COMPACT TOWNSHIP, RURAL MIGRATION AND

URBANIZATION

Professor Salim Rashid and Shakil Quayes

Department of Economics
Box 34,1206 S.Sixth St, Champaign, IL:61820, USA.

There is no denying the fact that population growth in Bangladesh is still deemed to be a
major problem despite enormous success in the area of family planning. Over the last
two decades the rural population growth has been almost stagnant not because inadequate
biological reproduction but as a consequence of large migration to the cities and on the
other hand the country lacks urban infrastructure to support even the prevailing urban
citizens. At the same time, very little has been done, either in terms of expansion and
development of existing urban centers to accommodate for increasing population or
development of new urban centers. Existing cities have had to absorb this excess
population without being in a position to provide even the bare minimum facilities, like
shelter and clean water. Although the migrant workers enjoy a relatively higher income
this is certainly at the expense of having to live in less healthy habitat, than their rural
abodes. Given the fact that the prevailing rural areas would not be able to support any
additional population the crucial problem over the next two decades is that even the most
optimistic vision of urban growth in line with World Bank projections cannot account for
the nearly 35 million people. We can either allow the country to gradually move towards
an impossible situation or fundamental views regarding urban policy needs to be totally
overhauled. Instead of allowing the prevailing urban centers to bear the brunt of
migration we could develop small urban centers in the form of compact townships that
would not only absorb willing migrants but also encourage other rural households to
move to these compact towns. This will not only ameliorate the forthcoming crisis of
excessive burden on existing urban centers but also release precious agricultural land for
cultivation of crops.

Compact Township
A Compact Township (CT) is an agglomeration of Houses, hospitals, schools, markets,
rural industries and local governmental units that provide all basic services to a
population of about 20,000. It is to be largely self-governing and self-financing. The size
is small enough for traffic within the CT to be conducted by non-motorized vehicles and
for motorized traffic to be thus isolated from the CT itself, making it well connected yet
environmentally friendly. As the size is small enough to provide effective protection from
floods, the CT’s will permit Bangladesh to do with many fewer embankments and thus
encourage the re-emergence of a wetlands environment as well as serve to stimulate the
renewal of freshwater fisheries, a critical source of nutrition. A tentative plan is to have
about 200-300 acres for each such CT, such as the Mirpur Staff College or BARD in
Comilla.

Eventually some 7000 of these will become the basic rural landscape of Bangladesh,
supplemented by a scattering of homesteads and some remaining number of the existent

1
68,000 villages. The general idea of gaining economies by agglomeration is an extremely
simple one and has been the staple of several branches of Economics and Regional
Planning. The concept put forth here is simple but important, necessary and practicable
and provides a concrete shape to suit the particular current situation of Bangladesh.

There are some partial antecedents for CT’s even within BD in the cluster villages (for
the homeless) and in growth centers based on market towns. There are several academic
precedents: Sarwar Jahan’s Masters Thesis, (BUET 1978), and The Role of Small Towns
in Rural Development, by Toufiq M Seraj, NILG 1989, (from a Liverpool thesis) are two
that I have seen. A.T.M. Nurul Amin of AIT is having a concept called “densification”
developed by his students, that Lutful Huq of the Geography Department had discussed
such a concept in 1974, and according to Professor Rehman Sobhan, Mahbubul Aalm
Chashi had broached the idea for the cyclone prone coast in 1970. A M A Muhith had
espoused the idea of 'habitation centers' almost twenty years ago, and he also raised the
issue again in the recent International Conference on the Environment (January 2000).
The general idea of a CT is simply a combination of Integrated Rural Development with
the idea of a ‘Growth Pole ’, or its spatial equivalent, a ‘Growth Center’. Insofar as there
is a difference, and the originality of the concept of CT is that the earlier proposals are
either focused upon poverty or as an ancillary to some other plan of economic growth.
The principal difference lies in the conjoined scale, scope and purpose of the CT.

Planning has come into considerable disrepute in recent years, has always been
welcomed with regard to the expansion of education or of roads and infrastructure—so,
to be careful, one has to say that suitable planning of inputs is still a desirable option. The
question is whether the CT’s will in fact be stimulating the appropriate inputs.
Furthermore, all moves into the CT’s will be voluntary. No force will be used. The only
inducement will be an indirect one. No Government can be asked to support that which is
beyond its means. It can be asked to provide education and health, but only in an
affordable way. Those who choose to move into the CT’s will have easy access to
education, health, banking, family planning etc. Those who do not so move would still
eligible for such benefits, but it will be so much harder for them to gain access.

Motivation
According to the World Bank 1 booklet, BANGLADESH 2020, by 2020 another 45
millions will be added to the population and another 50 million to the labor force. In 40
years we will have to employ another 80 million---or another Bangladesh. The WB
emphasizes that Bangladesh must urbanize and indeed, rural BD has not grown in the last
decade. But the WB proposal seems a non-starter on the basis of its own facts. Consider
the description of Dhaka City today:

When only 20 percent of Dhaka’s daily output of solid waste is collected everyday and
the country’s total output is likely to rise eight to ten times by 2020, it is not difficult to
imagine cities overwhelmed and even fatally polluted by their own refuse. Dhaka has
only a single sewage treatment plant – a completely inadequate one – and a sewage
1
Whenever possible World Bank (WB) figures are used since everyone seems to rely on WB figures.

2
system that covers only one third of its area. Thirty percent of the capital’s people either
use open latrines or none at all.(p10)Urban demand for water is due to rise from 645
million gallons a day to 3,300-4,200 million gallons. Already only one half of the
country’s city dwellers have access to clean, safe drinking water.

More and more of that future will take place in cities and towns and in construction and
in the informal and service sectors. As a final component of the setting for growth, the
process of urbanization and its attendant activities have to be examined both for the
promise they hold and the threats they raise. In 1996 the urban population of Bangladesh
stood at 24 million, one fifth of the country’s population. By 2020 it is projected to be
nearly 80 million, not much short of half the total. Together Dhaka (with an estimated
15-20 millions inhabitants in 2020, compared to over 8 million now) and Chittagong
(doubling to 9-12 million by 2020) will be the megacity homes to one of every four
Bangladeshis. Another one of four will live in urban areas that are already sizable
(Rajshahi and Khulna) and in towns that may now be sleepy rural centers but that are
destined, because of their location on road and rail routes that are axes of development, to
boom (p10).

Anyone who has seen the reality unfold in Dhaka City will know that keeping Dhaka
livable with 15 million people is a sufficient challenge. Add to this another 10 million in
Chittagong. Even if the smaller cities grow to a million each and there are ten of them
and if rural BD remains at 100 million, it leaves 35 million people unaccounted for. The
WB does not face this issue squarely after having set up the facts to make it painfully
evident.

The solution by default is that things will get done when the situation becomes
sufficiently unbearable. Even though the WB is resolutely against such a policy it does
not have any concrete suggestions much beyond the Washington Consensus of “sound
fundamentals” in macroeconomics and privatization. We strongly believe that
governments must help coordinate the activities of its citizens and enable the growth of
self-governing civic bodies. In Bangladesh we have a rice-based economy and all rice
economies are crucially dependent on water, a resource that is scarce, mobile and
indispensable. Since markets can fail to work well with such common resource goods,
some communal ethic is frequently necessary for practical and effective solutions. Even
though Taiwan is a Private enterprise economy, it has one of the most successful
community-based water sharing schemes.

The WB does not face the historical point that growth occurs in agglomerations. There is
also a second point that may be relevant--- growth is typically focused upon a few years,
say a critical decade. One concentrated effort then sets the stage for the future. The CT’s
seem to be part of a solution, they are certainly not a solution by themselves, but
currently stands out as the only one due to lack of other alternatives.

First of all it is necessary that we be able to envision a future. In reading many a report
and talking to many an expert, it became clear that for almost fifty years the overriding
concern has been to avoid famines. Obviously a vision of future development cannot be

3
based on simply avoiding unforeseen calamities. We must try to visualize what
Bangladesh will look like, may look like, in another 20 years or 50 years and even
beyond. The Water Resource projects are then justified only by the role they will play
within this vision of the future. There is no reason why Bangladesh has to be self-
sufficient in rice while international trade opportunities allow us produce industrial goods
and import rice from other countries. Instead of pouring billions of dollars on ambitious
embankment projects we would be better off making the CT flood resistant instead. The
engineering solution proposed by the French Plan after 1988 was to spend was an
outrageous 6 billion dollars.

Secondly, it is imperative that a an integrated vision be emphasized at the expense of

detailed expertise. Research on Bangladesh has always been overwhelmingly project
based. However good the results of the individual project, when the project folds, perhaps
because donors have other interests, the good work dies. There is no cumulative impact.

Feasibility
This brings up the question of financing. The important point to bear in mind is that all
the benefits proposed such as Education or Medicine are no innovations. These promises
have already been made to the people. The expenditures upon them are due. The real
question is--- how can we effectively do what has already been promised? The exact
numbers provided below are not of primary significance, only the orders of magnitude
are.

If a four story building takes about $150,000 to build and a school/community center
takes about $100,000,then a total of 200 houses in a CT will require a total of $20
million. Once we add another 5 million for the accompanying infrastructure, each CT
requires about $25million for its construction.

Let us now consider the savings in roads, electricity and marketing that will arise from
the CT’s. One will not have the obligation to take electric lines into the interior at
government cost, and smaller village roads can be left in an unfinished state and preserve
the environment and the countryside. Suppose we did not extend electricity to villages or
pave their roads but instead tried to use the money to make CT's. The average kilometer
of rural electric line costs about $11,000 2 and rural roads cost between Taka 2.5-3.5
million3 if fully paved.

The number of kilometers of electric line per village actually varies quite a bit, from 7km
per village in Chittagong, to 1.8 km per village in Kurigram4 , with a mean around 3.5 km
per village. By accounting for the villages yet to be connected about 178,390 km's of
lines need to be laid, the total cost of which stands about $1.962 billions.

Now if we assume the (minimal) additional cost of developing each rural road is $20,000
per km. Since the lines go into houses while roads do not, we should probably reduce the

2
Courtesy of Kamaruzzaman and Colin Jack of USAID
3
Courtesy of Kashem of LGED (RRIMP2)
4
Rural Electrification Board Booklet 1997

4
length of electric lines previously calculated by 25% to give 133,800 km's of roads yet to
be developed. If this money is saved by the development of the CT’s then we get an
additional $2.68 billions. Only two items give us a potential savings of $4.5 billion that
can used instead in developing CT’s, about 200 of them. The difference between the two
patterns of expenditures is that the CT’s look directly to the future and hold out more
hope than the current scheme. Also quite possibly due to dispersed village homesteads,
Bangladesh has the highest density of rural roads in the world. 5

The WB sees a total expenditure of $325 billion over the next 25 years, or about $13
billion per year. The WB sees BD as having to attract a good deal of FDI to finance this
goal. The CT proposal is a distinct alternative.

Table 1. Investment requirement to year 2020 (in billion $)*

Sectors Annual Average Total
Social sector 1.0 25
Urban housing & infrastructure 4.0 100
Physical infrastructure 3.0 75
Industry & Agriculture 4.0 100
Environment & other 1.0 25
TOTAL 13.0 325
* Source:WB(p102)

If we subtract an arbitrary amount of $5 billion from the numbers above to finance the
CT’s, for 4000 CT’s till 2020 we need $100 billion. As those who move into the
townships will be expected to pay for their benefits this is an expense but it may not be a
cost as much of it will be recovered over time. The land can perhaps be acquired in
exchange for the lands of those who are moving in.

Of the 68,000 villages, each of the 13,817,000 households lives on 0.07 acres of
homestead on an average. Suppose only 80% of the villagers agree to enter the CT’s. and
assuming they need 0 .02 acres each in the CT’s, this will save 500,000 acres for fruits
and vegetables. Out of a total of 1,949,000 ponds we can free up about 1,500,000 ponds
for scientific fisheries. These are orders of magnitude that surely require closer
examination.

First of all, the CT's will the 35 million who are unaccounted for in the WB plan.
Although the major problem is said to be population control, we believe population
growth will not stop overnight However, fertility rate is already declining very sharply in
due more to women’s awareness and to social changes than to the variables so often
posited by economists. Finally, of all the variables used to predict actual family size, the
most effective by far is ‘desired family size’. Real change occurs when people change
their minds. Insofar as attitudes are mostly changed by social interaction, the CT’s will
provide the maximal potential for reducing future population

5
Bangladesh Rural Infrastructure Strategy Study, World Bank, 1997

5
Secondly, entrepreneurship is the driving force for growth in a Market Economy. Nothing
stimulates this more than the perception of high profits. It is hoped that the self-
government of the CT’s will serve to stimulate this last prerequisite. Indeed, tax
collection and local expenditures will be under the direct control of each township, and
these CT's can be stimulated into a friendly rivalry. If the central government focuses
upon making the principal Roads fast and safe these CT’s will be able to provide
nurturing places for the growth of subcontracting and rural based industries. China will
serve as a useful model.

Thirdly, Technology is the great force leading to continual productivity. However, the
potential of the small CT’s is hard to appreciate because of some misconceptions about
Technology. It is assumed that Technology is largely dependent upon Science, especially
basic Science. However, according to such authorities as Solla Price, it is arguable that
Science depended more upon Technology than vice versa. Cumulatively, these small
improvements made an enormous difference. The reluctance to consider small,
specialized and flexible production units is based upon the belief that only large scale
production can be efficient. However, the secret of Japanese JIT production lies in the
ability to adapt the workplace to skilled, cooperative workers who take participatory
pride in what they do.

In other words, the CT’s can reasonably be expected to provide a stimulating nursery for
the most important inputs involved in economic growth---education, entrepreneurship,
and technology. The data to support such a proposal are yet to be gathered. Current data
either do not address these issues or fail to take into account the large scale nature of the
cumulative and agglomerative effects envisioned. There are significant differences in the
demand for marketed products between rural and urban households. This suggests that
the agglomeration produced by the Compact Townships has the potential of increasing
domestic demand considerably.

Using Household Expenditure Survey we now show the significant differences in the
pattern of demand for marketed products between rural and urban households. There are
numerous examples of studies analyzing as to why rural people migrate to the cities and
never come back. However the question which has seldom (if ever) been answered as to
what prevents all the other rural people to migrate to the cities also, and what would be
the consequence there of. Innate reluctance to leave the ancestor's home, uncertainty in
urban employment and lack to funds necessary to successfully migrate are obvious
reasons that come to mind. But the fact that adequate incentive in terms of better
economic and social conditions have been luring relatively wealthy and the not-so-well-
to-do towards the city is ample proof reasonable incentives can increase the rate of this
migration. The concept of CT tries to answer whether there should there be a policy of
actively encouraging migration to the cities, and how would these people get adequate
accommodation.

We assume that wages are higher in the cities, living standards are better in the cities, the
young worker in the countryside has a natural tendency to migrate to the cities, and urban

6
employment growth is much higher than the stagnant employment opportunities in the
rural area. Therefore mass transfer of rural population to the cities is a mere eventuality.
In this context we make an attempt to analyze the profiles of the rural household and the
urban household side by side to detect and identify noticeable differences.

Data and Variables

The data source is the Bangladesh Household Expenditure Survey (HES) 1995, collected
and compiled by Bangladesh Bureau of Statistics. The HES sample consists of 7420
households in 371 unions (there are twenty sample households taken from each union)
from all 22 regions of Bangladesh. We retained 3296 rural households and 1296 urban
households.

Consumption expenditures are computed for three types of commodities: farm goods
( C ), market goods ( M ), and leisure ( L ). There are 20 households in each union. First
calculate the average unit price of an item for each union. Now we use these prices to
compute a price index for own goods and market goods for each of the households. The
unit of measurement was one kilogram or one liter in most cases except for dozens for
eggs, yards for cloth and some other standard variations.

Rural Urban Differences

We use a Stone-Geary linear expenditure system (LES) to estimate the household
demand for farm goods, market goods and leisure. We assume that the household utility
function is identical for each member and additive across individuals and hence total
household consumption problem is to maximize per capita consumption of C, M, and L.
In addition to prices and wage, other variables which affect household decision are
household characteristics such as number of members in the labor force ( n1 ) , number of
dependents ( n 2 ) and education level of household head (e) .

Assuming that the model fulfills the standard assumptions of a regression model we
estimate the above nonlinear seemingly unrelated regressions equation system using
maximum likelihood estimation. Coefficient estimates that are not significant at the five-
percent level of were dropped from the model and hence all elasticities listed in the
following tables are significant. Elasticities of household consumption of farm goods and
market good, and supply of labor, are calculated at the arithmetic mean of the variables
on the basis of keeping total expenditure exogenous (i.e., expenditure is assumed to be
independent of w and n1 ). Constant expenditure mainly implies that we do not allow
family expenditure to reflect changes in wage rate or in the number of labor force in the
household. 6

Labor supply in the urban area is highly elastic to changes in the wage rate. A 1-percent
increase in the wage rate is accompanied by 1.31-% increase in labor supply by the urban
households and by 0.91% increase in the labor supply for the rural households. The effect
of a change in the wage rate and the change in the family labor force is better reflected as

6
Estimates of parameters and corresponding elasticities are reported in dissertation of Quayes.

7
in elasticities 1 when we allow total expenditure to vary. An increase in the wage rate has
very little effect on the household's consumption of farm goods but increases the
consumption of market goods by 0.10 for rural households and by 0.18 for urban
households. With an increase in wage rate there is a slight decrease in labor supply by the
rural households while urban households increase their labor supply by 0.03% as a result
of a 1-percent increase in wage rate. The urban households are characterized by
exhibiting a higher price elasticity of farm goods, lower price elasticity of market goods
and a more elastic labor supply compared to the rural households. Not surprisingly,
elasticities change when we allow total expenditure to be affected by wage rate and
number of members in the labor force in the household.

Rural
Households

Urban
Households

Over the short run household characteristics such as size, labor force and number of
dependents remain exogenous even at the macro level. At the micro level, wage rate is
the only variable, which can react to changes in these exogenous variables reflecting
interdependence at he macro level. Assumption of representative household at the micro
level allows us to obtain macroeconomic relationships, by multiplying the corresponding
microeconomic relationships by the total number of households. We also assume that the
interdependence of exogenous variables at the macro level is only reflected between
wage rate and the other exogenous variables. In a competitive labor market wage rate is
determined by the interaction of labor supply and labor demand. The equilibrium wage
rate for the agricultural labor can be determined by equating the demand for labor and
supply of labor for a representative household.

The rural market response and urban market response are calculated separately on the
basis of the assumption that changes in the pattern of demand for farm goods, market
goods and labor in one sector does not directly affect the equilibrium price in the other

8
sector. However changes in one sector will affect the other sector indirectly via
production and labor migration. We also assume that elasticity of labor demand is
identical for the all the industries (especially for the agriculture sector and the
manufacturing sector). Although using estimated demand elasticities for market goods
would be ideal we have made the assumption so that we may look at the differences
between urban and rural households mainly on the basis of their demand pattern.
Figure 1 depicts the pattern of market goods consumption of rural and urban households.
The horizontal axis represents price and the vertical axis represents quantity. The
smoothed curves represent plots of household consumption of market goods against the
respective price index for the household. The lower curve representing the rural
households shows that at each given price level, urban households consume a greater
amount of market goods than the rural households do on an average. The rural price-
market goods curve also indicates higher price elasticity than the urban demand curve.

The major difference between the consumption profile of rural households and the urban
households is in their consumption of market goods. Obviously we would expect the
relatively more affluent urban to indulge in higher consumption of market goods. While
the consumption of farm for the median urban household is only 16% higher than the
consumption of a median rural household; the median urban household consumes nearly
two and half times (233%) of the market goods consumption of the median rural
household. Needless to mention that the median wage rate for the median urban
household is more than two and half times that of the median rural household Hence the
only important transition for a migrant rural household to get gradually allocate a higher
fraction of income towards market goods. A simple implication of such migration is also
an impetus to the production of market goods, which generally have a large
manufacturing component. Figure 2 shows us what we already know, i.e., the average per
capita expenditure for urban households is higher than that of the rural households.
However we do notice that the variance of per capita expenditure for urban households is
comparatively larger than the variance for rural households.

9
Average daily consumption is 2659 in the rural area and 2558 in the urban areas. The
bottom 20-percent amongst the rural households has an average daily consumption of
2012 calories while the top 20 percent consumes 3349 calories. In the urban region the
bottom 20-percent consume a daily intake of 2069 calories while the top 20-percent
consume 3109. Therefore, if the rural households migrate to the cities their per capita
expenditure will go up but there won't be any significant change in their nutritional
intake. But we do notice that differences exist in the expenditure elasticity of calories
between urban and rural households. At the median rural households have a higher
elasticity (0.37) than the elasticity (0.23) for urban households. This means that when the
rural household migrates to the city its consumption of market goods will increase
substantially. The market for manufactured goods would expand and a side effect of the
migration would stimulate the market goods industry.

Traditional estimation of the cost of migration on the rural economy involves accounting
for the loss of marginal product due to the drop in labor supply. In fact trying to calculate
the loss in terms of lost productivity would underestimate the true cost since other family
members put in more effort following the departure of a family member. Ahmed (1974)
uses average family product to calculate the personal cost of migrating from the village.
For the rural areas a 1% decrease in labor force results in 0.83% reduction in the labor
supply, i.e., remaining members of the household put in extra effort to make up for 0.17%
of loss in productivity. So the loss incurred in terms of productivity in the rural sector due
to the outmigration of one worker would be equivalent to 83% of his marginal product of
labor. For a median wage earner in the rural sector this loss can be quantified in monetary
terms if we multiply this loss by the median wage rate prevailing in the rural market. The
median wage rate in the rural sector is approximately Taka 33 per day. In table 4 we have
expressed the wage rates as a percentage of average national wage rate, calling it wage
index. The wage index for the median wage earner in the rural sector is 66. If we multiply
66 by 0.83 it will give us the productivity loss in the rural sector due to the migration of a
median worker to the urban area. If this worker obtains employment in the urban sector
and earns equal to a wage index of 55, there is no loss in overall economic productivity.
However we have to account for the fact that every addition in the urban labor force does

10
not obtain employment. But since the median wage index for the urban sector is 175,
even if one third of the migrant workers obtain employment in the urban region, the
economy as a whole will have compensated for the loss in rural production (one third of
173 is still greater than 55). In fact a median age earner in the rural region migrates to the
cities and obtains employment at the bottom 10-percentile wage rate (wage index = 58) of
urban workers, there still would not be any loss of productivity for the economy as a
whole. In the urban sector, an addition to the family labor force increases the labor supply
by 0.68 units. So the addition of each rural migrant adds an amount of productivity equal
to 118 (173 multiplied by 0.68), if he is able to obtain employment at median urban wage
rate. If workers earning median wages in the rural sector migrate and obtain employment
at median urban wages, on an average productivity gains per migrant worker will be
equal to 63 in terms of wage index. As a percentage of the labor productivity of the
median migrant worker, this is a productivity gain of 95%. If rural workers earning wages
at 25-percentile migrate to the city and obtain urban employment at the 10-percentile
wage rate only, even then the net effect will be productivity gains equal to 18% of their
marginal product of labor in the rural sector.

We assume adequate availability of employment. Therefore the level of employment is

meant to depend solely on the labor supply of the household. Also the household labor
supply is an observed phenomenon, hence it already reflects the equilibrium in the labor
market. Notwithstanding the demand for labor and implicitly assuming that labor market
clears, the gain for the urban economy due to immigration of one unit of labor is 214% of
the loss incurred by the rural economy for the outmigration of the said unit of labor.

Table 2. Impact of Rural Urban Migration in terms of marginal product of labor.

Percentile Rural Wage Rural Loss Urban Wage Urban Gain Net Gain
10 25 21 58 39 75
25 39 32 96 65 85
50 66 55 173 118 95
75 119 99 331 225 105
90 248 206 571 389 74

Table 5 describes lateral comparison across various wage earner groups and shows that
productivity gains in urban regions far outweigh the loss in the rural economy. Wages are
expressed as an index, where the average wage rate of Bangladesh has an index of 100.
 Wage Rate 
Wage Index =   * 100
 Aberage Wage Rate for Bangladesh 
Rural loss is calculated by multiplying wage index by the labor force elasticity of labor
supply. Similarly urban gain is computed by multiplying the wage index by the elasticity
for the urban area. The net gain measures the net productivity gain as a percentage of the
rural wage of the migrating worker.
 Urban Gain - Rural Loss 
Net Gain =   *100
 Rural Wage Index 
Table 2 is constructed on the very simple assumption that migrant workers are able to
obtain employment maintaining their relative income status in the society. In other words

11
a worker earning wages at the 25-percentile in the rural sector will also be able to earn at
the 25-percentile urban wage rate. Although it is a fairly strong assumption it still serves
the simple task of finding the cost/benefit of migration in elementary terms. For the
bottom 10-percentile the productivity gains from migration is 75%. The productivity
gains are even higher for the bottom 25-percentile wage earners at 85% and highest for
the 75-percentile at 105%. Even for the top 10-percentile rural workers, migration will
yield a productivity gain of 74%.

We need to look at Table 2 to have a birds eye view of the change in productivity that
ought to follow after mass migration of rural people to the urban centers. The picture tells
us that the overall productivity of the economy can only go up.

The consumption pattern of the urban households and the rural households are quite
similar except for the market response in the demand for non-farm goods. This is an
indication of transition of the rural household in consuming more non-farm goods when
they migrate to the cities (existing city dwellers) consume more non-farm goods than the
rural households. There are a few areas where regional differences exist; however the
large productivity loss associated with rural workers leaving combined with small
productivity gains in urban employment deeming net gains to be substantially smaller
than all the other divisions was quite surprising.

The goal of this exercise of analyzing the consumption profile of the rural and urban
household and the prevailing market response in these two areas, was an attempt to
identify the differences. This would allow us to predict the future course of household
behavior when a large fraction of the rural population migrates to the cities.

To facilitate the absorption of the increasing population from the country side, newer
urban centers need to develop, either as a natural course of market-demand for such new
townships or by way of partial government intervention in the form of supplying such
new townships and creating the demand for such. As we have already mentioned in the
beginning, adequate incentives will induce rural people to migrate to urban centers. These
incentives have to be a lot less if these urban centers are in close proximity to their
ancestor's home.

If the government or private developers build small local townships with adequate
municipal facilities (like electricity, sanitation, schools and medical services) all over the
country it would reduce the burden of immigration to the existing large cities. This by
itself will free up resources that are being used in trying to facilitate the burden of excess
population on the existing towns. Second, it will release precious land (used for habitat
and small dirt roads needed for the current dwellers) and ponds (that are currently being
used for bathing and other purposes) if a large fraction of rural people move to these
newly developed townships. This will increase the production of corps and fish at the
same time. Third, it would allow for integration of smaller land holdings into larger plots
allowing for more efficient use of modern technology. This might have adverse effects
on the land ownership profile but eventual urbanization also means that a smaller and
smaller fraction of the population will be engaged in agricultural activity. Finally, it

12
improves the living conditions of the rural people and accelerates and completes the slow
but gradual transition of rural-urban migration. This also reduces the total social cost of
migration cost. This also is devoid of the cost of reduction in the agricultural labor force
(the source of a major concern) since all farm workers remain in the vicinity and are
actually able to participate in a larger labor market (efficient for both the employer and
the employee)

The analysis of the profiles of consumption and market response at the national level and
at the regional level show that some differences exist across rural and urban households,
and the new townships will possible exhibit a profile which is a convex combination of
the rural and urban profiles. At first glance the small existing differences would not cause
huge transitional costs on the moving households, although we shall only be able to
quantify such social costs after the implementation.

References
Ahmed, Fakhruddin. 1974, " Migration and Employment in a Multisector Model--An
Application to Bangladesh." Michigan University Microfilms.

Barnum, Howard, and Lyn Squire. 1979, “An Econometric Application of the Theory of
the Farm Household.” Journal of Development Economics, vol.6, pp. 79-102.

Barnum, Howard, and Lyn Squire. 1979, A Model of an agricultural Household.

Washington , D.C.: World Bank

Deaton, Angus. 1974, "A Reconsideration of the Implications of Additive Preferences,"

Economic Journal, vol. 84, pp. 338-348.

Jorgenson, Dale, and Lawrence Lau. 1969, "An Economic Theory of Agricultural
Household Behavior." Paper read at the 4th Far Eastern Meeting of the Econometric
Society.

Mortuza, Syed Ali. 1992, "Rural-Urban Migration in Bangladesh." Dietrich reimer

Verlag, Berlin.

Shukla, V, and O. Stark. 1985, "On Agglomeration Economics and Optimal Migration."
Economic Letters, vol. 18, pp. 297-300.

Williamson, Jeffrey. 1988, "Migration and Urbanization." Handbook of Development

Economics vol. I, pp. 425-465

13