World Development: Azreen Karim

World Development 103 (2018) 40–59
Contents lists available at ScienceDirect
World Development
journal homepage: www.elsevier.com/locate/worlddev
The Household Response to Persistent Natural Disasters: Evidence from

Bangladesh
Azreen Karim
Bangladesh Institute of Development Studies, Dhaka, Bangladesh
a r t i c l e i n f o s u m m a r y
Article history: Recent literatures examine the short-run effects of natural disasters on household welfare and health
Accepted 13 October 2017 outcomes. However, less advancement has been observed in the use of self-reported data to capture
the short-run disaster–development nexus in least developed countries’ with high climatic risks. This
self-identification in the questionnaire could be advantageous to capture the disaster impacts on
Key words: households more precisely when compared to index-based identifications based on geographical expo-
economic development sure. In this paper, we ask: ‘‘what are the impacts on household income, expenditure, asset, and labor
natural disasters
market outcomes of recurrent flooding in Bangladesh?” We examine the short-run economic impacts
persistent
measures of disaster risk exposure
of recurrent flooding on Bangladeshi households surveyed in year 2010. In 2010 Household Income
Asia and Expenditure Survey (HIES), households answered a set of questions on whether they were affected
Bangladesh by flood and its likely impacts. We identify treatment (affected) groups using two measures of disaster
risk exposure; the self-reported flood hazard data, and historical rainfall data-based flood risk index.
The paper directly compares the impacts of climatic disaster (i.e., recurrent flooding) on economic devel-
opment. We further examine these impacts by pooling the data for the years’ 2000, 2005, and 2010 and
compare the results with our benchmark estimations. Overall, we find robust evidence of negative
impacts on agricultural income and expenditure. Intriguingly, the self-reported treatment group experi-
enced significant positive impacts on crop income.
Ó 2017 Elsevier Ltd. All rights reserved.
1. Introduction with high climatic risks.2 In this paper, we ask: ‘‘what are the
impacts on household income, expenditure, asset, and labor market
Bangladesh has a long history with natural disasters due to its outcomes of recurrent flooding in Bangladesh?”
geography and its location on the shores of the Bay of Bengal. We examine the short-run economic impacts of recurrent flood-
Climate change models predict Bangladesh will be warmer and ing on Bangladeshi households surveyed in year 2010. In 2010
wetter in the future.1 This changing climate induces flood risk Household Income and Expenditure Survey (HIES), households
associated with the monsoon season each year (Gosling et al., answered a set of questions on whether they were affected by flood
2011). It is now widely understood that climate-induced increas- and its likely impacts. This self-identification in the questionnaire
ingly repeated risks threaten to undo decades of development efforts could be advantageous to capture the disaster impacts on house-
and the costs would be mostly on developing countries impacting holds more precisely when compared to index-based identifica-
existing and future development (Beg et al., 2002; McGuigan, tions based on geographical exposure. However, literatures have
Reynolds, & Wiedmer, 2002; OECD, 2003). Recent literatures exam- identified shortcomings in self-reporting and various determinants
ine the short-run effects of natural disasters on household welfare of flood risk perception.3 Therefore, this paper contributes the fol-
and health outcomes (Arouri, Nguyen, & Youssef, 2015; Lohmann lowing in the ‘‘disaster-development” literature: first, it identifies
& Lechtenfeld, 2015; Lopez-Calva & Ortiz-Juarez, 2009; Rodriguez- treatment (affected) groups using two measures of disaster risk
Oreggia, de la Fuente, de la Torre, Moreno, & Rodriguez, 2013; exposure – the self-reported flood hazard data and historical rainfall
Silbert & del Pilar Useche, 2012). However, less advancement has
been observed in the use of self-reported data to capture the
2
short-run disaster–development nexus in least developed countries Poapongsakorn and Meethom (2013) looked at the household welfare impacts of
2011 floods in Thailand (an upper-middle-income country by World Bank definition)
and Noy and Patel (2014) further extended this to look at spillover effects.
1 3
See Bandyopadhyay and Skoufias (2015). Limitations of self-reported data have been detailed in Section 3(a).
https://doi.org/10.1016/j.worlddev.2017.10.026
0305-750X/Ó 2017 Elsevier Ltd. All rights reserved.
A. Karim / World Development 103 (2018) 40–59 41
data-based flood risk index; second, it directly compares the impacts

of climate disaster (i.e., recurrent flooding) on four development
dimensions i.e., income, expenditure, asset, and on labor market out-
comes. Our novelty in this paper is the identification of flood treat-
ment households using self-reported flood hazard data and
historical rainfall-based flood risk index. The development responses
of the climatic disasters may therefore depend on the novel
approach i.e., accuracy in identifying the treatment groups using
self- and non-self-reported data. In this paper, we show that there
is inconsistency between self- and non-self-reported information-
based estimates with literature outcomes questioning the designa-
tion of survey questions (related to natural shocks) and their useful-
ness to capture development impacts.
The paper is designed as follows: Section 2 describes the theo-
retical framework between social vulnerability and community
resilience. Section 3 reviews the empirical evidences highlighting
recent insights to explore the nexus between climatic disasters Figure 1. Author’s elaboration of the theoretical framework based on Wisner et al.
and economic development in both developed and developing (2004) and IPCC (2014).
countries. Section 4 portrays our identification strategy while Sec-
tion 5 describes the data, provides detailed breakdown of our
methodological framework, identifies the key variables, and justi- ticularly mentioned. In a study on community resilience to coastal
fies the choice of the covariates with added descriptive statistics. hazards in the Lower Mississippi River Basin (LMRB) region in
In Section 6, we present and analyze the estimation results com- South-eastern Louisiana, the Resilience Inference Measurement
paring with previous literatures along with robustness checks in (RIM) Model has been applied to assess the resilience of higher
Section 7. Finally, in Section 8 we conclude with relevant policy and lower resilient communities (Cai, Lam, Zou, Qiang, & Li, 2016).
implications and also some insight for further advancements. Interestingly, the authors identified the location of the lower resili-
ent communities to be along the coastline and in lower elevation
area (in the context of developed country here) that has also been
2. Social vulnerability and community resilience: theoretical argued in the context of developing countries (e.g., Karim & Noy,
framework 2016a). Our aim in this paper is to understand this relationship
among hazard, vulnerability, and exposure and look at the impacts
Figure 1 displays the conventional way to consider disaster risk of climate-induced disaster risks (e.g., flood hazards) on various
as a function of the following factors: socio-economic dimensions (i.e., income, consumption, asset, and
labor market outcomes).
Risk=Disaster Risk ¼ f ðHazard; Exposure; VulnerabilityÞ
where a country’s pre-determined geo-physical and climatic char-

acteristics are part of its hazard profile compared to exposure which 3. Climate disasters and development: empirical evidences
is largely driven by poverty forcing people to live in more exposed
and unsafe conditions (e.g., living in flood plains).4 Poverty is both a The last few years have seen a new wave of empirical research
driver and consequence of disaster risk particularly in countries with on the consequences of changes in precipitation patterns, temper-
weak risk governance (Wisner, Blaikie, Cannon, & Davis, 2004). Vul- ature, and other climatic variables on economic development and
nerability in the above functional form depicts disaster risk not only household welfare. Climate-related natural disasters are expected
depends on the severity of hazards or exposure of urban living and to rise as the earth is getting warmer with prospect of significant
human assets but also on the exposed population’s capacities to negative economic growth mostly affecting the poor countries
withstand and reduce the socio-economic impacts of hazards.5 (Acevedo, 2014; Felbermayr & Gröschl, 2014). Vulnerable econo-
Therefore, disaster risk can be viewed as the intersection of hazard, mies for example, the Pacific islands could expect a growth drop
exposure, and vulnerability. Since resilience has often been defined by 0.7 percentage points for damages equivalent to 1% of GDP in
as the flip-side of vulnerability6, there seems to be a clear connection the year of the disaster (Cabezon, Hunter, Tumbarello, Washimi,
between disaster risk reduction efforts and enhancement of commu- & Wu, 2015). On the causality between catastrophic events and
nity resilience as occurrence and severity of natural hazards is long-run economic growth using 6,700 cyclones, Hsiang and Jina
uncontrollable. However, vulnerability is multi-dimensional and (2014) find robust evidence that national incomes decline com-
dynamic; hence it demands inter-disciplinary approaches to under- pared to pre-disaster trends and the recovery do not happen for
stand both the physical and socio-economic aspects. Literatures have twenty years for both poor and rich countries. This finding con-
attempted to put forth conceptual frameworks in various contexts trasts with the earlier work of Noy (2009) and Fomby, Ikeda, and
and identify global and community-level indicators to quantify vul- Loayza (2013)7 to some extent and carry profound implications as
nerability. Among them; the Hazard-of-Place Model of Vulnerability climate change-induced repeated disasters could lead to accumula-
(Cutter, Boruff, & Shirley, 2003), the Pressure and Release Model tion of income losses over time. Therefore, climate disasters have
(Blaikie, Cannon, Davis, & Wisner, 1994:23), the Social Vulnerability become a development concern with likelihood of rolling back years
Model (Dwyer, Zoppou, Nielsen, Day, & Roberts, 2004:5) and the of development gains and exacerbate inequality.
framework to approach social vulnerability (Parker & Tapsell, Climate resilience has become integral in the post-2015 devel-
2009; Tapsell, McCarthy, Faulkner, & Alexander, 2010) could be par- opment framework and recent cross-country ‘‘micro” literatures
explore the channels through which climate disasters impacted
4
See Karim and Noy (2016a).
5
poverty.8 Recent studies on rural Vietnam looked at the impacts
See Noy and duPont IV (2016).
6 7
See Crichton (1999) and Wilson (2012). However, Cutter, Ash, and Emrich (2014) These studies focus on the short-run effects of natural disasters.
8
found evidences that inherent resilience is not the opposite of social vulnerability Karim and Noy (2016a) provide a qualitative survey of the empirical literature on
using the Baseline Resilience Indicators for Communities (BRIC) metric. poverty and natural disasters.
42 A. Karim / World Development 103 (2018) 40–59
on climate disasters such as floods, storms, and droughts on house- in identifying the true impacts which is also one of the highlights
hold resilience, welfare, and health outcomes (Arouri et al., 2015; in this paper.9
Bui, Dungey, Nguyen, & Pham, 2014; Lohmann & Lechtenfeld,
2015). Arouri et al. (2015) pointed out that micro-credit access,
(a) Limitations of self-reported data
internal remittance, and social allowances could strengthen house-
hold resilience to natural disasters. However, high resilience might
Recent studies have identified various limitations of reported
not necessarily reflect low vulnerability as evident in a study con-
flood risk and showed that perceived flood exposure could be dif-
ducted on tropical coastal communities in Bangladesh (Akter &
ferent from actual risk. In a study conducted in Bray, Dublin city;
Mallick, 2013). Moreover, another study on the Pacific island of
O’Neill, Brereton, Shahumyan, and Clinch (2016) find that distance
Samoa by Le De, Gaillard, and Friesen (2015) suggests that differen-
to the perceived flood zone (perceived flood exposure) is a crucial
tial access to remittances could increase both inequality and vulner-
factor in determining both cognitive and affective components of
ability. Bandyopadhyay and Skoufias (2015) show that climate
flood-risk perception. Another recent study by Trumbo et al.
induced rainfall variability influence employment choices impacting
(2016) develops an interesting measure of risk perception (in the
lower consumption in flood-prone sub-districts in rural Bangladesh.
context of hurricanes) to understand how people make decisions
Agricultural specialization-based occupational choices are also
when facing an evacuation order. This literature found to validate
found to be negatively affected with high variations in rainfall in
previous works and justifies its approach to other contexts within
the Indian context (Skoufias, Bandyopadhyay, & Olivieri, 2017).
natural hazards, and elsewhere. Self-reporting in terms of being
Assessing relationship between household heterogeneity and vul-
affected could be subjective and might bring biased results due
nerability to consumption patterns to covariate shocks as floods
to sorting or selective reporting.10 Self-reported data could not only
and droughts, Kurosaki (2015) identified landownership to be a crit-
be a subject of recall error, but also to other forms of cognitive bias
ical factor to cope with floods in Pakistan. A recent study on the
like reference dependence (Guiteras et al., 2015).
Indian state of Tamil Nadu by Balasubramanian (2015) estimates
the impact of climate variables (i.e., reduction in ground water avail-
ability at higher temperatures than a threshold of 34.31 °C) on agri- 4. Identification strategy
cultural income impacting small land owners to get low returns to
agriculture. In one particular examination on occurrence and fre- Our objective in this paper is to analyze the short-run impacts
quency of typhoons and/or floods in Pasay City, Metro Manila by of recurrent flooding on household income, expenditure, asset,
Israel and Briones (2014) reveals significant and negative effects and labor market outcomes through identification of treatment
on household per capita income. (affected) groups using both self- and non-self-reported data (his-
This literature also explored vulnerability to natural disasters in torical rainfall data based flood risk index). We use the term ‘‘per-
the context of developed countries; for example, the case of hurri- sistent natural disasters” to refer to repeated natural disasters (e.g.,
cane Katrina in the US city of New Orleans. Evidences suggest that flood) that occurs almost every year and possess increase risks of
the pre-existing socio-economic conditions and racial inequality in occurrence due to rainfall variability.11 Our estimation strategy
New Orleans played a crucial role in exacerbating damages due to identifies affected households using two different measures of disas-
Hurricane Katrina in addition to the failure of flood protection ter risk exposure (i.e., flood hazard) and directly compares the
infrastructure and disaster anticipation combined with poor impacts on various socio-economic outcomes. Our primary focus is
responses management (Cutter et al., 2006; Levitt & Whitaker, the year 2010 as shock module was introduced in the 2010 House-
2009; Masozera, Bailey, & Kerchner, 2007). A recent study by hold Income and Expenditure Survey (HIES) with questionnaire
Martin (2015) used a grounded theory approach to develop the related to natural disasters and no new survey has been published
Social Determinants of Vulnerability Framework and applied on at the national level since then.12 The module on shocks and coping
responses was first introduced in HIES 2010 to identify households
the US city of Boston. The author found that those living with
affected by various idiosyncratic and covariate shocks. As our focus
low-to-no income are at the highest risk for negative post-
in this paper is on covariate shocks i.e., flood, we identify households
incident outcomes. Bergstrand, Mayer, Brumback, and Zhang
who have self-reported to be affected by floods only in 2010 survey.
(2015) adds to this social vulnerability-community resilience to
The earlier surveys – 2000 and 2005 – did not have any shock mod-
hazards literature by measuring these indices in counties across
ule and hence identification of self-reported affected groups was not
the United States and find a correlation between high levels of vul-
possible. However, Bangladesh as a disaster-prone country, disasters
nerability and low levels of resilience (indicating that the most vul- particularly flood are a repeated phenomenon every year. Here, we
nerable counties also tend to be the least resilient). The authors took flood as persistent natural disaster due to its repeated occur-
further identified that the Northern parts of the United States, par- rence every year mostly during the monsoon period (May–October).
ticularly the Midwest and northeast, were more resilient and less Due to limitations of the self-reported data (as evident in litera-
vulnerable than the South and West. This finding has also been tures), we identify two ‘‘treatment” groups – treatment group A
confirmed by Cutter et al. (2014) using an alternative resilience and treatment group B to compare the impacts using two different
metric. measures of disaster risk exposure.
This growing ‘‘Climate-Development” literature further The first treatment group i.e., treatment group A is identified
explores empirical patterns in risk, shocks and risk management through the self-reported information using the shock module in
by using shock modules in questionnaire-based surveys to com- year 2010. From 2010 survey, the treatment group are the respon-
plement existing risk management tools. This usage of self- dents who have said ‘‘Yes” as being affected by flood hazard only.
reported information on natural shocks motivated researchers to In 2010, the comparison groups are those households who have
develop different dimension of identification strategies and com-
pare impact findings using econometric models. Two recent stud- 9
See Guiteras, Jina, and Mobarak (2015) and Heltberg, Oviedo, and Talukdar
ies by Noy and Patel (2014) and Poapongsakorn and Meethom (2015).
10
(2013) investigate household welfare and spillover effects of the See Heltberg et al. (2015) for a discussion on how survey modules fall short of
expectations in several ways.
2011 Thailand flood identifying self-reported affected (treatment) 11
See Bandyopadhyay and Skoufias (2015) and Gosling et al. (2011).
group in a difference-in-difference modeling framework. 12
The HIES 2015 survey is currently underway according to the information
Nevertheless, evidences suggest careful use of self-reported data provided by the current Project Director.
responded ‘‘No” to being affected by flood hazard only. To identify with randomly selected households in designated primary sam-
our second treatment group i.e., treatment group B, we use a pling units (PSUs). Therefore, the strength of the dataset is the
rainfall-based flood risk probability index using historical rainfall large sample size covering a broad range of households. However,
dataset13 from the Bangladesh Meteorological Department (BMD) limitations are there in capturing the impacts over time. We fur-
to identify upazilas/thanas14 (in particular, the survey areas) which ther utilize HIES data spanning over a time period of 10 years con-
are affected by more than average rainfall over a long period sisting of years 2000, 2005, and 2010 to check robustness of our
(1948–2012).15 The rule of thumb is the survey areas (i.e., upazi- main results. The number of households in year 2000 is 7,440 with
las/thanas) which have experienced more than average rainfall com- 10,080 and 12,240 in year 2005 and in year 2010 respectively. We
pared to the benchmark of average rainfall of 64 years in the also use the Bangladesh Meteorological Department (BMD) rainfall
corresponding weather station in year 2010 only, the surveyed dataset from 1948–2012 (i.e., 64 years) for 35 weather stations
households in those upazilas fall under treatment group B. The com- across the country to identify flood-affected treatment group in
parison (not affected) group here are those households who resided respective survey years under consideration.
in survey areas that did not experience excessive rainfall compared
to the average rainfall of 64 years in the corresponding weather sta- (b) Methodological framework
tion in year 2010 only. The advantages of using different flood risk
measure in comparable contexts are twofold. First, it justifies homo- Our main aim here is to examine the short-run economic
geneous circumstances among affected households in terms of a impacts of recurrent flooding on households socio-economic out-
common natural shock i.e., flood. Second, we can directly compare comes i.e., income, consumption, asset, and on labor market out-
the development impacts on two different treatment groups and comes. We start by examining the most parsimonious specification:
the differences could refer to discrepancies in capturing the true
yij ¼ a þ b1 Aij þ b2 Bij þ b3 Cij þ cðX ij Þ þ uij ð1Þ
impacts using shock module. Also, it fits well with the distinction
between covariate and idiosyncratic shocks. Households located in where yij is the outcome variable for household (i) in sub-district (j)
the areas with rainfall shocks may not report that they are affected (i.e., income, expenditure, asset and labor market outcomes), b1 rep-
by floods or droughts e.g., if they are not engaged in agriculture. resents the coefficient for treatment group A (self-reported flood
Richer or more educated households may be able to smooth con- impacts only), b2 represents the coefficient for treatment group B
sumption and in this case might not report being affected by rainfall (flood-risk index based shocks only), b3 represents the coefficient
shocks.16 It is also possible that individuals with higher level of edu- for both self-reported disaster (flood) impact and index-based iden-
cation over-report their preparedness behavior in order to present tifications (C), Xij denotes the control variables indicating house-
themselves in a positive way following socially accepted standards holds socio-economic characteristics and infrastructural features,
(Hoffmann & Muttarak, 2017). and uij indicate the error term. We use robust standard errors for
Figure 2 represents the map showing the upazilas/thanas (i.e., our hypothesis tests. The distinction between treatment group A
sub-districts) in which the two treatment groups had been located. (self-reported) and treatment group B (flood-risk index based) will
The red symbol exhibits the self-reported treatment areas (i.e., treat- allow us to directly compare the differences in terms of impacts
ment group A) whereas the blue symbol locates the rainfall-based using these two different measures of disaster risk exposure on
treatment areas (i.e., treatment group B). There are some upazilas household welfare. The constant term, a in our benchmark model
which are found similar in terms of treatment (for both groups – A will define the impacts on the comparison groups i.e., households
and B) and have been identified using the box structure in Fig. 2. who are not affected by repeated flood hazards. The number of
households in treatment group A, B, and C is 271, 2,031, and 46 con-
secutively with 9,938 households being in the control group.
5. Data and methodology To further investigate whether household-level characteristics
(e.g., rural, landownership, and more education) has impacts on
(a) Data description disaster-risk identifications, we further estimate the following
equation:
We use the 2010 Household Income and Expenditure Survey
(HIES) of the Bangladesh economy in our main analysis. The HIES yij ¼ a þ b1 Aij þ b2 Bij þ b3 Cij þ c1 ðX 1ij Þ þ c2 ðX 2ij Þ þ d1 ðAij :X 2ij Þ
is the nationally representative dataset conducted by the þ d2 ðBij :X 2ij Þ þ d3 ðCij :X 2ij Þ þ uij ð2Þ
Bangladesh Bureau of Statistics (BBS) (in affiliation with the Min-
istry of Planning, Government of Bangladesh and technical and The coefficients of the interaction among the treatment groups
financial assistance from the World Bank) that records information – A, B, C and the household-level characteristics i.e., rural,
regarding income, expenditure, consumption, education, health, landownership, and formal education (d1, d2 and d3) will define
employment and labor market, assets, measures of standard of liv- the effect of these characteristics on the magnitude of the impacts
ing and poverty situation for different income brackets in urban on the outcome variables.
and rural areas. The BBS conducts this survey every five (5) years.
The latest HIES conducted in 2010 added four (4) additional mod- (c) Outcome variables and choice of the control variables
ules in which one refers to ‘‘Shocks and Coping” (Section 6B) in the
questionnaire. The BBS HIES is a repeated cross-section dataset Table 7 shows the lists of key outcome and the control variables
(continuous and categorical) and their descriptive statistics for two
different sets of treatment and control groups. Our outcome vari-
ables of interest include four sets of development indicators. They
are: income (income by category), expenditure (expenditure/con-
13
Guiteras et al. (2015) use satellite data for rainfall, but find that these data are sumption by category), asset types, and labor market outcomes.
poorly correlated with actual flooding. Income and expenditure are divided into various sub-groups with
14
Sub-districts are named as ‘‘Upazilas/Thanas” in Bangladesh.
15
statistics shown in per capita household measures. Asset and labor
A breakdown of the index construction has been provided in the Appendix. See
Karim and Noy (2015) for more details.
market outcomes are also sub-divided into various categories (also
16
We thank an anonymous reviewer for pointing out this interesting insight in our described in Table 7). The continuous (monetary) variables in each
analysis. category are inflation-adjusted using consumer price index (CPI)
Figure 2. Map showing the treatment areas (sub-districts) in the study.
data from the Bangladesh Bank17 to allow for comparisons across livelihoods (Fadeeva, 2014; Ferdousi & Dehai, 2014; Japan Bank
different years. for International Cooperation, 2007). Hence, we control for ‘‘rural”
Alleviating poverty is a fundamental challenge for Bangladesh that takes the value 1 if the household resides in a rural area and 0
with the majority of the extreme poor living in rural areas with if otherwise reported. The male member as household head is gen-
considerable flood risk bringing annual agricultural and losses to erally considered as ‘‘bread earner” and a good amount of literature
also highlighted the positive association between female-headed
17
households and poverty especially in developing countries
Bangladesh Bank is the Central Bank of Bangladesh.
(Aritomi, Olgiati, & Beatriz Orlando, 2008; Buvinic & Gupta, 1997; both treatment and comparison groups with marginal variation
Mallick & Rafi, 2010). Female-headed households are particularly (approximately 2.2% higher) observed in treatment in rainfall-
vulnerable to climate variability as well (Flatø, Muttarak, & based identifications. Observable differences could also be seen
Pelser, 2017). Therefore, a dummy variable has been created indi- in labor market outcomes between both treatment and comparison
cating 1 if the household head is male and 0, if reported otherwise. groups. Daily wages are found to be somewhat higher (almost
Household characteristics such as age structure and number of 0.7%) in treatment group B whereas households identified in
dependents are critical to analyze poverty status and one might treatment group A seem to earn more salaried wages (around 2%
expect larger number of dependents leads to greater poverty higher) compared to their respective comparison groups – B and
(Haughton & Khandker, 2009; Kotikula, Narayan, & Zaman, 2010; A. Intriguingly, the rainfall-based treatment households (B) are
Lanjouw & Ravallion, 1995). Education is also related with lower found to earn more daily wages compared to more salaried wages
poverty (Kotikula et al., 2010). Community-level characteristics earned by the self-reported treatment (A) cases. There are interest-
such as access to sanitation and access to safe drinking water are ing parallel trends in the mean results of the control variables
clearly associated with better health outcomes improving poverty (independent variables) between the two treatment groups. More
status (Duflo, Galiani, & Mobarak, 2012; World Bank, 2014) of self-reported households are found to reside in the rural areas
households with access to electricity also showing a positive trend showing their dependency in rain-fed agriculture.18 The rainfall-
in living standards (Kotikula et al., 2010). Therefore, three (3) bin- based flood treatment households (treatment group B) have more
ary variables are created indicating 1 to imply access to these ser- working adults i.e., fewer dependents (around 0.3%) compared to
vices, 0 otherwise. Ownership status of households such as house the self-reported identifications. However, the self-reported flood
and land has also been argued as important determinant of poverty treatment group owns more land (around 11%) and houses (almost
with owners of a dwelling place are found to be less vulnerable to 6% higher) compared to the non-self-reported ones. Community
flood risk (e.g., Gerstter, Kaphengst, Knoblauch, & Timeus, 2011; characteristics such as access to sanitation, safe drinking water,
Khatun, 2015; Meinzen-Dick, 2009; Rayhan, 2010; Tasneem & and electricity also show parallel trends in their mean outcomes in
Shindaini, 2013). A description of these variables including sum- both treatment groups – A and B.
mary statistics is also provided in Table 7.
6. Estimation results
(d) Descriptive statistics
We start by estimating our benchmark model (as specified in

We provide descriptive statistics for two different treatment
Eqn. (1)) with treatment groups identified using two measures of
and comparison groups (treatment group A and treatment group
disaster risk exposure: self-reported data (treatment group A)
B) in Table 7. We present mean and standard deviation for various
and historical rainfall-based flood risk index (treatment group B).
outcome categories and control variables for both rainfall-based
We estimate our model on development dimensions such as
and self-reported treatment (affected) and comparison (not
income, expenditure, assets, and labor market outcomes. We
affected) groups. Most of the income categories seem to be higher
therefore, compare our results for each category (in terms of aggre-
for the comparison group compared to treatment for treatment
gate and disaggregated outcome measures) with previous litera-
group A (self-reported) with exception in the ‘‘crop income” cate-
tures and extend our analysis by estimating our model specified
gory. The crop income per capita for treatment group A is on aver-
in Eqn. (2).
age, almost 11% higher compared to the comparison group. The
other treatment group i.e., treatment group B (rainfall-based flood
treatment households) also do not show too much variation in (a) Income
terms of mean income by categories. However, mean of the ‘‘other
income” turns out to be almost 10% lower for the comparison Table 1 reports the impacts of recurrent-flooding on different
group compared to treatment in treatment group B. The compar- income categories i.e., crop, non-crop, business, and other income
ison group also have around 1.2% less business income compared for self-reported treatment group (A) and rainfall-based flood
to treatment in contrary to most income categories in the non- affected treatment group (B). We find both treatment (affected)
self-reported case. The expenditure categories also reveal interest- households experience negative impacts on total income being
ing patterns in agricultural expenditure (i.e., crop and non-crop), in consistent with previous disaster literatures (e.g., Asiimwe &
particular. Non-crop expenditure in treatment group A is about Mpuga, 2007; De la Fuente, 2010; Thomas, Christiaensen, Do, &
4.5% higher with having a lesser variation in crop expenditure Trung, 2010). Our results indicate that total income reduces by
(around 1.5% higher) compared to the comparison group. Agricul- almost 1.1% more (estimated to be approximately BDT 11,665)
tural input also reveals a higher expenditure amount (i.e., approx- for treatment group B compared to the mean.19 A decline in crop
imately 2.5%) in treatment compared to comparison group A. income is significantly higher for treatment group B (by around
Interestingly, most of the expenditure categories in comparison BDT 3,456) whereas both treatment group (C) observe comparatively
group B seems to be higher than treatment with exceptions in greater reduction in non-crop income (by approx. BDT 23,601) being
‘‘non-crop” expenditures (around 0.4% lower). Interesting contrast consistent with evidences that show decline in agricultural income
could also be portrayed in educational and health expenditure cat- due to rainfall shocks (e.g., Baez & Mason, 2008; Skoufias,
egories for both treatment groups. Educational and health expendi- Katayama, & Essama-Nssah, 2012; UNISDR, 2012). We do not
tures are found to be less in comparison group B with exceptions in observe any significant negative impacts on business income (non-
comparison group A (in health expenditures) compared to their agricultural enterprise) and other income in both treatment cases.
respective treatment groups – B and A. However, on average, the These results could also be justified by previous works done by
educational and health expenditure are found to be higher in Attzs (2008) and Patnaik and Narayanan (2010).
self-reported treatment group (A) compared to the non-self- The rainfall-based affected group (treatment group B) experi-
reported one (B). It is here to note that, the proportion of house- enced a fall in both crop and non-crop income (although coefficient
hold members getting access to formal education exhibits almost of crop income is significant). Although the self-reported affected
a similar pattern in both treatment and comparison groups - A
and B. Parallel trends could also be observed in terms of total 18
See Haile (2005).
change in agricultural and other business asset categories between 19
1 US Dollar = 77.88 Bangladeshi Taka (BDT).
Table 1
Impact on household income per capita
Variables (1) (2) (3) (4) (5)

Total income Crop income Non-crop income Business income Other income
Treatment group A 1,566.17 21,884.21*** 4,933.50 16,937.20 1,329.77
(Self-report) (37,329.74) (8,328.57) (33,804.55) (15,403.25) (2,070.46)
Treatment group B 11,665.43 3,455.71* 17,499.10 6,634.50 1,780.13
(Rainfall-based) (12,850.91) (2,094.36) (10,849.45) (6,593.46) (1,564.46)
Both treatment group C 70,484.30 3,764.62 23,600.94 97,687.97 294.26
(70,701.24) (12,313.16) (32,642.82) (59,432.73) (4,747.52)
Constant 3898785.83*** 18,342.35 3925239.35*** 898.32 52,667.38
(151,241.77) (18,239.77) (134,173.44) (47,079.97) (39,908.79)
Observations 12,242 12,222 12,232 12,242 12,242
R-squared 0.19 0.12 0.12 0.21 0.05
Treatment group A 332,832.82** 142,294.50** 109,491.86 78,193.44 6,665.53

(Self-report) (130,063.55) (66,072.70) (93,443.48) (59,470.25) (7,903.15)
Treatment group B 32,407.34 13,883.61* 41,104.20* 2,247.68 1,127.58
(Rainfall-based) (40,347.94) (7,825.42) (23,314.76) (23,700.09) (4,797.41)
Both treatment group C 250,726.57 34,353.69 71,339.47 212,991.08 4,037.39
(250,456.37) (31,556.09) (108,105.23) (289,247.54) (9,569.99)
Constant 3909442.31*** 21,563.08 3924719.01*** 3,824.37 53,964.73
(151,012.71) (18,140.35) (134,578.19) (47,057.14) (40,194.24)
Observations 12,242 12,222 12,232 12,242 12,242
R-squared 0.20 0.13 0.12 0.21 0.05
Source: Author’s calculations.

Notes: aRobust standard errors in parentheses ***p < 0.01, **p < 0.05, *p < 0.1.
b
The upper part of the Table (above the middle line) shows estimation results using Eqn. (1) i.e., regression without interaction terms. The bottom part of the Table (below the
middle line) shows estimation results using Eqn. (2) i.e., regression with the interaction terms. All control variables are included in the models, but not displayed.
group (A) observed a fall in total income, there has been a signifi- Mpuga, 2007; Auffret, 2003; Dercon, 2004; Foltz, Gars, Özdogan,
cant increase in crop income. However, crop income decreases by Simane, & Zaitchik, 2013; Jha, 2006; Shoji, 2010). Our focal cate-
almost BDT 3,765 for both treatment groups (C). The interesting gories i.e., crop and agricultural input expenditures (as we assume
thing to note here is that persistent flooding seems to impact these categories are directly related to rainfall shocks and flood)
non-crop income in higher magnitude. Our results show that treat- show negative impacts for rainfall-based affected households.
ment group B (rainfall-based) experienced a drop of almost BDT This evidence, however demonstrates a significant decrease in
12,566 more in non-crop income compared to the treatment group agricultural input expenditure in particular. Food and non-food
A (self-reported). The other two categories of income we analyze expenditures are found to decrease in treatment groups – A and
are business and other incomes which are more indirectly affected B. However, although both categories show sign consistencies,
by flood hazards. Business and other income are found to decrease non-food expenditures are found to be statistically significant for
(not significant) for the self-reported affected households. How- treatment group B. This observation is found to aggravate in fur-
ever, in both of these categories, we observe positive coefficients ther investigations associated with the interactions. This decrease
for affected households who had been identified through rainfall- in non-food spending is particularly of concern as it implies the
based identifications. possibility that disasters prevent longer term investments and
We extend our analysis on households agricultural income (as therefore trap households in cycles of poorer education and health
assumed to have direct impact through repeated flooding) by fur- outcomes and persistent poverty (Karim & Noy, 2016b). These evi-
ther investigating their relationship with rural (defining reliance dences turn out to be interesting when we extend our analysis by
on agriculture), formal education and ownership of land.20 Agricul- further investigating the relationship with rural, formal education,
tural incomes (crop and non-crop) are found to drop significantly for and landownership of households.20 Interestingly, we find that
rainfall-based affected households (B). Crop income is also found to crop expenditure increases significantly for self-reported treat-
decrease in higher magnitude in both self and non-self-reported ment group (A) (estimated BDT 21,798) whereas non-crop expen-
cases (but not significant). The interesting thing to note here is that, diture per capita significantly increases (estimated approx. BDT
crop income had increased quite significantly (around 5.6% more) 37,026) for both rainfall-based and self-reported treatment group
compared to the mean for treatment group A impacting on total (C).
income as well. The various categories of expenditure – food, non-food, crop,
non-crop, agricultural input, educational and health expenditure
– could also be categorized based upon their time horizons e.g.,
(b) Consumption/expenditure short- and medium to long-run impacts. Expenditure categories
as food, non-food, and agricultural consumption indicate the
We report impact estimates of various expenditure categories short-term impacts whereas education and health expenditures
i.e., food, non-food, crop, non-crop, agricultural input, education, may lead to longer term impacts. The treatment households (A
and health for non-self- and self-reported treatment groups in only and B only) experienced significant contrast in terms of the
Table 2. Our results show a significant decline of around 1% com- direct impacts (food and non-food in estimated model with inter-
pared to the mean in total expenditure per capita (i.e., drop by actions).21 Positive estimates have been observed in education and
approx. BDT 14,742) for treatment group B (non-self-reported) health spending for treatment groups A and B as well. However,
being consistent with previous literatures (e.g., Asiimwe &
20 21
Full tables are shown in the appendix and also in an online appendix. Estimated using Eqn. (2).
Table 2
Impact on household expenditure per capita
Variables (1) (2) (3) (4) (5) (6) (7) (8)

Total exp Food exp Non-food exp Crop exp Non-crop exp Agri input exp Edu exp Health exp
Treatment group A 19,801.55 66.47 2,326.36 2,171.13 5,506.69 9,650.88 2,922.63 2,067.18
(Self-report) (18,141.33) (489.80) (10,839.23) (2,263.55) (3,974.32) (7,286.28) (3,523.85) (1,430.33)
A. Karim / World Development 103 (2018) 40–59

Treatment group B 14,742.40** 89.76 6,897.35* 773.76 624.42 6,963.83*** 752.76 164.74
(Rainfall-based) (6,432.06) (153.68) (4,145.87) (930.07) (1,063.60) (2,598.86) (1,160.53) (323.03)
Both treatment group C 10,980.94 1,262.74 14,489.95 181.87 1,460.01 20,035.51 1,641.43 3,364.20
(45,247.27) (886.03) (24,036.11) (6,207.58) (4,511.67) (21,969.91) (6,914.43) (3,867.18)
Constant 1114279.73*** 22,677.58*** 257,176.61*** 102,041.95*** 151,561.73*** 640,434.52*** 42,389.94*** 5,984.99***
(54,989.83) (1,386.18) (31,418.65) (8,324.48) (8,163.02) (23,707.98) (8,298.83) (1,786.27)
Observations 12,242 12,242 12,242 12,222 12,232 12,242 12,242 12,242
R-squared 0.63 0.92 0.52 0.36 0.25 0.31 0.27 0.04
Treatment group A 109,304.40 893.17 24,065.29 21,798.01*** 23,606.10 29,781.65 8,633.44 3,046.33
(Self-report) (66,457.58) (1,842.23) (38,785.57) (7,093.41) (32,813.66) (26,748.11) (11,254.61) (3,285.08)
Treatment group B 30,077.56 471.48 28,168.21** 3,412.53 9,460.94 3,632.92 3,452.94 204.28
(Rainfall-based) (24,804.11) (585.81) (14,352.09) (2,967.95) (8,839.87) (9,704.29) (3,460.80) (856.08)
Both treatment group C 110,647.41 975.08 8,624.64 11,033.88 37,025.74* 82,551.46 17,473.19 12,004.26
(190,050.55) (3,115.48) (90,739.61) (12,885.29) (21,240.47) (81,685.85) (20,801.70) (13,917.93)
Constant 1117883.19*** 22,543.11*** 257,255.92*** 102,577.12*** 153,919.47*** 641,350.11*** 42,321.43*** 5,901.83***
(55,285.22) (1,394.69) (31,353.61) (8,307.50) (8,140.76) (23,813.61) (8,316.00) (1,793.13)
Observations 12,242 12,242 12,242 12,222 12,232 12,242 12,242 12,242
R-squared 0.63 0.92 0.52 0.36 0.25 0.31 0.27 0.04

Notes: aRobust standard errors in parentheses ***p < 0.01, **p < 0.05, *p < 0.1. bThe upper part of the Table (above the middle line) shows estimation results using Eqn. (1) i.e., regression without interaction terms. The bottom part of
the Table (below the middle line) shows estimation results using Eqn. (2) i.e., regression with the interaction terms. All control variables are included in the models, but not displayed.
47
Table 3
Impact on total asset outcomes
Variables (1) (2) (3)

Total change in agricultural and other business asset Total agricultural input asset value Total consumer durable asset value
Treatment group A 1,990.73 15,691.41 17,421.92
(Self-report) (33,184.36) (14,723.87) (88,561.35)
Treatment group B 4,475.63 6,620.57 25,310.38
(Rainfall-based) (12,235.93) (5,435.37) (30,435.15)
Both treatment group C 11,009.94 16,831.74 154,958.27
(69,402.10) (21,241.71) (122,824.84)
Constant 633,711.13*** 184,938.08*** 1197123.52***
(103,531.85) (36,826.19) (234,716.60)
Observations 12,242 12,242 12,242
R-squared 0.02 0.07 0.23
Treatment group A 111,132.59 90,455.01* 256,836.22

(Self-report) (75,127.02) (47,600.49) (220,139.63)
Treatment group B 28,898.86 3,374.68 7,225.67
(Rainfall-based) (30,787.83) (15,521.03) (91,144.36)
Both treatment group C 178,097.20* 82,060.72 291,623.55
(101,758.84) (57,755.00) (469,256.70)
Constant 637,696.80*** 191,216.45*** 1195841.23***
(103,626.07) (36,935.21) (235,092.39)
Observations 12,242 12,242 12,242
R-squared 0.02 0.07 0.24

Notes: aRobust standard errors in parentheses ***p < 0.01, **p < 0.05, *p < 0.1. bThe upper part of the Table (above the middle line) shows estimation results using Eqn. (1) i.e.,
regression without interaction terms. The bottom part of the Table (below the middle line) shows estimation results using Eqn. (2) i.e., regression with the interaction terms.
All control variables are included in the models, but not displayed.
the rainfall-based affected households experienced a decrease in what has been justified in some previous empirical researches
educational expenditure (approximately by BDT 3,453) compared (e.g., Banerjee, 2007; Shah & Steinberg, 2012).23 However, real wages
to a sharp decrease by both flood treatment households (C) (esti- are found to decrease for flood-affected (self-reported) households in
mated approx. by BDT 17,473). Intriguingly, the total expenditure both estimations 1 and 2 (but in this case without statistical signif-
in the self-reported treatment group (A) increases (although not sig- icance). Interestingly, salaried wage seems 2.7% higher compared to
nificant) compared to a significant decline for the non-self-reported the mean (estimated approx. BDT 2,969) in treatment group A with
group in our benchmark estimation. 0.3% drop (compared to the mean) for treatment group B but without
statistical significance as well.24 This result is also partially found
consistent with the findings of Mueller and Quisumbing (2011).
(c) Asset
The other labor market outcomes are found to significantly improve
for flood-affected (rainfall-based) households when the estimated
Table 3 demonstrates the impacts of repeated-flooding on three
model (Eqn. (2)) interacted with rural, formal education and land
asset categories: changes in agricultural and other business asset,
ownership status.20 We also observe a contrast in estimates of yearly
agricultural input asset value and consumer durable asset value
benefits for treatment groups – A and B.
for both affected (treatment) groups. We do not observe much con-
trast in these categories though. Both treatment group (C) experi-
enced significant negative impacts (estimated by BDT 178,097) on (e) Control variables
change in agricultural and other business asset quite consistent
with previous evidences on asset categories (e.g., Anttila-Hughes We present the coefficients of the control variables for the main
& Solomon, 2013; Mogues, 2011). Intriguingly, the self-reported variables of interest in Table 5.20 The coefficients of the control
flood affected group (treatment group A) observe significant posi- variables do not vary substantially in terms of sign and significance
tive impacts (estimated by BDT 90,455) in the category represent- for treatment groups – A, B, and C. Among the controls; male-
ing agricultural input asset value. These evidences are particularly headed households, average age, and formal education seem to
valid when we incorporate interaction terms in our estimated have a stronger positive association (highly significant) with total
model.20 Nevertheless, the self-reported flood treatment house- income and total expenditure per capita in addition to community
holds (A) experienced a decline on change in agricultural and other characteristic such as access to sanitation. Ownership of land
business assets when the estimated model do not account for the demonstrates a stronger positive impact (highly significant) on
interaction terms. per capita total expenditure. It is more likely that the household
heads possess control over ownership of land and house.25 How-
ever, the number of dependents displays a stronger negative associ-
(d) Labor market ation with total expenditure as evident in the literatures as well. We
also anticipate similar reasoning as of previous literatures for
We present impacts on labor market for both treatment groups
– A and B in Table 4 and our results reveal contrasts in households
experiences. Daily wages are not found to be severely affected 23
Banerjee (2007) find that floods have positive implications for wages in the long
(positive impact) with statistical significance for rainfall-based run. Interestingly, Mueller and Osgood (2009) reveal that droughts have significant
negative impacts on rural wages in the long run. We are quite agnostic on the general
flood treatment households (estimated by BDT 146).22 This some-
implications of natural disasters on wages due to limitations in this study.
24
Estimated using Eqn. (1).
22 25
Estimated using Eqn. (2). See Zaman (1999).
Table 4
Impact on labor market outcomes
Variables (1) (2) (3) (4) (5) (6)

Total month per year Total days per month Total hours per day Daily wage Salaried wage Yearly benefits
Treatment group A 1.03 5.56 1.60 69.98 2,969.25 5,611.90
(Self-report) (5.50) (12.53) (4.02) (61.39) (2,771.30) (5,455.67)
Treatment group B 2.92 9.61** 1.74 26.60 356.90 2,381.25
(Rainfall-based) (1.94) (4.53) (1.52) (22.82) (936.56) (1,954.30)
Both treatment group C 15.43 16.51 10.98 77.83 243.26 7,051.62
(9.39) (24.12) (7.63) (125.25) (6,422.49) (12,622.45)
Constant 339.91*** 829.15*** 246.68*** 773.85*** 121,717.37*** 287,370.86***
(19.96) (44.49) (13.77) (183.22) (7,021.86) (14,980.65)
Observations 12,242 12,242 12,242 12,242 12,242 12,242
R-squared 0.87 0.87 0.87 0.54 0.35 0.21

(Self-report) (19.25) (42.47) (13.81) (190.89) (9,676.01) (17,064.58)
Treatment group B 16.67** 26.18* 10.35** 146.17* 1,858.83 6,788.29
(Rainfall-based) (6.91) (15.46) (5.07) (76.18) (3,296.66) (6,395.29)
(31.79) (79.09) (29.26) (416.49) (22,900.16) (42,232.08)
Constant 341.74*** 831.20*** 247.55*** 775.72*** 121,591.01*** 287,205.30***
(20.36) (45.04) (13.93) (186.37) (7,032.51) (14,982.38)
Observations 12,242 12,242 12,242 12,242 12,242 12,242
R-squared 0.87 0.87 0.87 0.54 0.35 0.21

Notes: aRobust standard errors in parentheses ***p < 0.01, **p < 0.05, *p < 0.1. bThe upper part of the Table (above the middle line) shows estimation results using Eqn. (1) i.e.,
regression without interaction terms. The bottom part of the Table (below the middle line) shows estimation results using Eqn. (2) i.e., regression with the interaction terms.
All control variables are included in the models, but not displayed.
Table 5
Effects of controls on outcome variables
Variables (1) (2) (3) (4) (5) (6) (7) (8)

Total income Total exp Asset stock Daily wage Total income Total exp Asset stock Daily wage
Rural 3,246.72 2,452.84 19,365.10* 21.61 10,588.13 1,802.73 14,252.62 10.61
(13,260.48) (5,785.10) (10,920.89) (20.06) (14,485.90) (6,285.55) (11,702.90) (21.82)
Male-headed HH 2,370,129.94*** 392,158.61*** 5,389.58 1,546.64*** 2,368,798.65*** 391,391.05*** 8,218.23 1,541.51***
(71,988.86) (18,834.92) (22,300.52) (70.46) (71,878.91) (18,671.89) (22,034.97) (71.77)
Average age 57,433.59*** 29,019.50*** 25,045.10*** 30.07*** 57,682.42*** 29,159.82*** 25,416.28*** 29.37***
(5,128.94) (2,439.89) (4,168.51) (7.92) (5,132.86) (2,415.98) (4,157.71) (8.00)
Dependent 96.50 2,413.53*** 5,188.91*** 10.99*** 90.86 2,399.32*** 5,228.63*** 10.94***
(728.14) (342.55) (791.12) (1.33) (727.59) (341.96) (790.41) (1.34)
Formal education 11,329.07*** 20,519.20*** 3,548.28*** 38.48*** 11,341.01*** 20,513.60*** 3,649.11*** 38.76***
(866.99) (421.58) (944.08) (1.65) (872.70) (420.18) (945.78) (1.66)
Access to sanitation 42,722.54*** 17,254.61*** 4,356.36 42.37** 43,113.45*** 17,293.25*** 4,299.30 42.27**
(11,057.01) (5,193.54) (9,517.48) (17.95) (11,035.25) (5,193.54) (9,519.76) (17.96)
Access to safe drinking 17,101.60 8,929.27 37,269.07 35.25 17,789.45 9,517.59 38,126.05 36.23
water
(28,182.97) (13,160.49) (28,588.08) (45.52) (28,262.85) (13,168.53) (28,610.75) (45.58)
Access to electricity 3,600.44 8,376.70 8,381.50 15.66 3,523.53 8,420.74 8,129.74 15.70
(12,373.04) (5,550.28) (10,063.28) (19.20) (12,379.56) (5,550.62) (10,016.54) (19.21)
House ownership 6,480.10 919.13 9,091.28 3.63 6,097.00 994.99 9,210.96 3.80
(14,178.52) (6,574.19) (12,448.63) (22.71) (14,170.55) (6,575.79) (12,464.21) (22.73)
Land ownership 56.47 126.65*** 36.87 0.20*** 35.75 105.74*** 11.72 0.16**
(37.49) (20.49) (46.59) (0.06) (41.86) (22.18) (43.81) (0.07)
Constant 3898785.83*** 1114279.73*** 633,711.13*** 773.85*** 3909442.31*** 1117883.19*** 637,696.80*** 775.72***
(151,241.77) (54,989.83) (103,531.85) (183.22) (151,012.71) (55,285.22) (103,626.07) (186.37)
Observations 12,242 12,242 12,242 12,242 12,242 12,242 12,242 12,242
R-squared 0.19 0.63 0.02 0.54 0.20 0.63 0.02 0.54

Notes: aRobust standard errors in parentheses ***p < 0.01, **p < 0.05, *p < 0.1. bEach column shows the effects of the control variables in the estimated regression results. The
first four columns (i.e., columns 1–4) show the estimation results using Eqn. (1) i.e., regression without interaction terms for the main variables of interest. The last four
columns (i.e., columns 5–8) show the estimation results using Eqn. (2) i.e., regression with the interaction terms. The variable ‘‘Asset Stock” represents total change in
agricultural and other business asset in both columns 3 and 7. All other variables are included in the models, but not displayed.
observing the control variables to be in expected directions for (f) Interaction terms
asset categories and labor market outcomes. The directions of
the control variables are also found quite similar when the To further investigate whether household characteristics e.g.,
model has been estimated by incorporating the interaction rural, formal education and landownership status have impacted
terms (Eqn. (2)). declines in the development dimensions; we estimate our model
Table 6
Coefficients of the interaction terms of main outcome variables of interest
Variables (1) (2) (3) (4)

Total income Total exp Asset stock Daily wage
Treatment group A*Education 3,426.55*** 1,586.93* 1,959.09** 2.71
(1,277.09) (843.68) (942.53) (2.46)
Treatment group B*Education 377.46 147.16 559.43 1.84*
(474.31) (318.76) (402.79) (0.97)
Treatment group A*Landownership 1.33 111.22 339.02* 0.34
(142.78) (155.19) (192.74) (0.56)
Treatment group B*Landownership 153.40* 131.33*** 228.12 0.25
(88.02) (48.84) (196.41) (0.18)
Treatment group A*Rural 105,290.92 36,706.85 90,296.50 4.07
(103,140.17) (42,104.29) (65,845.48) (138.43)
Treatment group B*Rural 28,266.56 6,592.90 37,888.60 60.27
(26,892.41) (13,508.65) (27,284.49) (48.49)
Both treatment C*Education 13.84 307.93 4,540.33 2.14
(3,046.68) (2,278.24) (3,247.31) (5.70)
Both treatment C*Landownership 216.98 177.94 187.55 0.42
(475.39) (205.67) (197.76) (0.52)
Both treatment C*Rural 285,759.28* 124,226.08 254,137.66 540.58*
(154,600.41) (95,121.54) (222,356.04) (290.29)
Constant 3909442.31*** 1117883.19*** 637,696.80*** 775.72***
(151,012.71) (55,285.22) (103,626.07) (186.37)
Observations 12,242 12,242 12,242 12,242
R-squared 0.20 0.63 0.02 0.54

Notes: aRobust standard errors in parentheses ***p < 0.01, **p < 0.05, *p < 0.1. bEach column shows the coefficients of the interaction terms in the estimated regression results of
main outcome variables of interest (i.e., estimated using Eqn. (2)). The variable ‘‘Asset Stock” represents total change in agricultural and other business asset in column 3. All
other variables are included in the models, but not displayed.
(Eqn. (2)) by incorporating the interaction terms. Table 6 present In the income category, we observe significantly negative
only the results of the interplay among the identified treatment impacts on non-crop income (drop by approx. BDT 8,497)28 due
groups – A, B, and C with rural, formal education, and landowner- to persistent flood hazard. The interesting aspect to note here is that
ship status.20 Interestingly, when the interaction terms are agricultural income (in particular, non-crop income) is found to
included in the model, they seem to increase both the main effects decline more (additional drop by around BDT 9,402) in our focal year
of the treatment groups and the respective control variables. When 2010 for flood-treatment households that are rainfall-based only.
interacted with rural, treatment group C experienced lower total However, business income is found to increase significantly when
income but higher daily wage with the latter indicating coping the estimated model interacted with rural, formal education, and
and/or adaptation strategy e.g., diversifying livelihoods and income landownership status (Eqn. (2)). Our findings also reveal a significant
source. The interaction terms between self-reported flood treat- positive increase in other income category with no interactions
ment households and education in total income and total expendi- being consistent with our prior estimations.
ture per capita are found to be negative and statistically significant. We find consistency in the robust coefficients in total expendi-
Alternatively, education has a positive influence on disaster pre- ture category compared to our baseline model specifications. Flood
paredness only for those who have not yet experienced a disaster treatment households experienced a significant decline in total
in the past (Hoffmann & Muttarak, 2017). Landownership seems expenditure, in particular non-food and agricultural input expendi-
to play a crucial role for the rainfall-based flood treatment house- ture (Eqn. (1)). These impact evidences are found to exacerbate
holds. The coefficients of the interaction terms for per capita total when food expenditures are also observed to decrease signifi-
income and expenditure between treatment group B and landown- cantly.29 The noticeable aspect here is that findings reveal an addi-
ership are found to be positive and statistically significant (not in tional decline in agricultural input expenditure (estimated around
higher magnitude) and are also consistent with previous litera- BDT 4,606) significantly contributing to an excess decline in total
tures (e.g., Kurosaki, 2015) (see Tables 8–13). expenditure (by approx. BDT 10,225) for flood treatment households
(rainfall-based). Non-food expenditures also seems to contribute to
this overall expenditure decline (an additional drop by approx.
7. Robustness checks BDT 4,238) and are found consistent with the benchmark estimation
results. Educational and health expenditures are also found to be
As robustness checks, we further examine these impacts by consistent with our prior estimations.
pooling the data for the years 2000, 2005, and 2010 and compare The impacts on agricultural input asset value display negative
the results with our benchmark estimations. As self-reported data impacts on treatment households (rainfall-based flood risk mea-
were unavailable for years’ 2000 and 2005; we therefore, estimate sure) that are found consistent with the benchmark results. Inter-
Eqns. (1) and (2) through identifications of flood treatment house- estingly, the impacts on changes in agricultural and other business
holds using rainfall-based disaster risk measure only to check asset category exhibit positive coefficient (not significant) com-
robustness of our main results.26 We also add year fixed effects in pared to a decline in prior estimation results.30 The category on
our estimated models.27 consumer durable asset value also illustrates consistency in the esti-
26 28
The full tables of the robustness check estimation results are shown in the Estimated using Eqn. (1) i.e., without interaction terms.
29
appendix and also in an online appendix. This is evident when the estimated model interacted with landownership, rural
27
We estimate the following equation and extend by adding the interaction terms and formal education of households.
30
(as of Eqn. (2)) to check robustness of our main results: yijt = at + b2 bijt + c (Xijt) + uijt. Estimated using Eqn. (2).
mated coefficients for flood treatment households. The various out- based on ‘‘yes/no” responses (i.e., close-ended) might not be suffi-
comes of the labor market do not seem to significantly vary with cient to identify the true development impacts. The selection of the
prior estimations as well. respondents (sample) in this particular set of questionnaire (shock
questions on natural disasters) is also questionable depending on
criteria.32 There is an obvious need to employ both qualitative and
8. Conclusion quantitative techniques to understand the degrees of experience in
impact analysis.33 One possible solution is, of course, more respon-
Our objective in this paper is to estimate the impacts of recur- dents and data availability in addition to incorporating degrees of
rent flooding on income, expenditure, asset, and labor market actual hazard awareness, experience, and preparedness questions
outcomes. We start with identification of the treatment (affected) to identify the real affected group in repeated natural shocks. There
groups adopting two measures of disaster risk exposure i.e., using is a need to thoroughly analyze the inconsistencies in the robust
self-reported flood hazard data and non-self-reported (historical research findings based on the shortcomings identified in the litera-
rainfall-based flood risk index) information in year 2010. We ture. However, the evidence and the novel approach that we adopt in
examine a parsimonious model to directly compare the short- this paper could justify future research in estimating welfare
run impacts of climatic disaster (i.e., repeated flood hazard) on impacts of climate-induced persistent natural events in developing
households socio-economic outcomes. Our results suggest a countries.
decline in agricultural income (crop and non-crop) for both
treatment groups – A (self-reported) and B (rainfall-based). This
significant decline in agricultural income, being consistent with Acknowledgments
previous literatures reveals a clear message on timely adoption
of insurance in the context of increased climatic threat to achieve I would like to gratefully thank the Editor-in-Chief and two
sustainable poverty goals especially in agriculture-based economy anonymous reviewers for providing me insightful comments and
like Bangladesh. As per expenditure in concerned, we also observe constructive suggestions that helped me to improve the draft ver-
a negative response to crop and agricultural input expenditure in sion of the paper. I am profoundly indebted to my Ph.D. supervi-
our focal categories (as we assume these categories are directly sors, Professor Ilan Noy and Dr. Mohammed Khaled who were
related to rainfall shocks and flood) and are found consistent with very generous with their time and knowledge to provide me guid-
our theoretical prior for rainfall-based flood treatment house- ance and thoughtful comments in the preliminary version of the
holds. In particular, this evidence demonstrates a significant paper. I am also grateful to Dr. Binayak Sen (Bangladesh Institute
decrease in agricultural input expenditure for treatment group of Development Studies) and M.G. Mortaza (Asian Development
B. A sharp decline in non-food spending for these treatment Bank, BRM) for providing useful inputs in the data collection pro-
households is also of policy concern as this suggests decreased cess. I thank the audiences of the 13th Australasian Development
spending in health and education impacting longer term Economics Workshop (in Sydney, Australia), 58th and 57th Annual
investment.31 Conference of the New Zealand Association of Economists (NZAE);
We extend our analysis by further interacting treatment groups in particular Arthur Grimes, Mark Holmes, Andrea Menclova, and
with household characteristics such as rural, formal education, and my Ph.D. thesis examiners’; Harold Cuffe, Asadul Islam and Profes-
ownership of land status. The interaction terms seem to increase sor David Fielding.
both the main effects of the treatment groups and the respective
control variables. Agricultural incomes (crop and non-crop) are
Appendix A
found to drop significantly for rainfall-based affected households
(B). Interestingly, we find that crop expenditure increases signifi-
cantly for self-reported flood treatment households whereas non- Construction of rainfall-based flood risk index
crop expenditure per capita significantly increases for households
who have both self-reported and identified through geographical To develop this index, we collected annual rainfall data of 64
exposure (C). We further strengthen our results pooling data from years for 35 weather stations covering the whole country from
the earlier years i.e., 2000, 2005, and 2010 as robustness checks the Bangladesh Meteorological Department (BMD).34 The BMD
and observe consistencies in most cases with our benchmark esti- records daily rainfall data since 1948 for all available weather sta-
mation results. We however, only use the rainfall-based index tions across the country. We first calculated total monthly rainfall
measure in our robustness check due to unavailability of self- for each year under each weather station. We next calculated the
reported data in years 2000 and 2005. mean and standard deviation for each month for each sub-district
The ‘‘disaster-development” literature has made considerably by matching weather stations with sub-districts.35 We develop
less progress on the use of shock modules to empirically estimate two indexes of low- and high-risk indices. For the low flood risk,
the impacts of natural disasters on development outcomes. The we count the number of months over the 64 years for which we have
recent addition of shock questionnaires in nationally representa- data with extreme rainfall using two thresholds: monthly rainfall
tive household income and expenditure surveys provides an ample exceeding 15% of average annual rainfall for this sub-district; and
scope to identify the self-reported affected groups in repeated nat- monthly rainfall exceeding one standard deviation above the mean
ural disasters. This self-identification in the questionnaire could be for that month throughout the available time period.36 We calculate
advantageous to capture the disaster impacts on households more the average number of months with extreme rainfall to obtain the
precisely when compared to index-based identifications based on 32
See Hawkes and Rowe (2008).
geographical exposure. However, literatures have identified short- 33
See Bird (2009).
comings in self-reporting and various determinants of flood risk 34
The available data were for the years 1948–2012.
perception. The dissimilarities in the results in terms of the devel- 35
In cases where a sub-district did not have a rainfall measurement station, we used
opment impacts on flood treatment households using different an average of the three nearest stations.
36
measures of disaster risk exposure might be due to the various The historical coverage of rainfall data in BMD weather stations varies depending
upon their establishment year. Therefore, we calculate the average number of
shortcomings identified in the literatures. Moreover, questions months with extreme rainfall by dividing with the total number of rainfall years
available to calculate the probability of annual flooding in that particular weather
31
See Karim & Noy, 2016b for a detailed analysis on this issue. station.
Table 7
Key variables with descriptive statistics (treatment and control group a, treatment and control group b)
Variables Type Mean standard deviation Description of variables

Outcome variables Treatment Control A Treatment Control B
A B
Per capita total income Continuous 911940.4 926187.1 910175.2 928971.5 Sum of per capita crop, non-crop, business and other incomes
606662.3 641924.1 581756.4 652139.9
Per capita crop income Continuous 194200.9 172257.1 169420 173286.1 Per capita income earned through selling of crops
120641.9 94183.72 91191.19 95445.03
Per capita non-crop income Continuous 233546.5 248931.6 230389.3 252173 Per capita income earned through selling of livestock and poultry,
537408.4 543124.3 456983.5 558031.4 livestock products, fish farming and fish capture and farm forestry
Per capita business income Continuous 468905.9 488696 493480 487336.6 Per capita net revenues earned from non-agricultural enterprises
255953.5 296302.3 303387.3 294090.6 and rental income from agricultural assets
Per capita other income Continuous 15287.03 16796.23 18149.76 16501.34 Per capita income earned from other assets (e.g., stocks, bonds,
31811.08 51837.51 66344.03 48152.93 jewellery etc.), rent, insurance, charity, gift, remittances, bank
interest and social safety net
Per capita total expenditure Continuous 1454900 1441364 1426657 1444506 Sum of per capita food, non-food, crop, non-crop, agricultural input,
431931.9 434467.7 431057.4 435013.9 education and health expenditures
Per capita food expenditure Continuous 85007.71 85364.12 85224.26 85383.35 Per capita daily and weekly food consumption
23346.4 22095.96 22023.35 22137.98
Per capita non-food Continuous 737893.7 742763.1 736179.2 743929.7 Per capita monthly and annual non-food consumption
expenditure 236691.1 242337.5 243034.8 242061.4
Per capita crop expenditure Continuous 107859.9 106216.5 105621.7 106367.3 Per capita crop consumption by household
41673.97 46624.14 47004.28 46447.4
Per capita non-crop Continuous 96695 92351.51 92745.11 92370.86 Per capita consumption of livestock and poultry, livestock products,
expenditure 62587.95 46292.79 48930.08 46192.2 fish farming and fish capture and farm forestry products by
household
Per capita agricultural input Continuous 292600.7 285233.4 278433.5 286710.1 Per capita expenses on agricultural inputs
expenditure 132586.7 129633 126025.8 130345.6
PER CAPITA EDUCATIONAL EXPENDITURE Continuous 121299.4 118282.5 117582.6 118484 Per capita expenditure for educational services
60326.1 56320.71 55490.5 56570.64
Per capita health expenditure Continuous 13543.29 11406.69 11530.68 11429.57 Per capita expenditure for health services
21518.79 12382.97 13545.26 12424.67
Total change in agricultural Continuous 174977.2 185579.5 188845 184715.4 Sum of agricultural assets households bought in the last 12 months
and other business asset 491300.3 497618.8 504566.3 496126.6 and expenditure in capital goods (in non-agricultural enterprises) in
(in real terms) the last 12 months
Total agricultural input asset Continuous 222404.2 237534.1 232030.2 238266.7 Value of owned equipment and asset used in agriculture
value (in real terms) 221172.2 248540 225549.4 252185
Total consumer durable asset Continuous 2976063 3019571 3037415 3015165 Total asset value of consumer durable goods
value (in real terms) 1500622 1413280 1413003 1415285
Total month per year worked Continuous 801.2356 806.6025 808.5736 806.1035 Total number of months per year worked
227.3714 217.8752 214.9811 218.6407
Total days per month worked Continuous 1784.529 1800.281 1807.041 1798.628 Total number of days per month worked.
512.2758 489.88 487.7013 490.7934
Total hours per day worked Continuous 617.1022 621.6139 622.7273 621.2996 Total number of hours per day worked
174.0195 166.3709 165.55 166.7001
Daily wage (in real terms) Continuous 3852.178 3941.571 3965.141 3935.052 Daily wage in cash (if paid daily)
1414.781 1354.073 1331.947 1359.669
Salaried wage (in real terms) Continuous 108128.6 105938.3 105521 106067.1 Total net take-home monthly remuneration after all deduction at
50919.97 47547.04 47314.31 47668.63 source
Yearly benefits (in real terms) Continuous 138585.8 134053.7 131762.1 134596.3 Total value of yearly in-kind or other benefits (tips, bonuses or
90501.7 88512.07 88682.74 88518.13 transport) from employment.
Covariates
Rural Binary 0.675556 0.63976 0.626008 0.643205 Whether living in a rural area = 1, otherwise 0
0.469211 0.48009 0.483984 0.479077
Head of household is male Binary 1.004444 1.003994 1.00252 1.004289 Whether head of the household is male = 1, otherwise 0
0.066667 0.066918 0.067322 0.066831
Average age Continuous 26.4378 26.67 26.63367 26.67193 Average age of household members
1.347331 1.386957 1.643582 1.331109
Dependent Continuous 90.41333 90.62736 90.70716 90.60723 Age of the household member is <15 and 65
24.58355 24.16744 24.26731 24.15725
Proportion of formal Continuous 76.81407 76.97305 76.95353 76.97334 Proportion of household members attended school, college,
education 19.83074 19.25972 19.29904 19.26478 university or madrasa
Access to sanitation Binary 0.488889 0.528168 0.524194 0.528076 Whether the household use sanitary or pacca latrines (water seal
0.500991 0.499227 0.49954 0.499236 and pit) = 1, otherwise 0
Access to safe drinking water Binary 0.991111 0.963968 0.970262 0.963346 Whether the household has access to supply water or tube well
0.09407 0.186378 0.169906 0.187921 water = 1, otherwise 0
Access to electricity Binary 0.608889 0.575934 0.571573 0.577501 Whether the household has got electricity connection = 1, otherwise
0.489087 0.494221 0.494976 0.493981 0
House ownership Binary 0.857778 0.809603 0.806452 0.811269 Whether the household own a house = 1, otherwise 0
0.350057 0.392631 0.395179 0.391314
Land ownership (in real Continuous 69.48 62.74894 61.85938 63.06863 Amount of total operating land (in acres)
terms)
128.0451 128.9109 121.0524 130.3597
Source: Author’s calculations and elaborations.

Table 8
Impact on household income per capita
Variables (1) (2) (3) (4) (5)

Treatment group A 332,832.82** 142,294.50** 109,491.86 78,193.44 6,665.53
(Self-report) (130,063.55) (66,072.70) (93,443.48) (59,470.25) (7,903.15)
Treatment group B 32,407.34 13,883.61* 41,104.20* 2,247.68 1,127.58
(Rainfall-based) (40,347.94) (7,825.42) (23,314.76) (23,700.09) (4,797.41)
Both treatment group C 250,726.57 34,353.69 71,339.47 212,991.08 4,037.39
(250,456.37) (31,556.09) (108,105.23) (289,247.54) (9,569.99)
Rural 10,588.13 2,848.28 11,912.35 2,120.64 3,122.61**
(14,485.90) (2,177.35) (12,997.85) (6,267.27) (1,264.07)
Male headed HH 2,368,798.65*** 11,914.41*** 2,490,203.56*** 78,471.56*** 36,387.45
(71,878.91) (1,669.76) (10,589.02) (4,771.17) (36,918.57)
Avg age 57,682.42*** 1,698.52*** 54,323.07*** 1,159.21 522.45*
(5,132.86) (604.68) (5,414.47) (1,798.51) (301.25)
Dependent 90.86 101.30 4,633.54*** 4,768.54*** 7.79
(727.59) (97.28) (557.06) (516.23) (52.47)
Proportion_formal education 11,341.01*** 1,581.53*** 2,684.13*** 12,438.64*** 9.55
(872.70) (122.12) (646.46) (646.73) (63.21)
Access_sanitation 43,113.45*** 5,138.40*** 17,808.24* 6,257.87 13,642.42***
(11,035.25) (1,749.42) (9,757.08) (5,074.56) (808.18)
Access_drinking water 17,789.45 4,721.25 2,483.01 16,305.28 3,296.77**
(28,262.85) (4,847.53) (25,741.41) (11,427.32) (1,373.56)
Access_electricity 3,523.53 2,720.74 7,322.33 4,293.78 11,884.05***
(12,379.56) (1,878.93) (11,128.56) (5,419.89) (827.41)
House ownership 6,097.00 1,485.33 1,106.72 2,103.23 6,006.25***
(14,170.55) (2,212.30) (12,431.79) (6,588.55) (1,335.07)
Land ownership 35.75 51.92*** 38.66 2.29 21.41***
(41.86) (10.05) (35.27) (19.71) (3.54)
Interaction_edu*self 3,426.55*** 1,541.98* 547.17 1,246.72 138.35
(1,277.09) (797.44) (528.88) (769.64) (102.86)
Interaction_edu*rain 377.46 73.70 277.13 143.83 89.13
(474.31) (100.56) (231.69) (314.11) (64.34)
Interaction_land*self 1.33 73.89 26.56 74.79 27.98**
(142.78) (51.88) (117.67) (85.74) (11.88)
Interaction_land*rain 153.40* 19.84 65.61 79.40 11.00*
(88.02) (18.30) (63.88) (53.63) (6.47)
Interaction_rural*self 105,290.92 4,794.74 109,936.88 6,766.19 6,706.57
(103,140.17) (17,866.76) (98,365.55) (36,778.21) (4,377.83)
Interaction_rural*rain 28,266.56 5,564.84 2,883.13 18,448.15 8,784.48**
(26,892.41) (4,424.02) (22,237.89) (14,453.90) (3,935.12)
Interaction_edu*both 13.84 600.20 534.65 112.16 19.12
(3,046.68) (523.88) (1,343.91) (3,118.24) (181.29)
Interaction_land*both 216.98 74.37** 151.12 143.02 4.19
(475.39) (37.38) (112.73) (433.10) (18.90)
Interaction_rural*both 285,759.28* 21,735.52 87,400.23 168,448.08 7,538.56
(154,600.41) (31,745.12) (87,197.53) (113,814.19) (14,271.53)
Constant 3909442.31*** 21,563.08 3924719.01*** 3,824.37 53,964.73
(151,012.71) (18,140.35) (134,578.19) (47,057.14) (40,194.24)
Observations 12,242 12,222 12,232 12,242 12,242
R-squared 0.20 0.13 0.12 0.21 0.05

***
Note: Robust standard errors in parentheses p < 0.01, **p < 0.05, *p < 0.1.
Table 9
Impact on household expenditure per capita
Variables (1) (2) (3) (4) (5) (6) (7) (8)

Treatment group A 109,304.40 893.17 24,065.29 21,798.01*** 23,606.10 29,781.65 8,633.44 3,046.33
(Self-report) (66,457.58) (1,842.23) (38,785.57) (7,093.41) (32,813.66) (26,748.11) (11,254.61) (3,285.08)
Treatment group b 30,077.56 471.48 28,168.21** 3,412.53 9,460.94 3,632.92 3,452.94 204.28
(Rainfall-based) (24,804.11) (585.81) (14,352.09) (2,967.95) (8,839.87) (9,704.29) (3,460.80) (856.08)
both treatment group c 110,647.41 975.08 8,624.64 11,033.88 37,025.74* 82,551.46 17,473.19 12,004.26
(190,050.55) (3,115.48) (90,739.61) (12,885.29) (21,240.47) (81,685.85) (20,801.70) (13,917.93)
Rural 1,802.73 242.36 3,407.63 373.74 1,560.82* 1,948.22 2,257.43** 189.02
(6,285.55) (150.46) (3,971.61) (883.11) (913.55) (2,598.06) (1,143.68) (267.27)
Male Headed HH 391,391.05*** 5,186.41*** 3,699.66 34,917.76*** 42,366.12*** 303,488.83*** 2,049.84* 2,951.47***
(18,671.89) (460.12) (9,928.50) (757.43) (709.18) (8,601.82) (1,108.28) (324.14)
Avg age 29,159.82*** 671.78*** 11,272.68*** 2,159.28*** 3,988.52*** 13,060.39*** 1,233.69*** 305.68***
(2,415.98) (61.62) (1,386.27) (311.37) (264.12) (933.72) (302.28) (59.24)
Dependent 2,399.32*** 166.58*** 2,661.61*** 316.63*** 234.31*** 1,625.64*** 1,087.52*** 2.60
(341.96) (10.28) (180.25) (99.47) (44.56) (134.08) (53.06) (11.00)
(continued on next page)

Table 9 (continued)
Variables (1) (2) (3) (4) (5) (6) (7) (8)

Proportion_formal education 20,513.60*** 1,292.96*** 12,034.98*** 1,806.85*** 928.99*** 1,549.82*** 2,741.30*** 128.32***
(420.18) (12.64) (222.66) (124.28) (63.13) (163.31) (62.75) (12.46)
Access_sanitation 17,293.25*** 117.35 6,720.77** 1,167.83 1,946.89** 6,901.97*** 741.57 347.54
(5,193.54) (125.61) (3,316.55) (725.81) (796.11) (2,103.76) (932.48) (237.55)
Access_drinking water 9,517.59 273.53 3,805.46 2,711.73 1,026.78 1,442.23 124.06 141.81
(13,168.53) (316.18) (8,210.24) (1,843.18) (2,367.63) (5,550.99) (2,551.95) (532.75)
Access_electricity 8,420.74 376.49*** 8,771.70** 56.34 414.91 1,847.10 241.60 338.17
(5,550.62) (133.52) (3,524.65) (763.14) (837.94) (2,274.92) (998.96) (262.60)
House ownership 994.99 166.91 806.85 1,243.84 718.72 662.14 1,317.05 372.74
(6,575.79) (158.83) (4,179.84) (950.78) (960.10) (2,642.09) (1,261.40) (269.49)
Land ownership 105.74*** 1.98*** 15.31 19.06*** 9.50*** 53.11*** 7.21* 0.46
(22.18) (0.50) (12.64) (3.48) (3.03) (10.65) (3.96) (0.76)
Interaction_edu*self 1,586.93* 7.00 587.34 240.48*** 314.78 354.30 69.40 36.78
(843.68) (22.81) (477.28) (85.77) (434.43) (337.07) (115.75) (38.18)
Interaction_edu*rain 147.16 1.96 249.64 34.98 104.70 67.41 35.44 8.49
(318.76) (7.56) (184.61) (43.20) (104.74) (122.33) (42.05) (9.42)
Interaction_land*self 111.22 3.04 101.79 18.53 18.21 42.81 25.66 11.37
(155.19) (3.91) (81.16) (14.57) (20.41) (90.36) (19.28) (7.65)
Interaction_land*rain 131.33*** 0.78 68.11* 8.54 15.11* 21.03 17.64** 1.50
(48.84) (1.30) (34.78) (10.04) (8.14) (22.58) (8.04) (2.26)
Interaction_rural*self 36,706.85 117.90 17,354.04 3,606.60 10,884.55 6,127.46 2,088.09 3,912.85
(42,104.29) (1,053.46) (24,960.06) (5,193.34) (8,662.20) (15,892.27) (8,463.01) (2,734.17)
Interaction_rural*rain 6,592.90 443.68 3,455.71 941.26 2,704.26 879.13 1,783.21 827.14
(13,508.65) (323.38) (8,710.50) (2,028.72) (2,307.33) (5,467.69) (2,477.04) (635.08)
Interaction_edu*both 307.93 10.76 255.17 197.26 425.43 469.93 261.43 139.77
(2,278.24) (40.58) (1,209.62) (179.59) (267.90) (964.24) (272.62) (177.90)
Interaction_land*both 177.94 9.56* 295.31** 18.03 14.48 135.72** 11.18 15.51
(205.67) (4.89) (127.41) (25.91) (16.14) (61.45) (62.09) (21.35)
Interaction_rural*both 124,226.08 1,021.28 86,641.55* 8,501.37 2,720.75 26,390.71 4,019.46 5,161.58
(95,121.54) (1,850.98) (50,266.06) (11,330.75) (10,575.23) (55,480.81) (11,952.45) (8,642.57)
Constant 1117883.19*** 22,543.11*** 257,255.92*** 102,577.12*** 153,919.47*** 641,350.11*** 42,321.43*** 5,901.83***
(55,285.22) (1,394.69) (31,353.61) (8,307.50) (8,140.76) (23,813.61) (8,316.00) (1,793.13)
Observations 12,242 12,242 12,242 12,222 12,232 12,242 12,242 12,242
R-squared 0.63 0.92 0.52 0.36 0.25 0.31 0.27 0.04

***
Table 10
Impact on total asset outcomes

Asset stock Agri input asset value Durable asset value
Treatment group A 111,132.59 90,455.01* 256,836.22
(Self-report) (75,127.02) (47,600.49) (220,139.63)
Treatment group N 28,898.86 3,374.68 7,225.67
(Rainfall-based) (30,787.83) (15,521.03) (91,144.36)
Both treatment group C 178,097.20* 82,060.72 291,623.55
(101,758.84) (57,755.00) (469,256.70)
Rural 14,252.62 71.63 61,639.28**
(11,702.90) (5,656.31) (29,687.66)
Male headed HH 8,218.23 40,935.76*** 504,722.21***
(22,034.97) (3,894.96) (36,784.03)
Avg age 25,416.28*** 4,417.15*** 42,026.79***
(4,157.71) (1,346.94) (8,571.44)
Dependent 5,228.63*** 2,944.75*** 6,514.56***
(790.41) (229.34) (1,579.32)
Proportion_formal education 3,649.11*** 183.31 42,871.98***
(945.78) (281.99) (1,981.59)
Access_sanitation 4,299.30 1,113.03 62,287.27***
(9,519.76) (4,791.74) (24,043.79)
Access_drinking water 38,126.05 3,462.39 112,338.24
(28,610.75) (11,583.60) (72,140.41)
Access_electricity 8,129.74 3,359.32 2,443.59
(10,016.54) (5,031.24) (25,963.45)
House ownership 9,210.96 11,980.88** 22,809.59
(12,464.21) (5,625.60) (31,497.00)
Land ownership 11.72 43.39* 104.46
(43.81) (23.22) (108.47)
Interaction_edu*self 1,959.09** 1,482.33** 53.84
(942.53) (597.10) (2,847.95)
Table 10 (continued)

Asset stock Agri input asset value Durable asset value
Interaction_edu*rain 559.43 69.95 239.14
(402.79) (212.63) (1,135.26)
Interaction_land*self 339.02* 1.56 78.07
(192.74) (121.02) (664.81)
Interaction_land*rain 228.12 11.03 319.24
(196.41) (43.53) (357.72)
Interaction_rural*self 90,296.50 11,563.94 369,207.02**
(65,845.48) (29,702.03) (162,794.83)
Interaction_rural*rain 37,888.60 6,252.62 9,190.84
(27,284.49) (11,606.95) (66,058.85)
Interaction_edu*both 4,540.33 476.40 3,673.89
(3,247.31) (801.11) (5,963.49)
Interaction_land*both 187.55 93.08 1,698.96***
(197.76) (135.75) (405.34)
Interaction_rural*both 254,137.66 32,259.61 364,719.94
(222,356.04) (37,424.74) (267,690.81)
Constant 637,696.80*** 191,216.45*** 1195841.23***
(103,626.07) (36,935.21) (235,092.39)
Observations 12,242 12,242 12,242
R-squared 0.02 0.07 0.24

Notes: aRobust standard errors in parentheses ***
p < 0.01, **p < 0.05, *p < 0.1. bThe variable ‘‘Asset Stock” represents total change in agricultural and other business asset in
column 1.
Table 11
Impact on labor market outcomes
Variables (1) (2) (3) (4) (5) (6)

Month per year_total Days per month_total Hours per day_total Daily wage Salaried wage Yearly benefits
(Self-report) (19.25) (42.47) (13.81) (190.89) (9,676.01) (17,064.58)
Treatment group B 16.67** 26.18* 10.35** 146.17* 1,858.83 6,788.29
(Rainfall-based) (6.91) (15.46) (5.07) (76.18) (3,296.66) (6,395.29)
(31.79) (79.09) (29.26) (416.49) (22,900.16) (42,232.08)
Rural 2.44 7.19* 2.72* 10.61 746.98 1,602.13
(1.87) (4.28) (1.42) (21.82) (915.78) (1,865.95)
Male headed HH 80.25*** 235.55*** 93.15*** 1,541.51*** 40,308.60*** 108,647.86***
(6.85) (14.89) (4.21) (71.77) (1,825.63) (4,795.39)
Avg age 9.75*** 21.68*** 5.92*** 29.37*** 2,804.36*** 5,869.62***
(0.89) (1.89) (0.55) (8.00) (271.08) (542.37)
Dependent 1.95*** 3.40*** 1.59*** 10.94*** 363.76*** 881.66***
(0.21) (0.48) (0.16) (1.34) (50.29) (87.44)
Proportion_formal education 8.27*** 19.65*** 6.18*** 38.76*** 1,013.08*** 977.55***
(0.26) (0.60) (0.19) (1.66) (61.96) (109.81)
Access_sanitation 3.91** 5.70 1.61 42.27** 642.50 3,830.55**
(1.54) (3.51) (1.16) (17.96) (749.87) (1,542.10)
Access_drinking water 4.00 0.98 1.44 36.23 3,331.43* 4,005.18
(4.02) (9.27) (3.19) (45.58) (1,819.89) (3,773.47)
Access_electricity 1.86 3.36 1.17 15.70 2,218.86*** 5,551.05***
(1.64) (3.74) (1.24) (19.21) (807.38) (1,658.07)
House ownership 4.60** 12.83*** 4.33*** 3.80 2,863.53*** 3,400.65*
(1.94) (4.47) (1.48) (22.73) (969.74) (1,966.14)
Land ownership 0.02** 0.03* 0.01* 0.16** 0.43 6.40
(0.01) (0.02) (0.01) (0.07) (2.82) (5.51)
Interaction_edu*self 0.08 0.05 0.04 2.71 5.69 150.66
(0.25) (0.56) (0.18) (2.46) (123.02) (223.76)
Interaction_edu*rain 0.15 0.17 0.10 1.84* 17.09 16.48
(0.09) (0.21) (0.07) (0.97) (43.89) (83.51)
Interaction_land*self 0.03 0.09 0.01 0.34 33.03 109.46*
(0.04) (0.09) (0.02) (0.56) (21.64) (58.92)
Interaction_land*rain 0.02 0.02 0.01 0.25 8.52 36.01**
(0.02) (0.04) (0.01) (0.18) (7.21) (17.08)
Interaction_rural*self 1.96 7.12 0.17 4.07 1,947.38 22,003.46*
(11.68) (26.72) (8.60) (138.43) (6,181.85) (12,776.98)
Interaction_rural*rain 1.58 2.77 0.17 60.27 545.85 1,437.62
(4.25) (9.94) (3.32) (48.49) (1,995.69) (4,105.48)
(continued on next page)

Table 11 (continued)
Variables (1) (2) (3) (4) (5) (6)

Month per year_total Days per month_total Hours per day_total Daily wage Salaried wage Yearly benefits
Interaction_edu*both 0.08 0.11 0.34 2.14 5.49 211.64
(0.39) (1.03) (0.37) (5.70) (332.72) (573.90)
Interaction_land*both 0.09** 0.06 0.03 0.42 17.67 63.62
(0.04) (0.10) (0.03) (0.52) (34.77) (97.04)
Interaction_rural*both 1.31 19.23 10.85 540.58* 447.52 17,898.27
(23.68) (54.57) (17.32) (290.29) (14,984.88) (25,735.38)
Constant 341.74*** 831.20*** 247.55*** 775.72*** 121,591.01*** 287,205.30***
(20.36) (45.04) (13.93) (186.37) (7,032.51) (14,982.38)
Observations 12,242 12,242 12,242 12,242 12,242 12,242
R-squared 0.87 0.87 0.87 0.54 0.35 0.21

***
Table 12
Impact on household income per capita (robustness checks)
Variables (1) (2) (3) (4) (5)

Rainfall-based treatment group 10,225.32 1,670.12 2,504.74 8,684.94* 4,942.52
(7,873.66) (2,287.10) (8,157.23) (5,218.46) (3,241.37)
Rural 3,826.76 1,843.09 6,906.42 2,742.50 2,366.68**
(9,563.84) (1,771.04) (9,458.35) (4,844.04) (962.72)
Male headed HH 106,190.04*** 4,279.77*** 156,315.88*** 5,036.55* 15,297.99***
(15,454.73) (498.48) (20,270.46) (2,910.89) (2,542.05)
Avg age 1,229.06*** 123.71*** 1,341.71*** 303.98*** 352.41**
(184.92) (25.66) (122.21) (75.12) (153.94)
Dependent 4,684.33*** 13.21 44.39 4,684.31*** 89.25*
(527.99) (88.52) (306.13) (464.95) (52.48)
Proportion_formal education 16,788.64*** 1,699.21*** 2,730.44*** 12,318.84*** 105.94*
(638.38) (116.66) (370.16) (577.17) (61.70)
Access_sanitation 28,361.03*** 3,757.93*** 9,622.42* 6,950.32** 12,395.65***
(5,839.73) (1,116.63) (5,636.50) (3,141.17) (661.46)
Access_drinking water 9,251.90 2,119.32 3,675.13 9,253.10 1,255.88
(14,324.94) (3,048.48) (14,606.39) (7,150.69) (1,009.04)
Access_electricity 8,490.24 2,503.22** 4,780.99 883.74 11,592.06***
(6,403.19) (1,181.74) (6,285.21) (3,342.68) (554.27)
House ownership 11,404.29 1,779.07 7,090.97 2,421.60 3,523.03
(8,474.06) (1,955.37) (9,531.14) (4,900.32) (2,249.82)
Land ownership 48.88 52.14*** 28.24 4.27 21.87***
(41.01) (9.84) (34.77) (19.19) (3.49)
Interaction_edu*rain 309.70* 59.93* 320.86** 54.45 14.54
(183.39) (32.74) (150.39) (100.45) (29.87)
Interaction_land*rain 122.67 16.54 64.23 66.93 19.69***
(87.21) (17.75) (64.00) (51.85) (7.59)
Interaction_rural*rain 9,898.18 3,256.91 3,842.42 11,769.63* 4,200.86
(10,316.34) (2,531.01) (10,339.58) (6,190.84) (2,863.08)
Year_2005 20,867.16* 35,071.41*** 25,279.97*** 56,376.63*** 1,749.41
(11,710.74) (4,010.14) (7,315.96) (8,798.92) (1,637.47)
Year_2000 14,153.68 34,754.37*** 28,506.92*** 50,629.27*** 4,099.21
(12,073.87) (4,025.72) (7,217.83) (8,719.09) (3,214.16)
Constant 122,649.37*** 28,543.27*** 169,272.72*** 33,435.61*** 5,134.42
(25,963.04) (5,833.17) (27,998.96) (11,865.19) (5,655.63)
Observations 26,151 19,859 23,448 21,278 26,138
R-squared 0.56 0.59 0.10 0.61 0.03

***
Table 13
Impact on household expenditure per capita (robustness checks)
Variables (1) (2) (3) (4) (5) (6) (7) (8)

Rainfall-based treatment group 4,656.43 367.75*** 3,689.06* 277.05 1,191.28 752.44 287.93 70.93
(3,414.86) (80.62) (2,151.86) (952.88) (874.69) (2,382.13) (755.07) (187.49)
Rural 2,543.96 221.98** 4,091.06 479.68 1,250.45* 1,318.86 1,681.34** 279.81
(4,087.69) (97.71) (2,554.35) (710.74) (651.19) (2,082.45) (833.69) (203.10)
Male headed HH 21,844.85*** 206.50*** 461.06 5,854.54*** 3,521.09*** 37,313.59*** 808.63 246.11***
(2,628.08) (41.00) (513.14) (646.85) (388.41) (4,436.65) (504.88) (47.47)
Avg age 353.18*** 1.66* 51.29** 73.47*** 85.84*** 481.79*** 78.95*** 4.05*
(37.79) (0.97) (22.48) (13.26) (6.16) (32.38) (16.48) (2.07)
Dependent 4,256.53*** 199.48*** 3,248.27*** 456.44*** 37.46 660.52*** 998.22*** 19.53**
(306.31) (8.72) (166.45) (84.40) (42.60) (121.75) (51.35) (8.71)
Proportion_formal education 22,643.13*** 1,329.00*** 12,714.95*** 1,969.87*** 1,217.45*** 2,667.67*** 2,645.65*** 107.05***
(377.51) (10.72) (202.46) (106.54) (65.13) (148.95) (59.62) (10.10)
Access_sanitation 9,030.51*** 64.67 3,775.33** 767.22* 1,111.87** 4,752.73*** 710.58 184.06
(2,645.33) (63.91) (1,672.77) (464.68) (441.19) (1,324.98) (569.22) (152.42)
Access_drinking water 4,777.74 73.72 1,536.55 1,559.91 684.82 1,623.33 111.43 167.40
(6,551.77) (158.35) (4,055.80) (1,165.84) (1,302.12) (3,462.36) (1,587.77) (359.83)
Access_electricity 4,328.96 68.88 4,614.16*** 207.58 175.76 795.33 488.89 254.78
(2,776.73) (66.84) (1,744.36) (481.25) (453.94) (1,419.30) (596.77) (165.61)
House ownership 4,176.69 91.33 1,098.85 1,395.98* 1,074.89 1,387.33 1,503.08* 211.59
(3,730.24) (90.11) (2,354.23) (839.25) (716.46) (2,212.08) (868.78) (190.17)
Land ownership 113.09*** 2.04*** 19.81 19.71*** 9.83*** 56.58*** 6.35* 0.68
(22.04) (0.49) (12.23) (3.43) (2.99) (10.82) (3.86) (0.74)
Interaction_edu*rain 289.87*** 0.69 113.43* 19.28 24.39* 129.77*** 13.58 1.31
(97.75) (2.31) (61.33) (15.60) (14.52) (40.55) (16.04) (3.75)
Interaction_land*rain 109.84** 1.06 56.84* 6.90 16.17** 15.36 14.80* 2.14
(48.04) (1.25) (33.34) (9.63) (7.78) (22.02) (7.87) (2.21)
Interaction_rural*rain 8,122.84* 494.64*** 4,156.50 166.07 784.83 1,482.95 13.89 2.88
(4,603.88) (109.59) (2,900.43) (1,102.87) (921.31) (2,757.98) (1,003.39) (263.53)
Year_2005 70,563.76*** 808.93*** 53,445.46*** 9,499.50*** 807.72 14,337.08*** 2,075.65* 1,153.24***
(7,880.37) (200.33) (4,858.11) (1,106.87) (2,519.42) (3,056.51) (1,187.31) (324.95)
Year_2000 56,010.06*** 711.07*** 45,309.70*** 9,798.67*** 483.29 13,396.09*** 1,599.52 1,295.20***
(7,916.14) (201.73) (4,879.17) (1,114.63) (2,510.56) (3,062.66) (1,190.32) (338.06)
Constant 35,004.12*** 931.42*** 52,908.32*** 16,619.81*** 7,383.85** 38,031.29*** 9,285.31*** 751.86
(11,318.09) (277.82) (6,802.81) (1,763.96) (2,991.51) (6,876.54) (2,304.29) (503.38)
Observations 26,155 26,155 26,141 19,859 23,448 20,750 21,219 20,034
R-squared 0.94 0.99 0.92 0.77 0.72 0.75 0.74 0.26

***
probability of flooding occurring annually in that particular weather Attzs, M (2008). Natural disasters and remittances: Exploring the linkages between
poverty, gender and disaster vulnerability in Caribbean. SIDS No. 2008.61,
station (and consequently sub-district). The mean probability is 0.93
Research paper/UNU-WIDER.
with 0.16 standard deviation. The second index, high flood risk, is Auffret, P (2003). High consumption volatility: The impact of natural disasters? World
constructed similarly, but in this case the two thresholds are 20% Bank Policy Research working paper 2962, The World Bank.
of average annual rainfall and more than two standard deviation Bangladesh Bank. Central Bank of Bangladesh. https://www.bb.org.bd/.
Bangladesh Bureau of Statistics (BBS). Government of Bangladesh. http://www.
above the monthly mean. For the high-risk measure, the mean prob- bbs.gov.bd/home.aspx.
ability is 0.26 with 0.08 standard deviation. Bangladesh Meteorological Department (BMD). Dhaka. www.bmd.gov.bd.
Bandyopadhyay, S., & Skoufias, E. (2015). Rainfall variability, occupational choice,
and welfare in rural Bangladesh. Review of Economics of the Household, 13(3),
1–46.
Appendix B. Supplementary data
Baez, J., & Mason, A. (2008). Dealing with climate change: Household risk management
and adaptation in Latin America. Available at SSRN 1320666.
Supplementary data associated with this article can be found, in Balasubramanian, R. (2015). Climate sensitivity of groundwater systems critical for
the online version, at https://doi.org/10.1016/j.worlddev.2017.10. agricultural incomes in South India. SANDEE Working Papers, ISSN 1893–1891;
WP 96–15.
026. Banerjee, L. (2007). Effect of flood on agricultural wages in Bangladesh: An
empirical analysis. World Development, 35(11), 1989–2009.
Beg, N., Morlot, J. C., Davidson, O., Afrane-Okesse, Y., Tyani, L., Denton, F., et al.
References (2002). Linkages between climate change and sustainable development. Climate
Policy, 2(2–3), 129–144. https://doi.org/10.3763/cpol.2002.0216.
Acevedo, S. (2014). Debt, growth and natural disasters a caribbean trilogy. IMF Bergstrand, K., Mayer, B., Brumback, B., & Zhang, Y. (2015). Assessing the
Working paper no. WP/14/125. relationship between social vulnerability and community resilience to
Akter, S., & Mallick, B. (2013). The poverty-vulnerability-resilience nexus: Evidence hazards. Social Indicators Research, 122(2), 391–409.
from Bangladesh. Ecological Economics, 96, 114–124. Bird, D. K. (2009). The use of questionnaires for acquiring information on public
Anttila-Hughes, J. K., & Solomon, H. M. (2013). Destruction, Disinvestment, and Death: perception of natural hazards and risk mitigation–a review of current
Economic and Human Losses Following Environmental Disaster. Available at SSRN: knowledge and practice. Natural Hazards and Earth System Science, 9(4),
http://ssrn.com/abstract=2220501 or https://doi.org/10.2139/ssrn.2220501. 1307–1325.
Aritomi, T., Olgiati, A., & Beatriz Orlando, M. (2008). Female Headed Households and Blaikie, P., Cannon, T., Davis, I., Wisner, B. (1994). At risk. London: Routledge.
Poverty in LAC: What are we measuring? Downloaded from: http://paa2008. Bui, A. T., Dungey, M., Nguyen, C. V., & Pham, T. P. (2014). The impact of natural
princeton.edu/papers/81458. disasters on household income, expenditure, poverty and inequality: Evidence
Arouri, M., Nguyen, C., & Youssef, A. B. (2015). Natural disasters, household welfare, from Vietnam. Applied Economics, 46(15), 1751–1766.
and resilience: evidence from rural Vietnam. World Development, 70, 59–77. Buvinic, M., & Gupta, G. R. (1997). Female-headed households and female-
Asiimwe, J. B., & Mpuga, P. (2007). Implications of rainfall shocks for household maintained families: Are they worth targeting to reduce poverty in
income and consumption in Uganda. AERC Research Paper 168, African developing countries? Economic Development and Cultural Change, 45(2),
Economic Research Consortium. 259–280.
Cabezon, E., Hunter, M. L., Tumbarello, M. P., Washimi, K., & Wu, M. Y. (2015). Jha, R. (2006). Vulnerability and natural disasters in Fiji, Papua New Guinea, Vanuatu
Enhancing macroeconomic resilience to natural disasters and climate change in the and the Kyrgyz Republic. Available at: doi: 10.2139/ssrn.882203.
small states of the pacific. IMF Working paper no. WP/15/125. Karim, A., & Noy, I. (2016a). Poverty and natural disasters–a qualitative survey of
Cai, H., Lam, N. S. N., Zou, L., Qiang, Y., & Li, K. (2016). Assessing community the empirical literature. The Singapore Economic Review, 61(1). https://doi.org/
resilience to coastal hazards in the Lower Mississippi River Basin. Water, 8(2), 10.1142/S0217590816400014.
46. Karim, A., & Noy, I. (2016b). Poverty and natural disasters: A regression meta-
Crichton, D. (1999). The risk triangle. In J. Ingleton (Ed.), Natural disaster analysis. Review of Economics and Institutions, 7(2), 26.
management (pp. 102–103). London: Tudor Rose. Karim, A., & Noy, I. (2015). The (mis)allocation of public spending in a low income
Cutter, S. L., Boruff, B. J., & Shirley, W. L. (2003). Social vulnerability to country: Evidence from disaster risk reduction spending in Bangladesh. School of
environmental hazards. Social Science Quarterly, 84(1), 242–261. Economics and Finance Working Paper no. 4194, Victoria University of
Cutter, S. L., Emrich, C. T., Mitchell, J. T., Boruff, B. J., Gall, M., Schmidtlein, M. C., et al. Wellington, New Zealand.
(2006). The long road home: Race, class, and recovery from Hurricane Katrina. Khatun, R. (2015). The impact of micro-level determinants of poverty in
Environment: Science and Policy for Sustainable Development, 48(2), 8–20. bangladesh: A field survey. International Journal of Research in Management &
Cutter, S. L., Ash, K. D., & Emrich, C. T. (2014). The geographies of community Business Studies, 2(2). Downloaded from: http://ijrmbs.com/vol2issue2/dr_
disaster resilience. Global Environmental Change, 29, 65–77. razia1.pdf..
De la Fuente, A. (2010). Natural disaster and poverty in Latin America: Welfare Kotikula, A., Narayan, A., & Zaman, H. (2010). To what extent are Bangladesh’s recent
impacts and social protection solutions. Well-Being and Social Policy, 6(1), 1–15. gains in poverty reduction different from the past? World Bank Policy Research
Dercon, S. (2004). Growth and shocks: Evidence from Rural Ethiopia. Journal of Working Paper Series, No. WPS5199.
Development Economics, 74(2), 309–329. Kurosaki, T. (2015). Vulnerability of household consumption to floods and droughts
Duflo, E., Galiani, S., & Mobarak, M. (2012). Improving access to urban services for the in developing countries: evidence from Pakistan. Environment and Development
poor: Open issues and a framework for a future research agenda J-PAL Urban Economics, 20, 209–235. https://doi.org/10.1017/S1355770X14000357.
Services Review Paper. Cambridge, MA: Abdul Latif Jameel Poverty Action Lab. Le De, L., Gaillard, J. C., & Friesen, W. (2015). Poverty and disasters: Do remittances
Dwyer, A., Zoppou, C., Nielsen, O., Day, S., & Roberts, S. (2004). Quantifying social reproduce vulnerability? The Journal of Development Studies, 51(5), 538–553.
vulnerability: A methodology for identifying those at risk to natural hazards. https://doi.org/10.1080/00220388.2014.989995.
Geoscience Australia Record 2004/14. Lanjouw, P., & Ravallion, M. (1995). Poverty and household size. The Economic
Fadeeva, A. (2014). A comparative study of poverty in China, India, Bangladesh, and Journal, 1415–1434.
Philippines. University of Southern California, Department of Sociology Working Levitt, J. I., & Whitaker, M. C. (2009). Truth crushed to earth will rise again Katrina
paper. Downloaded from: https://dlc.dlib.indiana.edu/dlc/bitstream/handle/ and its aftermath. In Hurricane Katrina: America’s Unnatural Disaster (pp. 1–21).
10535/9613/Fadeeva. percent20A percent20comparative percent20study University of Nebraska Press.
percent20of percent20poverty.pdf?sequence=1&isAllowed=y. Lohmann, S., & Lechtenfeld, T. (2015). The effect of drought on health outcomes and
Felbermayr, G., & Gröschl, J. (2014). Naturally negative: The growth effects of health expenditures in rural Vietnam. World Development, 72, 432–448.
natural disasters. Journal of Development Economics, 111, 92–106. Lopez-Calva, L. F., & Ortiz-Juarez, E. (2009). Evidence and policy lessons on the links
Ferdousi, S., & Dehai, W. (2014). Economic growth, poverty and inequality trend in between disaster risk and poverty in Latin America: Summary of regional studies
Bangladesh. Asian Journal of Social Sciences & Humanities, 3, 1. RPP LAC – MDGs and Poverty – 10/2008. New York: RBLAC-UNDP. Downloaded
Flatø, M., Muttarak, R., & Pelser, A. (2017). Women, weather, and woes: The from: http://www.preventionweb.net/english/hyogo/gar/background-papers/
triangular dynamics of female-headed households, economic vulnerability, and documents/Chap3/LAC-overview/LAC-Oveview.pdf.
climate variability in South Africa. World Development, 90, 41–62. https://doi. Mallick, D., & Rafi, M. (2010). Are female-headed households more food insecure?
org/10.1016/j.worlddev.2016.08.015. Evidence from Bangladesh. World Development, 38(4), 593–605.
Foltz, J., Gars, J., Özdogan, M., Simane, B., & Zaitchik, B. (2013). Weather and welfare Martin, S. A. (2015). A framework to understand the relationship between social
in Ethiopia. In 2013 Annual Meeting, August 4–6, 2013, Washington DC, No. factors that reduce resilience in cities: Application to the City of Boston.
150298, Agricultural and Applied Economics Association. International Journal of Disaster Risk Reduction, 12, 53–80.
Fomby, T., Ikeda, Y., & Loayza, N. V. (2013). The growth aftermath of natural Masozera, M., Bailey, M., & Kerchner, C. (2007). Distribution of impacts of natural
disasters. Journal of Applied Econometrics, 28(3), 412–434. disasters across income groups: A case study of New Orleans. Ecological
Gerstter, C., Kaphengst, T., Knoblauch, D., & Timeus, K. (2011). An assessment of the Economics, 63(2), 299–306.
effects of land ownership and land grab on development-with a particular focus on McGuigan, C., Reynolds, R., & Wiedmer, D. (2002). Poverty and climate change:
small holdings and rural areas. European parliament ad-hoc briefing EXPO/DEVE/ Assessing impacts in developing countries and the initiatives of the international
2009/Lot 5/13. Downloaded from: http://www.ecologic.eu/sites/files/project/ community. The Overseas Development Institute. Downloaded from: http://
2013/lot5_13_land_grabbing_en.pdf. www.odi.org/sites/odi.org.uk/files/odi-assets/publications-opinion-files/3449.
Gosling, S. N., Dunn, R., Carrol, F., Christidis, N., Fullwood, J., Gusmao, D. D., et al. pdf.
(2011). Climate: Observations, projections and impacts: Bangladesh. Climate: Meinzen-Dick, R. S. (2009). Property Rights for Poverty Reduction? DESA Working
Observations, Projections and Impacts. Downloaded from: http://eprints. Paper No. 91. New York: United Nations Department of Economic and Social
nottingham.ac.uk/2040/6/Bangladesh.pdf. Affairs. Downloaded from: http://www.un.org/esa/desa/papers/2009/wp91_
Guiteras, R. P., Jina, A. S., & Mobarak, A. M. (2015). Satellites, self-reports, and 2009.pdf.
submersion: Exposure to floods in Bangladesh. American Economic Review: Mogues, T. (2011). Shocks and asset dynamics in Ethiopia. Economic Development
Papers and Proceedings, 105(5), 232–236. and Cultural Change, 60(1), 91–120.
Haile, M. (2005). Weather patterns, food security and humanitarian response in Mueller, V. A., & Osgood, D. E. (2009). Long-term impacts of droughts on labor
sub-Saharan Africa. Philosophical Transactions of the Royal Society B: Biological markets in developing countries: Evidence from Brazil. The Journal of
Sciences, 360(1463), 2169–2182. Development Studies, 45(10), 1651–1662.
Haughton, J., & Khandker, S. R. (2009). Handbook on poverty and inequality. Mueller, V., & Quisumbing, A. (2011). How resilient are labor markets to natural
Washington DC: The World Bank. disasters? The case of the 1998 Bangladesh Flood. The Journal of Development
Hawkes, G., & Rowe, G. (2008). A characterisation of the methodology of qualitative Studies, 47(12), 1954–1971.
research on the nature of perceived risk: Trends and omissions. Journal of Risk Noy, I. (2009). The macroeconomic consequences of disasters. Journal of
Research, 11, 617–643. Development Economics, 88(2), 221–231.
Heltberg, R., Oviedo, A. M., & Talukdar, F. (2015). What do household surveys really Noy, I., & duPont IV, W. (2016). The long-term consequences of natural disasters—A
tell us about risk, shocks, and risk management in the developing world? The summary of the literature. School of Economics and Finance Working Paper no. 2,
Journal of Development Studies, 51(3), 209–225. https://doi.org/10.1080/ Victoria University of Wellington, New Zealand.
00220388.2014.959934. Noy, I., & Patel, P. (2014). After the flood: Households after the 2011 great flood in
Hoffmann, R., & Muttarak, R. (2017). Learn from the past, prepare for the future: Thailand. Victoria University SEF Working Paper 11/2014.
Impacts of education and experience on disaster preparedness in the O’Neill, E., Brereton, F., Shahumyan, H., & Clinch, J. P. (2016). The impact of
Philippines and Thailand. World Development, 96, 32–51. perceived flood exposure on flood-risk perception: The role of distance. Risk
Hsiang, S. M., & Jina, A. S. (2014). The causal effect of environmental catastrophe on Analysis, 36(11), 2158–2186.
long-run economic growth: Evidence from 6,700 cyclones. National Bureau of Organisation for Economic Co-operation and Development, OECD (2003). Poverty
Economic Research Working Paper 20352. http://www.nber.org/papers/ and climate change: Reducing the vulnerability of the poor through adaptation.
w20352. Downloaded from: http://www.oecd.org/env/cc/2502872.pdf.
IPCC (2014). Summary for policymakers. In: C. B. Field, et al., (Eds.), Climate change Parker, D., & Tapsell, S. (2009). Deliverable 2.1. Relations between different types of
2014: Impacts, adaptation, and vulnerability. Part A: Global and sectoral aspects. social and economic vulnerability. Final draft report submitted to EU project
Contribution of Working Group II to the Fifth Assessment Report of the ‘Enhancing resilience of communities and territories facing natural and na-tech
Intergovernmental Panel on Climate Change [online] (pp. 1–32). Cambridge: hazards’ (ENSURE).
Cambridge University Press. Available from: http://ipcc-wg2.gov/AR5/. Patnaik, U., & Narayanan, K. (2010). Vulnerability and coping to disasters: A study of
Israel, D. C., & Briones, R. M. (2014). Disasters, poverty, and coping strategies: The household behaviour in flood prone region of India. Munich Personal RePEc
framework and empirical evidence from micro/household data-Philippine case. Archive.
Philippine Institute for Development Studies. Discussion Paper Series No. 2014– Poapongsakorn, N., & Meethom, P. (2013). Impact of the 2011 floods, and flood
06. management in Thailand, ERIA Discussion Paper Series, ERIA-DP-2013-34.
Japan Bank for International Cooperation, JBIC (2007). Poverty profile: People’s Rayhan, M. I. (2010). Assessing poverty, risk and vulnerability: A study on
Republic of Bangladesh. Downloaded from: http://www.jica.go.jp/activities/ flooded households in rural Bangladesh. Journal of Flood Risk Management, 3
issues/poverty/profile/pdf/bangladesh_e.pdf. (1), 18–24.
Rodriguez-Oreggia, E., de la Fuente, A., de la Torre, R., Moreno, H., & Rodriguez, C. Thomas, T., Christiaensen, L., Do, Q. T., & Trung, L. D.(2010). Natural disasters and
(2013). The impact of natural disasters on human development and poverty at household welfare: Evidence from Vietnam. World Bank Policy Research Working
the municipal level in Mexico. The Journal of Development Studies, 49(3), Paper Series 5491, The World Bank.
442–455. Trumbo, C. W., Peek, L., Meyer, M. A., Marlatt, H. L., Gruntfest, E., McNoldy, B. D.,
Shah, M., & Steinberg, B. M. (2012). Could droughts improve human capital? Evidence et al. (2016). A cognitive-affective scale for hurricane risk perception. Risk
from India. Unpublished manuscript. Davis: University of California. Analysis, 36(12).
Shoji, M. (2010). Does contingent repayment in microfinance help the poor during United Nations International Strategy for Disaster Reduction & UNISDR (2012).
natural disasters? The Journal of Development Studies, 46(2), 191–210. Disaster risk—Poverty trends in Jordan, Syria, Yemen: Key findings and policy
Silbert, M., & del Pilar Useche, M. (2012). Repeated natural disasters and poverty in recommendations. Cairo: UNISDR Regional Office for the Arab States.
Island nations: A decade of evidence from Indonesia. University of Florida, Wilson, G. (2012). Community resilience and environmental transitions. Routledge.
Department of Economics, PURC Working Paper. Wisner, B., Blaikie, P., Cannon, T., & Davis, I. (2004). At risk, natural hazards.
Skoufias, E., Bandyopadhyay, S., & Olivieri, S. (2017). Occupational diversification as Routledge, London, UK: People’s Vulnerability and Disasters.
an adaptation to rainfall variability in rural India. Agricultural Economics, 48(1), World Bank (2014). Climate change and health impacts: how vulnerable is Bangladesh
77–89. and what needs to be done? Disaster risk and climate change unit. Sustainable
Skoufias, E., Katayama, R. S., & Essama-Nssah, B. (2012). Too little too late: Welfare Development Department, South Asia Region. Downloaded from: http://
impacts of rainfall shocks in rural Indonesia. Bulletin of Indonesian Economic siteresources.worldbank.org/INTSOUTHASIA/Resources/223497-1378327471830/
Studies, 48(3), 351–368. HealthImpactClimateChangeerfCLEAN2DraftFinalReport_final.pdf.
Tapsell, S., McCarthy, S., Faulkner, H., & Alexander, M. (2010). Social vulnerability and Zaman, H. (1999). Assessing the Poverty and Vulnerability Impact of Micro-Credit in
natural hazards. CapHaz-Net WP4 Report, Flood Hazard Research Centre – FHRC, Bangladesh: A case study of BRAC. Policy Research Working Paper 2145, The
Middlesex University, London. World Bank. Downloaded from: http://www1.worldbank.org/prem/poverty/ie/
Tasneem, S., & Shindaini, A. J. M. (2013). The effects of climate change on agriculture dime_papers/260.pdf.
and poverty in coastal Bangladesh. Journal of Environment and Earth Science, 3
(10), 186–192.

World Development: Azreen Karim

Uploaded by

Copyright:

Available Formats

World Development: Azreen Karim

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

World Development: Azreen Karim

Uploaded by

Copyright:

Available Formats

World Development 103 (2018) 40–59

Contents lists available at ScienceDirect

The Household Response to Persistent Natural Disasters: Evidence from

data-based flood risk index; second, it directly compares the impacts

where a country’s pre-determined geo-physical and climatic char-

Figure 2. Map showing the treatment areas (sub-districts) in the study.

We start by estimating our benchmark model (as specified in

Variables (1) (2) (3) (4) (5)

Treatment group A 332,832.82** 142,294.50** 109,491.86 78,193.44 6,665.53

Source: Author’s calculations.

Variables (1) (2) (3) (4) (5) (6) (7) (8)

A. Karim / World Development 103 (2018) 40–59

Source: Author’s calculations.

Variables (1) (2) (3)

Treatment group A 111,132.59 90,455.01* 256,836.22

Source: Author’s calculations.

Variables (1) (2) (3) (4) (5) (6)

Treatment group A 1.85 3.30 6.16 256.80 2,432.84 24,475.78

Source: Author’s calculations.

Variables (1) (2) (3) (4) (5) (6) (7) (8)

Source: Author’s calculations.

Variables (1) (2) (3) (4)

Source: Author’s calculations.

Variables Type Mean standard deviation Description of variables

Source: Author’s calculations and elaborations.

Variables (1) (2) (3) (4) (5)

Source: Author’s calculations.

Variables (1) (2) (3) (4) (5) (6) (7) (8)

(continued on next page)

Variables (1) (2) (3) (4) (5) (6) (7) (8)

Source: Author’s calculations.

Variables (1) (2) (3)

Variables (1) (2) (3)

Source: Author’s calculations.

Variables (1) (2) (3) (4) (5) (6)

(continued on next page)

Variables (1) (2) (3) (4) (5) (6)

Source: Author’s calculations.

Variables (1) (2) (3) (4) (5)

Source: Author’s calculations.

Variables (1) (2) (3) (4) (5) (6) (7) (8)

Source: Author’s calculations.

You might also like

Treatment group A 332,832.82 142,294.50 109,491.86 78,193.44 6,665.53