Re-examining the Resettlement Question in the Context of Migrant Women's...
Re-examining the Resettlement Question in
the Context of Migrant Women's Security
and Urban Poverty in Bangladesh
RITA AFSAR
The growth of shanties, slums and squatter settlements is inextricably linked with
rapid urbanisation and the emergence of mega cities in the developing countries.
Simultaneously, resettlement featured prominently in the agenda for poverty
alleviation programmes in many of these countries, particularly in Bangladesh.
Draconian measures are also taken to remove the migrant poor from the centre of
the city, both by the government and the vested interest groups with a view to
taking control of the land. Such measures could neither resettle the poorer people,
nor reduce urban poverty. Rather, it has more negative implications for poorer
women than men, as it threatens their security and job opportunities, more than
anything else. The paper argues that to be meaningful, poverty reduction pro-
grammes in urban areas must address the issues of women's security, identity and
the protective umbrella of social networks, work opportunities and must involve
poorer communities as partners while implementing housing and shelter pro-
grammes.
Dr. Rita Afsar is a Research Fellow at the Bangladesh Institute of Development
Studies, Dhaka, Bangladesh.
Introduction
Rapid urbanisation and widespread urban poverty are the emerging
features of many third world countries. The number of people living
in the urban areas of the developed regions doubled from 452 million
to about 894 million and in developing world, it increased six times
from 285 million to 1.7 billion over the last four and a half decades
(1950-1995) (Afsar, 1995 :15; United Nations, 1995a). Nearly half of
the urban population is living in slum and squatter settlements in many
third world cities. More than a quarter of the inhabitants in most mega
cities live in poverty (Oberai,1993:xi). Some 300 million urban dwell-
ers in developing countries currently live in poverty. The number of

Re-examining the Resettlement Question 825
poor urban households living in absolute poverty in developing coun-
tries is projected to increase from 40 to 72 million between 1980 and
2000 and that of poor rural households to fall from 80 to 56 million
during the same period (United Nations Development Programme,
1990). For the developing world as a whole, internal migration and
reclassification of urban areas accounted for a large part of urban
growth in the 1980s (54 per cent), but its contribution varied consid-
erably from one region to the another. In South Asia and Southeast
Asia for example, the rural-urban migration and reclassification of
urban areas together contributed to the urban growth which increased
from 44 and 51 per cent in 1980-85 to 46 and 55 percent in 1990-95,
respectively. It is likely to increase further to 53 and 58 per cent
respectively for the two regions between 2000 and 2005 (Afsar,
1995:34). Often rural-urban migration is seen as a means by which
rural poverty is transferred to urban areas thus posing a threat to urban
stability. Draconian measures are adopted to remove migrant poor
from the city centre to the periphery in order to discourage migration.
Resettlement of slum dwellers emerges as one of the most commonly
adopted strategies in many developing countries and Dhaka, the capital
city of Bangladesh, is no exception.
Dhaka can be described as a city of migrants like any other cosmo-
politan city, as more than three quarters of its population constitute
migrants predominantly from rural areas. Here too poverty co-exists
with prosperity (Afsar, 1995:72). Nearly half of Dhaka city's popula-
tion live below the poverty line income and this is manifested in a
number of relative deprivation indicators such as large proliferation of
slum and squatter settlements; co-existence of kutcha types of houses
along with skyscrapers; a stream of half naked, barefoot, illiterate and
malnourished children compared to highest concentration of well-
groomed children going to English medium schools. Poorer migrants,
who came predominantly in search of job or better employment
opportunities, are housed in more than an estimated 3007 slum and
squatter settlements in the Dhaka City (Asian Development Bank,
1996:4). Ever since the emergence of Dhaka as the capital of an
independent country, a series of resettlement and eviction attempts
were made under almost all political regimes, starting from 1974-75
till 1996.
The ideology behind the resettlement strategy was essentially de-
rived from the theory of 'culture of poverty' that perceived poor
migrants as evils, misfits and intruders to the city. Hence, instead of

826 Rita Afsar
removing poverty, poor people were uprooted from their workplaces
and from the protective umbrella of their kinship networks. Such a
measure has more negative implications for poorer women than men,
as it threatens their security the most.
The present paper examines, critically, the gender dimension of
urban poverty in the migration and settlement process in Bangladesh
in the first part. The whole issue of the growth of slum and squatter
settlements, their functional role in housing poor men and women is
analysed in the second part. Finally, the third part is devoted to the
analysis of emerging problems of eviction and resettlement and how
it affects men and women. This is followed by examining the impli-
cations of the findings on policy options in the context of rapid
urbanisation, challenges of urban poverty and women's security.
Data, Sample and Methods
A wide range of methods and sources are used for drawing observa-
tions for the present paper. These include a thorough review of the
existing Asian Development Bank (ADB) findings (1996) and census
data collected by the author in 1996, as well as relevant survey/data,
rapid appraisal of selected bastis (settlements), in depth interviews
with individuals and institutions, group meetings with leaders and
resettled migrants. Initially two bastis, Kallayanpur and Nilkhet, hav-
ing 1420 and 522 households were selected for an in depth study.
Nilkhet is one of the oldest bastis established for resettlement of
Biharis in the 1950s and the Kallayanpur settlement was established
nearly 10 years ago. It is believed that two different durations would
help to generate insights on the dynamics of slum formation in a better
way. Moreover, these two bastis experienced fire accidents. The
choice of these locations is likely to provide an opportunity to examine
the issues of decay of basti and resettlement better. Findings from these
two bastis were also compared with other bastis more particularly,
Khilgaon Bagicha and T.T Para, as the author had baseline information
on these bastis.
During the field survey, it was observed that a few affected residents
of Nilkhet moved to Kamrangir Char at the southwestern part of Dhaka
City, for resettlement. Subsequently, a tracer survey was also con-
ducted with those migrants, who moved from Nilkhet basti to Kam-
rangirchar after the eviction attempts. This is a unique component
added to the study of migrants at destination, which helped to
re-examine the issues of resettlement in a better way.

Re-examining the Resettlement Question 827
The focus on two categories of poor women migrants, namely
female headed households and female garment factory workers has
provided insights for this paper. Whilst the female headed households
are recognised to be the poorest and the most disadvantageous category
among the poor, the female garment factory workers are often the first
generation migrant workers in Dhaka city, who deserve serious atten-
tion. Not only the sheer size of nearly 800,000 female migrant workers
in the garment factories in this city alone, but the problems they face
in the migration and settlement process in a most rapidly expanding
but least equipped metropolis, demand urgent attention of the planners
and policy makers. Hence, these two categories of women in Dhaka
city are covered to highlight the multidimensional problems and the
special needs in women's migration, settlement and the resettlement
process.
The female headed households for the study were drawn from the
census data often slum and squatter settlements from seven randomly
selected thanas of Dhaka city conducted by the author in 1996. Thanas
are the largest administrative units having a population between
100,000 and 650,000 (Afsar, 1995:59). About 4037 households were
enumerated in 1996, of which 17 per cent are headed by women. There
are 14 thanas in Dhaka city of which seven are selected randomly from
random number tables. Out of seven, four thanas namely, Mirpur,
Mohammadpur, Lalbagh and Demra have the highest concentration of
slum and squatter populations (ADB, 1996:8, Bangladesh Bureau of
Statistics, 1988:20). Generally one slum or a squatter settlement is
selected from one thana excepting Mirpur and Sabujbagh from where
four sample study areas were drawn. The selection of slum or squatter
settlements were based on the following criteria:
• having a minimum 200 to 1000 households;
• poor living, environmental and economic conditions; and
• existing for at least five years.
A subset of gender disaggregated data on male and female headed
households was derived by stratifying households on the basis of
occupations of heads and a proportional random sample of 200 was
drawn. Data on income, expenditure, savings and indebtedness, gender
division of labour, hygienic practices, physical and environmental
conditions, migration and mobility, were covered through household
surveys.
The observations for the garment factory workers, is drawn from
the data collected from a recently conducted survey by the author

828 Rita Afsar
(Afsar, 1998a) covering 11 garment factories and 508 randomly se-
lected workers, representing 15 per cent of the total workers. This is
supplemented from the insights generated with the help of focus group
discussions.
Gender Dimension of Urban Poverty in the Migration and
Settlement Process
Women comprise 70 per cent of the world's poor (United Nations,
1995b) and poverty is over represented among women and children.
A female headed household in an urban area earns three-quarters of
the income earned by a male head in Bangladesh (Bangladesh Bureau
of Statistics, 1995:47). Poor settlements are often characterised by a
predominance of children and greater number of female headed house-
holds, compared to other areas. Children under 15, constitute nearly
45 per cent in those agglomerations compared to 39 per cent for urban
areas (Afsar, 1996). Similarly, the magnitude of female headed house-
holds increases from well below 10 per cent in urban areas to 17 per
cent in those slums (Mitra, Nawab Ali, Islam, Cross and Saha, 1994).
The correlation between the magnitude of poverty and female headed
households is globally evidenced (International Labour Organisation,
1995; United Nations, 1995). In Bangladesh, the magnitude of female
headed households under ultra poverty (1600 kilo calorie per person),
the worst manifestation of poverty, is significantly higher among
females (nearly a quarter) than male headed households (15 per cent)
(Bangladesh Bureau of Statistics, 1995:50).
Women are more affected by the migration process than men, both
at the place of origin and destination. At the place of origin, often the
husband migrates, leaving behind the wife and children, without either
much familial help or financial support. Sometimes they do not come
back or come with another wife. In slum and squatter settlements, the
emergence of female headed households is invariably linked with
erosion of familial support, forms of marital instability and a lower
number of male than female adults. Nearly three-quarters of female
heads are divorced or widowed compared to only two per cent of the
male heads (Table 1). In such a situation of abandonment, women bear
the burnt of poverty along with the responsibility for child rearing, with
little or no support from adult male members. Female headed house-
holds have lesser number of working age members — 1.0 and 0.95
earners on an average, compared to 1.14 and 1.04 members, respec-
tively of male headed households. Moreover, they have only 0.6 male

Re-examining the Resettlement Question 829
earners as opposed to 1.3 for male headed households (Table 2).
Elsewhere, based on a review of the empirical evidence of Bangladesh
and a few other cities, the (Afsar, 1996:7) argued that:
It is important to have male earning members because adult
males nearly always work and earn more than a woman
worker. Due to labour market discrimination against women,
women's larger share of household and child bearing respon-
sibilities, lack of familial support, women in general and
female heads in particular, are worse off than their male
counterparts.
TABLE 1
Percentage Distribution of Household Heads by Age, Sex and
Current Marital Status in Ten Slums of Dhaka City
Source: Slum Census of Dhaka city, 1996

830 Rita Afsar
TABLE 2
Household Composition by Gender of the Household Head in
Ten Slum Areas of Dhaka City
Source: Slum census of Dhaka city, 1996.
At the urban end, be it with spouse or alone, women are continu-
ously waging war against want, adversities, deprivations and insecu-
rity, towards achieving a better future (Afsar, 1992). Unlike men, who
can sleep in the open, women irrespective of their marital status, need
a secure shelter. Hence, with an increasing independent migration of
women, who come predominantly to work in export-oriented Ready
Made Garment (RMG) factories, new types of families such as branch
families, and single member families are emerging. It is interesting to
observe that, 42 per cent of the non-slum households in Dhaka city
have joint and extended types of families, whereas slum households
are characterised predominantly by nuclear types of family (between
66 per cent and 75 per cent). A quarter and a third of female and male
garment factory workers respectively, live in branch and single mem-
ber families (Table 3).

Re-examining the Resettlement Question 831
TABLE 3
Distribution of the Sample Garment Factory Workers by Sex and Types of
Families they are Currently Living with at the Destination
Note : *Branch families include mainly siblings, cousins and co-workers living
together.
Source : Survey of 14 garment factory workers in Dhaka city, 1997.
In the main part of the city, four-five other family members and
relatives live largely in one room accommodations sharing a floor area
ranging between 50 and 200 metres. Outside Dhaka City Corporation
and in fringe areas, people live in slightly bigger floor areas of about
250 metres with more rooms than in the main city and at a cheaper
rate. Thus, whilst a garment factory worker has to spend nearly US$
15 on house rent in the city proper, it reduces by US$ 3-4 for those
living in fringes or outside Dhaka City Corporation areas. The amount
they pay for houserent constitutes the whole salary of the unskilled
production workers and two-fifths of the skilled production workers.
As a result, they cannot afford to live alone and have to live with either
parents or brothers and sisters or relatives. Female married garment
workers live with their husbands and are in a more disadvantageous
situation than the former, in terms of domestic work. Table 4 shows
that nearly four out of every five currently married females have to
manage the domestic chores alone compared to nearly 13 per cent of
their male spouse and 28 per cent of the unmarried female workers.
Married women face tremendous problems. On the one hand, they
generally manage household chores almost single handledly and, on
the other hand, they face the problems of suspicion if they come late
or do not hand over the expected salary. Often they are late to reach
the factory which results in a salary cut. Some husbands beat their
wives, if they do not get their food ready after returning from work.

832 Rita Afsar
Husbands, on the other hand, often come late and are not bound either
to give explanations or to hand over their salary to their wives. Some
of them establish extra-marital relations and abandon their first wives.
The abandoned first wife joins a garment factory and works hard to
maintain herself and her children. They do not get any maintenance
nor any legal protection in the case of threat by the ex-husbands.
TABLE 4
Percentage Distribution of Poor Urban Women by Time
Spent on Fetching Water/Fuel in Selected Slum Areas in
Metropolitan Manila, 1990 and Dhaka, 1996
It is widely believed that in the Indian sub-continent, including
Bangladesh, women's autonomous migration from rural to urban areas
is a low, compared to Sri Lanka and Southeast Asian countries, largely
due to low economic growth and the rigid normative prescription of
the patriarchal society (Fidley and Williams, 1991; Yousef, Buvinic
and Kudat, 1979). Notwithstanding such a negative image, with a rapid
expansion of RMG factories in the capital city in the 1980s, there has
been a boost in independent migration by young women in Bangla-
desh. This is reflected in the sex-ratio of migrants (Afsar, 1995) and
even in the age specific (15-30 years) urban sex ratio, which became
more feminine in 1994 (Mitra et al, 1994) from male dominated in the
previous census years 1981 and 1991 (Afsar, 1995). In Bangladesh,
seclusion of women was observed in the case of landed families only.
For the large majority of landless families, women always contributed
to family's income, whether directly through harvesting or indirectly
through post-harvest and other income generating activities. Thus,
when the demand for female labour was generated through the RMG
sector, as a result of export promotion by the private entrepreneurs, it
was met with the help of a loyal kinship network support from villages.

Re-examining the Resettlement Question 833
Through in depth interviews with employers of garment factories,
it was revealed that they preferred migrant female labourers over urban
residents because of the former's loyalty, docility and law abiding-
ways. These factories generally produce ready-made garments such as
shirts, pants, jackets, jogging sets, blouses. A few also produce knit-
wear, with the raw materials and orders, supplied by external buyers,
mainly from North American and the European countries. Women
work here mostly as unskilled production workers and operators. Some
of them also work as supervisors, but very few were quality inspec-
tors.
Elsewhere, the author has demonstrated that more than half of the
garment factory workers had connections with employer's agents and
nearly 90 per cent of them had acquaintances in Dhaka, prior to
migration (Afsar, 1998b). Thus, a large number of rural women, almost
equally divided between ever-married and currently unmarried, are
migrating to Dhaka city in their teens. Four out of every five of these
women belong to functionally landless families as opposed to a half
of their male colleagues. A female garment factory worker has four
median years of schooling, which is comparable with her age cohort
(15-39 years) in urban areas, but clearly much greater than her rural
counterpart (Afsar, 1998a).
With a heavy concentration of working age population and unmar-
ried workers, the dependency ratio is very low in the families at the
destination of migrant workers. Conversely, with the migration of
young and active adult members from villages, the dependency ratio
is a very high 76 per cent for those households in rural areas. This
suggests that migration is adopted as a family strategy for income
diversification, where young and adult working men and women are
sent to the areas of economic opportunities. Network support of family
members, relatives, friends, neighbours and acquaintances from the
same districts facilitate migration of both men and women, prior to and
after migration. An individual's coping capacity improves with the
availability of aid and assistance in the new location and place of
residence.
However, though living with relatives or siblings can be a cost-
saving strategy, it does not always ensure security to the female
garment factory workers. Discussions at the residential locations
disclosed the following problems related to their settlement in urban
areas which threaten the security of the female garment factory
workers :

834 Rita Afsar
• Women workers work late into the night and if there is load
shedding at the time of departure, they are approached and
teased by mastaans (muscle men) who sometimes kidnap a few
of them. Two such cases in Kalapani area of Mirpur thana, were
cited, where girls were kidnapped and raped by mastaans.
• In the city, they are often forcefully evicted and their belongings
are damaged or thrown and even confiscated when they fail to
give the house rent in time. This is the situation in many areas,
such as Mahakhali, Banani, Chairmanbari, and so on. The work-
ers often cannot pay house rent in time because, generally, they
are not paid on a fixed day or even a fixed week.
• Outside the city corporation area, there are vacant places bor-
dered by a canal or river and the danger of rape and kidnap by
mastaans looms large there. In the Arichpur Union of Tongi
thana, where such locations exist, six cases of kidnap and sub-
sequent rape of garment factory workers have been reported in
the last two months.
• In the southern fringe areas of Dhaka city such as Madhya Badda
and Hajipara, semi-pucca and semi-kutcha types of houses are
built by owners over marshy land, ditches and drains with the
help of bamboo poles. There have been occasions when women
workers have fallen from those poles in the drains, at night.
• Cases of women garment factory workers quitting their jobs
because of sexual abuse by mechanics or other staff, also exist.
It is difficult to estimate the magnitude of the incidence of the above
instances cited above. The quantitative techniques used to determine
the magnitude are not found suitable to generate such insights.
Gaining access to basic services such as water, bathroom and toilets
through informal or often combined sources, involves not only a
significantly high cost but also more time and energy of women and
girls. Generally, urban poor households incur more expenditure to
procure basic services such as water, fuel and lighting (10 per cent)
than non-poor households (6 per cent) (Afsar,1996:12; Bangladesh
Bureau of Statistics, 1995:90-91). Here too, women heads spent more
on housing, fuel and lighting (21 per cent) than their male counterparts
(16.6 per cent) (ADB, 1996:15). Since poor men and women procure
those services from non-formal sources, they pay more compared to
non-poor households. As most of these service providers are men, poor
men have greater accessibility than their female counterparts. They are
also likely to master the trick of bribing a few influential agents to save

Re-examining the Resettlement Question 835
the bulk of the expenditure. Women might not have acquired that skill
yet. Table 5 reveals another dimension, that is, nearly two-thirds of
women spend more than an hour and travel a distance of about one
kilometre to fetch water from informal and or collective sources. This
finding is consistent with Fernail (1990, cf. Oberai, 1993:151) and
Mukherjee and Ramesh (1991).
TABLE 5
Distribution of Respondents who were Sick last Week and
last Month by Sex and Types of Diseases Suffered
Note : Others include, Menstrual Problems/ Leucorrhoea, Dental problem, Pox/
Measles, Bleeding from nose.
Source : Survey of 14 garments factory workers in Dhaka City, 1997.
Had they got running water in their compound, at home they could
have used the time fetching water more productively in income earning

836 Rita Afsar
activities. However, daughters who help their mothers in those activi-
ties are hampered in their development with regard to education and
job opportunities. Nearly 90 per cent of boys and girls in the age group
between 11-19 years are not going to school in poorer agglomerations
of Dhaka city. Many such boys are working, whereas a majority of
girls not attending school, are engaged in household work and child
care (56 per cent) (Afsar, 1996).
Nearly 90 per cent of the female and 80 per cent of the male garment
factory workers, have access to combined bathrooms in the commu-
nity, which are shared by 22 and 16 persons respectively, on an
average. Very few women workers (7 per cent) as opposed to 20 per
cent of their male counterparts, have access to private bathrooms and
toilets each. A dominant majority (80 per cent male and 73 per cent
female) of the respondents share slab latrine and septic tank/modern
toilets with about 20 families, on an average. Here, access to bathrooms
did not necessarily mean well-equipped bathrooms with running tap-
water. Often bathrooms are just a space covered with jute curtains on
the four sides under the open sky. The workers rely for their bath
either on tube-wells or fetch water from other sources such as nearby
ponds, lakes and canals, which often get contaminated by the adjacent
sewerage pipes or drains. Often at night, they urinate or defecate in the
open space, aggravating the deterioration in the neighbourhood envi-
ronment. Thirty to forty families share one or two taps or go to their
neighbour's or landlord's houses for filling up the pitcher which costs
additional money. They also dig wells and drink that water and in the
winter when the wells dry up, women bear the burden of collecting
water from other sources.
Similarly, two to three gas burners or electric heaters are shared
between with 20-30 families. Subsequently, they queueup in the
morning for cooking and as a large majority (three-quarters) of the
ever-married women do it single handedly, they are often late to reach
their workplace which makes them more vulnerable to pay cuts than
their male colleagues. During focussed group discussion sessions,
women workers complained that often when they were late by 10-15
minutes, they were marked absent for the whole day. Men, however,
did not have such complaints. Hence, women garment factory workers
get up earlier (between 3.00 a.m. to 5.00 a.m.) compared to men, who
get at least one or two hours extra sleep. Women have to cook rice and
curry, which they and other family members eat for breakfast and
lunch. If basic services were available at their doorstep, women could

Re-examining the Resettlement Question 837
have avoided such pay cuts and even have a little rest, which they
deserve badly after an average of 10 long hours of work (Afsar,
1997:6). It also suggests that women make more adjustments to
surmount odds they face in their new settlement than men.
Nonetheless, heavy work pressure and environmental pollution
exacerbate health problems more than their male counterparts. More
than 40 per cent of the female garment workers suffered from various
diseases during the time interview, compared to 26 per cent of men.
The sickness patterns of the workers suggest that a large majority (61
per cent) suffered from water borne and malnutrition related diseases,
followed by general weakness, aches and pains all over the body (see
Table 5). In every community visited by the author, there are instances,
where a few garment factory workers went back to their villages after
prolonged sicknesses. As gathered through community based focussed
group discussions of the garments factory workers, it was found that
more women than men were return migrants.
The above findings clearly show how women bear a disproportion-
ate brunt of poverty and insecurity at the urban end to cope with the
demands of settlement in poor agglomerations. As slum and slum types
of structures house a large majority (nearly a half) of the urban poor
in Dhaka city, it is important to focus on the formation of slum and the
issues of sustenance and resettlement, and also how they affect men
and women of those settlements.
Formation of Slums and Squatter Settlements in Dhaka City
Bastis are generally developed along railway tracks, bus/launch termi-
nals, commercial centres and factories. It is logical for poorer migrants,
who generally come in search of a job or better employment opportunities
to settle in those places that are well-connected with transportation
networks and sources of employment. Thus, out of ten bastis for
which census surveys were conducted, three namely Khilgoan
Bagicha, T.T. Para and Jurain were developed along railway tracks.
Agargaon, Mirpur and Mohakhali squatter settlements came into
existence as a result of land-litigation and the availability of em-
ployment opportunities. Construction of the flood protection em-
bankment also accelerated the growth of slums along the western
periphery of the embankment.
Private slums are generally developed with a profit motive, where
the landowners build temporary low-cost structures on low or marshy
land or controversial land for rental purposes. Squatter settlements

838 Rita Afsar
often grow with the patronage of political leaders or party men, and
the staff of some government/semi-government departments on vacant
public land or on those places where there is some controversy with
regard to ownership.
Field data revealed that apart from party affiliation, the occupational
status of the leaders is an important source of their power. All the leaders
of the sample slums are engaged in businesses, be it charging ground rent
or owning rickshaws or hotels or shops. Employment in urban infrastructure
and services related government and semi-government departments such
as, Public Works, Power Development Board, Water and Sewerage
Authority, and so on, is another important source of power.
Women are more secure in the private slums and low-cost semi-
permanent houses, rather than on squatter settlements. In the private
slums, they face the threat of eviction from one or a few recognised
persons. On the contrary, in the squatter settlements where there are
several de facto leaders, not only do they have to pay ground rent, to
the ever changing rent collectors, they are also more susceptible to
physical and sexual harassment by goons, and party men, who often
come from outside, as a result of fights to establish supremacy between
two or more defacto leaders. However, it is difficult to have purely
non-controversial slums in Dhaka city.
A majority of the garment factory workers predominantly live in
rented accommodation and they do not face problems of eviction, but
they face other problems due to inadequate urban amenities, discussed
earlier. Half of the female heads, on the other hand, are the long-term
migrants, who have been able to establish their claims as de facto
owners, like male heads. The other half of female heads, who live as
tenants, are more susceptible to such threats of eviction and extortion.
Hence, threats of settlement are multidimensional in nature, which
affect different categories in different ways. Notwithstanding the
multi-dimensionality involved, women suffer more because often it
endangers their security the most. This will be more clear from our
discussion on eviction issues in the following section.
Sustenance and Evictions in the Poorer Agglomerations
Contrary to the existing ideas that migrant population in poorer urban
agglomerations are floating masses and also highly mobile people, a
quarter of slum and squatter residents are found to have not changed
their residences at all since their arrival to Dhaka city. Census of ten
slum and squatter settlements in Dhaka city also estimated an average

Re-examining the Resettlement Question 839
tenure of seven years for any ordinary slum dweller. Average tenure
in those agglomerations also varies by the longevity of the settlement.
Nearly two out of every three dwellers of the slums have been staying
in the respective slums and squatter settlements for five and more years
and nearly a fifth of them are living for more than ten years. The
remaining one-third of the inhabitants have lived less than five years
in those settlements (Afsar, 1997:12-13).
The average tenure in the poor agglomerations should not be seen
as a gender neutral phenomenon. It is interesting to observe that
women garment factory workers' average tenure in the current resi-
dence is nearly one year lesser than their male counterparts. Whilst the
duration of a female garment factory worker is three years in the
current residence on an average, for their male counterparts it is nearly
four years. Notwithstanding the fact that more than four-fifths of the
male garment factory workers are currently unmarried and nearly a
fifth of those workers live in mess or boarding houses, they still have
a longer duration of residence than their female counterparts. It not
only reflects the problem of security but also of forced eviction by the
landlords and relatives. This was observed during field visits to their
communities. As the women workers do not receive payment in time
and get lower wages than their male counterparts, they are more
vulnerable to eviction. Generally, the male-female income ratio is
0.87, suggesting that a male worker earns 13 per cent higher gross
income than his female colleagues. It increases to a quarter in the case
of an unskilled production worker, suggesting that on an average, she
earns three-quarters of the income earned by her male co-workers
(Afsar, 1998a: 9). Being young, more than two-fifths of the female
garment factory workers are currently unmarried. They have greater
options for mobility and networking with cousins, siblings and co-
workers as compared to female heads of households, who are on an
average older than their male counterparts by one year. On the other
hand, female heads have a longer tenure that is 7.4 years in those
settlements than male heads (6.3 years). It suggests that, not only the
female heads are less mobile than male heads, but possibly also
indicates their age, lack of support from adult male members and
concern with the security question, which makes them less mobile.
Whilst settlement is an important aspect of migration, evacuation
and resettlement are other issues which need to be addressed for policy
formulation and programme implementation. In Dhaka, there are two
widely practised means of eviction: bulldozing and fire. In squatter

840 Rita Afsar
settlements of Agargaon, Drainpar, and so on, time and again, there
was bulldozing to evacuate the settlers. Kallayanpur, Islambad and
Nilkhet, on the other hand, witnessed fires on a number of occasions
caused by arson by some vested interest groups, with a view to taking
control of the land (Wahid, 1997: 7).
Question of Resettlement Re-examined
The range of policy options available for controlling migration is quite
limited. A review of global policies shows that attempts to restrict
movement of people in China, Poland, Cuba, Indonesia, the United
Republic of Tanzania and Zaire, among other countries, through
administrative and legal controls, have been unsuccessful to a great
extent (Oberai, 1993:184). They have also raised serious questions
relating to human rights. Other options pursued by the government
such as land-settlement schemes, administrative and industrial decen-
tralisation and rural development programmes designed to make rural
living more attractive, have met with only limited success as rural-urban
migration continued on a significant scale. Similarly, in the large cities,
attempts to evict the poor population through resettlement schemes
have, by and large, been a failure. In the case of Bangladesh, resettle-
ment policies since 1974-1975 did not achieve the basic objectives
because they focused on symptoms rather than on causes. In this
section, an indepth analysis of the situation of the resettled families,
particularly women, has been made to generate greater insights on the
problems and causes of resettlement.
After a fire incident in the Babupura basti, a few of the affected
families moved to Kamrangir Char and Nawabganj, but the bulk
remained in Nilkhet. Those who moved to Kamrangir Char, did so
because of the fear of possible fire/eviction attempt by the government.
Tracer survey of those 20 migrant families at the destination showed
that although eviction was the most imminent reason for their move-
ment, there were other reasons too for such inter-slum movement. One
such reason was that the house rent was very high in Nilkhet basti,
much higher than Kamrangir Char. Whilst the house rent in Nilkhet is
around Tk 600 ($ 13 approximately), it is nearly Tk 400 ($8.5 approxi-
mately) in Kamrangir Char. Another reason was that the mastaans of
Nilkhet basti often disturbed their young daughters. One of the evicted
women reported that her daughter was kidnapped by those mastaans
and hence she decided to move to the Kamrangir Char for the safety
and security of her young daughters.

Re-examining the Resettlement Question 841
As a result of resettlement from Nilkhet to Kamrangir Char, the
dwellers (especially the females) are facing job displacement prob-
lems. Many of them, either stopped working or work only when it is
available in Kamrangir Char. It should be noted that a large number
that is, half of the female earning members of the slum and squatter
settlements, work within a distance of one kilometre from their homes,
compared to nearly a quarter of their male counterparts. On the other
hand a little over three-fifths of the male earning members work all
over the city, as opposed to a fifth of the female earners. Similarly,
nine out of ten female garment factory workers live in the same wards
as the factory and walk to their workplace (Afsar, 1998b: 31, 38)
Hence, whilst distance of the workplace from home is the major
determining factor in the case of inter-slum and inter-job mobility,
particularly for poor women, establishing a network of acquaintances
and security of younger women are among the other important factors
in this regard. Being new, the well-to-do families in the Kamrangir
Char do not hire them as domestic servants. Hence, female migrants
cannot find work so easily in their new place of settlement. As for male
dwellers, they travel to their previous workplace. But residing in the
Kamrangir Char becomes costly for them as the income has remained
the same or has reduced (if the female spouse has stopped working),
while their transport cost had increased. There are a few NGOs having
micro credit programmes to finance income generating activities of
the urban poor such as Proshika, Shakti Foundation and another
international organisation working in Kamrangir Char, but the new
migrants are unaware of their activities. Husbands of the currently
married female respondents are either rickshaw/van pullers or decora-
tors, and so on. The female respondents themselves used to work as
domestic helps, helpers in hotels/restaurants and also in garment
factories, prior to their migration to the new location.
All the respondents reported that even though they are not helped
by any of the slum leaders or their landlords, they often get help from
their workplace or from nearby houses or their fellow slum dwellers.
It was revealed during group meetings that the respondents had rela-
tives or acquaintances in some other slums of Dhaka city, who often
provided necessary help and support. In their time of crisis, they also
pawn/mortgage their ornaments to pawn-brokers, in lieu of which they
get money at a high interest rate (20-30 per cent per month or even
more). Women strongly felt that in the case of rehabilitation or
resettlement by the government, they would decide whether they

842 Rita Afsar
would go or not go, depending on job availability, security in the new
place, and so on. Female headed households, who do not have adult
male companions and earning members, feel insecure to move to a
new/unknown place, more so when they have young daughters whose
security is important. Most of the dwellers who shifted to the Kam-
rangir Char are deserted/widows with young daughters. They had
moved to there mainly because security of their young daughters was
often threatened by mastaans in Nilkhet basti and not necessarily for
eviction only. Majority of the dwellers who moved are migrants from
Barisal. It became clear by interviewing many slum dwellers that they
generally do not move as a result of eviction. They make every attempt
to stay in a familiar place for job and security reasons and seldom move
to far off places without the support of their social networks. Even if
they move out from one slum, they do maintain close contacts with the
previous place of residence.
Implications for Urban Housing and Poverty Alleviation
Projects

Dynamics of slum formation, settlement, evacuation and resettlement
have several implications for the urban development and poverty
alleviation projects. Squatter settlements are formed with the help of
muscle power, strong social networks based on kinship and district
based affinity and with the patronage of political leaders. All these big
squatter settlements studied have already confronted eviction attempts
made by both legal authorities and vested interest groups on several
occasions. The fire incident in Kallayanpur and Nilkhet bastis and
subsequent strategies adopted by squatters to face those challenges can
be cited as the case in point. On both occasions, squatters were able to
identify and unveil the masks of the real culprit behind the incidence.
They also attracted public sympathy and mobilised massive relief both
from the government and NGOs. They also continued to stay in the
same squatter settlement although not exactly in the same place, where
they had been residing so far. Only a fraction of the affected families
moved to other areas, as a result of partial eviction attempts made in
Kallayanpur and Nilkhet bastis. Hence forced eviction, either partial
or total, does not seem to be a cost-effective strategy for removing poor
people or poverty.
One important issue that make both the garment factory workers
and women in slum and squatter settlements, including female heads,
more vulnerable in the settlement process than their male counterparts,

Re-examining the Resettlement Question 843
is the extreme scarcity of adequate civic amenities at home compound.
What they have at present is the shared sources of basic necessities
namely, water, bathrooms, latrines and cooking gas burners or electric
heaters. Lack of running water, proper drainage, waste collection and
sewer facilities, poor lighting and accessable roads make them highly
susceptible to environmental degradation, health hazards, income loss and
sexual harrassment. As long as these women live in overcrowded housing
with inadequate access to civic amenities and limited access to preventive
health care, they are likely to be seriously affected by ill health, as
demonstrated in this paper. Absenteeism due to illness raises production
costs in general, thereby reducing surpluses available for investment.
Hence, apart from the intrinsic welfare reasons for providing safe water,
sanitation and health services, the generation and maintenance of employ-
ment requires these services (Oberai, 1993: 191).
Experience in Asian cites suggest that service extension affordable
by the poor and low income groups is workable, only when there is an
in-built mechanism of self-help and local maintenance. Studies in Sri
Lanka (Fernando, Gomage and Dharmawansa, 1987) document sig-
nificant upgrading of water, sanitary and solid waste collection serv-
ices because of a high involvement of people in the decisions regarding
the standards and locations of services. Similarly, by involving the
local government bodies and squatters in community upgradation, San
Martin De Porres in Manila represents a good example of a close
relationship in affordability, replicability and cost recovery.
It should be realised that the root cause of housing problems in most
developing countries, including Bangladesh, are the low level of produc-
tivity and low incomes of the urban poor. To solve the problem effectively,
efforts have to be directed towards improving productivity and incomes
of the urban poor and towards reducing the cost of housing. Whilst
employment, income generation and skills development, especially for
women, through micro-finance and other devices are needed to ensure the
former, efficient management of land, control of land speculation and de
facto land ownership are critical elements in improving the functioning
of housing markets in large metropolitan cities like Dhaka. It needs a
strong political will and effective taxation measures which should aim to
suppress land speculation but not discourage investment on housing.
Taxes on housing units should be levied more heavily on holdings than
on transactions (Oberai, 1993: 194).
Availability of low-cost housing or easy access to credit for hous-
ing, where the poorer people have an option to buy or build low-cost

844 Rita Afsar
housing, on the basis of their perceived priority, is considered to be a
more practical strategy in the market dominated economy. Leadership
from the municipal government can be more effective if it is based on
a system of accountability and participatory approach, involving en-
ergy and creativity of poor men, women and children and community
groups to improve their lives and situations. Security of young girls
and proximity to and availability of low-skilled jobs are important
considerations of the migrants, which was revealed through group
meetings with displaced migrant families from Babupura basti. Hence,
any urban poverty reduction programme should emphasise these fac-
tors in implementing its housing and shelter component. It should give
priority to the need of a secure shelter for younger women (aged
between 15-34 years), who outnumber their male counterparts in the
urban poor agglomerations. Simultaneously, security of adolescent
daughters, who constitute a fifth of the total populations of those
settlements demands serious attention. A sizeable number (17 per cent)
of slum families headed by women have very few adult male members
and those which have adolescent daughters, are not likely to move to
a new or unknown place, if security cannot be ensured.
Similarly, availability of work and the question of identity are very
important in the competitive urban market. For example, women who
moved to Kamrangir Char complained about job displacement. Being
unknown in the new place, they were not trusted by housewives of the
nearby well-to-do families and hence they lost their jobs as domestic
helps, which was available in Nilkhet basti. Moreover had they not
known the co-workers from Kamrangir Char, they would not have
moved to this place. Hence any poverty reduction programme in urban
areas must address these issues of security, identity, work opportuni-
ties and the protective umbrella of social networks of the urban poor
communities, particularly women, while implementing housing and
shelter components.
Moreover, as eternal vigilance is the price of liberty, a strong
organisation of poorer women is needed to safeguard their own rights
and those of the adolescent girls.
REFERENCES
Afsar, R.
: Gender Based Urban Poverty Alleviation Strategies,
1992
Grassroots, 1(4), 7-17.

Re-examining the Resettlement Question 845
Afsar, R.
: Causes, Consequences and Challenges of Rural-Urban
1995
Migration in Bangladesh. Unpublished Doctoral Disser-
tation, Adelaide:University of Adelaide.
1996
Application of Demographic Techniques in the Analysis
of Gender Biased Urban Poverty Situation in Bangladesh.
Paper Presented at the Australian National Population
Association, Adelaide, December 2-6, 1996.
1997
A Study on Dynamics of the Formation, Settlement and
Sustenance of bastis in Dhaka City, Working Paper No. 2,
Dhaka: Bangladesh Institute of Development Studies.
1998a
Too Early or too Inadequate: A Situation Impact Analysis
of the UTPs Programmes on the Garment Factory Work-
ers, their Families and the Garments Factory Manage-
ment. Draft Report, Dhaka: Action Aid.
1988b
A Case Study of Gender Dimension of Labour Migration
in the Formal Manufacturing Sector in Dhaka City and
Sociological Implications (mimeo), Dhaka: United Na-
tions Research Institute for Social Development.
Asian Development Bank Final TA Report on the Urban Poverty Reduction Project,
1996
Dhaka.
Bangladesh Bureau
Report on the Household Expenditure Survey, 1991-92,
of Statistics
Dhaka: Ministry of Planning.
1995
1988
A Case Study of the Floating Population of Dhaka City,
Dhaka: Ministry of Planning.
Fernando, S.,
Nawagampura and Armaya Place: Two Urban Case Stud-
Gomage, W. and
ies of Support Based Housing and Basic Services. Paper
Dharmawansa, P.
Presented at the Expert Group Meeting on Shelter and
1987
Services for the Poor Metropolitan Regions, Nagoya,
January, 12-16.
Fidley, S. and
Women Who Go and Women Who Stay: Reflections of
Williams, L.
Family Migration Processes in a Changing World, Work-
1991
ing Paper No. 176, Geneva: International Labour Office.
International Labour
Gender, Poverty and Employment, Geneva.
Office
1995
Mitra, S.N., Nawab Ali,
Bangladesh: Demographic and Health Survey, 1993-94,
M, Islam, S., Cross, A.R. Calverton: National Institute of Population Research and
and Saha, T.
Training.
1994
Oberai, A. S.
Population Growth, Employment and Poverty in Third-
1993
World Mega-Cities: Analytical and Policy Issues, New
York: St. Martin's Press.
United Nations
World Urbanisation Prospects: The 1994 Revision, New
1995a
York: United Nations Publications.
1995b
Platform for Action and Beijing Declaration, New York:
Oxford University Press.

846 Rita Afsar
United Nations
: Human Development Report, 1990, New York: Oxford
Development
University Press.
Programme
1990
Wahid, S.
Fighting Fire, Weekend Independent, February 7, 4—8.
1997
Yousef, N., Buvinic, M.
Women in Migration: A Third World Focus. Report of the
and Kudat, A.
International Centre for Advancement of Women, Wash-
1979
ington D.C.:United States Agency for International De-
velopment.