crowded
cities, excessive migration also imposes external costs on the rural areas emptied of better-
educated, more venturesome young people as well as external costs on urban infrastructure and lost
output. Relevant migration and employment policies include an emphasis on rural development, rural
basic needs strategies, elimination of factor price distortions, and appropriate technology choice as
well as appropriate education. Each is intended to increase the incentives for rural residents to
remain in rural areas rather than migrate to cities.

Migrants tend to be younger and better educated than non-migrants, and come from all
socioeconomic strata. Almost by definition they seem to show above-average entrepreneurial drive.
Thus, paradoxically, migrants seem to have slightly better prospects in the rural areas than those
who do not migrate.

75
 India provides an interesting setting for a case study, because future urban migration is
potentially so vast, and because a number of interesting studies have been undertaken there.
Botswana offers a good counterpoint, because it has better published data and more advanced
statistical analysis of that data has been undertaken there than for most developing regions.

One of the best detailed studies of the economics of rural-urban migration, providing some
tests of the Todaro migration models and depicting the characteristics of migrants and the migration
process, is Biswajit Banerjee's Rural to Urban Migration and the Urban Labor Market: A
Case Study of Delhi.

Everyone who has been to a major city in a developing country has noticed the sharp
inequality between those with modern sector jobs and those scratching out a living in the informal
sector. But can the informal sector be seen as a temporary way station to the formal sector, or can
the barriers between these sectors be explained by education and skill requirements that informal
sector workers cannot hope to meet? Banerjee found that the idea of segmented formal-informal
rural labor markets could be substantiated statistically. After carefully controlling for human capital
variables, Banerjee was still left with earnings in the formal sector 9% higher than in the informal
sector that were not explained by any standard economic factor.

We should not automatically assume that when we observe those with more schooling
receiving higher earnings that this is a direct reflection of more skills learned in school. It may be that
employers simply hire more educated workers as a way of selecting among their many job
applicants. It should be noted that there are differences in wages across industries even in a
developed country such as the United States. Even so, the earnings differences found in India were
not nearly so dramatic as implied in some of the migration literature.

In the literature on urbanization, and in the anecdotes told about the informal sector by
lecturers on development the world over, the typical laborer is characterized as self-employed or
working on some type of piecework basis. But Banerjee found that only some 14% of his informal
sector sample worked in nonwage employment. Interestingly, average monthly income of nonwage
workers were some 47% higher than those of formal sector workers.

Banerjee argued that entry into nonwage employment was not easy in Delhi. Some activities
required significant skills or capital. Those that did not were often controlled by cohesive
"networks" of operators that controlled location of activities in various enterprises. Entry barriers
to self-employment in petty services are probably lower in other Developing country cities.

76
 Consistent with these findings, Banerjee found that mobility from the informal to the formal
sector was low: there was little evidence that more than a very small minority of informal sector
workers were actively seeking jobs in the formal sectors; and only 5 to 15% of rural migrants into
the informal sector had moved over to the formal sector within a year.

Moreover, the rate of entrants into the formal sector from the informal sector was just one-
sixth to one-third that of the rate of direct entry to the urban formal sector from outside the area.
Informal sector workers tended to work in the same job almost as long as those in the formal
sector; the average informal sector worker had worked 1.67 jobs over a period of 61 months in
the city, while formal sector workers averaged 1.24 jobs over an urban career of 67 months.

Banerjee's survey data suggested that a large number of informal sector workers had
migrated to the city attracted by the informal rather than the formal sector, coming to work in such
occupations as domestic servants, informal construction laborers, and sales workers. Of those who
began nonwage employment upon their arrival, 71% had expected to do so. The fact that only a
minority of informal sector workers continued to search for formal sector work was taken as further
evidence that migrants had come to Delhi expressly to take up informal sector work.

One reason for this focus on the informal sector was concluded to be the lack of contacts
of informal sector workers in the formal sector. About two-thirds of direct entrants into the formal
sector and some 60% of those switching from the informal to the formal sector found their jobs
through "contacts." This overwhelming importance of contacts was taken to explain why some 43%
of Banerjee's sample migrated after receiving a suggestion from a contact, which suggests that job
market information is often available to potential migrants without their being physically present in
the city. An additional 10% of the sample had a prearranged job in the city prior to migration.

Finally, there is on average a very short spell of unemployment following migration. Within
one week, 64% had found employment, and although a few were unemployed for a long period,
even averaging these in, the average waiting time to obtain a first job was just 17 days.

Banerjee also found that migrants kept close ties to their rural roots. Some three-quarters
of the migrants visited their villages of origin and about two-thirds were remitting part of their urban
incomes, a substantial 23% of income on average. This indicates that concern for the whole family
appeared to be a guiding force in migration. It also suggests a source of the rapid flow of job
market information from urban to rural areas.

77
 In a separate study, A. S. Oberai, Pradhan Prasad, and M. G. Sardana examined the
determinants of migration in three states in India—Bihar, Kerala, and Uttar Pradesh. Their findings
were consistent with the ideas that migrants often have a history of chronic underemployment before
they migrate, migrate only as a measure of desperation, and with the expectation of participating
in the informal urban sector even in the long run. Remittances were found to be substantial and
considerable levels of return-migration were also documented, among other evidence of continued
close ties of migrants to their home villages.

Of course, Delhi in particular and India in general may be different from other developing
areas. As a whole, these two studies have been taken as a challenge to the applicability of Harris-
Todaro or other "probabilistic migration models," at least in the case of India, and suggest that they
need to be extended to accommodate the apparently common pattern of migrating with the ultimate
aim of urban informal sector employment.

But Banerjee's findings, fascinating as they are, do not necessarily represent as strong a
case against the Todaro migration models as some (including Banerjee) have made them out to be.
As Ira Gang has noted, one can modify the model to include in the urban area not only a formal
sector, but a high paid informal sector, as well as a low-paid (or unemployed) sector. In this case,
people will migrate looking for either a formal sector job or a high-paid informal sector job. This
seems to be consistent with Banerjee's evidence. The assumption that keeps the essence of the
Todaro models intact is that the wage of the formal urban sector exceeds the high-paid informal
wage, which in turn exceeds the agricultural wage, and in turn exceeds the low-paid informal (or
unemployed) wage. The particular formulations of the Todaro models are really no more than
examples of a general principle—that migrants go where they expect in advance to do better, not
where they do better after the fact. The essence of the ideas of the Todaro models do not depend
on a particular notion of an informal or a formal sector.

Oded Stark's ideas on a family's use of migration can be a useful supplement to the Todaro
models, and may apply to some of Banerjee's findings. In this view, a family will send members to
different areas as a "portfolio diversification" strategy, to reduce the risk that the family will have no
income. This approach is useful to explain any observed migration from higher- to lower-wage
areas, and into higher-wage areas, but not necessarily the area with the highest expected wage. The
basic idea of the Todaro models still applies, but this approach looks at families rather than
individuals, and stresses risk aversion. Some evidence for this for the case of Thailand has been
identified by Anna Paulson.

78
 In other studies, the Todaro migration models have held up well without modification in
other parts of the world. The survey by Dipak Mazumdar showed that the evidence is
overwhelming that migration decisions are made according to rational economic motivations. But
most of these have used macro-level data not really designed to address the Todaro hypotheses
directly.

A study of migration behavior conducted by Robert E. B. Lucas in Botswana addressed

Rural migrants in Botswana move to five urban centers (they would be called towns rather
than cities in many parts of the world) as well as to neighboring South Africa. Lucas found that
unadjusted urban earnings are much higher than rural earnings—68% higher for males—but these
differences become much smaller when schooling and experience are controlled for.

Lucas' results confirm that the higher a person's expected earnings and the higher the
estimated probability of employment given a move to an urban center, the greater the chances that
the person will migrate. And the higher the estimated wage and probability of employment for a
person in his or her home village, the lower the chances that the person will migrate. This result was
very "robust"—not sensitive to which subgroups were examined or the way various factors were
controlled for—and statistically significant. It represents clear evidence in support of Todaro's
original hypothesis.

Moreover, Lucas estimated that at current pay differentials, the creation of one job in an
urban center would draw more than one new migrant from the rural areas, thus confirming the
Harris-Todaro effect. Earnings were also found to rise significantly the longer a migrant had been
in an urban center, holding education and age constant. But the reason was because of increases
in the rate of pay rather than in the probability of modern sector employment.

Taken together, the best-conducted studies of urbanization confirm the value of the Todaro
migration models as the best place to start in seeking explanations of rural to urban migration in
developing countries. But these studies underscore the need to expand these explanations of

79
migration considering that many today migrate to participate in the informal rather than the formal
urban sector.

Workers who appear underemployed may not consider themselves as such, may perceive
no possibility of moving into the modern sector, may be unable to effectively search for modern
sector work while busy employed in the informal sector, and thus do not create as much downward
pressure on modern sector wages as it would first appear. This may be one factor keeping modern
sector wages well above informal sector wages for indefinite periods of time despite high measured
urban underemployment.

80
Sources

Banerjee, Biswajit. Rural to Urban Migration and the Urban Labor Market: A Case Study of
Delhi. Bombay: Himalaya Publishing House, 1986.

Banerjee, Biswajit. "The Role of the Informal Sector in the Migration Process: A Test of
Probabilistic Migration Models and Labour Market Segmentation for India." Oxford Economic
Papers 35 (1983): 399-422.

Cole, William E., and Sanders, Richard D. "Internal migration and urban employment in the Third
World." American Economic Review. 75 (1985): 481-94.

Corden, W. Max, and Findlay, Ronald. "Urban Unemployment, Intersectoral Capital Mobility and
Development Policy." Economica. 42 (1975): 37-78.

Gang, Ira N., and Shubhashis Gangopadhyay. "A Model of the Informal Sector in Development.
Journal of Economic Studies 17 (1990): 19-31.

Gang, Ira N., and Shubhashis Gangopadhyay, “Optimal Policies in a Dual Economy with Open
Unemployment and Surplus Labour,” Oxford Economic Papers, 39 (1987) 378-387

Habitat (UNCHS), State of the World’s Cities, 2001, http://www.unchs.org/Istanbul+5/86.pdf

Harris, John, and Michael P. Todaro. "Migration, Unemployment and Development: A Two-Sector
Analysis. American Economic Review 60 (1970): 126-42 .

Lucas, Robert E. B. "Emigration to South Africa's Mines." American Economic Review 77

Lucas, Robert E. B. "Migration amongst the Batswana." Economic Journal 95, (1985): 358-82.

Mazumdar, Dipak. "Rural-urban migration in Developing Countries." Handbook of Regional and

Oberai, A. S., Pradhan Prasad, and M. G. Sardana. Determinants and Consequences of

81
University Press, 1989.

Paulson, Anna, "Insurance Motives for Migration: Evidence from Thailand," manuscript,
Northwestern Univ., 2000. ).

Oded, Stark. The Migration of Labor. Cambridge, MA: B. Blackwell, 1991.

Todaro, Michael P. "A Model of Labor Migration and Urban Unemployment in LDCs," American
Economic Review 59, no. 1 (1969): 138-48.

United Nations. An Urbanizing World: Global Report on Human Settlements. Report presented
to the Habitat II conference, Istanbul, 1996.

United Nations Population Division, World Urbanization Prospects: The 1999 Revision, New
York: United Nations Population Division, 2000. Accessed at:
http://www.un.org/esa/population/publications/wup1999/urbanization.pdf

82