Coping strategies with floods in Bangladesh: An empirical study

Abstract

In the time of the flooding, rural people in Bangladesh suffer from the lingering effects of labor market disruption and income deficiency. This study shows a model based analysis on the research question, ‘what coping strategies are followed by the flooded households in Bangladesh and how?’. Data are collected through a three stage stratified random sampling technique on 595 flooded and rural households’. Survey is carried out aftermath of the flood in the year 2005, from four different districts in Bangladesh. A major proportion of households are found to borrow money or resources from informal sources, such as nearby shops or the pharmacy, friends or relatives, or local money lenders, to buy food items and other essentials. A censored tobit model analysis shows that households initiate coping with borrowing money after the realization of floods, and gradually lead to cope with savings and selling assets as the duration of flood increases.

Full text

1 23
Natural Hazards
Journal of the International Society
for the Prevention and Mitigation of
Natural Hazards
ISSN 0921-030X
Volume 64
Number 2
Nat Hazards (2012) 64:1209-1218
DOI 10.1007/s11069-012-0291-5
Coping strategies with floods in
Bangladesh: an empirical study
Nayeem Sultana & Md.Israt Rayhan

1 23
Your article is protected by copyright and
all rights are held exclusively by Springer
Science+Business Media B.V.. This e-offprint
is for personal use only and shall not be self-
archived in electronic repositories. If you
wish to self-archive your work, please use the
accepted author’s version for posting to your
own website or your institution’s repository.
You may further deposit the accepted author’s
version on a funder’s repository at a funder’s
request, provided it is not made publicly
available until 12 months after publication.

ORIGINAL PAPER
Coping strategies with floods in Bangladesh: an empirical
study
Nayeem Sultana •Md. Israt Rayhan
Received: 14 January 2012 / Accepted: 6 July 2012 / Published online: 19 July 2012
Springer Science+Business Media B.V. 2012
Abstract In the time of the flooding, rural people in Bangladesh suffer from the lingering
effects of labor market disruption and income deficiency. This study shows a model based
analysis on the research question, ‘what coping strategies are followed by the flooded
households in Bangladesh and how?’. Data are collected through a three stage stratified
random sampling technique on 595 flooded and rural households’. Survey is carried out
aftermath of the flood in the year 2005, from four different districts in Bangladesh. A major
proportion of households are found to borrow money or resources from informal sources,
such as nearby shops or the pharmacy, friends or relatives, or local money lenders, to buy
food items and other essentials. A censored tobit model analysis shows that households
initiate coping with borrowing money after the realization of floods, and gradually lead to
cope with savings and selling assets as the duration of flood increases.
Keywords Coping Flood Divestment Borrow Tobit
1 Introduction
Bangladesh consists mostly of a low-lying river delta with over 230 rivers and tributaries
situated between the foothills of the Himalayas and the Bay of Bengal. The country lies
within the catchment areas of the Ganges, Brahmaputra, and Meghna rivers which mainly
drain through Bangladesh into the Bay of Bengal. In Bangladesh, floods are usually defined
as the submerge of land by water which can damage crops and property, disrupt people’s
normal living conditions, communities infrastructures, household’s communications and
economic activities, and endanger the lives of people and their livestock. About 60 % of
the country’s land mass is less than six meters above the mean sea level (Paul and Routray
2010) and floodwater inundates around 20.5 % of the country every year (Mirza et al.
N. Sultana
Department of Development Studies, University of Dhaka, Dhaka, Bangladesh
Md. I. Rayhan (&)
Institute of Statistical Research and Training (ISRT), University of Dhaka, Dhaka, Bangladesh
e-mail: israt@isrt.ac.bd
123
Nat Hazards (2012) 64:1209–1218
DOI 10.1007/s11069-012-0291-5
Author's personal copy

2001). High-magnitude floods inundate large areas causing widespread damage to crops,
human beings, livestock and property, as well as devastation to life and livelihoods (Paul
1997; Few 2003). The extent and depth of flooding vary from year to year depending on
rainfall and river levels. Damages of floods also differ both in time and places. There may
be a local flood affecting only a relatively small area in a particular part of the country.
Or the floods may be extensive, as in the years 1988, 1998, and 2004, affecting large parts
of the country’s major floodplains. Flood damages are reported in one or more parts of
Bangladesh almost every year. Even in years with average rainfall, large areas of low-lying
floodplains go under water for several weeks or months, as in the year 2005.
Flood risks are multidimensional and complex issues (Scheuer et al. 2011). The frequent
occurrence of disastrous floods results in losses for both human life and property values in
Bangladesh. According to the report of GOB (2006), the 1974 flood was followed by a
famine and as a result 30,000 people died. In 1987, about 40 % of the country was flooded,
affecting 30 million people and causing about 1,800 deaths. Loss of the main crop (paddy)
was estimated to be 0.8 million tons. The floods in 1988 were even more serious, covering
about 62 % of the land area, affecting about 45 million people, and causing more than
2,300 deaths. In 1998, Bangladesh experienced the worst flood in its history. Over 68 % of
the country was inundated, there were about 2,380 deaths, 1.56 million hectares of crops
were lost, and over 900,000 houses destroyed. In the year 2004, during July and August,
devastating monsoon floods submerged two-thirds of the country, resulting in 35.9 million
affected people, 726 deaths, 160,000 cases of disease and millions of homeless people.
About 12 % of the country’s area was flooded in the year 2005. The characteristics and
enormity of risks that households face, the access to risk management mechanisms, and the
surroundings in which households operate their activities, play a significant role in poverty
dynamics—these findings are supported by some theoretical analyses and empirical evi-
dences (Holzmann and Jørgensen 2000; Heitzmann et al. 2002). According to the World
Disaster Report 2011, about 8,408 people were died in the world only for flood, and about
186,894 people were affected in the year 2010. In the absence of adequate assets and
insurance to smooth income or consumption, risks emerged from disasters may lead to
irreversible losses, such as damage of productive assets, the fall in a vicious cycle of debt,
reduced nutrient intake, or disruption of education that eternally reduces human capital
(Jacoby and Skoufias 1997). Therefore, a study on coping strategies to flood disasters in
Bangladesh could be an inherent aspect of well-being.
Frankenberger (1992) defines coping strategies as the fallback mechanisms when
habitual means of meeting needs are disrupted. If households suffer from a shock like a
flood, they utilize the resources and options they have to survive. The actions for survival
strategies are mentioned as coping strategies. Initially, households try to minimize risks
and maintain some minimal level of sustenance. Gradually, the households start the dis-
posal of assets in several phases (divestment) as a coping strategy. According to Rashid
et al. (2006), coping strategies of flooded households in Bangladesh are categorized into
three stages: ‘current adjustment’, ‘unsecured borrowing’, and ‘secured borrowing/
divestment’. Current adjustment strategies include reducing household food consumption,
shifting to less preferred foods with lower cash cost, and reallocating household labor to
increase current income. Unsecured borrowing refers to borrow against expected future
incomes from relatives, moneylenders, merchants, and NGOs. Finally, households may
cope with flood shocks by divestment or borrow against their liquid and productive assets.
However, there is very limited literature available on the model based analysis of rural
people’s coping strategies against floods in Bangladesh (Haque and Zaman 1993; Hutton
and Haque 2003; Ninno et al. 2001; Rashid et al. 2006; Rayhan 2008; Paul and Routray
1210 Nat Hazards (2012) 64:1209–1218
123
Author's personal copy

2010). This study mainly focus on the research question, based on the flood in rural
Bangladesh during 2005, ‘what coping strategies are followed by the flooded households
and how?’.
2 Data and methodology
This study uses primary data from a sample survey conducted in rural Bangladesh just after
the flood in 2005. During that year, Bangladesh was affected by two types of floods: a
monsoon flood which occurred during mid August to September in the east and west parts
of the country, and a flash flood which occurred in the northern areas during November. A
cross-sectional household survey was carried out 2 weeks after the floods. Four districts
were randomly chosen from four different divisions to cover the diversity. A three stage
stratified random sampling technique was applied to the survey, where the first stage was
the district, the second one the mouza (the smallest administrative unit in the rural area),
and the third stage the household level. Flooded households were detected if at least the
home or homestead was submerged by flood water. A total of 300 households per district
was targeted in order to have representative figures for that particular area and also justified
by the formula (Cochran 1977, p. 75) for determining the sample size by estimated
proportion:
n¼t2pq
d2;PrðpP
jj
dÞ¼a
where, t=abscissa of normal curve =1.96, P=population proportion, p=estimated
proportion =.433 (poverty rate in rural areas, BBS 2005), a=probability of type I error,
or, level of significance, d=some margin of error in p(sampling error) =.07,
n=192.48 9design effect (1.5) =288.72 ¼
:300. Stratified random sampling (flooded
and non-flooded strata) was used to select the population listing or sampling frame for this
study, and systematic sampling was used for the households’ interviews from the frame.
Three districts, Jamalpur, Sirajganj, and Sunamganj, were randomly chosen from three
different divisions after the monsoon flood, so a total of 900 households (flooded and non-
flooded) were surveyed. Shortly after that a flash flood affected the northern part of the
country. Then a second survey of 150 flooded households was conducted from the ran-
domly chosen area, Nilphamari. The total number of surveyed households from different
flooded rural areas amounted to 600. Five outliers on the dependent variable (log per capita
income and each district level) are verified with the box plot approach, standardized
approach, and Cook’s distance and are deleted from the sample to get the robust estimates
of the regression coefficients. The estimated poverty line for the rural areas is Taka 594.60
per person per month according to the Food Energy Intake (FEI) method (BBS 2004). In
this study, a household is considered as poor if the per capita income is below Taka 594.60.
Households’ heads were asked about the coping strategies they followed during flood. Out
of 595 flooded households, all of them followed at least one strategy to cope with flood and
aftermath. Some of the households followed more than one coping strategy.
There are multiple coping strategies that are simultaneously determined by the flooded
households. As not all coping mechanisms are chosen by each of the flooded households,
this study uses some reduced form of regression analyses to identify the determinants for
borrowing, savings, and selling assets. If the dependent variable is dichotomous or cate-
gorical, four models: Linear Probability Model (LPM), probit, logit, and tobit are proposed
Nat Hazards (2012) 64:1209–1218 1211
123
Author's personal copy

for data analysis. LMP is criticized for estimating constant marginal effects and ignoring
heteroscedasticity. The logit model is assumed cumulative logistic probability function for
errors, and the probit is associated with the cumulative normal probability function, these
two models are usually used to circumvent the non-linearity problem. The logit and probit
models are appropriate for non-truncated dichotomous dependent variable in regression
analysis. As our responses from the households are not restricted only for borrowing,
savings, and selling assets, some of them did not mention any of three coping strategies, so
the dependent variable is truncated. The regression models suitable for this type of trun-
cated sample, where there are significant zero values in the dependent variable (for
households who did not use any coping amount from borrowing, savings, or selling items),
is known as the censored regression model or tobit model (Greene 2003, p. 764), proposed
by Tobin (1958). The general formulation is given in terms of an index function
y
i¼x0
ibþeið1Þ
where y
i
=0ify
i0 and yi¼y
iif y
i[0
The random variable y
i
is transformed from the original dependent variable y
i,x
i
is
denoting the exogenous variables, and e
i
is the error term. For the standard case with
censoring at zero (households do not cope with borrowing, savings, or selling assets) and
normally distributed error terms, coefficients would be estimated as
oE½yixi
j
oxi¼bUb0xi
r
 ð2Þ
where U(.) denotes the cumulative density of the standard normal distribution. From the
above equation, it is depicted that least square estimates of the coefficients in a tobit model
usually resemble the MLE (Maximum Likelihood Estimate) times the proportion of non-
limited observations in the sample. Some specification issues may arise in tobit estimates,
namely heteroscedasticity and non-normality. As heteroscedasticity varðeÞ 6¼ r2

emerges
as a serious problem for MLEs, this study measures tobit estimates with robust standard error
for each of the coping strategies and finds that the significance levels of the independent
factors are not changed from the estimated models with normal standard errors.
3 Empirical results
Table 1shows that flooded households have multiple responses for coping with floods,
2005. The highest frequency (193 households, about 32 %) is observed for borrowing
goods and cash from the nearby shop or pharmacy. The coping strategies are classified into
six broader groups, namely borrowing, using savings, selling items, changing habits, taking
aids, and others. The empirical results in Table 1show that flooded households relied more
on taking loan from different sectors compared to the other five general classifications. The
highest amount used for coping with flood and aftermath is Taka 293,956,
1
which was
borrowed from neighbors or relatives (about 20 % of households with interest and 19 %
without interest). Ninno et al. (2001) examines the impact of disastrous floods in the year
1998 using 757 rural households in seven flood affected regions in Bangladesh and shows
that borrowing is the major coping mechanism of the sampled flooded households, in terms
of both the value of borrowing and number of households that borrowed. Whereas, Table 1
1
Taka is the national currency of Bangladesh, also named as BDT, about Taka 80 equals $1.
1212 Nat Hazards (2012) 64:1209–1218
123
Author's personal copy

depicts that only 33 households responded having received aid from Government or NGO
sectors, and the amount (Taka 3,674) was significantly lower than from borrowing loan or
selling of assets. Some households (20 in numbers) also consume or use their savings to
cope with flood. 154 flooded households reported that they reduced their number of meals
and amount of consumption in a day, or sometimes bought cheap food items to cope with
flood and aftermath. In total Taka, borrowing from neighbors or relatives with interest was
the highest source of coping. Total amount of loan taken by flooded households was Taka
1,073,011. Out of the 595 flooded households, 413 households (about 69 %) mentioned to
cope with borrowing or taking loans from any of the eight different sources. It appears
from the Table 1that informal sources of credit were more effective than traditional micro-
credit programs (as noted in Zaman 1999). Zaman examines the extent to which micro-
credit reduces poverty and vulnerability through a case study on 1,072 households in rural
Bangladesh, comes up with a conclusion that loan installment repayment takes place
through a reduction in consumption and informal sources of credit likely to have more
significant effect than formal credit alone.
Table 1 Types of coping strategies and frequency of flooded households
Type of coping Number of
household
Amount
(in Taka)
Borrowing/
taking loan
Loan from neighbors/relatives (with interest) 119 293,956
Loan from bank 12 64,000
Loan from NGO 23 107,000
Loan from employer/master 3 6,000
Loan from money lender 57 154,200
Loan from neighbors/relatives (without interest) 115 240,870
Loan food grain from kin 47 49,572
Loan from nearby shop/pharmacy 193 157,413
Use savings Use/consume savings 20 62,800
Selling Sell home/homestead/cultivable land 2 5,700
Sell domestic assets (furniture, utensils, clothing,
means of transport)
3 6,450
Sell livestock (cattle, buffalo, chicken, ducks, hens) 42 73,400
Sell jewelry 1 2,000
Sell agricultural products in advance at lower price 1 2,400
Sell trees 2 5,000
Sell rice stocks 15 31,580
Changing habits Change frequency of meals, food items and reduce
consumption
154
Change occupation or working pattern 8
Taking aid Aid from Government or NGO 33 3,674
Aid from any organization 1 3,000
Aid from neighbors/relatives 4 1,500
Others Begging 3 1,300
Mortgage land 1 5,200
Nat Hazards (2012) 64:1209–1218 1213
123
Author's personal copy

While the frequency of households and amount of coping are illustrated in Table 1,itis
worthwhile to know about the percentage distribution of rural households using coping
strategies against floods in the year 2005 among different districts and poverty levels.
Figure 1shows that borrowing plays a major role for both poor and non-poor households in
each district (as noted in Ninno et al. 2001). The highest proportion of borrowing is taken
by the poor households (89.7 %) in Jamalpur district, and the same group of people sells
their assets at a high proportion (16.1 %) for coping after monsoon flood. The flash flood in
Nilphamari district also forces people, especially the poor households, to take loan
(78.5 %) and sell their assets (15.9 %) at high proportion.
Table 2shows average weekly interest rates mentioned by the flooded households who
borrowed money or goods from different sources. The highest interest rate (8.25 %) is
taken by the money lenders. The average weekly interest rate taken by Government or
private banks (1.4 %) is much lower than those taken by NGOs or money lenders (7.5 %),
but rural people have limited access to claim for the loan during and aftermath of flood.
From focus group discussion, it is depicted that people take loans from NGOs or money
lenders despite high interest rates at an emergency basis. The finding of this study on
interest rates taken by NGOs contradicts with the result of Ninno et al. (2001), who found
from the flood survey in Bangladesh that NGOs took the lowest interest rates, based on the
flood scenario in 1998 on different locations.
0 20 40 60 80 100 120
Poor
Non-poor
Poor
Non-poor
Poor
Non-poor
Poor
Non-poor
Sirajganj
Jamalpur
Sunamganj
Nilphamari
Borrowing
Savings
Selling
Fig. 1 Percentage distribution of households’ coping strategies among districts and poverty levels
Table 2 Weekly interest rates
among borrowing categories
Figures are in percentages
Types of loan Average weekly
interest rate
Loan from neighbors/relatives (with interest) 2.32
Loan from bank (Government or private) 1.38
Loan from NGO 7.54
Loan from money lender/employer/master 8.25
Loan from neighbors/relatives (without interest) 0
Loan from nearby shop/pharmacy/grain from kin 0.28
1214 Nat Hazards (2012) 64:1209–1218
123
Author's personal copy

Table 3shows that about 78 % of the households used the loan for buying food that
amounts to 25 % of the total loan (Taka 1,073,011). Extensive borrowing for purchasing
food, aftermath of flood in 1998, is also depicted the highest frequency (45.2 %) in the
study of Ninno et al. (2001). Two hundred and thirty households spent the borrowing
money for farming which was 20 % of the total amount of loan. About 19 and 18 % of the
total loan were used for business and house repair after flood, respectively. In rural areas,
spending for marriage ceremony and paying dowry are customary events. 6 % of the total
loan was used for marriage disbursement. The category ‘Others’ stands for spending
money on judiciary cases, transition cost of temporary migration, or repaying loans.
Among flooded households (595), only 20 households responded to use their savings for
coping with flood. Table 3also depicts that 75 % of households, out of 20 households who
used savings for coping, utilize their money for food. About 33 % of the saving amount is
spent on food items. About 18 % of savings is used for repaying loans, and about 15 % of
savings is utilized for the agriculture sector.
Table 4shows the estimated results using tobit models for the three main coping
strategies using maximum likelihood estimator (MLE). A household’s decision to borrow
as coping strategy has a positive and statistically significant relationship with the height of
flood; if the flood height increases by one foot then households would borrow on average
Taka 165 more. Male-headed households would borrow more than their female counter-
parts. It may be the reason that 89 % of the surveyed flooded households are male-headed,
and female-headed households are addressed by de facto system. Households were asked
about their coping sources within the last month, as female-headed households have their
easy access to micro-credit in rural areas (Amin et al. 1999), this could be another reason
for the higher rate of borrowing among male-headed households. Interestingly, the amount
of borrowing is negatively correlated with household size but positively correlated with the
dependency ratio that means the more active members a household has the less amount
they borrow for coping. The age dependency ratio for this study is 0.68 that is very near to
the national figure in 2005.
2
Therefore, an increase in the household size indicates addi-
tional working-age people in the sample, and the empirical results in Table 4depicts that
the amount of borrowing is negatively correlated with household size but positively related
Table 3 Utilization of loans and savings by flooded households
Sectors Coping with borrowing/loan Coping with savings
Number of
households
Percentage of total
amount of loan spent
Number of
households
Percentage of total
amount spent from
savings
Agriculture/farming 230 20.4 8 14.64
Business 45 18.9 2 12.72
Health/education 161 7.1 8 08.84
Food 322 25 15 32.69
Housing (repairing) 79 18 2 11.21
Marriage/dowry 27 6 4 17.55
Others 24 4.6 3 2.35
2
In 2005, the national age dependency ratio is 0.64, that is there are about 6 dependents (younger than 15 or
older than 64 years) for every 10 working-age (15–64 years old) people. http://www.nationmaster.com/
compare/Bangladesh/India/People. Last access 15th April, 2012.
Nat Hazards (2012) 64:1209–1218 1215
123
Author's personal copy

with dependency ratio. The loss of assets is positively related to the amount of borrowing
but households with more income are less likely to borrow. Amount of savings spent by the
households shows a significant positive relation with duration of floods, that is if flood
water stays at their homestead one more day then the average spending from savings would
increase by Taka 128. Dependency ratio and loss of working days are also positively
interacted with the money spent from savings. Households with higher income are sup-
posed to have higher savings and can afford to spend from savings in crisis time. Data of
Table 4 Determinants of coping strategies: tobit model estimates
Factors Variables Borrowing
coefficients
Savings
coefficients
Selling assets
coefficients
Shock factors:
Flood height 164.51 (53.06)** -766.97 (572.37) 88.14 (116.41)
Flood duration 1.08 (7.28) 127.93 (67.89)** 25.99 (8.16)***
Loss of crops 0.006 (0.003)** 0.36 (0.39) 0.07 (0.06)
Loss of assets 0.60 (0.11)*** 0.72 (0.83) 0.25 (0.15)*
Loss of working days 8.72 (4.67)* 73.39 (38.39)* 8.40 (9.62)
Demographic factors:
Household size -39.88 (17.37)** -25.62 (151.91) -34.20 (32.59)
Dependency ratio 12.25 (39.26) 209.96 (122.06)* 3.50 (0.54)***
Age of household head -4.80 (2.58)* -2.47 (26.22) -8.05 (5.55)
Average education of
working members
-2.96 (13.82) -120.03 (70.19)* -23.44 (31.11)
Gender of household head
(male =1)
209.64 (116.19)* 159.13 (152.15) -39.51 (49.56)
Occupation
(agriculture =1)
111.52 (63.09)* 476.37 (975.09) 73.95 (70.20)
Economic and community factors:
Per capita income before
flood
-0.12 (0.04)** 1.73 (0.52)** -0.34 (0.25)
Membership in
cooperatives
51.63 (68.89) 761.37 (688.72) 12.54 (6.53)**
District dummy
1(Sirajganj =1)
-117.54 (127.10) 713.58 (628.68) -12.99 (55.88)
District dummy
2(Jamalpur =1)
184.14 (107.86) 192.11 (279.50) 249.20 (205.42)
District dummy 3
(Sunamganj =1)
-133.66 (128.09) -258.40 (238.06) -122.43 (175.13)
Number of uncensored observations 413 20 57
Log
likelihood
-89.55 -116.99 -149.14
LR v
2
112.72 35.29 22.24
Probability [v
2
0.000 0.008 0.001
Pseudo R
2
0.577 0.382 0.413
Dependent variables: amount of borrowing, savings, and selling assets (truncated from the lower limit zero,
i.e. zero amount of borrowing, spending from savings or selling assets), robust standard errors are in
parentheses, values are statistically significant at *** =at 1 %, ** =at 5 %, * =at 10 % level, all the
models include constant terms not reported in the table
1216 Nat Hazards (2012) 64:1209–1218
123
Author's personal copy

this study illustrate a similar pattern. Amount of selling assets is highly significant and
positively correlated with flood duration and dependency ratio. Any membership of
households in the local cooperatives would help to sell their assets during floods. In the
Table 4, it is depicted that households start borrowing when they realize that a flood shock
is taking place. Gradually, they instigate spending money from savings and selling assets
with the extended period of flood.
4 Conclusion
Flooded households have multiple responses for coping with floods in the year 2005. The
highest frequency of households was to borrowing goods and cash from a nearby shop or
pharmacy. The highest proportion of borrowing is taken by the poor households of
Jamalpur district affected by monsoon flood. About 78 % household used the borrowing
for buying food that amounted to 25 % of the total borrowing. Ninno et al. (2001) also
found a similar kind of result from the flood survey in 1998, that a high percentage of
flooded households took loan especially for food consumption. One-fourth of flooded
households reported that they reduced their number of meals and amount of consumption
in a day, or sometimes bought cheap food items to cope with flood and aftermath
(assimilate with the study of Maxwell et al. 1999).
The estimates show that a household’s decision to borrow as coping strategy has a
positive and statistically significant relationship with the height of flood. Male-headed
households would borrow more than their female counterparts. Interestingly, the amount of
borrowing is negatively correlated with household size but positively correlated with the
dependency ratio that means the more active members a household has the less amount
they borrow for coping. The loss of assets is positively related with the amount of bor-
rowing but households with more income are less likely to borrow. The amount of savings
spent by the households shows a significant positive relation with duration of floods.
Dependency ratio and loss of working days are also positively interacted with the money
spent from savings. Amount of selling assets is highly significant and positively correlated
with flood duration and dependency ratio. Any membership of households in the local
cooperatives would help to sell their assets during floods.
It is depicted from tobit model estimates that households start borrowing when they
realized that a flood shock is taking place. Gradually, they start divestment strategy or
spending money from savings and selling assets with the extended period of flood.
Frankenberger (1992) and Rashid et al. (2006) have revealed similar pattern of coping
strategies of flooded households.
A major proportion of households have been found in this study to borrow money or
resources from informal sources, such as nearby shops or the pharmacy, friends or rela-
tives, or local money lenders, to buy food items and other essentials. Another interesting
outcome is that households preferred NGOs over banks in borrowing (twofold in numbers),
despite higher interest rate. The findings indicate that better access to public financial
services like banks could enhance households’ ability to cope with floods and help them to
recover from the debt cycle. Haque and Zaman (1993) and Paul and Routray (2010) have
made similar argument that formal supports and informal coping strategies significantly
improve the capacity of people and reduce vulnerability to floods. More targeted credit
programs would be useful where formal bank credit programs are limited in scope. Pro-
viding access to income-generating sources for the most vulnerable households can both
help to reduce poverty as well as increase their coping capacity against floods.
Nat Hazards (2012) 64:1209–1218 1217
123
Author's personal copy

References
Amin S, Rai AS, Topa G (1999) Does microcredit reach the poor and vulnerable? Evidence from northern
Bangladesh. Working paper 28, Harvard University, Center for International Development
BBS (Bangladesh Bureau of Statistics) (2004) Poverty monitoring survey. Planning Division, Ministry of
Planning, Government of the People’s Republic of Bangladesh, Dhaka
BBS (Bangladesh Bureau of Statistics) (2005) Household income and expenditure survey. Planning Divi-
sion, Ministry of Planning, Government of the People’s Republic of Bangladesh, Dhaka
Cochran WG (1977) Sampling techniques, 3rd edn. Wiley, New York
Few R (2003) Flooding, vulnerability and coping strategies: local responses to a global threat. Prog Develop
Stud 3(1):43–58
Frankenberger T (1992) Indicators and data collection methods for assessing household food security. In:
Maxwell S, Frankenberger T (eds) Household food security: concepts, indicators, measurements. A
technical review. UNICEF and IFAD, New York and Rome
GOB (Government of Bangladesh) (2006) Flood forecasting and warning centre report. Dhaka, Bangladesh
Greene WH (2003) Econometric analysis, 5th edn. Prentice Hall International, New York
Haque CE, Zaman MQ (1993) Human responses to riverine hazards in Bangladesh: a proposal for sus-
tainable floodplain development. World Develop 21(1):93–107
Heitzmann K, Canagarajah RS, Siegel PB (2002) Guidelines for assessing the sources of risk and vulner-
ability. Social protection discussion paper 0218. World Bank, Washington, DC
Holzmann R, Jørgensen S (2000) Social risk management: a new conceptual framework for social pro-
tection, and beyond. Social protection discussion paper no. 0006. World Bank, Washington, DC
Hutton D, Haque CE (2003) Patterns of coping and adaptation among erosion-induced displaces in Ban-
gladesh: implications for hazard analysis and mitigation. Nat Haz 29(3):387–403
Jacoby H, Skoufias E (1997) Risk, financial markets and human capital in a developing country. Rev Econ
Stud 64(3):311–335
Maxwell D, Ahaideke C, Levin C, Armar-Klemesu M, Zakariah S, Lamptey GM (1999) Alternative food-
security indicators: revisiting the frequency and severity of coping strategies. Food Policy 24:411–429
Mirza QMM, Warrick RA, Ericksen NJ, Kenny GJ (2001) Are floods getting worse in the Ganges, Brah-
maputra and Meghna basins? Environ Haz 3(2):37–48
Ninno CD, Dorosh PA, Smith LC, Roy DK (2001) The 1998 floods in Bangladesh: disaster impacts,
household coping strategies, and response. Research report 122, IFPRI, Washington, DC
Paul BK (1997) Flood research in Bangladesh in retrospect and prospect: a review. Geoforum 28(2):
121–131
Paul SK, Routray JK (2010) Flood proneness and coping strategies: the experience of two villages in
Bangladesh. Disasters 34(2):489–508
Rashid DA, Langworthy M, Aradhyula S (2006) Livelihood shocks and coping strategies: an empirical study
of Bangladesh households. Selected paper prepared for presentation at the American Agricultural
Economics Association Annual Meeting, Long Beach, California, July 23–26, (http://ageconsearch.
umn.edu/handle/21231). Last access 23rd Dec 2011
Rayhan MI (2008) Assessing household vulnerability and coping strategies to floods: a comparative study of
flooded and non-flooded areas in Bangladesh, 2005. Cuvilliar Verlag, Goettingen
Scheuer S, Haase D, Meyer V (2011) Exploring multicriteria flood vulnerability by integrating economic,
social and ecological dimensions of flood risk and coping capacity: from a starting point view towards
an end point view of vulnerability. Nat Haz 58:731–751
Tobin J (1958) Estimation of relationships for limited dependent variables. Econometrica 26:24–36
World Disaster Report (2011) International Federation of Red Cross and Red Crescent Societies (http://
www.ifrc.org/PageFiles/…/307000-WDR-2011-FINAL-email-1.pdf). Last access 23rd Dec 2011
Zaman H (1999) Assessing the impact of microcredit on poverty and vulnerability in Bangladesh. Policy
research working paper 2145. World Bank, Washington, DC
1218 Nat Hazards (2012) 64:1209–1218
123
Author's personal copy