FSUP Baseline Report Final 23 June

FUNDED BY THE EUROPEAN UNION

IMPLEMENTED BY

BASELINE STUDY BY

Food Security for the Ultra Poor – Haor Baseline Study Report

FSUP-H Baseline Report, June 2010 ii

Principal Authors

Richard Caldwell

Executive Director

TANGO International, Inc.

Bruce Ravesloot

Asia Representative


Md. Abdul Quddus

Team Leader, FSUP-H Baseline Study

Data Management Aid

Maqbul H. Bhuiyan

Executive Director

Data Management Aid

FSUP-H Baseline Report, June 2010 iii

Table of Contents

List of Tables, Figures and Pictures ................................................................................................... v

Acknowledgements ........................................................................................................................... viii

List of Abbreviations ........................................................................................................................... ix

Glossary of Bengali Terms................................................................................................................... x

Glossary of English Terms................................................................................................................... x

Executive Summary ............................................................................................................................. xi

1.0 INTRODUCTION .................................................................................................................... 1

1.1 Food security and poverty context in Bangladesh ................................................................ 1

1.2 Background of the FSUP-H project ....................................................................................... 2

1.3 Implementation framework of the FSUP-H project ............................................................... 3

1.4 FSUP-H site and impact group selection .............................................................................. 3

2.0 FSUP-H BASELINE STUDY .................................................................................................. 4

2.1 Rationale of the study ........................................................................................................... 4

2.2 Objectives of the study .......................................................................................................... 5

2.3 Scope of the study ................................................................................................................ 5

3.0 STUDY METHODS ................................................................................................................. 5

3.1 Study design.......................................................................................................................... 5

3.2 Quantitative study design ...................................................................................................... 5

3.2 Qualitative study design ...................................................................................................... 10

4.0 DEMOGRAPHIC CHARACTERISTICS ............................................................................... 12

5.0 LIVELIHOODS AND ECONOMIC SECURITY .................................................................... 15

5.1 Occupational patterns ......................................................................................................... 15

5.2 Household employment and income/expenditure ............................................................... 20

5.3 Income in peak and lean seasons....................................................................................... 24

5.4 Coping strategies for lean seasons ..................................................................................... 25

5.5 Migration .............................................................................................................................. 27

5.6 Loans ................................................................................................................................... 27

5.7 Assets .................................................................................................................................. 31

5.8 Housing characteristics ....................................................................................................... 36

6.0 FOOD SECURITY ................................................................................................................ 38

6.1 Food consumption score ..................................................................................................... 38

6.2 Food intake .......................................................................................................................... 40

6.3 Coping strategies ................................................................................................................ 43

6.4 Trend analysis ..................................................................................................................... 44

7.0 WATER AND SANITATION ................................................................................................. 49

7.1 Drinking, cooking and washing water sources .................................................................... 49

7.2 Arsenic testing ..................................................................................................................... 53

7.3 Sanitation ............................................................................................................................ 53

8.0 HEALTH PRACTICES AND ILLNESS ................................................................................ 57

8.1 Hand washing...................................................................................................................... 57

FSUP-H Baseline Report, June 2010 iv

8.2 Illness among adults and health-seeking behavior ............................................................. 58

9.0 PARTICIPATION AND ACCESS ......................................................................................... 60

9.1 Participation in development ............................................................................................... 60

9.2 Access to GoB services ...................................................................................................... 62

9.3 Access to other services ..................................................................................................... 66

9.4 Access to common property ................................................................................................ 68

10 DISASTERS AND CRISES .................................................................................................. 70

10.1 Natural disasters: effects and coping strategies ................................................................. 70

10.2 Household crises: effects and coping strategies................................................................. 72

10.3 Climate change ................................................................................................................... 74

11 FAMILY AUTHORITY AND DECISION MAKING ............................................................... 76

11.1 Household decision making ................................................................................................ 76

11.2 Family life attitudes ............................................................................................................. 80

11.3 Daily time patterns of men and women ............................................................................... 81

12 CHILD NUTRITION, ANTENATAL CARE AND FAMILY PLANNING ................................ 83

12.1 MCHN characteristics ......................................................................................................... 83

12.2 Anthropometric measurements ........................................................................................... 89

13 STATUS OF FEMALE-HEADED HOUSEHOLDS ............................................................... 91

14 CONCLUSION AND RECOMMENDATIONS ...................................................................... 92

FSUP-H Baseline Report, June 2010 v

List of Tables, Figures and Pictures

TABLES

Table 1: Four main pillars of food security 1

Table 2: Illustrative sample sizes for stratified random sampling using variants of deff and standard

error

8

Table 3: Sampling statistics of 1,892 households, by District and Haor type 8

Table 4: Qualitative techniques applied for the baseline survey 10

Table 5: Key demographic characteristics of the population, by District and Haor type 12

Table 6: Demography and dependency ratios, by District and Haor type 14

Table 7a: Primary and secondary occupations for individuals aged 8 years and older, by Haor type 16

Table 7b: Primary and secondary occupations for individuals aged 8 years and older, by District 17

Table 8: Income sources for previous 30 days, by Haor type 20

Table 9: Income sources for previous 30 days, by District 21

Table 10a: Key income and expenditure data for households, by District and Haor type 22

Table 10b: Detailed expenditure data, by District and Haor type 23

Table 11: Top ten ways of coping with lean periods (multiple response), by District 26

Table 12: Top ten ways of coping with lean periods (multiple response), by Haor type 26

Table 13: Type of work performed by those migrating out of the household within the last 12 months,

by District and Haor type

27

Table 14a: Key loan data for households, by District and Haor type 29

Table 14b: Detailed interest rate data for loans, by District and Haor type 29

Table 14c: Loan source for women, by District and Haor type 30

Table 15: Reasons for taking out a loan, by Haor type 30

Table 16: Reasons for taking out a loan, by District 31

Table 17: Average number of domestic assets owned, by District and Haor type 32

Table 18: Average number of productive assets owned, by District and Haor type 33

Table 19: Average number of land assets owned, by District and Haor type 33

Table 20: Average number of animal assets owned, by District and Haor type 34

Table 21: Average number of resource assets owned, by District and Haor type 34

Table 22: Average financial assets owned, in Taka, by District and Haor type 35

Table 23: Housing characteristics, by District and Haor type 36

Table 24: Food consumption score 38

Table 25: Proportion of sampled households by FCS threshold values 39

Table 26: Seasonal calendar 44

Table 27: Drinking water sources, by District and Haor type 49

Table 28: Cooking water sources, by District and Haor type 50

Table 29: Washing water sources, by District and Haor type 51

Table 30: Tube wells/tara pumps tested for arsenic, by District and Haor type 53

Table 31: Types of latrines used by adult men and women, by District and Haor type 54

Table 32: Types of latrines used by boys and girls 5-15 years of age, by District and Haor type 55

Table 33: Hand-washing behaviors among the FSUP baseline study households, by District and

Haor type (1)

57

Table 34: Hand-washing behaviors among the FSUP baseline study households, by District and

Haor type (2)

57

Table 35: Top ten illnesses experienced by adults in households during the previous 12 months, by

District and Haor type

58

Table 36: Usual treatment source for household members, by District and Haor type 58

Table 37: Household members involved in development processes 60

Table 38 Type of development institution/person that HH members were involved with 61

FSUP-H Baseline Report, June 2010 vi

Table 39: Type of collective action that households have participated in, by District and Haor type 61

Table 40: Proportion of households using various types of Government service providers, by District

and Haor type

63

Table 41: Types of services received by GoB service providers, by District and Haor type 65

Table 42: Proportion of households that have various property types available in their household

area, by District and Haor type

68

Table 43: Proportion of available property that is accessible by households, by District and Haor type 69

Table 44: Disasters experienced by households in the last 12 months, by District and Haor type 70

Table 45: Proportion of households experiencing various consequences of a natural disaster in the

last 12 months, by District and Haor type

71

Table 46: Household decision making, by Haor type (1) 76

Table 47: Household decision making, by Haor type (2) 77

Table 48 Household decision making, by District (1) 78

Table 49: Household decision making, by District (2) 79

Table 50: Attitudes about family life, by Haor type 80

Table 51: Attitudes about family life, by Haor type 81

Table 52: MCHN characteristics, by District and Haor type 85

Table 53: Weaning foods used, by District and Haor type 86

Table 54: Who attended last delivery, by District and Haor type 86

Table 55: Child health and immunization, by District and Haor type 87

Table 56: Health issues of mothers with children under 2, by District and Haor type 88

Table 57: Health issues of children under 2, by District and Haor type 89

Table 58 Anthropometric measurements 90

Table 59: Key variables for female-headed households, by district and Haor type 91

Table 60: Baseline values and recommendations for FSUP-H logframe indicators 92

FIGURES

Figure 1: Age distribution of study population, by sex 13

Figure 2a: Mean values of annual per capita income, by District 22

Figure 2b: Median values of monthly household cash income and expenditures per capita, by District

and Haor type

24

Figure 3: Average monthly incomes during peak and lean seasons, by District and Haor type 25

Figure 4: Mean FCS values, by District and Haor type 39

Figure 5a: Proportion of households reporting enough food, by month and Haor type (1) 40

Figure 5b: Proportion of households reporting enough food, by month and Haor type (2) 41

Figure 6 Mean number of lean months, by District and Haor type 41

Figure 7: Frequency of three 'square meals' taken a day in 12 months, by District 42

Figure 8: Frequency of three 'square meals' taken a day in 12 months, by Haor type 42

Figure 9: Coping Index‟ for households, by District and Haor type 43

Figure 10: Distances to sources of drinking water, by District and Haor type 51

Figure 11: Distances to sources of cooking water, by District and Haor type 52

Figure 12: Distances to sources of washing water, by District and Haor type 52

Figure 13: Types of service providers accessed, by District 63

Figure 14: Types of service providers accessed, by Haor type 64

Figure 15: Level of satisfaction with selected GOB services 65

Figure 16: Mean asset loss from households experiencing asset loss in a natural disaster in the last

12 months, by District and Haor type

71

Figure 17: Mean number of working days lost from households experiencing a natural disaster in the

last 12 months, by District and Haor type

72

FSUP-H Baseline Report, June 2010 vii

Figure 18: Loss of assets among households experiencing household crises in the last 12 months, by


73

Figure 19: Loss of work days among households experiencing household crises in the last 12 months,


73

Figure 20: Average number of days lost due to illness for those households with an ill member

designated as a household crises in the last 12 months, by District and Haor type

74

Figure 21: Daily time use of men and women 82

Figure 22: Comparison between SHOUHARDO and FSUP-H malnutrition levels 90

PICTURES

Picture 1: Baseline enumerator using the PDA for a household interview 11

Picture 2: Agricultural day labor - males 15

Picture 3: Agricultural day labor - females 18

Picture 4: Cow rearing by ultra-poor households 19

Picture 5: Grameen Bank office 28

Picture 6: Jack fruit trees 35

Picture 7: Housing made of jute and straw 37

Picture 8: Housing made of corrugated iron 37

Picture 9: Women supporting household income through produce sales 45

Picture 10: Men fishing in the peak season 46

Picture 11: Non-agricultural day labor 47

Picture 12: Non-agricultural day labor - mat making 48

Picture 13: Woman uses hand tube well as the water source for washing 50

Picture 14: Ring slab latrine 54

Picture 15: Open defecation facilities 56

Picture 16: Village medicine shop 59

Picture 17: Union Health Center 59

Picture 18: Community collective action to improve road infrastructure 62

Picture 19: Women engaged in alternative livelihood activities 66

Picture 20: Government-owned Khas land 69

Picture 21: Damage to buildings as a result of natural disasters 70

Picture 22: Social mobilization around community issues 75

Picture 23: Balanced meal taken by a pregnant woman 84

FSUP-H Baseline Report, June 2010 viii

Acknowledgements

The baseline study was organized by TANGO International in partnership with Data Management Aid

(DMA). Special thanks go out to Mr. Md. Abdul Quddus, DMA Team Leader, Mr. Maqbul H. Bhuiyan,

Executive Director of DMA, and the DMA study team for their fantastic work in organizing the field

data collection, and their contributions to the study design and report.

The members of the baseline study team wish to thank the staff of CARE Bangladesh and their local

NGO partners for their time and effort during this study, in particular CARE‟s Social Development Unit

for their guidance to and facilitation of the qualitative data collection. We thank Mr. Zakir Khan, the

Team Leader of FSUP-H, and Mr. M. Zakaria, M&E Coordinator for FSUP-H, for all the support and

guidance they provided to this study.

We would also like to acknowledge Ms. Khaleda Khanom, Deputy Team Leader, and Ms. Salma

Akter, F&A Manager, the FSUP-H Technical Coordinators and the many CARE, ASD, POPI and SUS

field staff that did a great job in facilitating the baseline survey needs. This study would not have been

possible without their efforts.

Finally, we want to acknowledge the Government of Bangladesh professional staff, FSUP-H program

participants, the FSUP Program Coordinating Unit and EC Delegation officials who gave freely of their

time throughout the baseline process.

TANGO International

8 June 2010


406 S. Fourth Ave.

Tucson, Arizona 85701

Tel: (1)-520-617-0977

Fax: (1)-520-617-0980

[email protected]

FSUP-H Baseline Report, June 2010 ix

List of Abbreviations

ASD Assistance for Slum Dwellers

BADC Bangladesh Agricultural Development Corporation

BBS Bangladesh Bureau of Statistics

BDHS Bangladesh Demographic and Health Survey

BMI Body Mass Index

CBO Community Based Organization

DAE Department of Agricultural Extension

DMA Data Management Aid

FoSHoL Food Security for Sustainable Household Livelihoods

FSUP-H Food Security for the Ultra-Poor in the Haor Region

FGD Focus Group Discussion

GoB Government of Bangladesh

KI Key Informant

MDG Millennium Development Goals

NCHS National Center for Health Statistics

NIPORT National Institute of Population Research and Training

NGO Non-Government Organization

POPI People‟s Oriented Program Implementation

PPS Probability Proportional to Size

SHOUHARDO Strengthening Household Ability to Respond to Development Opportunities

SUS Sabalamby Unnayan Samity

TBA Traditional Birth Attendant

UN United Nations

WB World Bank

WFP World Food Programme

VGD Vulnerable Group Development

VGF Vulnerable Group Feeding

FSUP-H Baseline Report, June 2010 x

Glossary of Bengali Terms

Ana Local unit of measuring gold/silver

Beel Open water body

Dadon Advance sale of crops/products

Khas Government-owned land or water bodies

Logni High interest loans

Madrasha Religious education center

Masjeed Mosque

Mohajan Informal moneylender

Salish Informal village court/arbitration

Upazila A geo-administrative unit under a district comprising several Unions

Union Parishad Lowest local government unit

Bengali Calendar:

Apr-may

May-Jun

Jun-Jul

Jul-Aug

Aug-Sep

Sep-Oct

Oct-Nov

Nov-Dec Dec-Jan

Jan-Feb

Feb-Mar

Mar-Apr

Baishak Jaisti Ashar Sravan Bhadra Ashin Kartik Agrahayan Payush Magh Falgun Chaitra

Glossary of English Terms

Decimal Decimal (100 decimals is equal to 1 acre)

Homestead The yard or compound of a household

Household A family unit, who share common resources for cooking and eating

FSUP-H Baseline Report, June 2010 xi

Executive Summary

Background

According to the 2005 Joint UN/GoB MDG report, Bangladesh was home to over 60 million food

insecure people (GoB-UN 2005). Income inequality and chronic poverty1 are the primary causes for

wide-spread food-insecurity, which is one of the most pressing crises facing Bangladesh today. To

respond to this challenge, CARE Bangladesh operates a longstanding and reputable program on food

security. The Food Security for the Ultra Poor - Haor (FSUP-H) Project, funded by the European

Union, was initiated in 2009. The design of the FSUP-H Project has taken into account the lessons

learnt from SHOUHARDO and other projects like FOSHoL, SETU etc. and is aligned with CARE-B‟s

long-term programming strategy for the Haor region.

The overall objective of FSUP-H Project is to reduce extreme poverty and food insecurity in the Haor

region of Northeast Bangladesh. The specific objective is to sustainably improve food access and

utilization and reduce vulnerability for women and their dependents in ultra poor households in

Sunamganj, Netrokona and Kishoreganj Districts

The project has four specific results:

a) Increase inclusion and capacity of 55,000 ultra poor HHs with focused attention to women

headed ultra poor HHs and their dependents, to actively engage with development processes

with greater support from their communities and local level institutions,

b) 55,000 ultra poor households (particularly women) have additional economic opportunities

and income, improving their access to food and household food security round the year,

c) 55,000 ultra poor households have reduced vulnerability to food insecurity and poverty and

improved resilience to quick and slow onset disasters, and

d) Improve and equitable utilization of food as well as reduced malnutrition among women and

their dependents in 55,000 ultra poor households.

Baseline study methodology

The objective of the baseline study is to better understand the current food insecurity, poverty and

vulnerability situation of the program impact group, and to establish baseline values of indicators for

intended outcomes against which future change can be measured in terms of: behavior, systemic

capacity and impact on the socio-economic conditions of target households such as number of food

insecure months, income and expenditure. In addition to tracking impact-level changes and livelihood

trends over time, the information and data generated by the survey will be useful in: designing future

similar projects and scaling up the current project.

The baseline survey was undertaken in January – February 2010, and utilized a combination of

quantitative and qualitative methods. The quantitative methods involved a detailed household-level

survey using Personal Digital Assistants (PDAs) for data collections instead of paper questionnaires;

and collection of anthropometric data from children aged 6-23 months, for which standard weight and

height scales were used. The survey utilized a multi-stage sample design stratified by district and haor

type. After data cleaning, the final sample size was 1892 respondents.

1 49.6% people live in poverty (below US$ 1 per day).

FSUP-H Baseline Report, June 2010 xii

The qualitative methods utilized mixed tools: male/female focus group discussions (24), key informant

interviews (30), in-depth interviews (12) and trend analysis (4 villages). Qualitative data collection

was organized by 3 teams made up of CARE and partner staff. Each team organized data collection

in 2 villages in the same district: one in the deep Haor, and one in the moderate Haor. Field research

was guided by CARE colleagues from the Social Development Unit (SDU) under overall coordination

by the FSUP-H M&E Coordinator and with inputs from the CARE Program Quality Unit in CBHQ

Demographic characteristics

The FSUP baseline survey included basic demographic information on 1,892 households and 8,957

individuals. The average age of the study population was 22.2 years old. On average, survey

households had about 2.3 adults of working age (15–60 years); 0.8 children under age 5; 1.4 children

between the ages of 5 and 14 years, and only 0.2 elderly persons above 60 years. For the FSUP-H

respondents, the total dependency ratio is 114.5%, which can be considered high. Adult members of

working age have more children to support than aged household members. Only about 2 percent of

individuals are reported to be disabled.

The gender ratio found in the study is 96, which means that there are 96 males for each 100 females.

For the overall population, the average age of Head of Household is 42.7 years. Almost 15 percent of

sampled households are female-headed, and female heads of households are significantly older

(p=.000) than their male counterparts averaging 48.2 and 41.7 years old, respectively. In terms of

household size the average for the study population is 4.8 people per household. Female-headed

households are significantly smaller than male-headed households (3.1 and 5.0, respectively. Eighty

five percent of household heads are married.

Livelihoods and economic security

The primary occupation of surveyed household members reflects the principle livelihood strategies of

households in the Haor region. Closely linked to occupations and livelihoods are economic indicators

of households, such as income, other cash sources, asset ownership, debt and savings. Together

these elements of economic security reveal how resilient households are to economic shocks and

natural disasters.

There are few distinct differences in occupational trends in the Haor region sampled. Agricultural and

non-agricultural labor are the two main livelihood opportunities, and together account for half of

primary occupations. Other important livelihood activities, in order of predominance, are fishing, petty

trade, and housemaid/servant. Few households engage in agriculture on their own fields; less than 1

percent overall. This is a direct result of the extensive degree of landlessness among the ultra poor in

Bangladesh. The majority of women are housewives. Non-agricultural laborer and housemaid/servant

are the most common primary occupations for women but account for less than 10% across all

Districts. The majority of respondents, including most women, reported having no secondary

occupation. Agricultural and non-agricultural day labor opportunities are the most common

occupational categories for those whose primary activities are in areas such as petty trade,

sharecropping, and fishing.

Each household had on average only 1.35 income earners. This can be considered quite low but

there could be seasonality factors due to the timing of the survey. The data by Haor type and District

show that the main household income sources align closely with the primary occupations, as was

FSUP-H Baseline Report, June 2010 xiii

expected based on number of income earners and main occupations. Casual agriculture labor is the

main income source across Haor areas and Districts, followed by casual non-agricultural labor.

Income from crop and animal sales is low overall. The mean monthly per capita income is 800 Taka

and the mean annual per capita income is 9,599 Taka, overall. The per capita income is significantly

lower in Sunamganj than in the other districts.

Mean monthly expenditure per capita is 1,419 Taka, and the median monthly expenditure per capita is

1,099 Taka. The majority of daily expenditure is on food purchases (72% of daily expenditure). The

purchase of tobacco products is the next highest daily expenditure (8%), followed by hiring manual

labor from others and purchasing fuel (including gasoline, kerosene, and fire wood). It is interesting to

note that the single highest monthly expenditure item is cell phone cards (48% of monthly

expenditure). The second highest monthly expenditure item is medical expenses (including fees,

medicine and travel) (36%). The single highest item annual expenditure is clothing for household

members (51% of annual expenditure), followed by social/religious events (14%) and household

goods (11%), on average. Expenditure on fishing or fish raising, agricultural equipment/input, and

livestock and poultry rearing accounted for less than 10% of annual expenditure each.

Median per capita monthly expenditures (including daily, monthly and annual expenditures) are

significantly higher than median per capita income. This is likely due to several factors. First of all,

there is the seasonality of the data collection; February falls in a lean period, which is characterized by

lower income and high lending. Secondly, respondents have the tendency to overestimate

expenditure and underestimate income.

There are significant differences between mean monthly income levels during peak and lean seasons.

The main coping strategies to deal with the lower income during lean seasons were adjusting meals

(60.1%), taking loans from friends/relatives (50.6%), and taking loans from money lenders (35.6%).

Selling labor in advance at reduced wage levels is another coping strategy used by households during

lean periods. Overall, 7% of households had at least one member who sold labor in advance.

Temporary migration was not a common coping strategy to deal with lean periods. However,

migration for employment purposes is relatively common in areas with a high degree of seasonal

work, such as the FSUP-H project area. In moderate Haor areas, 38.5% of households had

somebody migrate in the last 12 months for employment purposes; in deep Haor areas this was

32.3%. About 75% of those who migrated were heads of household, while about 20% were

sons/daughters. About 70% migrated to urban areas and 30% to other rural areas. While migration

went on throughout the year, there was more migration for employment purposes from August to

October. Agricultural contract labor and agricultural day labor are by far the most common types of

work performed by migrants.

78% of households overall held at least 1 current loan over the last 12 months. The average number

of loans per household overall was 1.4 and 35% of loans were taken by women. The average loan

amount was 6,652 Taka and overall interest rates were 51%. Overall, the outstanding loan amount at

the time of the interview was 5,393 Taka, which is about 81% of the average loan amount - indicating

a very high debt burden on households. The majority of loans (41%) were taken from money lenders,

NGOs (24%), and friends/relatives (23%). Only 6% of loans were taken from Grameen Bank and 4%

from clubs/CBOs. Informal money lenders give loans without collateral but instead charge higher

interest rates; the high level of lending from informal sources such as money lenders largely explains

the high interest rates found in this survey. The most common reasons for taking out a loan were

consumption purposes (food, clothing etc), followed by medical treatment and non-agricultural

purchases. Very few households reported taking out a loan for productive purposes such as the

FSUP-H Baseline Report, June 2010 xiv

purchase of agricultural tools/equipment, purchase of agricultural inputs, land leasing or mortgaging or

livestock purchases.

Assets are an integral component of livelihoods, and the accumulation and sale of assets reflect

important economic characteristics of households. Overall the ownership of productive assets in the

survey population was very low. Generally, far less than one out of ten households owned any of the

productive assets. Land ownership varied greatly among sampled households. Agricultural land

ownership was the highest and averaged 4.05 decimals per household, or less than 1/20th of one

acre. Ownership of homestead land averaged at 2.88 decimals per household. Chickens were the

most common animal asset owned, averaging 1.48 per household, followed by ducks and cows.

Ownership of resource assets such as timber and fruit trees, bamboo, and medicinal plants, was fairly

common in surveyed households. Cash with NGOs was the most common financial asset measured

and averaged 495 Taka per household.

The majority of all houses have floors made of mud (99.9% and 0.1% made of brick), walls made of

straw/jute or corrugated iron sheets/tin/wood, and roofs made of corrugated iron. Less than 1% of all

houses have brick walls and only 1 house in Kishoreganj had a concrete roof. Average total square

feet of living space is 175ft and the average number of rooms is 2 across all strata. About 10 percent

of households share their living space with their cattle, mostly for safety of the animals in absence of

more than one housing structure.

Food security

The survey used the Food Consumption Score (FCS), which is widely used by the World Food

Program and endorsed by FANTA, as a measure of diet diversity and diet quality. The FCS is derived

by weighting various food groups based on their protein value and assigning a score for each food

group consumed by the household during the recall period. For the FSUP survey population, 52.3% of

total households had an acceptable FCS, 31.5% had a borderline FCS and 16.2% had a poor FCS.

Overall, the mean number of lean months per year is 4.3 and the mean value for households that take

3 meals per day „most of the time‟ is 14%. The combined mean values for „most of the time‟ and

„often‟ is 56.3 %. There are two distinct annual lean periods in terms of insufficient food. The first lean

period is from April to June, with the leanest period in April-May (13%), the month of Baishak in the

Bengali calendar. The second lean period is from November to February with the leanest period in

Dec-Jan (12%), the month of Payush in the Bengali calendar. In both periods, almost 90% of

households in the sample reported insufficient food. The recovery from the insufficient food period in

April to June is notable longer than for the second lean period - with another smaller decrease in

August-September (31%) before reaching a peak at 63 percent in Oct-Nov. The highest number of

households report sufficient food in March-April (82%), with a very sharp decrease between the

Bengali months of Chaitra and Baishak.

The survey utilized a index to measure how households deal with food insecurity. A high score

indicates that households in specified areas avail themselves of a broad range of coping strategies to

deal with food insecurity; the higher the index value is - the higher the assumed stress on households.

Overall, the coping index score of almost 24 indicates a moderately-high level of stress on households

due to food insecurity. Bulk purchases of rice‟, „running out of food‟ and „reducing personal food

intake‟ were the top 3 coping strategies.

FSUP-H Baseline Report, June 2010 xv

Water and sanitation

Hand tube wells are the most common drinking water source followed by shallow tube wells and deep

tube wells. Overall, 97% of households depend on the various types of tube wells for drinking water.

Almost no households draw drinking water from open water sources such as ring wells, ponds and

rivers/canals. Hand tube wells are also the most common cooking water source followed by

rivers/canals and shallow tube wells. Most households reported open water sources for washing.

River/canals are the most common washing water source followed by hand tube wells and ponds. The

average distance to water sources for drinking, cooking and washing purposes is around 200 meters.

Of the households that reported tube wells or tara pumps as a source for drinking, cooking or washing

water, 50.5% of households reported that the tube wells / tara pumps were tested for arsenic. Of the

tube wells/tara pumps that were tested, 14.1% were found to contain arsenic.

The most common type of latrines used by adult men and women are ring slab/offset latrines (with the

seal broken) and hanging/open latrines, followed by uncovered pit latrines and then open defecation.

Overall, the use of hygienic latrines such as ring slab/offset latrines (with the seal intact), septic

latrines, covered pit latrines and locally adapted hygienic latrines is very low in the project area.

Similar to adults, the most common types of latrines used by boys and girls 5-15 years of age are ring

slab/offset latrines (with the seal broken) and hanging/open latrines. For boys and girls, this is

followed by open defecation and then uncovered pit latrines – the opposite to adults. Overall, the use

of hygienic latrines such as ring slab/offset latrines (with the seal intact), septic latrines, covered pit

latrines and locally adapted hygienic latrines is very low.

Health and illness

The majority of respondents wash their hands before eating but less than half do so before preparing

food and only one-third wash their hands before feeding children. The majority of respondents wash

their hands after defecation but only one-third of respondents do so after cleaning a baby‟s bottom.

The use of ash or clay for hand washing is most common followed by use of only water. The use of

soap is least common.

The average number of illnesses cited per household was 2.4. The most common illness experienced

by adults during the previous 12 months was a cold attack, followed by gastric illness and diarrhea.

Only 3 percent of households reported no illnesses at all. Medicine shops and village doctors are the

most common treatment sources for household members.

Participation and access

Participation in community development processes is low at 4.5% of all households, which was too

low for meaningful analysis. Among the only 186 responses received, household head was mentioned

as the most common household member involved in development processes. Females (spouses plus

female heads of household) accounted for 15.1% the responses. The Masjeed or religious committee

was the most common type of institution that household members were engaged with for

development purposes in the last 12 months (24%), followed by engagement with NGOs (19%) as

village group members, which is often a prerequisite to receiving microcredit. Half (49.3%) of

respondents reported that their engagement with development institutions consisted of receiving

services, followed by being a volunteer (27.9%); committee member (19.3%); participant in activities

(19.3%); and recipient of training (1.9%). Only 5% of households had experience with collective action

FSUP-H Baseline Report, June 2010 xvi

in last 12 months, primarily consisting of road construction/repair and Mosque construction/repair.

Over two-thirds of households (68.7%) had accessed one or more GoB service providers in the

previous year. The most common service providers used were Union Parishad and Government

Immunization Services, followed by Government Family Planning, Upazilla Health Services and Union

Health Services. The most common services received from Union Parishad were categorized as

„other‟, which likely refers to government safety net programs. The most common service received

from Union and Upazila Health Services is medication followed by suggestions. The most common

service received from Government Family Planning are suggestions, medicines and vaccinations. For

Government Immunization Services, vaccinations are the main services received, as was to be

expected. Overall, training provided by GoB service providers is very low. At present, the main

sources of knowledge and skills for economic/livelihood activities are knowledge transfer from

previous generations, and from relatives and neighbors. The little external assistance that ultra-poor

households do receive comes primarily from NGOs. For all service providers, the majority of

respondents indicated they were satisfied or highly satisfied.

Only 33% of households reported receiving services from other non-government service providers.

The three most common non-government service providers were NGOs (76%), Grameen Bank

(16%), and Local Service Providers (18%). Less than 1% of households reported receiving services

from Commercial Banks, CBOs, input retailers/dealers and non-Government Vocational

Education/Training, respectively. The most common services received from NGOs were credit (68%),

suggestions (16%), and relief/aid (4%). The most common services received from Grameen Bank

were credit (99%) and suggestions (13%). The most common services received from Local Service

Providers were suggestions (75%), credit (65%), suggestions (16%), medicines (71%) and relief/aid

(12%).

Regarding access to common property, the highest proportion of households has access to

river/canals, followed by roadside sloping and beels/haors. Access to Khas land is lowest.

Disasters and crises

Overall, 78% of households reported that they did not experience a natural disaster in the previous

year. The highest proportion of disasters experienced in the last 12 months was wind damage, floods,

excessive rain and storms. Among those who experienced a natural disaster, the highest proportion of

households reported partial damage to their house, followed at a distance by loss of working days and

full damage to their house. For those affected, the mean asset loss per disaster was reported at Taka

3,017, and the mean number of working days lost reported by the 26.5% of affected households was

10. The most common coping strategies used by respondents to recover from a natural disaster were:

taking out a loan from friend/neighbor (41%), taking loan from a moneylender (31%), adjusting meals

(25%), using savings (25%), accepting help from others: (24%), purchasing on credit (21%) and

taking a loan from NGO (11%).

Respondents were also asked the same range of effect and coping strategy questions for a range of

household crises, not caused by natural disasters. Only 16.7% reported the occurrence of such crises

in the last 12 months. The most common types of household crises reported were illness of income

earners (57.2% of cases where a household crisis was reported) and illness of other household

members (32% of cases where a household crisis was reported). The main effects of the household

crises were asset/income loss and working days lost. The mean loss of assets for the survey

population was just under Taka 5,000. The mean number of working days lost for the survey

FSUP-H Baseline Report, June 2010 xvii

population was 36. The mean number of working days lost due to illness was also 36 (mode=15). The

most common coping strategies used by respondents to cope with household crises were: took out a

loan from friend/relative (42.7%), took out a loan from moneylender (36.0%), made adjustment to

meals (27.8%), accepted help from others (20.6%), purchased goods on credit (18.4%), used savings

(11.7%), took out a loan from an NGO (11.1%), took a grain loan (10.4%), ate famine foods (8.2%),

and accepted aid (5.4%)

Family authority and decision making

The highest proportion of decisions is made by the husband after discussion with the female

household member. It is also apparent that women have greater involvement in certain household

decisions such as minor household purchases, children‟s clothing and education, medical expenses

and in spending money that they have directly earned. Women have less involvement in expenditures

that relate to livelihoods, higher value assets, loans/savings and events such as weddings and

ceremonies and shelter in case of disasters. The proportion of decisions made without any

involvement by the female is low for almost all decision types, except salish decision making.

Overall, a higher proportion of women agree that the husband should help with household chores if

the female is working; and that they have the right to express their opinion, even when they disagree

with their husband. The proportion of women overall who disagree with the statement that it is better

to send a son to school than a daughter is also significantly higher. However, it is interesting to note

that despite the more liberal attitudes about family life expressed by women, the proportion of women

who agree that a wife should tolerate being beaten is significantly higher that the proportion who

disagrees.

Child nutrition, antenatal care and family planning

Of the total number of respondents, 70 percent did not have any children < 2 years of age. Of those

who did, 29 percent had one and 1 percent had two < 2 children. Virtually every mother breastfed her

child (99.5%) and 45% of overall mothers initiated breastfeeding with the first hour of birth. The

average age for introducing solid/semi-solid foods (weaning) was just over 5 months age. Overall,

35.5% of mother‟s took iron or folic acid supplements. The majority of mother‟s did not change the

amount of food that they consumed during their last pregnancy; 16% increased their food intake and

33% decreased their food intake. The majority of women also did not change the amount of rest they

took after the last birth. Only 23% took more rest than usual. Overall, mothers attended on average 1

ANC session. Qualitative data shows that there are very few periodic medical checkups during

pregnancy due to lack of knowledge, lack of money, and difficulties in communicating with the medical

centers. Overall, the majority of births were attended by Traditional Birth Assistants. Less than 1% of

births were attended by a doctor.

Only 7.4% of mothers reported suffering no illnesses in the last 12 months. The highest proportion of

women suffered from cold attacks, followed by gastric complications and anemia. The lowest

proportion of women suffered from Typhoid. Only 5.8% of children did not suffer any illnesses in the

last 12 months. The highest proportion of children suffered from cold attacks, diarrhea and

pneumonia. The lowest proportion of children suffered from skin diseases and other illnesses. For

households currently with a child 2 years of age or under, 81.8 % of the oldest child in this age group

has received at least one immunization. For those children who did receive immunizations, 72.9%

have immunization cards. For those children who needed antihelmintics, 47% received them. 57.7%

FSUP-H Baseline Report, June 2010 xviii

of < 2 children are underweight: 39.4% are moderately underweight and 18.3% are severely

underweight. Underweight is a composite index of height-for-age and weight-for-height. A child can be

underweight for his/her age because s/he is stunted, wasted or both. In general, an underweight

prevalence of > 30% is considered to be „very high‟.

Status of female-headed households

Almost 15% of the households sampled had female heads of household. When comparing female-

and male-headed households across Districts and Haor type, the following observations can be

made:

- female-headed households have significantly lower per capita monthly income levels than

male-headed households

- there are no significant differences in per capita expenditures between female- and male-

headed households, except in deep Haor areas where expenditures in female-headed

households are significantly lower

- female-headed households have significantly lower food consumption scores than male-

headed households

- female-headed households have a significantly higher coping strategy index score in

Kishoreganj and Sunamganj, and in deep Haor areas

*****

FSUP-H Baseline Report, June 2010 1

1.0 INTRODUCTION

1.1 Food security and poverty context in Bangladesh

“Food security exists when all people, at all times, have physical and economic access to sufficient,

safe and nutritious food that meets their dietary needs and food preferences for an active and healthy

life”2. Food security is said to have four main pillars: availability, access, utilization and stability with

regards to the availability and access dimensions of food security.

Table 1: Four main pillars of food security

Availability The availability of sufficient quantities of food of appropriate quality, supplied through domestic

production or imports (including food aid).

Access Access by individuals to adequate resources (entitlements) for acquiring appropriate foods for a

nutritious diet. Entitlements are defined as the set of all commodity bundles over which a person

can establish command given the legal, political, economic and social arrangements of the

community in which they live (including traditional rights such as access to common resources).

Utilization Utilization of food through adequate diet, clean water, sanitation and health care to reach a

state of nutritional well-being where all physiological needs are met. This brings out the

importance of non-food inputs in food security.

Stability To be food secure, a population, household or individual must have access to adequate food at

all times. They should not risk losing access to food as a consequence of sudden shocks (e.g.

an economic or climatic crisis) or cyclical events (e.g. seasonal food insecurity). The concept of

stability can therefore refer to both the availability and access dimensions of food security

In Bangladesh, there has been significant progress in improving the gross food availability, in

particular through cereal self sufficiency and improvements in land productivity. However, food access

and utilization continue to remain critically low, especially among the poorest and disaster affected.

According to the 2005 Joint UN/GoB MDG report, Bangladesh was home to over 60 million food

insecure people (GoB-UN 2005). Income inequality and chronic poverty3 are the primary causes for

wide-spread food-insecurity. This is compounded by the population growth of around 2 million

individuals annually combined with a reduction of around 82,900 hectares of tillable land annually due

to infrastructure and housing development, and industrialization. About a third of the population lives

below the lower poverty line with seriously imbalanced diets and extremely inadequate intake of fats,

protein and micronutrients. While poverty is one of the main underlying causes of food insecurity of

many people, it has manifested in wide scale malnutrition of various types.

In recent decades, malnutrition has become a major public health concern in Bangladesh, affecting

the well being of the majority of the population, particularly the children, adolescent girls and

pregnant/lactating women in the ultra-poor households. The 2005 Joint UN/GoB MDG report states

that nearly half the children are underweight or stunted, with 13 to 19 percent being severely

underweight or stunted in terms of being more than three standard deviations below the relevant

NCHS standards. Another 2005 report by Hellen Keller International4 states that almost 40 percent of

under-5 children in rural Bangladesh are reported as stunted and 46 percent are reported as

underweight, indicating that chronic under-nutrition is widespread. A 2009 study found that prevalent

macro malnutrition problems (NIPORT, 2009), particularly in under-5 children include underweight

(46%), stunting (36%) and wasting (16%) and maternal under nutrition measured by BMI (32%). This

2 World Food Summit, 1996

3 49.6% people live in poverty (below US$ 1 per day).

4 Bangladesh in Facts and Figures: 2005 Annual Report of the Nutritional Surveillance Project, Hellen Keller


suggests that children in Bangladesh suffer from short-term acute shortfall in food intake as well as

longer-term under-nutrition. It is important to note that there are also large differences in child

malnutrition rates across economic groups. Child malnutrition is pervasive among the poor. More than

60 percent of the children 6-71 months old suffering from stunting, belong to the bottom consumption

quintile.

In Bangladesh, there is an important spatial dimension to poverty and food insecurity creating

disproportionate affects on people in disaster risk prone areas, such as char lands, Haors and coastal

areas. In 2009, the GoB/WFP/WB undertook a joint vulnerability assessment to prioritize development

initiatives and resources in areas of highest food security needs (BBS, 2009), based on upazila-level

population estimates of individuals living below the lower poverty line, which is defined as food calorie

consumption of less than 1805 Kcal/person/day. The assessment identified six geographical

areas/clusters with 145 highly food-insecure and poverty-prone upazilas. The six identified clusters

were (i) The North-West disaster area (ii) The North-Central Chars (iii) The Drought Zone (iv) The

Haor Basin (v) The Coastal Zone and (vi) Chittagong Hill Tract.

1.2 Background of the FSUP-H project

CARE Bangladesh has a longstanding and reputable program on food security. After the successful

completion of the Integrated Food Security Program (IFSP) in 2003, CARE initiated the Strengthening

Household Ability to Respond to Development Opportunities (SHOUHARDO), which was

implemented during 2005-2009. The SHOUHARDO design was the largest development program in

Bangladesh at the time, and was designed to be consistent with CARE‟s Unifying Framework for

Poverty Eradication & Social Justice.5

FSUP-H is part of the program approach of CARE-B and the design of the project is based on the

analysis of underlying causes of poverty (UCP) and social injustice at multiple levels and, theories of

change (ToC) around three long term programming areas of CARE-B: marginalized women, extreme

poor people, and people living in environmental and geographical vulnerable areas. Design of FSUP-

H has taken into account the lessons learnt from SHOUHARDO and other projects like FOSHoL,

SETU etc. and is aligned with CARE-B‟s long-term programming strategy for the Haor region.

The FSUP-H project aims to make a sustainable impact on the lives of the four CARE Bangladesh

„impact groups‟: (a) most socially, economically and politically marginalized women, (b) lowest

category of the wellbeing ranking especially those people trapped in a set of unequal power relations,

(c) most marginalized groups in urban areas (the project will indirectly contribute to this by reducing

migration), and (d) most vulnerable people and communities prone to disasters and environmental

changes.

FSUP-H‟s overall objective aligns with the objectives outlined in the EC Country Strategy Paper for

Bangladesh (2007-2013) where poverty, gender inequality and access to food are prioritized. Project‟s

overall focus on the reduction of poverty and food insecurity fits well with the Millennium Development

Goals, especially Goal One: to eradicate extreme poverty and hunger, Goal Three: to promote

gender equality and empower women, and it also addresses Goal Five: to improve maternal health.

5 SHOUHARDO – a Title II program of USAID, Final Evaluation Report, December 2009, TANGO International


1.3 Implementation framework of the FSUP-H project

The FSUP-H project constitutes an intra-CARE partnership between CARE International UK and

CARE Bangladesh. CARE International UK is the formal project lead and holds overall contract

responsibility and accountability to the European Commission. CARE Bangladesh is responsible for

day-to-day management of project implementation. CARE Bangladesh implements the project with

three national partners: Assistance for Slum Dwellers (ASD); People‟s Oriented Program

Implementation (POPI); and Sabalamby Unnayan Samity (SUS).

The overall objective of the project is to reduce extreme poverty and food insecurity in the Haor region

of Northeast Bangladesh. The specific objective is to sustainably improve food access and utilization

and reduce vulnerability for women and their dependents in ultra poor households in Sunamganj,

Netrokona and Kishoreganj Districts

The project has four specific results:

e) Increase inclusion and capacity of 55,000 ultra poor HHs with focused attention to women

headed ultra poor HHs and their dependents, to actively engage with development processes

with greater support from their communities and local level institutions,

f) 55,000 ultra poor households (particularly women) have additional economic opportunities

and income, improving their access to food and household food security round the year,

g) 55,000 ultra poor households have reduced vulnerability to food insecurity and poverty and

improved resilience to quick and slow onset disasters, and

h) Improve and equitable utilization of food as well as reduced malnutrition among women and

their dependents in 55,000 ultra poor households.

1.4 FSUP-H site and impact group selection

Among the six clusters of highly food insecure and poverty prone areas stated in section 1.1, CARE

Bangladesh identified the Haor basin as the target area for its FSUP-H project. The Haor is a wetland

ecosystem in northeastern Bangladesh that is a saucer shaped shallow depression in the land that is

also known as a back swamp. The Haor is a remote and difficult area that is flooded every year during

the monsoon. It remains under water for 6-8 months of the year, turning Haor settlements mostly built

on earthen mounds into islands. Villages are regularly washed away, which plays a large role in

driving people to migrate to urban centers. Some of the most extensive seasonally flooded areas in

South Asia are located in the Bangladesh Haor region. During the dry season most of the water drains

out, leaving small shallow lakes or may completely dry out by the end of dry season. This exposes

rich alluvial soil, extensively cultivated for rice.

In selecting the Haor region, CARE Bangladesh took into account the seasonal dimensions and socio-

political factors that particularly increase the vulnerability of the Haor population. There are two related

seasonal dimensions to food insecurity in the Haor: the high exposure to cyclic climatic shocks such

as flooding, flash flooding and erosion; and the single-season food production and consequent

seasonal variation in food availability and pricing as there is only one annual rice harvest in the Haor

area. In the hoar area, there are traditionally two food-insecure lean seasons, January to mid-April

(between rice planting and harvest) and mid-July to September (during the monsoon). The first is

particularly severe for rural landless people because it coincides with the pre-harvest period of low

employment opportunities in agriculture.

In addition, socio-political factors such as gender, age, ethnicity and religion determine people‟s


position in society, their relationships to those in power, and access to resources and services, and

are equally important determinants of food insecurity in Bangladesh. In recent years, poor people

have increasingly lost their fishing rights in the Haors or rivers which had been their only source for

livelihoods and food-security for decades. These water bodies have now been taken over by powerful

people with political connections who control the majority of water bodies and only allow poor people

to fish for a payment of daily fees and a percentage of sales proceeds for a certain period. Moreover

the Haor region is considered socially conservative and imposes strict restrictions on women‟s

mobility. The harsh physical environment further impedes their movement. Poor women are

marginalized because of male-dominated systems and structures, unequal gender power relations,

and limited choices and opportunities. Wage discrimination is a significant contributor to food

insecurity for women in the Haors who earn approximately half the daily wage of their male

counterparts for the equal work, and even then are severely affected by the seasonality of work

availability. The absence of health services and transportation facilities affects especially pregnant

women severely

The Haor region covers Sunamganj, Habiganj and Moulvibazar districts and Sylhet Sadar Upazila, as

well as Kishoreganj and Netrokona districts outside the core Haor area. Based on a vulnerability

assessment of all the Union Parishads in the Haor region6, CARE found Kishoreganj, Netrokona and

Sunamganj had the largest number of communities in the highest vulnerability categories, and

selected these 3 districts for project implementation. Due to the remote location and difficult physical

conditions, government services are almost absent.

Within the three project districts, CARE Bangladesh undertook a rigorous selection process during the

start up phase of the FSUP-H project to identify 55,000 ultra poor households as the main project

target group, with an important focus on the women in these households. By October 2009, the

project team had completed 672 community WBAs and selected 645 communities in 94 unions of 17

upazilas in the three project districts. Based on this selection, the project team collected information

from 55,000 households and developed individual household profiles. These profiles captured key

information such as: sex of household heads, household size, primary and secondary occupation of

household heads, occupation of women in the households, homestead and cultivable land size,

number of livestock and poultry, types of latrine used, tube well ownership, NGO involvement and

loan status.

2.0 FSUP-H BASELINE STUDY

2.1 Rationale of the study

The main purpose of the survey was to generate baseline information and data on food security

status, poverty and vulnerabilities of the impact groups. By providing a benchmark the baseline survey

provided an opportunity to collect follow-up data and information over the life of the project to measure

effect and impact of project interventions/activities. This allowed FSUP-H staff to understand to what

extent the project contributes to improving food security and poverty level.

The information and data generated by the survey will be useful in: (1) designing future similar

projects; and (2) scaling up the current project; and (3) to track impact-level changes and livelihood

trends over time.

6 SHOUHARDO Haor Region: Union Selection Process, CARE 2005; Md, Raquibul Hasan, Village Selection

Survey; An Elaboration of the Process, CARE 2005.


2.2 Objectives of the study

The objective of the study is to better understand the current food insecurity, poverty and vulnerability

situation of the program impact group, and to establish baseline values of indicators for intended

outcomes against which future change can be measured in terms of: a) behavior, b) systemic capacity

and c) impact on the socio-economic conditions of target households such as number of food

insecure months, income and expenditure.

The specific objectives of the survey were to:

a) Assess socio-economic characteristics of the households;

b) Identify the level of food insecurity, diversity of food consumptions and prevalence of

malnutrition (including infant & child feeding practices) of the households;

c) Assess current ability of program HHs to participate in the development process and access

to different services;

d) Understand the natural crisis/shocks experienced by the households and coping mechanisms

(resilience);

e) Validate the needs and priorities of project participants, communities and institutions identified

in the project proposal.

f) Gather and analyze information for the purpose of in-depth learning and to assist the project

in modifying appropriate interventions, refining the Logframe and M&E plan.

2.3 Scope of the study

The scope of the survey is not limited to indicator measurement requirements of the project. The

survey will also seek to better understand livelihood issues of the ultra poor households of the hoar

regions. The study will also explore different aspects of household food security (availability, access

and utilization patterns), households‟ exposure to development processes and their ability to negotiate

for services and rights, vulnerability to climate changes etc. The study will produce household-level

analysis by district and Haor type.

3.0 STUDY METHODS

3.1 Study design

The baseline survey utilized a combination of quantitative and qualitative methods. These methods

were in part complementary, so that each type of information could contribute to an overall

understanding of households. The quantitative methods involved a detailed household-level survey,

while the qualitative methods utilized mixed tools.

3.2 Quantitative study design

The study collected data on a variety of subjects and issues by administering a structured

questionnaire at the household level with the key woman of the household and/or her spouse as

respondent. For collecting anthropometric data from children aged 6-23 months, standard weight and

height scales were used.

The household questionnaire was divided into ten sections, each covering a different aspect of

livelihoods or subjects relevant to CARE FSUP programming objectives. The following topics were

covered:


Section A: Identification – area identification, religion and ethnicity.

Section B: General Information on household members – includes elements of household

demographics, education, disabilities, marital status, primary and .secondary occupations

Section C: Economic Security – includes housing characteristics, ownership of assets,

household expenditures, income and employment, and loans.

Section D: Food Security – includes information on food consumption, months of food

sufficiency, and household food access.

Section E: Water and Sanitation – access to clean water and latrines.

Section F: Health Practices and Illness – data on hand-washing behaviors, illnesses.

Section G: Participation – Information on household participation in development processes

and access to services and common property

Section H: Natural Disasters – types of disasters that have impacted the household in the

previous year and their effect on the household.

Section I: Family Authority and Decision-making – decision-making at the household level

and attitudes about family life.

Section J: Child nutrition, Antenatal Care and Family Planning – information on breastfeeding

practices, food consumption during antenatal care, child food consumption, antenatal care

and family planning, immunizations, and anthropometrics of children 6-24 months.

The quantitative methods employed random selection criteria in order for the results to be generalized

at the household level to both District and Haor type (moderate and deep; discussed below). The

baseline study was not designed to be generalized by Haor type within Districts, as this would have

resulted in an unmanageable sample size for the household survey.

The questionnaire for the household survey was developed jointly by CARE Bangladesh, DMA and

TANGO staff, and was based in part on questions posed in similar food security baseline surveys in

Bangladesh and elsewhere. Technical input by DMA and CARE Bangladesh both before and during

training ensured that questions were relevant, culturally appropriate, well-translated, and the listed

response codes were correct. Draft instruments were pre-tested in approximately fifty households

during enumerator training, which took place in Mymensingh in January 2010. The final questionnaire

is attached as Annex 1.

A multi-stage sample design was used for the household survey. The first stage was a stratification

based on three districts – Kishoreganj, Netrokona, and Sunamganj - where FSUP is implementing its

program. The justification for using stratification at this level was to use the baseline to inform CARE

Bangladesh of the major differences among Districts in order that program adjustments could be

made and that future studies could disaggregate changes by location. The first sampling stage also

included a second stratification of the entire project are into two types of Haor – moderate and deep.

Each of the three districts has both types of Haor area, and they are found in different Upazilas (sub-

districts). The third stage of sampling was the selection of clusters (villages) within each Haor type in

each District using probability proportional to size selection methods. A limited number of clusters (20

per strata) were selected due to two factors – the relatively small geographical variation of Haor areas

within each district and the expected magnitude of intra-cluster variation being relatively large

compared to inter-cluster variation, which made it more reasonable to sample more households within

clusters to reduce error.

The fourth and final stage of the sampling process was the selection of households. In each cluster a

fixed number of households were randomly selected from a sampling frame of the households.


Systematic random sampling was used with a randomly selected starting point (household number on

the list) and a sampling interval when lists or maps were not available. The survey used no

replacement and instead sample sizes were upwardly adjusted to account for non-replacement

assuming a non-response rate of 5 percent.

The formula used to calculate the baseline sample size was the following:

n = deff(z/standard error)² (p) (1-p)

Where:

n = sample size

deff = design effect

z = standard score corresponding to a given confidence level (z = 1.645 for the 95%

confidence level)

Standard error = acceptable error level

p = expected proportion of the population expressing a particular characteristic

(1-p) = expected proportion without the characteristic

This formula is only appropriate for baseline measurements of multi-variable surveys. It establishes

variation and expected proportions of key variables which subsequent surveys can use to base

sample sizes required for estimating differences in means or proportions. In applying this formula p

was given a value of .5, as this maximized the influence the proportion of the population with any

given characteristic had on the size of the sample. For the baseline survey, z was fixed at 1.645 (95%

confidence limit) and p was set at 0.5. The two remaining variants were the design effect and the

standard error.

The term “deff” is the design effect. This provides a correction for the loss of sampling efficiency

resulting from the use of cluster sampling instead of simple random sampling, and the gain of

sampling efficiency resulting from stratification. It is the factor by which the sample size must be

multiplied by in order to produce survey estimates with the same precision as a simple random design

would. Ideally, an estimate of deff for the indicators of interest could be obtained from a prior survey in

a given setting, providing some insight on the similarity or homogeneity among households in the

clusters. Short of this, typical values from surveys conducted elsewhere are normally used. When

clustering is the only sampling stage prior to the random selection of households, a default value of

2.0 is commonly used. However, for this survey there was an additional stage of stratification.

Stratification actually increases the efficiency of sampling by accounting for variation in the sample

even before the sample is drawn. Thus, it usually has a design effect of less than 1.0. Combining the

two stages – stratification and clustering – usually results in a design effect between 1.0 and 2.0. In

the case of Bangladesh households, it is assumed a priori that inter-household variation is small

compared to that of population-based surveys that are district-wide. Thus a design effect (deff) of 1.6

was used, mainly due to the fact that two stages of stratification were employed.

Table 2 provides estimated sample sizes using variants of deff and standard error. As can be seen,

for a given standard error the design effect raises the required sample size by modest amounts.

However, at a given deff, changes in the standard error have a profound effect on sample size

(because it is the denominator of a squared term). For the CARE Bangladesh FSUP survey it was

recommended that a sample size of 316 households per strata, or 1,896 total households be selected

(there are six strata – 3 districts by 2 Haor types). Data was collected from 16 randomly selected

households in each of twenty clusters within each stratum. Data was collected from 1,920

households. After data cleaning, 28 households were removed from the sample due to incomplete


interviews or other inadequacies, for a final tally of 1,892 households. Table 3 provides a breakdown

of the sample by District and Haor type.

Table 2: Illustrative sample sizes for stratified random sampling using

variants of deff and standard error.

Design

Effect Z

Standard

Error TERM P 1-P

Sample

Size

Sample

Size

*1.05

1 1.645 0.05 1082.41 0.5 0.5 271 284

1.2 1.645 0.05 1082.41 0.5 0.5 325 342

1.4 1.645 0.05 1082.41 0.5 0.5 380 400

1.6 1.645 0.05 1082.41 0.5 0.5 434 455

1.8 1.645 0.05 1082.41 0.5 0.5 488 512

2 1.645 0.05 1082.41 0.5 0.5 542 570

1.6 1.645 0.04 1691.26 0.5 0.5 676 710

1.6 1.645 0.05 1082.41 0.5 0.5 434 455

1.6 1.645 0.06 751.67 0.5 0.5 302 316

1.6 1.645 0.07 552.25 0.5 0.5 221 232

1.6 1.645 0.08 422.81 0.5 0.5 170 178

Subsequent surveys to estimate change from the baseline survey will utilize the following formula:

n = deff [(Z1 + Z2)2 * (sd1

2 + sd2

2) / (X2 - X1)

2]

This formula takes into account the magnitude of change that can be detected with 95 percent

confidence given the expected standard deviations for the indicators of interest.

Table 3: Sampling statistics of 1,892 households by District and Haor type

Sample Sizes

District

Kishorega

nj

Netrokona Sunamganj

# of Households Surveyed 628 634 630

% of Households Surveyed 33.2 33.5 33.3

Number of <2‟s Measured 146 124 132

By Haor Type Moderate Deep

# of Households Surveyed 947 945

% of Households Surveyed 51.1 49.9

Number of <2‟s Measured 227 175

The study collected data from 16 randomly selected households from each village/community. The

households included in the sampling frame, and thus eligible to be sampled, represented 55,000

extreme poor households in the Upazilas and who are the ultimate beneficiaries of FSUP (about 25%

of all households in the area). The Field Researchers collected data directly from the one of the

female participants from the household and from her spouse (if available). If the wife was not available

on the day of interview, the enumerator went on to the next randomly selected household. For

households with children between the ages of 6 and 23 months anthropometric data (height, weight


and age) was also collected on the same day using enumerators trained specifically for this function

and using appropriate scales and measuring boards.

Quantitative data collection took place from January 13 through

February 16, 2010 using Personal Digital Assistants (PDAs);

small hand-held computers that provide facilities for taking notes,

storing data and retrieving information, and running survey

software. PDA‟s are pen-based and use a stylus to tap selections

on menus and enter printed characters. TANGO International

used The Survey System (TSS) software package to build the

FSUP-H baseline questionnaire because it is one of the few

software packages that will accommodate non-Roman

characters. The Bangla TSS questionnaires were then

transferred to the PDAs using a series of .xml files.

When doing the interview, enumerators read each question off

the screen, just as if they were using paper questionnaires. Each

household was stored as an independent record. The Team

Leader or Supervisors downloaded data from enumerators each

day and stored that day‟s data on an SD memory card or laptop

using a unique file name. A copy of the data remained on the

PDA for the entire survey as a back-up.

The use of PDAs gives significant benefits over traditional paper-based surveys. Using PDA-based

questionnaires greatly reduces survey error, especially data entry error. In a PDA-based survey, as

data is being entered it is subject to validation, i.e. it is controlled on the spot for possible errors

(numbers out of range, percentages that do not add up to 100%, etc.) and some questions may be

enabled or disabled on the basis of replies to the previous questions. In addition, all logic rules, such

as skips to other questions or avoids, are controlled by the PDA and not by the enumerator. Such

rules are automatic, taking the enumerator to the next relevant question. This is particularly useful in

complex, multi-indicator surveys such as the FSUP-H baseline.

Using PDA-based survey instead of paper based questionnaire reduces the time needed for data

collection and processing. Once data is collected with paper-based surveys, each questionnaire then

has to be entered into a database by a data-entry clerk, and then cleaned of errors by a data analyst.

With a PDA-based survey there is no need for a subsequent data entry process since the data is

entered directly into a data file. This greatly reduces data entry errors (these are the largest single

source of error for a survey). Finally, PDA based surveys don‟t need to use paper and ink. This is

environmentally friendlier. In addition it reduces the logistics burden of carrying many questionnaires

around the field and then storing them in the office.

Data cleaning and analysis was undertaken as a 2-step process. The first round of data cleaning and

analysis was undertaken by DMA in March 2010, the second and final round of cleaning and analysis

was undertaken by TANGO International in April 2010. During the data cleaning process, 28

respondent files (1.5 percent) were discarded because of incompleteness/inadequacies. The final

sample size was 1892 respondents.


3.2 Qualitative study design

Qualitative data collection was organized by 3 teams made up of CARE and partner staff. Each team

included at least one female facilitator. Field research was guided by CARE colleagues from the

Social Development Unit (SDU) under overall coordination by the FSUP-H M&E Coordinator and with

inputs from the CARE Program Quality Unit in CBHQ.

The qualitative work focused on three main themes:

1. Maternal Child Health and Nutrition (MCHN)

2. Participation in Economic Activities

3. Participation in Development Activities

Data collection started on 10 February 2010 and lasted for 10 days. During this time, each team

undertook qualitative field work in 2 villages in the same district: one in the deep Haor, and one in the

moderate Haor. Villages were selected based on a degree of convenience in accessing the village,

and representativeness of the project objectives and main intervention areas.

The teams spent 5 days on data collection in each village; 4 days of qualitative work and 1 day to

finalize reporting. Data was recorded as hand-written notes and was transferred to structured

reporting templates on the same day it was collected.

Data cleaning and analysis was undertaken as a 2-step process. The first round of data cleaning and

analysis using top-line methodology was undertaken by CARE Bangladesh in February 2010,

additional analysis was undertaken by TANGO International in March 2010 for integration of findings

into this report. The data collection tools and guidelines, and qualitative findings are attached as

Annex 2.

Table 4: Qualitative techniques applied for the baseline survey

A. Focus group discussions (FGD) x 24

6 per village, different combinations possible

Time: 1-1.5hr

Semi-structured group discussions with 5-10

participants; male or female, no mixed; 1 facilitator with

same gender as the FGD participants; 2 note

takers/observers, ideally also same gender as FGD

participants; use of participatory mapping (Venn

diagram, social mapping) and ranking (problem,

preference and wealth ranking) techniques

B. Key informant interviews (KII) x 30

5 per village

Time: 1hr

In each village, 4 semi-structured interviews with

individuals in the village and minimum of 1 external KII

at union/upazilla level; 1 facilitator and 1 note taker

C. In-depth interview (IDI) x 12

2 per village

Time: 1-2hrs

Unstructured interviews. Individuals selected from

FGDs; 1 interviewer only. Focus was to collect “rich”

human interest stories to complement/add to FGD

findings.

D. Trend analysis (TA) x 4

Organized in 4 villages only (2 deep, 2

moderate)

Time: 2-3hrs

TA included two techniques: seasonal calendar (1-

1.5hrs) and daily pattern mapping (1-1.5hrs). Both

techniques were undertaken with females.


Picture 1: Baseline enumerator using the PDA for a household interview


4.0 DEMOGRAPHIC CHARACTERISTICS

The FSUP baseline survey included basic demographic information on 1,892 households and 8,957

individuals. The vast majority of respondent households in the survey are Muslim (85.7%), followed by

Hindu (14.1%) and „Other‟ (0.2%). Kishoreganj has a significantly higher proportion (p=.000) of

Muslims (91.2%) than Netrokona (80.6%) and Sunamganj (85.4%). There is also a significant

difference by Haor type (p=.000), with Moderate Haor having a higher proportion of Muslim

households (90.4%) than Deep Haor (81.1%). Virtually all households (99.9%) are Bengali.

The gender ratio found in the study is 96, which means that there are 96 males for each 100 females,

and the proportion of the sample that was 51.1 percent. This ratio portrays the opposite picture

regarding the male female ratio compare to the national level statistics for same category (i.e.

nationally male-female ratio is 104 males for each 100 females. Source: Statistical Pocketbook, BBS,

2004) but matches exactly the ratios found in the FoSHoL-CARE baseline study 2nd

cycle.

For the overall population, the average age of Head of Household is 42.7 years. Table 5 shows that in

Kishoreganj heads of household are significantly younger, but there is no difference by Haor type.

Almost 15 percent of sampled households are female-headed, but Netrokona has a significantly

higher proportion of female-headed households (17.4%). Again there is no difference by Haor type.

Female heads of households are significantly older (p=.000) than their male counterparts averaging

48.2 and 41.7 years old, respectively. In terms of household size the average for the study population

is 4.8 people per household, however Sunamganj has a significantly higher household average size at

5.3. Female-headed households are significantly smaller than male-headed households (3.1 and 5.0,

respectively).

The smaller household size for female headed households is a function of several factors. First, they

are older than their male counterparts, so the chance of having younger children is smaller, and the

chance that at least some of their children have married and moved out of the household is greater.

Also, they are less likely to have a counterpart male adult in the household. As a result, even though

many are still of reproductive age, their chances of having more children of their own are small.

Table 5: Key demographic characteristics of the population, by District and

Haor type

By District Kishoreganj Netrokona Sunamganj

Average age HHH

(yrs)

Overall 40.0*** 43.7 44.3

Moderate 40.5 44.1 44.3

Deep 39.5 43.3 44.2

Household Size

Overall 4.6 4.4 5.3***

Moderate 4.5 4.3 5.3

Deep 4.6 4.4 5.3

% female-headed

HHs

Overall 14.6 17.4** 12.1

Moderate 15.2 18.5 11.4

Deep 14.0 16.3 12.8

By Haor Type Moderate Deep

Average age HHH (years) 42.3 43.0

Household Size 4.8 4.7

% female-headed households 14.4 15.0

*** denotes p=.000; ** denotes p=.050


Only about 2 percent of individuals are reported to be disabled, and there are no significant

differences among the three Districts or by Haor type.

Eighty five percent of household heads are married; there is no significant difference between Haor

types. Only about 1 percent of the household heads was never married; 12 percent were widowed;

and 2 percent were divorced/ separated female household heads. The proportion of widowed

household heads is significantly higher in deep Haors in Netrokona.

Figure 1 shows the age distribution for the study population. The average age of the study population

is 22.2 years old. The median age is 16 (half of the population is below 16 years of age and half of the

population is above 16 years of age). There is a slight but obvious skewness in favor of females

between the ages of 18 and 30, and a slight bias of males in the ages between 35 and 45. Other than

these two anomalies ages are fairly equally distributed by gender. The modal (most common age) age

is 0 (infants between birth and one year old). Just over 21% of the population is under 5 years old,

and about 4 percent is over 60 years old.

Figure 1: Age distribution of study population, by sex

Ag

e, in

years

100

80

60

40

20

0

Frequency

400 300 200 100 0

Ag

e, in

years

100

80

60

40

20

0

Frequency

4003002001000

Sex

FemaleMale

Table 6 shows the demography and dependency ratios of FSUP-H households. On average, survey

households have about 2.3 adults of working age (15–60 years);7 0.8 children under age 5; 1.4

children between the ages of 5 and 14 years, and only 0.2 elderly persons above 60 years.

Three types of dependency ratios are presented in the table: child aged and total. The total

dependency ratio is defined as the ratio of the number of members in the age groups 0–14 years and

7 This is the notion of working age commonly used by demographers (see, for instance, Shryock et al. 1976). The

actual working age of individuals of course depends in part on their standard of living and can often be lower, especially for the poor.


above 60 years to the number of members of working age (15–60 years). The ratio is expressed in a

percentage. The total dependency ratio has strong and negative correlation with household income.

High dependency ratios mean a higher burden on household income. For the FSUP-H respondents,

the total dependency ratio is 114.5%, which can be considered high. Adult members of working age

have more children to support than aged household members.

Table 6: Demography and dependency ratios, by District and Haor type

Characteristic

District Haor Type Total

Kishoreganj Netrokona Sunam-

ganj Deep Moderate

Number of household members in the age group

0-4 years 0.8 0.7 0.9 0.8 0.8 0.8

5-14 years 1.3 1.3 1.6 1.4 1.4 1.4

15-60 years 2.2 2.2 2.5 2.4 2.3 2.3

Over 60 years 0.2 0.2 0.2 0.2 0.2 0.2

Demographic composition (percent)

0-4 years 18.2 16.4 17.0 17.1 17.3 17.2

5-14 years 28.9 28.7 30.8 29.7 29.4 29.5

15-60 years 49.4 49.9 48.2 49.4 48.8 49.1

Over 60 years 3.5 4.9 4.0 3.8 4.5 4.1

Total 100.0 100.0 100.0 100.0 100.0 100.0

Dependency ratio (percent)

Child (0-14) dependency ratio 104.2 100.9 112.4 104.8 105.2 105.0

Aged (>60) dependency ratio 8.7 8.0 9.3 8.1 11.0 9.5

Total dependency ratio 112.9 108.9 121.7 112.9 116.2 114.5


5.0 LIVELIHOODS AND ECONOMIC SECURITY

5.1 Occupational patterns

The primary occupation of surveyed household members reflects the principle livelihood strategies of

households in the Haor region. The survey collected data on the primary and secondary occupations

of all household members eight years of age and older, with a recognition that many have access to

multiple occupations in rural Bangladesh.

Tables 7a and 7b show the primary and secondary occupations for adults aged 8 years and older by

Haor type and District, respectively. What the data by Haor type and District show is that there are few

distinct differences in occupational trends in the Haor region sampled. As expected for the FSUP sub-

population, few households engage in agriculture on their own fields; less than 1 percent overall. This

is a direct result of the extensive degree of landlessness among the ultra poor in Bangladesh.

Sharecropping is more prevalent in the moderate Haor (3.8% versus 1.6%).

Agricultural and non-agricultural labor are the two main livelihood opportunities for FSUP households,

and together account for half of primary occupations, with no significant difference between Haor

types (table 7a). Other important livelihood activities, in order of predominance, are fishing, petty

trade, and housemaid/servant. Together, these five livelihood activities account for almost 65 percent

of primary occupations for men and women together. In terms of differences between moderate and

deep Haor areas, table 7a shows that sharecropping and fishing opportunities differ, with

sharecropping being more prevalent in the moderate Haor areas and fishing being more prevalent in

the deep Haor areas.

Picture 2: Agricultural day labor - males

The majority of respondents reported having no secondary occupation. This would be expected for

livelihoods such as salaried employees, business owners, many skilled laborers, etc., but not for

those who rely heavily on day labor opportunities. However, over 46 percent of agricultural and non-

agricultural day laborers have no secondary occupation/activity, suggesting that these individuals and


households have very low resiliency and could benefit greatly from diversifying their basic livelihood

skills.

Among those with secondary occupations, agricultural and non-agricultural day labor opportunities are

the most common occupational categories for those whose primary activities are in areas such as

petty trade, sharecropping, and fishing. Livestock husbandry is slightly more important as a secondary

occupation in the moderate Haor areas, but again there is very little difference in livelihood patterns

between the two Haor types.

Table 7a: Primary and secondary occupations for individuals aged 8 years and older, by Haor type.

Occupational Categories

Adults 8 years and older Deep Haor

Adults 8 years and older Moderate Haor

Primary occupation

Secondary occupation

Primary occupation


Men Women Men Women Men Women Men Women

No secondary occupation 64.8 89.2 66.0 85.7

Own agriculture (crop production) 1.1 0.1 0.3 0.0 1.5 0.4 0.6 0.1

Sharecropper 1.7 0.1 1.0 0.1 4.6 0.0 1.3 0.1

Own agriculture and sharecropper 0.6 0.0 0.3 0.0 0.9 0.0 0.1 0.0

Livestock husbandry 0.9 1.0 0.1 0.4 0.6 1.2 0.5 1.7

Agricultural laborer 36.5 1.3 12.6 0.2 37.9 0.3 10.4 0.1

Non-agricultural laborer 14.9 3.0 8.8 0.9 15.1 3.7 8.6 1.4

Housemaid/servant 3.0 6.0 0.1 1.1 2.5 5.0 0.1 1.3

Skilled labor8 2.5 0.5 0.3 0.2 2.6 0.2 0.5 0.1

Salaried employment (GOB-NGO) 3.6 1.0 0.1 0.0 4.2 1.0 0.2 0.0

Business 2.1 0.1 0.8 0.1 2.9 0.8 1.3 0.1

Petty business 7.8 1.2 1.3 0.2 7.0 1.5 0.8 0.3

Rickshaw/van pulling 3.6 0.0 0.6 0.0 6.5 0.1 1.5 0.1

Fishing (including fish culture) 10.5 0.2 7.5 0.1 3.4 0.1 6.8 0.0

Fishing laborer 0.9 0.1 0.4 0.0 0.2 0.0 0.4 0.0

Natural resource collection 0.5 0.6 0.1 0.1 0.2 0.4 0.2 0.2

Housewife 0.0 69.8 0.0 6.4 0.0 69.6 0.0 8.0

Beggar 0.5 0.7 0.0 0.1 0.5 0.8 0.1 0.1

Unemployed 3.8 7.1 0.2 0.1 4.2 7.6 0.1 0.1

Other 1.3 1.5 0.5 0.7 0.9 1.0 0.2 0.5

Unable to work 4.2 5.8 0.1 0.1 4.3 6.3 0.2 0.2

There are also many similarities in terms of primary and secondary opportunities by District, as Table

7b shows. Agricultural laborer is less important in Sunamganj compared to Kishoreganj and

Netrokona, but non-agricultural opportunities appear greater. Together, however, these two forms of

day laborer comprise approximately half of primary occupations for men and women combined in


each of the three districts. Sharecropping is significantly more common in Sunamganj and salaried

employment is more common in Netrokona. Petty business is a more common primary occupation in

Kishoreganj. In all three Districts about nine percent of respondents over 16 years of age are unable

to work, but not surprisingly the mean age of this category is 67 years.

As with the disaggregation by Haor type, agricultural and non-agricultural day labor opportunities

represent the most common secondary occupational categories for those whose primary activities are

in areas such as petty trade, sharecropping, and fishing. Petty business is slightly more important as a

secondary occupation in Kishoreganj, but again there are very few important differences in livelihood

patterns among the three Districts.

Table 7b: Primary and secondary occupations for individuals aged 8 years and older by District.

Occupational Categories

Adults 8 years and older Kishoreganj

Adults 8 years and older Netrokona

Adults 8 years and older Sunamganj

Primary occupation


Primary occupation


Primary occupation


Men Wom-

en Men

Wom-en

Men Wom-

en Men

Wom-en

Men Wom-

en Men

Wom-en

No secondary occupation 62.8 88.1 67.2 86.7 66.1 87.5

Own agriculture 1.2 0.2 0.2 0.1 1.0 0.4 0.5 0.0 1.4 0.1 0.6 0.0

Sharecropper 1.2 0.0 0.6 0.2 1.5 0.0 1.8 0.0 6.1 0.1 1.1 0.0

Own agr and sharecropper 0.9 0.0 0.1 0.0 1.0 0.0 0.2 0.0 0.4 0.0 0.3 0.0

Livestock husbandry 0.1 1.3 0.0 1.4 1.1 1.1 0.3 1.8 0.9 1.0 0.6 0.1

Agricultural laborer 38.5 1.6 12.3 0.2 46.9 0.5 8.7 0.2 27.9 0.4 13.2 0.2

Non-agricultural laborer 13.8 3.7 9.3 1.3 6.9 2.4 6.8 0.5 23.2 4.1 9.7 1.6

Housemaid/servant 1.4 4.4 0.0 1.3 3.3 7.9 0.2 1.4 3.4 4.3 0.2 0.8

Skilled labor 2.7 0.6 0.1 0.1 2.6 0.1 0.8 0.0 2.4 0.3 0.3 0.3

Salaried employment 3.3 0.9 0.1 0.0 5.5 1.9 0.2 0.0 3.0 0.3 0.1 0.0

Business 2.9 0.9 1.7 0.2 3.3 0.6 0.7 0.0 1.4 0.0 0.8 0.1

Petty business 12.0 1.6 2.2 0.4 4.3 1.9 0.8 0.1 6.4 0.6 0.3 0.3

Rickshaw/van pulling 7.6 0.1 2.4 0.1 4.3 0.0 0.6 0.0 3.7 0.0 0.2 0.0

Fishing (inc. fish culture) 5.5 0.1 7.2 0.1 6.8 0.1 9.8 0.0 8.5 0.2 4.9 0.0

Fishing laborer 0.3 0.0 0.5 0.0 0.5 0.0 0.5 0.0 0.9 0.1 0.2 0.0

Natural resource collection 0.0 0.6 0.0 0.0 0.4 0.9 0.0 0.4 0.6 0.0 0.4 0.1

Housewife 0.0 69.5 0.0 5.7 0.0 68.1 0.0 8.1 0.0 71.3 0.0 7.8

Beggar 0.4 0.9 0.0 0.1 0.9 1.2 0.1 0.0 0.2 0.3 0.0 0.3

Unemployed 3.8 6.1 0.1 0.1 4.1 6.8 0.3 0.1 4.0 8.9 0.1 0.0

Other 1.3 0.6 0.3 0.4 0.8 0.4 0.2 0.7 1.2 2.6 0.5 0.8

Unable to work 3.1 7.0 0.1 0.2 5.0 5.9 0.1 0.1 4.4 5.4 0.2 0.2

When comparing between men and women, table 7a and 7b show similarities by District and Haor.

The majority of women are housewives. Non-agricultural laborer and housemaid/servant are the most

8 Includes blacksmith, potter, porter, cobbler, carpenter, weaver etc.


common primary occupations for women but account for less than 10% across all Districts. Most

women reported having no secondary occupation. Qualitative data showed that while men primarily

earn income through agricultural and non-agricultural day labor, fishing and petty trade; the majority of

women do so through homestead activities such as fish processing / preparation of goods for market,

making handicrafts, and livestock and poultry rearing.

When asked about preferences for income generating activities, some important differences can be

noticed with the main occupations reported. Men indicated agriculture, cow rearing, fishing and

bamboo handicraft. Women indicated poultry and livestock rearing, shop keeping, agriculture and

small business. Furthermore, there were indications that ultra poor households have high rates of

child labor. Children were found to assist with fishing, livestock and poultry rearing, brick making,

paddy harvesting and vegetable cultivation.

In addition to the monetary benefits, community members identified other benefits of participation in

economic activities at the individual, household and community levels. At the individual level,

community members identified skills improvement, and improved social status as a key benefit.

Women stated increased mobility and communication as key benefits. Greater involvement of women

in economic activities was also seen as a way to better deal with lean periods when the males migrate

to sell labor. It was also mentioned that involvement of women in economic activities improved

education of children and reduce child labor, although others stated that it in fact contributed to child

labor. At the household level, reduced dependency on money lenders, improved household status,

increased food intake and diversity, better loan repayment, better home maintenance, reduced family

conflict over money and improved education for children were all seen as important benefits. At the

broader community level, increased employment opportunities for neighbors/friends and others,

increased access to essential goods due to newly established shops, improved women‟s decision-

Picture 3: Agricultural day labor - females


making in community issues, establishment of community schools, reduced dependency on

middle/rich class, reduced dependency on selling labor to other communities, and increased

community dignity all seen as important benefits

Community members also many identified costs of participation in economic activities including: travel

costs (in some cases as high as 150-200 Taka), wage days lost for skill training, and the need for

initial investments require high interest loans (perpetuating the debt cycle). There were also costs

identified that were specific to women, such as: wives are beaten if household chores are not

completed, children are not properly cared for and do not attend school regularly in absence of

mothers, and women are robbed of wages while traveling home from work. The time spent at

meetings/training and undertaking economic activities reduces time available for the traditional

household responsibilities of women. Although women participating in economic activities attempt to

distribute some of their household duties to other household members such as their husband or

relatives, this is rarely successful and often children end up doing the work.

Community members indicated that they would like to expand their economic activities, particularly in

the areas of cow rearing, other poultry and livestock rearing, fish culture, small shop keeping, small

business and homestead gardening. Males indicated additional preferences for cash-for-work,

rickshaw pulling and rice cultivation. Females indicated additional preferences for handicrafts (such as

bamboo products) and nursery development. In expanding their economic activities, community

members identified the following main barriers: lack of technical knowledge, support from GoB such

as livestock and fishery departments, capital, credit, production materials such as quality seeds,

access to water bodies; social kinship; and - for women - the prevailing social structures that prohibit

many women from participating in economic activities without the husband‟s consent, particularly with

respect to participation in agriculture and fishing.

Picture 4: Cow rearing by ultra-poor households


5.2 Household employment and income/expenditure

Closely linked to occupations and livelihoods are economic indicators of households, such as income,

other cash sources, asset ownership, debt and savings. Together these elements of economic

security reveal how resilient households are to economic shocks and natural disasters.

One of the first indicators of economic resiliency is the number of income earners per household. For

the survey population, each household had on average only 1.35 income earners. This can be

considered quite low but there could be seasonality factors due to the timing of the survey.

Kishoreganj and Netrokona had 1.31 and 1.29 income earners per household, respectively, while

Sunamganj had a small but statistically higher average (p=.000) of 1.45 income earners. Statistically,

the deep Haor areas had more income earners than the moderate Haor areas (1.39 versus 1.32,

respectively).

The data by Haor type and District show that the main household income sources align closely with

the primary occupations, as was expected based on number of income earners and main

occupations. Casual agriculture labor is the main income source across Haor areas and Districts,

followed by casual non-agricultural labor. Income from crop and animal sales is low overall. There are

few distinct differences in income sources between deep and moderate Haor regions sampled. The

main exception is the sale of fish/aquatic animal, which is significantly higher in the deep Haor. In turn,

sale of agricultural produce is higher in moderate than deep Haor areas. Income sources such as

salaried work, small business, petty trade and rickshaw/van pulling are more prevalent in deep Haor

areas.

Table 8: Income sources for previous 30 days, by Haor type

Income Sources

(multiple response)

Deep Haor Moderate Haor

N % of Responses N % of Responses

Selling vegetables 2 0.2 0 0.0

Selling livestock/poultry/birds 5 0.5 5 0.5

Selling agricultural produce 13 1.4 28 3.0

Selling fish/aquatic animals 106 11.4 26 2.8

Self-employed (carpenter, barber, etc.) 35 3.8 37 4.0

Salaried 63 6.8 83 9.0

Casual labor (agriculture) 544 58.3 501 54.3

Casual labor (non-agriculture) 288 30.9 276 29.9

Rickshaw/van pulling 45 4.8 82 8.9

Small business 129 13.8 140 15.2

Petty trade 6 0.6 7 0.8

Remittances/Pensions/Savings 5 0.5 10 1.1

Renting/leasing out property 1 0.1 2 0.2

Relief assistance from GoB or NGO 5 0.5 0 0.0

Selling HH assets 13 1.4 15 1.6

Begging 56 6.0 38 4.1

Total 1316 141.1% 1250 135.6%


When comparing income sources among districts, sale of fish/aquatic animals is higher in Sunamganj

than in the other districts. Casual non-agriculture labor and begging are also significantly higher in

Sunamganj, while casual agriculture labor is higher in Netrokona. Small business and rickshaw/van

pulling are highest in Kishoreganj.

Table 9: Income sources for previous 30 days, by District

Income Sources

(multiple response)

Kishoreganj Netrokona Sunamganj

N % of

Responses N

% of

Responses N

% of

Responses

Selling vegetables 0 0.0 1 0.2 1 0.2

Selling livestock/poultry/birds 6 1.0 3 0.5 1 0.2

Selling agricultural produce 15 2.4 12 2.0 14 2.2

Selling fish/aquatic animals 39 6.3 39 6.4 54 8.6

Self-employed (carpenter, barber, etc.) 22 3.5 28 4.6 22 3.5

Salaried 41 6.6 55 9.0 50 8.0

Casual labor (agriculture) 318 51.3 399 65.4 328 52.5

Casual labor (non-agriculture) 157 25.3 132 21.6 275 44.0

Rickshaw/van pulling 60 9.7 31 5.1 36 5.8

Small business 134 21.6 65 10.7 70 11.2

Petty trade 3 0.5 6 1.0 4 0.6

Remittances/Pensions/Savings 2 0.3 5 0.9 8 1.3

Renting/leasing out property 0 0.0 2 0.3 1 0.2

Relief assistance from GOB or NGO 2 0.4 2 0.3 1 0.2

Selling HH assets 6 1.0 16 2.6 6 1.0

Begging 24 3.9 22 3.6 48 7.7

Total 829 133.7 818 134.1 919 147.0

Table 10a shows the mean and median monthly income and expenditure per capita. The mean

monthly per capita income is 800 Taka overall. The per capita income is significantly lower in

Sunamganj than in the other districts. When comparing between Haor types, the mean per capita

income in the deep Haor is significantly higher than in the moderate Haor, but the median per capita

income in the deep Haor is significantly lower than in the moderate Haor.

Mean monthly expenditure per capita is 1,419 Taka, and the median monthly expenditure per capita is

1,099 Taka. When comparing across haor type, the median monthly per capita expenditure is

significantly lower in the moderate Haor.


Table 10a: Key income and expenditure data for households, by District and Haor type

Income/expenditure

Variable



ganj Deep Moderate

Mean monthly per capita

income (Taka) 845 880 674c 821

a 779 800

Median monthly per capita

income (Taka) 750 750 614

b 700

b 750 717

Mean monthly household

expenditures per capita

(Taka)

1,255b 1,583 1,478 1,487 1,353 1,419

Monthly median household

expenditures per capita

(Taka)

1,115 1,097 1,071 1,136 1,059b 1,099

Letters denote significant differences among Districts or between Haor types for a given variable.

Significance levels for comparisons: a = .10; b = .05; c = .00

Figure 2a shows the mean values of annual per capita income. For the survey population overall, the

mean annual per capita income is 9,599 Taka. Annual per capita income in Kishoreganj and

Netrokona are 10,136 Taka and 10,567 Taka, respectively – with no significant differences. However,

nnual per capita income in Sunamganj is significantly lower at 8,090 Taka.

Figure 2a: Mean values of annual per capita income, by District

Table 10b shows detailed expenditure data, by District and Haor type. Data is presented in three

categories, namely: daily, monthly and annual expenditure. The majority of daily expenditure is on

food purchases (72% of daily expenditure). The purchase of tobacco products is the next highest daily

expenditure (8%), followed by hiring manual labor from others and purchasing fuel (including gasoline,

kerosene, and fire wood). The remainder of the daily expenses goes to daily allowance for children,

transportation and beverages.


It is interesting to note that the single highest monthly expenditure item is cell phone cards (48% of

monthly expenditure). The second highest monthly expenditure item is medical expenses (including

fees, medicine and travel) (36%). The single highest item annual expenditure is clothing for household

members (51% of annual expenditure), followed by social/religious events (14%) and household

goods (11%), on average. Expenditure on fishing or fish raising, agricultural equipment/input, and

livestock and poultry rearing accounted for less than 10% of annual expenditure each.

Table 10b: Detailed expenditure data, by District and Haor type

Expenditures in Taka



ganj Deep Moderate

N 628 634 630 947 945 1892

Daily Expenditures

Food purchases 99.8 107.5 131.7c 117.8

c 108.2 113.0

Daily allowance for children 5.3 b

3.8 4.4 4.7 4.4 4.5

Transportation 4.2 3.8 6.4 4.8 4.7 4.8

Cigarettes 7.3 7.8 19.4 8.4 14.6 11.5

Beverages 0.2 0.5 1.0 c 0.6 0.6 0.6

Fuel (livelihood) 0.5 0.1 2.2 1.3 0.3 1.0

Manual labor 4.7 b

11.3 16.6 11.5 10.2 10.9

Kerosene oil 4.2 15.6 c

4.4 4.5 11.7 c 8.1

Wood fuel 6.0 c 1.7 3.4 4.6 2.8 3.7

Total: 132.2 152.1 189.5 158.2 157.5 158.1

Monthly Expenditures

Shelter rental 18.1 a 2.9 0.4 9.5 4.8 7.1

Insurance 4.1 3.2 4.5 2.8 5.1 a 3.9

Utilities 17.2 a 12.6 6.1 7.4

b 16.6 12.0

Education 45.5 52.8 55.7 62.2 a 40.4 51.3

Child care 17.2 18.5 38.7 c 21.0 28.6

b 24.8

Medical/dental care 100.5 c 45.4 41.8 61.8 63.1 62.5

Medicine 333.5 c 228.6 171.8 276.1

c

212.8 b

244.5

Medical travel 27.0 c 15.4 16.3 18.4 20.7 19.5

Cell phone cards 433.6 372.3 a

486.4 500.2 363.0 c 430.6

Transportation 14.2 20.7 20.4 16.8 20.1 18.4

Loan payment 28.7 c 4.2 32.2 22.5 20.9 21.7

Total 1,039.6 776.8 874.3 998.7 796.1 896.3

Yearly Expenditures

Clothing 1,765.0 1845.1 c 1577.7 1716.2 1742.8 1729.5

Livelihood equipment 244.2 162.8 b

266.6 214.4 234.3 224.4

Agr equipment/inputs 127.0 166.8 273.1 b

155.7 222.3 189.0

Fishing/fish-raising 265.3 a 340.1 407.1 470.8 204.1

c 337.6

Household goods 250.1 583.8 c

234.1 a 330.7 382.6 356.6

Livestock/poultry 94.3 c 83.9 23.0

b 55.8 78.4 67.1

Social/religious events 487.2 456.4 428.6 508.4 b

406.2 457.3

Dowry payment 57.8 93.3 49.4 90.4 43.3 66.9

Total: 3290.9 3734.2 3259.6 3542.4 3314.0 3428.4

Letters denote significant differences among Districts or between Haor types for a given variable. Significance levels for comparisons: a = .10; b = .05; c = .00


Figure 2 shows that median per capita monthly expenditures (including daily, monthly and annual

expenditures) are significantly higher than median per capita income. This is likely due to several

factors. First of all, there is the seasonality of the data collection; February falls in a lean period, which

is characterized by lower income and high lending. More information on lending and other coping

strategies is provided in section 5.4. Secondly, respondents have the tendency to overestimate

expenditure and underestimate income. It is important to note here that accurate income and

expenditure data collection requires very detailed questioning. The income and expenditure data

presented is this report, while indicative of income and expenditure levels, is most useful for

assessing trends over time and relative change between baseline and endline measurements.

Figure 2b: Median values of monthly household cash income and expenditures per capita, by District

and Haor type

5.3 Income in peak and lean seasons

Figure 3 clearly shows that there are significant differences between mean monthly income during

peak and lean seasons. When comparing peak season income levels across districts and between

Haor types, there are no significant differences in mean monthly income levels among districts, and

between Haor types. When comparing lean season income levels across districts and between Haor

types, the mean monthly income level in Sunamganj is significantly lower than in the other two

districts.

When asked for the main reasons that cause a lean period of income for a household, many

respondents found it difficult to clearly express the reasons for this. After probing, 82 percent of those

mentioned no opportunity for other/alternative work as a reason, 81 percent mentioned poor health of

the main income earner, 72 percent identified seasonal work, and 18 percent identified inability to

work due to bad weather/disaster. There were very few differences among districts and between Haor

types.


Figure 3: Average monthly incomes during peak and lean seasons, by District and

Haor type

5.4 Coping strategies for lean seasons

Table 11 and table 12 show the top ten ways of coping with lean periods by District and Haor type, as

per respondents‟ answers from a multiple choice list of 31 possible responses. Overall, adjusting

meals is the main coping mechanism (60.1%), followed by taking loans from friends/relatives (50.6%),

and taking loans from money lenders (35.6%).

When comparing across districts (table 11), a higher number of households take loans from

friends/relatives in Kishoreganj than in the other Districts. Percentage of households taking loans from

a moneylender is highest in Sunamganj. Adjusting meals is lower in Kishoreganj than in the other two

Districts, while eating famine foods is higher in Sunamganj. Accessing savings is higher in Netrokona

than in the other two districts.


Table 11: Top ten ways of coping with lean periods, by District

Coping Mechanisms

(multiple response)


N % of

Responses N

% of

Responses N

% of

Responses

Adjusting meals 342 54.5 404 63.7 391 62.1

Taking loans from friends/relatives 407 64.8 275 43.4 275 43.7

Taking loans from a money lender 188 29.9 203 32.0 283 44.9

Purchasing goods on credit 187 29.8 197 31.1 217 34.4

Accessing savings 146 23.2 257 40.5 129 20.5

Taking loans from an NGO 86 13.7 120 18.9 141 22.4

Relying on relief/aid 62 9.9 128 20.2 105 16.7

Eating famine foods 37 5.9 19 3.0 137 21.7

Reducing treatment costs 55 8.8 66 10.4 17 2.7

Temporarily migrating 52 8.3 40 6.3 44 7.0

Total 1721 274.0 1819 286.9 1888 299.7

When comparing between deep and moderate Haor types, table 12 shows few distinct differences.

Adjusting meals and informal lending are the main coping mechanisms in both Haor areas.

Table 12: Top ten ways of coping with lean periods, by Haor type

Coping Mechanisms

(multiple response)

Deep Haor Moderate Haor Overall

N % of

Responses N

% of

Responses N

% of

Responses

Adjusting meals 594 62.7 543 57.5 1137 60.1

Taking loans from friends/relatives 516 54.5 441 46.7 957 50.6

Taking loans from a money lender 313 33.1 361 38.2 674 35.6

Purchasing goods on credit 286 30.2 315 33.3 601 31.8

Accessing savings 267 28.2 265 28.0 532 28.1

Taking loans from an NGO 220 23.2 127 13.4 347 18.3

Relying on relief/aid 155 16.4 140 14.8 295 15.6

Eating famine foods 94 9.9 99 10.5 193 10.2

Reducing treatment costs 78 8.2 60 6.3 138 7.3

Temporarily migrating 62 6.5 74 7.8 136 7.2

Total 2794 295.0 2634 278.7 5428 286.9

Respondents were also asked about selling advance labor, separately from the multiple choice

question described in tables 11 and 12 above. Selling labor in advance is another coping strategy

used by households during lean periods. Qualitative data shows that advance labor is usually sold at

reduced wage levels. Overall, 7% of households had at least one member who sold labor in advance.

There was no significant difference between Haor types but there were significant differences among

Districts: Kishoreganj 5.4%, Netrokona 4.9%, Sunamganj 10.8% (p=.000).


5.5 Migration

Tables 11 and 12 show that temporary migration was not a common coping strategy to deal with lean

periods. However, migration for employment purposes is relatively common in areas with a high

degree of seasonal work, such as the FSUP-H project area. In moderate Haor areas, 38.5% of

households had somebody migrate in the last 12 months for employment purposes; in deep Haor

areas this was 32.3%. Moderate Haor areas also had a significantly higher average number of

household members migrating out of the village in the previous 3 months: 0.43 persons versus 0.36

persons for Deep; (p=.018). There were no differences when comparing among districts.

About 75% of those who migrated were heads of household, while about 20% were sons/daughters.

About 70% migrated to urban areas and 30% to other rural areas. There were no differences when

comparing these values among district or Haor type. While migration went on throughout the year,

there was more migration for employment purposes from August to October.

Table 13 shows the types of work performed by those migrating out of the household. Agricultural

contract labor and agricultural day labor are by far the most common types of work. When comparing

across districts, agricultural contract labor is higher for migrant workers from Sunamganj than in the

other two districts. In Netrokona, agricultural day labor is higher. In Kishoreganj, salaried employment

is higher than in Netrokona and Sunamganj.

While it is still mainly men who migrate for agricultural contract and day labor, qualitative data showed

that an increasing number of women also migrate for economic purposes, with many young females

migrating to work in garment factories.

Table 13: Type of work performed by those migrating out of the household within the last 12 months,


(N=672, multiple response) District Haor Type Total

Kishoreganj Netrokona Sunamganj Deep Moderate

Agricultural contract labor 37.1 39.5 57.8 42.3 46.0 44.3

Agric. day labor 31.9 53.3 26.1 32.6 40.3 36.8

Non-agric. day labor 1.6 3.8 2.4 3.6 2.7 3.1

Salaried empl - fixed business 17.1 2.4 4.7 11.1 6.6 8.6

Salaried employee 9.2 1.0 0.5 5.2 2.7 3.9

Maid/servant 4.8 8.6 2.8 3.9 6.6 5.4

Other 4.3 9.5 7.1 7.5 5.2 6.3

5.6 Loans

Table 14a shows that 78% of households overall held at least 1 current loan over the last 12 months.

When comparing across Haor types, a significantly higher number of households in moderate Haor

areas (80%) held loans than in deep Haor areas (75%). There were no significant differences when

comparing among districts.

The average number of loans per household overall was 1.4. When comparing across Haor types, the

average number of loans per household was significantly lower in deep Haor areas (1.2) than in

moderate Haor areas (1.5). There were no significant differences when comparing among districts.


The average loan amount was 6,652 Taka. When comparing across Haor types, the average loan

amount per household was significantly lower in deep Haor areas (6,346 Taka) than in deep Haor

areas (6,938 Taka). When comparing across districts, the average loan amount in Sunamganj was

significantly lower than in the other districts. Overall, the outstanding loan amount at the time of the

interview was 5,393 Taka, which is about 81% of the average loan amount - indicating a very high

debt burden on households. In Sunamganj, the outstanding loan amount was significantly lower than

in the other districts. There were no significant differences between Haor types.

Picture 5: Grameen Bank office

There was no significant difference in loan source among moderate and deep Haor areas, with the

majority of loans (41%) taken from money lenders, NGOs (24%), and friends/relatives (23%). Only 6%

of loans were taken from Grameen Bank and 4% from clubs/CBOs. Informal money lenders give

loans without collateral but instead charge higher interest rates. The high level of lending from

informal sources such as money lenders largely explains the high interest rates found in this survey.


Table 14a: Key loan data for households, by District and Haor type

Loan Variable



ganj Deep Moderate

N

Households with a loan (%) 80 74 79 75 80 b 78

Average number of loans per HH 1.5 1.1 b

1.4 1.2 c 1.5 1.4

Average loan amount (Taka) 7,148 6,944 5,880

b 6,346

a 6,938 6,652

Outstanding loan amount (Taka) 5,732 6,103 4,448 c 5,284 5,482 5,393

Outstanding as a % of average

loan amount 80.2 87.9 75.6 83.3 79.0 81.1



Table 14b shows that overall interest rates were 51%. The overall interest rates in Sunamganj were

significantly higher than in the other districts, and the rates in deep Haor areas were significantly

higher than in moderate Haor areas. Interest rates of money lenders were the highest, followed by

friends/family, NGOs and the Grameen bank. It is interesting to note that while Grameen bank

maintains a unified interest rate of 20% throughout the country, the survey data shows a range of 14-

20%.

Table 14b: Detailed interest rate data for loans, by District and Haor type

Interest Rate



ganj Deep Moderate

N 628 634 630 947 945 1892

Overall Interest rate (%) 46 45 62 c 59

a 44 51

Moneylenders 79 55 112 75 94 84

NGOs 13 15 25 17 23 19

Friends/relatives 63 55 37 65 43 56

Banks 11 8 13 11 8 10

Grameen Bank 14 20 19 14 21 18

GOB 12 9 13 10 14 11

Clubs/CBOs 39 57 55 43 50 16

Table 14c shows that, overall, 35% of loans over the 12-month recall period were taken by women.

Almost all women (98%) had taken a loan from the Grameen Bank, which reflects the Grameen

Bank‟s policy of lending to women. The proportion of women who took a loan from NGOs is also high

(88%), for similar reasons. The proportion of women taking loans from moneylenders is the lowest

among all loan sources.


Table 14c: Loan source for women, by District and Haor type

Loan source

District Haor Type


ganj Deep Moderate

Loans to Women (%) 29 32 43 39 30 35

Moneylenders 11 7 12 12 9 10

NGOs 84 86 92 90 82 88

Friends/relatives 20 24 21 21 21 21

Banks 33 29 50 33 33 33

Grameen Bank 96 100 98 98 97 98

GOB 0 0 50 50 50 50

Clubs/CBOs 12 10 33 13 12 12



Tables 15 and 16 show that the most common reasons for taking out a loan were consumption

purposes (food, clothing etc), followed by medical treatment and non-agricultural purchases. Lending

for consumption purposes was higher in deep Haor areas than in moderate Haor areas. Very few

households reported taking out a loan for productive purposes such as the purchase of agricultural

tools/equipment, purchase of agricultural inputs, land leasing or mortgaging or livestock purchases.

Table 15: Reasons for taking out a loan, by Haor type

Reason for Loan

(multiple response)

Deep Haor Moderate Haor Overall

N % of

Responses N

% of

Responses N

% of

Responses

Purchase agricultural tools/equipment 18 2.4 15 2.2 33 2.3

Purchase agricultural inputs 60 8.1 76 11.0 136 9.5

Land leasing or mortgaging 27 3.6 8 1.2 35 2.4

Livestock purchases 11 1.5 10 1.5 21 1.5

Non-agricultural purchases 249 33.6 158 23.0 407 28.5

Medical treatment/medicine 436 58.8 358 52.0 794 55.5

Consumption (food, clothes, etc.) 975 131.4 819 119.0 1794 125.5

Education 31 4.2 16 2.3 47 3.3

House repair/construction 115 15.5 96 14.0 211 14.8

Marriage/social 43 5.8 22 3.2 65 4.5

Total 1965 264.8 1578 229.4 3543 247.8


Table 16: Reasons for taking out a loan, by District

Reason for Loan

(multiple response)


N % of

Responses N

% of

Responses N

% of

Responses

Purchase agricultural tools/equipment 16 3.3 13 2.8 4 0.8

Purchase agricultural inputs 56 11.6 26 5.7 54 11.0

Land leasing or mortgaging 17 3.5 11 2.4 7 1.4

Livestock purchases 9 1.9 9 2.0 3 0.6

Non-agricultural purchases 138 28.6 125 27.3 144 29.4

Medical treatment/medicine 330 68.5 252 55.0 212 43.3

Consumption (food, clothes, etc.) 604 125.3 541 118.1 649 132.4

Education 14 2.9 21 4.6 12 2.4

House repair/construction 59 12.2 73 15.9 79 16.1

Marriage/social 22 4.6 20 4.4 23 4.7

Total 1265 262.4 1091 238.2 1187 242.2

Qualitative data clearly showed the impact that high interest rates are having on households. These

high rates perpetuate the household debt cycle, which leads to use of loans for day-to-day

consumption purposes and prevents productive investments – as can be seen from the very high debt

burden in Table 14 and loan uses described in Tables 15 and 16.

Many community members specifically mentioned the high interest rates of NGOs and even called it

exploitative. Households that have no choice but to take loans at these high interest rates, often end

up taking additional loans and selling land to pay their weekly installments. It was stated that the credit

provided by NGOs is not suitable for the needs of ultra poor, who instead require soft or even interest-

free loans. Soft loans are preferred over current NGO credit arrangements that require weekly

installments, which are hard to maintain.

It was also mentioned that there is an important gender dynamic to NGO credit. Although loans are

given to women, decisions regarding loan use and repayment are frequently made by men who are

not properly trained to optimize business opportunities or manage household income/expenditures.

5.7 Assets

Assets are an integral component of livelihoods, and the accumulation and sale of assets reflect

important economic characteristics of households. Each respondent was questioned about ownership

of fifty-four different assets, divided into six asset classes – domestic, productive, land, animal,

resource and financial. Asset ownership is a powerful economic indicator to monitor over time as it

reflects household-level decision-making regarding where to invest additional resources.

Table 17 shows results for 16 domestic assets. Relatively few assets differed significantly by Haor

type, but there was greater ownership of cupboards, lanterns and mobile phones in deep Haor, and


greater ownership of showcases in moderate Haor. Kishoreganj had significantly greater ownership of

six assets, including both gold and silver jewelry, suggesting that household domestic asset

ownership is greater in this District. In contrast, domestic asset ownership is least in Netrokona.

Table 17: Average number of domestic assets owned, by District and Haor type

Domestic assets District Haor Type Total


N 628 634 630 947 945 1892

Chairs 0.27 0.35 0.48b 0.38 0.35 0.37

Beds 1.07 1.01 0.91b 0.98 1.01 1.00

Cupboards 0.16c 0.03 0.07 0.10

b 0.07 0.09

Tables 0.13b 0.15

b 0.19

b 0.16 0.15 0.16

Showcases 0.17c 0.07 0.09 0.10 0.13

b 0.11

Dressing tables 0.01 0.00 0.00 0.00 0.01 0.00

Watches 0.08 0.06 0.07 0.08 0.06 0.07

Clocks 0.04 0.02 0.03 0.03 0.03 0.03

Lanterns 0.80c 0.38 0.48 0.64

c 0.47 0.55

Radios 0.01 0.01 0.03 0.02 0.01 0.02

TVs 0.02 0.00 0.01 0.02 0.00 0.01

Cassette players 0.01 0.00 0.01 0.01 0.01 0.01

Electric fans 0.04b 0.02 0.00 0.03 0.02 0.02

Mobile phones 0.13 0.10 0.13 0.14b 0.10 0.12

Gold jewelry (ana) 1.05b 0.77 0.82 0.88 0.88 0.88

Silver jewelry (ana) 7.50c 4.00 4.34 5.47 5.08 5.27

Letters denote significant differences among Districts or between Haor types for a given asset. Significance levels for comparisons: a = .10; b = .05; c = .00

Productive assets include various types of transportation and livelihood equipment, and are an

important indicator of a household‟s investment in livelihood opportunities. Overall the ownership of

productive assets in the survey population was very low. Generally, far less than one out of ten

households owned any of the productive assets (Table 18). Productive assets related to fishing (boats

and nets) were significantly more common in deep Haor, while boats and bicycles were more

commonly owned in Kishoreganj. Aside from these differences there was little differentiation among

Districts.


Table 18: Average number of productive assets owned, by District and Haor type

Productive assets District Haor Type Total


N 628 634 630 947 945 1892

Boat 0.04c 0.08 0.10 0.09

c 0.05 0.07

Motorcycle 0.00 0.00 0.00 0.00 0.00 0.00

Rickshaw/van 0.02 0.01 0.01 0.01 0.02 0.02

Bicycle 0.04b 0.01 0.00 0.00 0.04

b 0.02

Sewing machine 0.01 0.00 0.00 0.01 0.01 0.01

Shallow/hand-tube well 0.07 0.01 0.00 0.02 0.04a 0.03

Power tiller 0.00 0.00 0.00 0.00 0.00 0.00

Paddle thresher 0.01 0.00 0.00 0.00 0.00 0.00

Spray machine 0.00 0.00 0.00 0.00 0.00 0.00

Plough 0.01 0.02 0.02 0.01 0.02 0.02

Fishing nets 0.21 0.32 0.45 0.48c 0.18 0.33

Other 0.03 0.25 0.07 0.20 0.03 0.12


Land assets, measured in decimals, are provided in Table 19. Land ownership varies greatly among

sampled households, so differences between Haor types or among Districts have to also be large to

be significantly different. Between Haor types, only homestead land differs significantly and ownership

is greater in moderate Haor than in deep Haor (3.44 and 2.32 decimals per household, respectively).

Significantly less homestead land is owned in Kishoreganj compared to Netrokona and Sunamganj.

Netrokona has more land leased in, while Kishoreganj has more land leased out. Netrokona

households also averaged 1.48 decimals of „other‟ land thought to be different from the six categories

of land pre-coded in the survey. Overall ownership of agricultural land is highest and averaged 4.05

decimals per household, or less than 1/20th of one acre.

Table 19: Average number of land assets owned, by District and Haor type

Land assets (in decimals*)



N 628 634 630 947 945 1892

Homestead land 1.98c 3.49 3.16 2.32 3.44

c 2.88

Agricultural land 3.49 4.94 3.73 3.53 4.58 4.05

Land lease - IN 3.21 1.72b 4.85 3.07 3.45 3.26

Land lease - OUT 2.23b 0.26 0.92 1.51 0.76 2.80

Haor land 0.14 0.33 0.44 0.36 0.25 0.30

Pond/ditch 0.02 0.06 0.15 0.03 0.13 0.08

Other land 0.04 1.48c 0.11 0.54 0.55 0.55


*100 decimals is equal to 1 acre


Chickens were the most common animal asset owned, averaging 1.48 per household (Table 20).

Ownership of chickens was also significantly higher in Kishoreganj where it averaged 1.70 per

household. Ducks were the second most common animal asset and averaged 0.68 per household,

but were significantly more common in deep Haor, and significantly less common in Sunamganj than

in Kishoreganj or Netrokona. Cows were the third most commonly owned animal asset but were least

common in Kishoreganj.

Table 20: Average number of animal assets owned, by District and Haor type

Animal assets District Haor Type Total


N 628 634 630 947 945 1892

Cows 0.20b 0.27 0.30 0.25 0.27 0.26

Buffalo 0.00 0.00 0.00 0.00 0.00 0.00

Goats 0.12 0.08 0.09 0.09 0.10 0.10

Sheep 0.00 0.00 0.02 0.01 0.01 0.01

Chickens 1.70 b

1.43 1.30 1.53 1.42 1.48

Ducks 0.81 0.73 0.49 b

0.80b 0.56 0.68

Pigs 0.00 0.00 0.00 0.00 0.00 0.00

Other 0.02 0.00 0.00 0.01 0.01 0.01


Ownership of some resource assets, which included timber and fruit trees, bamboo, and medicinal

plants (mostly for use against cough and fever, used in lieu of adequate health care service), was

fairly common in surveyed households. Bamboo trees were the most commonly owned resource

asset and averaged just over three trees per household, but were significantly more common in

moderate Haor areas, and significantly less common in Kishoreganj, where ownership of timber and

fruit trees was also significantly less compared to the two other Districts.

Table 21: Average number of resource assets owned, by District and Haor type

Resource Assets District Haor Type Total


N 628 634 630 947 945 1892

Timber trees 0.42 0.63 b

0.99 c 0.65 0.71 0.68

Fruit trees 0.87 c 1.40 1.30 1.16 1.23 1.20

Bamboo trees 1.65 b

3.55 3.87 1.74 4.34 c 3.04

Medicinal plants 0.15 0.02 0.03 0.01 0.12 0.07

Others 0.01 0.06 0.15 b

0.10 0.15 0.08


The last asset category was financial assets and results are shown in Table 22. Cash with NGOs

averaged 495 Taka per household and was the most common financial asset measured. Households

in deep Haor had significantly more cash with NGOs (587 Taka compared to 403 Taka in moderate

Haor), but significantly less loans or credits given to others. Very few households had any cash at

banks but there was a slightly higher amount in Kishoreganj, where cash on hand was also

significantly higher. Cash with NGOs was highest in Sunamganj.


Table 22: Average financial assets owned, in Taka, by District and Haor type

Financial Assets District Haor Type Total


N 628 634 630 947 945 1892

Cash at bank 2.39 b

0.00 7.32 5.91 a 0.54 3.23

Cash w/ NGO 416.86 456.53 612.67 a 587.14 403.38

a 495.36

Insurance 62.49 57.82 45.64 38.80 71.86 55.32

Cash on hand 203.40 c 90.89 130.85 121.79 161.33 141.54

Loan/credit to others 187.42 c 75.21 2.96 56.90 119.96

a 88.40

Other 24.20 59.76 14.23 20.85 44.78 32.80


Picture 6: Jack fruit trees


5.8 Housing characteristics

Table 23 shows the housing characteristics of households. The majority of all houses have floors

made of mud (99.9% and 0.1% made of brick), walls made of straw/jute or corrugated iron

sheets/tin/wood, and roofs made of corrugated iron. Less than 1% of all houses have brick walls and

only 1 house in Kishoreganj had a concrete roof. Total square feet of living space is 175ft and the

average number of rooms is 2 across all strata. About 10 percent of households share their living

space with their cattle, mostly for safety of the animals in absence of more than one housing structure.

When comparing across districts, the proportion of houses using corrugated iron/tin/wood building

materials is significantly higher than in the other districts. The proportion of houses with mud walls is

significantly higher in Sunamganj. Total living area was significantly higher in Sunamganj and sharing

of living space with cattle was significantly higher in Netrokona than in the other districts. When

comparing across Haor region, the proportion of houses with mud or straw/jute walls was significantly

higher in moderate Haor than in deep Haor.

Table 23: Housing characteristics, by District and Haor type

House characteristics District Haor Type Total


N 196 166 203 220 242 565

Wall

Material

Brick 0.3 0.5 0.6 0.4 0.6 0.5

CI sheet/tin/wood 56.4c 24.4 19.8 34.1 32.9 33.5

Mud 0.5 5.5 16.3 c 4.2 10.7

c 7.5

Bamboo 5.9 7.9 11.7 8.7 8.4 8.5

Straw/jute/etc. 36.9 b

61.7 51.4 52.7 47.4 a 50.1

Roof

Material

CI sheet/tin 91.4 b

83.1 83.7 87.1 85.0 86.0

Straw/jute/etc. 7.8 b

16.2 15.7 12.4 14.1 13.3

Other 0.7 0.6 0.7 0.5 0.8 0.7

Total area (square feet) 173.6 170.2 179.7 a 172.6 176.3 174.5

Average number of rooms 2.0 2.1 2.0 2.0 2.0 2.0

Share with cattle (%) 9.1 11.8 b

7.6 8.7 10.4 9.5




Picture 7: Housing made of jute and straw

Picture 8: Housing made with corrugated iron


6.0 FOOD SECURITY

6.1 Food consumption score

The Food Consumption Score (FCS) is widely used now by the World Food Program and endorsed

by FANTA9 as a measure of diet diversity and quality, and is derived by weighting various food groups

based on their protein value and assigning a score for each food group consumed by the household

during the recall period. Points for the FSUP baseline study are assigned as follows:

Table 24: Food consumption score

Food Group Score

Cereals: 2 points

Pumpkin, squash carrots, sweet potatoes : 2 points

White potatoes, white yams: 2 points

Dark green leafy vegetables: 3 points

Other vegetables: 1 point

Papayas, mangoes: 3 points

Other fruits: 1 point

Meat: 4 points

Eggs: 4 points

Fresh or dried fish/shellfish: 4 points

Legumes/pulses: 3 points

Milk/Dairy: 4 points

Oil/fats: 0.5 points

Sugar/honey: 0.5 points

Total Possible: 34.0 points

The thresholds used for the FSUP study are: 0-4 is poor, 4-8 is borderline food security and 9+ is

acceptable food security. These are modified from the World Food Program‟s Comprehensive Food

Security and Vulnerability Assessment Guidelines and are specific for the FSUP study. Future

measurements of the FCS within FSUP should use the same food group weights and the same

thresholds.

Table 25 shows the responses organized by thresholds. The highest proportion of sampled

households with acceptable FCS values is located in Kishoreganj, and the highest proportion of

sampled households with poor FCS values is located in Sunamganj. When comparing among deep

and moderate Haor areas, the highest proportion of sampled households with acceptable FCS is

located in deep Haor areas, and the highest proportion of households with poor and borderline FCS

values is located in moderate Haor areas. One of the reasons for the higher score in deep Haor areas

is the higher consumption of fish (which scores 4 points) in those areas. However, it is also important

to note that the scores provided in Table 25 relate to the timing of the data collection, particularly the

difference in food consumption in peak and lean seasons, as will be elaborated on below.

9 Food Aid and Nutritional Technical Assistance Project of USAID.


Table 25: Proportion of sampled households by FCS threshold values

FCS Thresholds District Haor Type Total


Poor (0-5) 8.3 16.9 23.5 14.1 18.3 16.2%

Borderline (6-8) 23.4 36.1 34.8 30.0 32.9 31.5%

Acceptable (9+) 68.3 47.0 41.7 55.9 48.8 52.3%

The mean FCS values are shown in Figure 4. Overall, FCS values are higher than 8 and can be

considered acceptable. When comparing among Haor areas, the deep Haor area has a significantly

higher FCS value than the moderate Haor area. When comparing among districts, Kishoreganj has a

significantly higher FCS value than the other two districts; with Sunamganj having the lowest FCS

value overall.

Figure 4: Mean FCS values, by District and Haor type

When analyzing by food group, the responses show that almost all household members (99%)

consumed „cereals‟; mostly rice and in few cases wheat flour/puffed rice. The second most frequent

(71%) food group consumed was „fresh and dried fish‟. This high level of fish consumption can be

partly attributed to the timing of data collection, which was undertaken at a time when the water levels

were dropping, and fish catch and drying was high. The third most frequent food group was „other

vegetables‟ (55%). This can be explained by the fact that data collection was undertaken in the

harvesting season of various types of indigenous vegetables, which were then available at relatively

lower prices. The fourth most frequent food group (54%) was „white potatoes and white yams‟,

followed by „oil/fats‟ (42%), „dark green leafy vegetables‟ (30%) and „pumpkin, carrots, squash, or

sweet potatoes‟ (20%). The remaining food groups were all < 5%.


6.2 Food intake

Figure 5a shows the households reporting enough food by month. It is important to focus on the

overall shape of the curve here, as there will be some respondent error in terms of their recall relative

to the mid-points between the months, as shown in the figure below.

The figure shows two distinct lean periods in terms of insufficient food. The first lean period is from

April to June, with the leanest period in April-May (13%), the month of Baishak in the Bengali

calendar. The second lean period is from November to February with the leanest period in Dec-Jan

(12%), the month of Payush in the Bengali calendar.

In both periods, almost 90% of households in the sample report insufficient food. The recovery from

the insufficient food period in April to June is notable longer than for the second lean period - with

another smaller decrease in August-September (31%) before reaching a peak at 63 percent in Oct-

Nov. The highest number of households report sufficient food in March-April (82%), with a very sharp

decrease between the Bengali months of Chaitra and Baishak.

It is important to note that the lean period shown here slightly differs from lean seasons in other food

insecure areas in Bangladesh, because the harvesting season of the boro rice in Haor areas takes

place slightly earlier than in other areas.

Figure 5a: Proportion of households reporting enough food, by month and Haor type (1)

Figure 5b below shows that in the period June-July to Oct-Nov, the proportion of households reporting

enough food is lower in deep Haor areas than in moderate Haor areas. The figure also shows a

higher number of households in deep Haor areas reporting sufficient food in the period January to

May. This matches the FCS value findings, which show that households in the deep Haor areas have

a significantly higher FCS value for the period in which the data was collected: January to February.


Figure 5b: Proportion of households reporting enough food, by month and Haor type (2)

Figure 6 shows the mean number of lean months, by District and Haor type. Overall, the mean

number of lean months is 4.3. When comparing across Districts, there are significant differences

among all Districts, whereby Sunamganj has the highest mean number of lean months and

Kishoreganj has the lowest number. When comparing across Haor types, the number of lean months

in moderate Haor areas is significantly higher than in deep Haor areas.

Figure 6: Mean number of lean months, by District and Haor type


Figure 7 compares frequency of three square meals taken among Districts. Overall, the mean value for

households that take 3 meals per day „most of the time‟ is 14%. The mean values for „most of the time‟ and often

combined is 56.3 %. Households with the highest frequency of taking three square meals per day are

located in Kishoreganj. Households with the lowest frequency of taking three square meals per day

are located in Sunamganj.

Figure 7: Frequency of three 'square meals' taken a day in 12 months, by District

Figure 8 compares frequency of three square meals taken among deep and moderate Haor areas.

Households with the highest frequency of taking three square meals per day are located in deep Haor

areas. Households with the lowest frequency of taking three square meals per day are located in

moderate Haor areas.

Figure 8: Frequency of three 'square meals' taken a day in 12 months, by Haor type


6.3 Coping strategies

Households were asked to indicate how they dealt with food insecurity. Questions D4 - D11 with a 12

month recall asked respondents whether households had to replace rice with grains, skip meals,

reduce food intake, run out of food, worry about where food would come from, purchase rice in bulk to

use it sparingly, purchase food on credit and/or borrow food/take donated food. Response categories

for these questions ranged from „Most of the time‟ to „Never‟.

Presenting all the data in a table would make meaningful interpretation difficult so a coping index was

created for questions. The index was computed by giving „Most of the time‟ a value of 5, „Often‟ a

value of 4, etc. The highest possible score would be 40. A high score indicates that households in

specified areas avail themselves of a broad range of coping strategies to deal with food insecurity; the

higher the index value is - the higher the assumed stress on households. This Index is suggested as a

useful monitoring tool for FSUP-H.

Overall, the coping index score of almost 24 indicates a moderately-high level of stress on households

due to food insecurity. Comparison across Haor types shows that the coping index score is

significantly higher in deep Haor areas than in moderate Haor areas. Coping index scores in

Kishoreganj and Sunamganj are statistically the same but Netrokona shows a significantly lower score

than the other two Districts. „Bulk purchases of rice‟, „running out of food‟ and „reducing personal food

intake‟ were the top 3 coping strategies, both overall and when disaggregated by district and Haor

type10

.

Figure 9: ‘Coping Index’ for households, by District and Haor type

.

10

Excluding question D8; although indicative of household stress and, therefore, included in the Index, worrying is not a meaningful coping strategy


6.4 Trend analysis Table 26 shows the results of a trend analysis/seasonal calendar undertaken in 4 villages:

1) Chorpara village, Itna Upazilla, Kishoreganj (deep Haor),

2) Sutarpara village, Sutarpara Union, Karimganj Upazilla, Kishoreganj (moderate Haor),

3) Boali village, Khaliajhury Upazilla, Netrokona (deep hoar),

4) Khurshimul village, Mohanganj Upazilla, Netrokona (moderate Hoar).

As part of this qualitative exercise, community members were asked to describe selected occurrences

and activities during the 12 months of the year, and to score the intensity of occurrences/activities. A

higher number indicates higher intensity as perceived by community members. In Kishoreganj, scores

were assigned on a scale from 0-10; in Netrokona on a scale from 0-5, which were subsequently

multiplied by 2 for the purpose of this analysis. As a result, lower intensity in the 0-1 range on the 10-

scale may not be properly reflected for the Netrokona villages.

Table 26 : Seasonal calendar Months Apr-may May-

Jun Jun-Jul

Jul-Aug

Aug-Sep

Sep-Oct

Oct-Nov

Nov-Dec

Dec-Jan Jan-Feb

Feb-Mar

Mar-Apr

Bangla month

Baishak Jaisti Ashar Sravon Bhadra Ashin Kartic Agrahayan

Payush Magh Falgun Chaitra

Rainfall

Village 1 ♦♦ ♦♦♦ ♦♦♦♦ ♦♦♦♦ ♦♦ ♦

Village 2 ♦♦♦ ♦♦♦ ♦♦♦♦♦♦♦ ♦♦♦♦♦♦♦ ♦♦♦♦ ♦♦ ♦♦ ♦♦ ♦♦♦

Village 3 ♦♦

♦♦

♦♦♦♦♦♦♦♦♦♦

♦♦♦♦♦♦♦♦♦♦

♦♦♦♦♦♦♦♦ ♦♦♦♦♦♦

♦♦

♦♦

Village 4 ♦♦ ♦♦ ♦♦♦♦♦♦♦♦♦♦

♦♦♦♦♦♦♦♦♦♦

♦♦♦♦♦♦♦♦ ♦♦♦♦♦♦ ♦♦ ♦♦

Food crisis

Village 1 ♦♦ ♦♦♦♦♦ ♦♦♦

♦♦♦♦♦♦

Village 2 ♦♦ ♦♦♦ ♦♦♦ ♦♦♦♦♦ ♦♦♦♦♦♦ ♦♦♦♦♦♦♦

Village 3 ♦♦♦♦♦♦♦♦♦♦

♦♦♦♦♦♦♦♦♦♦

♦♦♦♦♦♦♦♦♦♦ ♦♦♦♦♦♦♦♦ ♦♦♦♦♦♦♦♦♦♦

♦♦♦♦

Village 4 ♦♦♦♦♦♦ ♦♦♦♦ ♦♦♦♦♦♦♦♦♦♦

♦♦♦♦♦♦♦♦♦♦

Disease

Village 1 ♦ ♦♦ ♦♦ ♦♦♦♦♦♦ ♦♦ ♦♦

Village 2 ♦♦♦♦♦♦ ♦♦ ♦♦ ♦♦ ♦♦♦ ♦♦ ♦♦♦♦♦ ♦♦♦♦ ♦♦♦♦♦ ♦♦♦♦♦

Village 3 ♦♦♦♦♦♦♦♦♦♦

♦♦♦♦

♦♦♦♦

♦♦♦♦♦♦

Village 4 ♦♦♦♦ ♦♦♦♦ ♦♦♦♦♦♦♦♦ ♦♦♦♦ ♦♦♦♦

Migration

Village 1 ♦ ♦ ♦ ♦♦♦♦ ♦♦

Village 2 ♦♦♦♦♦♦♦ ♦♦♦♦♦♦♦ ♦♦♦♦♦♦♦ ♦♦♦♦♦♦♦ ♦♦♦♦♦♦♦ ♦♦♦♦♦♦♦

Village 3 ♦♦♦♦♦♦ ♦♦♦♦♦♦♦♦♦♦

♦♦♦♦♦♦♦♦

♦♦♦♦

♦♦♦♦♦♦♦♦

Village 4 ♦♦ ♦♦♦♦♦♦ ♦♦♦♦♦♦ ♦♦♦♦ ♦♦♦♦♦♦

Male level of work

Village 1 ♦♦♦♦♦ ♦♦♦♦♦

♦♦ ♦ ♦ ♦♦ ♦♦♦♦ ♦♦♦♦♦ ♦♦♦♦♦ ♦♦

♦♦♦♦♦ ♦

Village 2 ♦♦♦♦♦♦♦♦♦♦ ♦♦♦♦♦♦♦

♦♦♦ ♦♦♦ ♦♦♦ ♦♦♦♦♦ ♦♦♦♦♦ ♦♦♦♦ ♦♦♦♦♦♦♦♦♦♦ ♦♦♦♦♦♦♦♦ ♦♦ ♦♦

Village 3 ♦♦♦♦♦♦♦♦♦♦ ♦♦♦♦♦♦♦♦

♦♦

♦♦

♦♦

♦♦ ♦♦

Village 4 ♦♦♦♦♦♦♦♦♦♦ ♦♦♦♦♦♦♦♦

♦♦♦♦ ♦♦♦♦ ♦♦♦♦♦♦ ♦♦♦♦♦♦ ♦♦♦♦♦♦ ♦♦♦♦♦♦♦♦ ♦♦♦♦♦♦ ♦♦♦♦

Female level of work

Village 1 ♦♦♦♦♦♦ ♦♦ ♦ ♦

Village 2 ♦♦♦♦♦♦♦♦♦♦ ♦♦♦♦♦♦♦♦

♦♦ ♦ ♦ ♦♦

Village 3 ♦♦♦♦♦♦♦♦♦♦ ♦♦♦♦♦♦♦♦

♦♦

♦♦

♦♦

♦♦ ♦♦

Village 4 ♦♦♦♦♦♦♦♦♦♦ ♦♦♦♦♦♦♦♦

♦♦ ♦♦ ♦♦ ♦♦♦♦♦♦ ♦♦♦♦ ♦♦ ♦♦

The descriptions provided by community members for the selected activities/occurrences over a 12-

month period match quite closely across the 4 villages, and are reported below based on qualitative

data collected.


The lean season ended in the first half or April and April-May is a busy period for households; during

this time they earn most of the income to pay back informal loans taken during the lean season, so-

called „logni‟. Men undertake agricultural day labor and are very busy completing the rice harvest

before the monsoon begins. Other work for men includes reaping, thrashing, and straw drying.

Women undertake the collection of rice from the fields, rice winnowing, boiling, drying and storing

(including making pulp rice – muri); and cow paddy/straw drying and storing. Children often help with

the collection of rice. At the same time, women are also harvesting ground nuts, sweet potatoes and

are making cow dung coils as fuel for cooking – in addition to their regular household chores. In Boali

village, community members stated that men get paid 10-12 mounds of rice for harvesting per season

and women get 3-4 mounds of rice, one saree and 2 meals of food at end of the season. During this

period, storms start increasing in intensity and frequency.

Picture 9: Women supporting household income through produce sales

During the period May- June, men continue the reaping and harvesting of rice and women continue

collection of rice from the fields, rice winnowing, boiling, drying and storing, and drying/storing of cow

dung and straw. At this time, men also do earth work and homestead raising to protect their homes,

and repair their houses, boats and nets. In this month, fishermen in Chorpara village take dadon,

conditional informal loans. This is a period of heavy rainfall. Many people suffer from colds, fevers,

coughs and influenza.

In the period June-July, men are mainly involved with fishing and the ongoing reparation of their

fishing nets and boats. In Sutarpara, women also help with these reparations. Fish catches are not

good and because of high waves in this period they cannot go out onto the water every day; fishing


only earns the men about 150-200 Taka per day. To meet the income shortfall, many men take „logni‟

and „dadon‟ from local Mohajonee for 5-6 months at average interest rates for 50%. Some men also

undertake short migration to Sylhet and Dhaka for contract labor on earth work and rickshaw pulling.

Men spend a lot of time playing cards and gossiping while women sew Kantha and make bamboo

handicrafts. In this period, community members report heavy rainfall and many suffer fevers,

headaches and influenza.

In July-August, there is sufficient fish to catch and men intensify their fishing in rivers and Haors.

However, Chorpara village reports that there is less fish than before. Fishing earns the men on

average 250-300 Taka per day. Women primarily sew Kantha. There is heavy rainfall in this period

and community members suffer from influenza, fevers and coughs.

In the period August-September, most of the men continue fishing the rivers and Haors. In Khurshimul

village, men also work on separating jute fibres. If men observe that there are sufficient fish to be

caught, they take „dadon‟ and try their fortune – with the potential of earning 250-300 Taka per day. If

they observe limited number of fish, men migrate to Sylhet or Dhaka for contract labor. During the

period, women are not involved in income-generating work. The rains are decreasing – community

members report diarrhea and dysentery.

September-October marks the start of the peak fishing season and men are very busy. Income from

fishing is reported to be same as previous months: 250-300 Taka per day. Women do no income-

related work. Diarrhea and dysentery cases are increasing. There are also some cases of jaundice

reported.

Picture 10: Men fishing in the peak season


In October-November, the fishing season is winding down. Men remain busy fishing but fish

availability decreases; women do not do income-related work. Average income from fishing is

reported as lower than previous months: 200-250 Taka per day. The ultra poor people in the villages

migrate to Dhaka, Chittagong, Sylhet, Bhairab, Ashuganj, Aliganj, and Volaganj to do contract labor.

To meet transportation and other expenses, villagers from Boali village report having to take 6-month

loans from money lenders at 200% annual interest rates. In some cases, the entire household

migrates to do work such as brick making – earning 60-70 Taka per day / household member. This

period marks the beginning of a food crisis and the ultra poor reduce start reducing food intake –

eating two rice meals/day and one roti meal. There is increased incidence of water-borne diseases

such as diarrhea, dysentery and jaundice. Fever, coughs and malaria are also reported. As a result of

water levels in the Haor dropping in this period, open latrines are becoming separated from water

bodies.

Picture 11: Non-agricultural day labor

In the period November-December, there are very limited opportunities for income-generating work in

the villages. Many households migrate to do contract labor such as brick making. Men and women

that remain behind start preparing paddy seed beds and planting seeds as agricultural day. Many

ultra poor are forced to take „logni‟. The villages in Netrokona report no income-related activities at all.

Overall, the food crisis continues and the ultra poor reduce continue to reduce food intake – eating

two rice meals/day and one roti meal.

In the period December-January, the main source of income is agricultural day labor, if available. Men

are becoming increasingly busy transplanting rice, sowing sweet potato and groundnuts. The women

work on uprooting rice seedlings and preparing them for transplantation. Community members in


Chorpara village also report that some ultra –poor women catch fish. The villages in Netrokona report

no income-related activities at all. The food crisis continues and food intake is reduced to one or two

meals per day. Many people suffer from fevers and colds, and community members in Sutarpara

report pneumonia among children.

In the period January-February, the main source of income is still agricultural day labor, if available.

Men continue rice transplantation, work on developing paddy irrigation, and weeding; earning them

about 150 Taka/day. Women harvest potatoes and receive 5kg out of every 40kg that they harvest as

payment. In Chorpara village some women continue to catch fish. Some households migrate to areas

where they can sell labor. The food crisis continues and most of the ultra poor households do not take

more than 2 meals per day.

In the period February-March, there are no opportunities for men to earn income in the villages. Most

of the households migrate to areas where they can sell their labor. Women in Netrokona are reported

to harvest groundnuts, chili, and sweet potato, and to do earth work. Food rationing to a maximum of

two meals per day continues and some households have to go without any meals on some days.

Community members report that there are an increasing number of storms in this period.

In the first half of March-April, there are still no income-generating opportunities with the exception of

some earth work that men are involved in. Migration to sell day labor remains common. The food

crisis continues for another 2-3 weeks and reduced food intake during this time is still very common.

There are heavy rainfall and storms. April marks the end of the lean season.

Picture 12: Non-agricultural day labor - mat making

The patterns shown in table 26 and the accompanying qualitative descriptions of the different periods

match the peak and lean periods shown in figures 5 and 6. The most severe food crisis occurs in the

period October to February, after which there is a relatively quick recovery in April when the rice

harvest starts. Adjusting meals and informal lending are the most common coping strategies in lean

periods, which correspond with tables 11 and 12 in Section 5, and the top 3 coping strategies in

described in Section 6.3.


7.0 WATER AND SANITATION

7.1 Drinking, cooking and washing water sources

Table 27 shows the drinking water sources by District and Haor region. Hand tube wells are the most

common water source followed by shallow tube wells and deep tube wells. Overall, 97% of

households depend on the various types of tube wells for drinking water. Almost no households draw

drinking water from open water sources such as ring wells, ponds and rivers/canals.

When comparing across Haor types, significantly more households in deep Haor areas use hand tube

wells than in moderate Haor areas. In turn, in moderate Haor areas, significantly more households

use shallow tube wells that in deep Haor areas.

When comparing across Districts, the proportion of households using hand tube wells in Kishoreganj

is significantly higher than in the other two districts. The proportion of households using shallow tube

wells is significantly higher in Netrokona; the proportion of households using deep tube wells is

significantly higher in Sunamganj.

Table 27: Drinking water sources, by District and Haor type

Drinking Water

(% of HHs)



N 628 634 630 947 945 1892

Hand tube well 72.8c 59.9 55.7 64.7

c 60.8 62.8%

Tara pump 0.5 0.2 0.0 0.1 0.3 0.2%

Deep tube well 7.2 8.5 19.5c 12.7 10.8 11.7%

Shallow tube well 19.1 29.5c 20.2 20.4 25.5

b 22.9%

Ring well/ indara 0.3 0.0 1.6 0.1 1.2 0.6%

Pond 0.2 0.0 1.0 0.1 0.6 0.4%

River/canal 0.0 1.9 2.1 1.9 0.7 1.3%

Letters denote significant differences among Districts or between Haor types for a given water

source. Significance levels for comparisons: a = .10; b = .05; c = .00

Table 28 shows the cooking water sources by District and Haor region. Hand tube wells are the most

common water source followed by rivers/canals and shallow tube wells. Deep tube wells and ponds

are the next most common water sources for cooking. Almost no households draw cooking water from

tara pumps and ring wells.

When comparing across Haor types, significantly more households in moderate Haor areas use hand

and shallow tube wells than in deep Haor areas. In turn, in deep Haor areas, significantly more

households use river/canal water for cooking that in moderate Haor areas.

When comparing across Districts, the proportion of households using hand tube wells in Kishoreganj

is significantly higher than in the other two districts. The proportion of households using shallow tube

wells is significantly higher in Netrokona; the proportion of households using deep tube wells is

significantly higher in Sunamganj.


Table 28: Cooking water sources, by District and Haor type

Cooking Water

(% of HHs)



N 628 634 630 947 945 1892

Hand tube well 56.7 c 35.5 32.1 34.7 48.0

c 41.4%

Tara pump 0.5 0.2 0.0 0.0 0.3 0.2%

Deep tube well 4.3 5.8 12.9 c 7.3 8.0 7.7%

Shallow tube well 16.9 20.7 a 14.4 12.2 22.4

c 17.3%

Ring well/ indara 0.3 0.0 2.1 0.4 1.2 0.8%

Pond 1.6 11.7 8.1 7.2 7.1 7.1%

River/canal 19.7 26.2 30.5 38.0 b

12.9 25.5%

Letters denote significant differences among Districts or between Haor types for a given water source.


Table 29 shows the washing water sources by District and Haor region. Most households reported

open water sources for washing. River/canals are the most common water source followed by hand

tube wells and ponds. Almost no households use water from tara pumps, ring wells and deep tube

wells for washing.

Picture 13: Woman uses hand tube well as the water source for washing


When comparing across Haor types, significantly more households in moderate Haor areas use hand,

deep and shallow tube wells, and ponds than in deep Haor areas. In turn, in deep Haor areas,

significantly more households use river/canal water for washing that in moderate Haor areas –

presumably, due the almost year round access to this water source. When comparing across Districts,

the proportion of households using hand tube wells in Kishoreganj is significantly higher than in the

other two districts. The proportion of households using ponds is significantly higher in Netrokona; the

proportion of households using river/canals is significantly higher in Sunamganj.

Table 29: Washing water sources, by District and Haor type

Washing Water

(% of HHs)



N 628 634 630 947 945 1892

Hand tube well 45.1 c 17.5 17.1 22.1 31.0

c 26.5%

Tara pump 0.3 0.0 0.0 0.0 0.2 0.1%

Deep tube well 2.9 1.6 4.0 2.1 3.5 b

2.8%

Shallow tube well 12.6 14.5 7.3 7.8 15.1 c 11.5%

Ring well/ indara 0.3 0.2 1.7 0.0 1.1 0.6%

Pond 12.7 29.2 b

18.7 17.6 22.9 a 20.2%

River/canal 26.0 37.1 51.0 c 49.8

c 26.1 38.0%

Other 0.2 0.0 0.2 0.1 0.1 0.1%

Letters denote significant differences among Districts or between Haor types for a given water source.


Figure 10 shows the distances to various drinking water sources by District and Haor type. Overall,

the mean distance to water sources is slightly higher than 200 meters (205m). There is no significant

difference between the distance to drinking water in deep and moderate Haor areas. When comparing

across districts, the distance to drinking water is significantly higher in Sunamganj than in the other

districts.

Figure 10: Distances to sources of drinking water, by District and Haor type

(N=1892)


Figure 11 shows the distances to various cooking water sources by District and Haor type. Overall,

the mean distance to water sources is higher than 200 meters (216m), and slightly higher that the

distance to drinking water sources. There is no significant difference in the distance to cooking water

among districts. When comparing across Haor types, the distance to cooking water sources is

significantly higher in deep Haor areas.

Figure 11: Distances to sources of cooking water, by District and Haor type

(N=1892)

Figure 12 shows the distances to various washing water sources by District and Haor type. Overall,

the mean distance to water sources is less than 200 meters (185m). When comparing across districts,

there is no significant difference between distance to washing water sources in Kishoreganj and

Netrokona. However, the distance to washing water is significantly higher in Sunamganj than in the

other two districts. (dry season) When comparing across Haor types, the distance to washing water

sources is significantly higher in deep Haor areas.

Figure 12: Distances to sources of washing water, by District and Haor type

(N=1892)


7.2 Arsenic testing

Of the households that reported tube wells or tara pumps as a source for drinking, cooking or washing

water, 45% of households in Kishoreganj reported that the tube wells / tara pumps were tested for

arsenic, which is significantly lower than in Netrokona and Sunamganj where 53% of households

reported that the tube wells / tara pumps were tested for arsenic. When comparing across deep and

moderate Haor type, 55% of households in moderate Haor areas reported that the tube well/tara

pumps were tested for arsenic, versus a significantly lower 45% in moderate Haor areas.

Of the tube wells/tara pumps that were tested, 12%, 18% and 11% were found to contain arsenic in

Kishoreganj, Netrokona and Sunamganj, respectively. The 18% in Netrokona is significantly higher

than the percentages in the other two districts. When comparing across Haor types, the percentage of

tube wells/tara pumps that contained arsenic was significantly higher in deep Haor areas at 17% than

the 11% in moderate Haor areas.

Table 30: Tube wells/tara pumps tested for arsenic, by District and Haor type

District Haor Type Total N


Tested (%) Yes 45.2c 53.4 52.9 45.8

c 55.1 50.5 933

No 36.4 34.2 30.3 35.0 32.4 33.7 623

Do not know 18.4 12.4 16.8 19.2 12.5 15.8 293

Has arsenic (%) 12.7 18.7c 10.7 17.6

c 11.2 14.1 132

No arsenic (%) 87.3 81.3 89.3 82.4 88.8 85.9 814

Significance levels for comparisons among Districts/across Haor type: a = .10; b = .05; c = .00

7.3 Sanitation

The most common type of latrines used by adult men and women are ring slab/offset latrines (with the

seal broken) and hanging/open latrines, followed by uncovered pit latrines and then open defecation.

Overall, the use of hygienic latrines such as ring slab/offset latrines (with the seal intact), septic

latrines, covered pit latrines and locally adapted hygienic latrines is very low in the project area.

Qualitative data confirms that few households have a sanitary latrine. Shoes are seldom worn when

visiting the latrine.

When comparing across districts, the use of hanging/open latrines and ring slab/offset latrines (with

the seal broken) is significantly higher in Kishoreganj than in the other districts. There are no

significant differences across Haor types, and there are no significant differences in latrine use by

adult men and women.


Table 31: Types of latrines used by adult men and women, by District and Haor type

Latrine Type District Haor Type Total


N Men 562 504 459 782 743 1892

Women 581 529 467 810 767 1577

Ring-slab/offset

latrine (water seal)

Men 2.7 1.6 0.4 1.0 2.3 1.6

Women 2.8 1.5 0.4 0.9 2.5 1.6

Ring-slab/offset

Latrine (seal broken)

Men 49.3 c 30.6 29.8 34.9 39.7 37.2

Women 49.1 c 31.4 29.8 35.6 39.4 37.4

Pit latrine (covered)

Men 1.2 0.2 0.2 0.4 0.8 0.6

Women 1.5 0.2 0.2 0.4 1.0 0.7

Pit latrine

(uncovered)

Men 7.3 12.5 14.4 11.3 11.0 11.1

Women 7.2 13.0 14.6 11.0 11.7 11.4

Septic latrine

Men 0.4 0.0 0.2 0.0 0.4 0.2

Women 0.3 0.0 0.0 0.0 0.3 0.1

Hanging/open latrine

Men 34.5 b

52.0 48.8 47.3 41.7 44.6

Women 34.8 b

52.2 49.0 47.3 42.2 44.8

Locally adapted

hygienic latrine

Men 0.0 0.4 0.0 0.0 0.3 0.1

Women 0.2 0.4 0.0 0.1 0.3 0.2

Open defecation

Men 4.6 2.8 6.1 5.1 3.8 4.5

Women 4.1 1.3 6.0 4.8 2.6 3.7


Similar to adults, the most common types of latrines used by boys and girls 5-15 years of age are ring

slab/offset latrines (with the seal broken) and hanging/open latrines. For boys and girls, this is

followed by open defecation and then uncovered pit latrines – the opposite to adults. Overall, the use

of hygienic latrines such as ring slab/offset latrines (with the seal intact), septic latrines, covered pit

latrines and locally adapted hygienic latrines is very low.

Picture 14: Ring slab latrine


When comparing across districts, the use of ring slab/offset latrines (with the seal broken) by boys

and girls is significantly higher in Kishoreganj than in the other districts. For hanging/open latrines, the

use by girls is significantly lower in Kishoreganj than in the other districts, and significantly higher for

boys in Sunamganj than in the other districts. Open defecation by boys is significantly higher in

Netrokona than in the other districts. There are no significant differences across Haor types.

Table 32: Types of latrines used by boys and girls 5-15 years of age, by District and Haor type

Latrine Type District Haor Type Total


N Boys 348 282 325 482 473 955

Girls 366 280 305 476 475 951

Ring-slab/offset

latrine (water seal)

Boys 2.9 1.1 0.3 1.0 1.9 1.5

Girls 2.7 1.8 0.3 0.8 2.5 1.7

Ring-slab/offset

Latrine (seal broken)

Boys 45.1 c 24.5 27.4 30.7 35.3 33.0

Girls 46.7 c 24.6 29.8 33.4 36.2 34.8

Pit latrine (covered)

Boys 1.4 0.0 0.3 0.2 1.1 2.1

Girls 1.4 0.0 0.3 0.4 0.8 0.6

Pit latrine

(uncovered)

Boys 6.9 15.2 14.8 13.7 10.4 12.0

Girls 6.6 12.9 15.7 11.3 11.4 11.4

Septic latrine

Boys 0.6 0.0 0.3 0.0 0.6 0.3

Girls 0.5 0.0 0.0 0.0 0.4 0.2

Hanging/open latrine

Boys 33.0 35.8 46.2 a 39.4 37.2 38.3

Girls 32.8 a 43.6 44.9 41.0 38.7 39.9

Locally adapted

hygienic latrine

Boys 0.0 0.4 0.0 0.0 0.2 0.1

Girls 0.3 0.0 0.0 0.0 0.2 0.1

Open defecation

Boys 10.1 23.0 b

10.8 14.9 13.3 14.1

Girls 9.0 17.1 8.9 13.0 9.7 11.4


Enumerators were also asked to personally verify that the latrines used by respondents were

functioning and to describe their condition and cleanliness. Ninety percent of latrines observed were

found to be functional, all showed signs of use, 63 percent of latrines were considered relatively clean,

and for 55 percent latrines the surrounding area was considered clean. However, for these questions

there were only 40 responses/observations each, which is not in any way representative for the study

population. Reasons for this could be the distance from the interview location to the latrine area,

which may have been inconvenient for the enumerator to cover in the allotted interview time.


Picture 15: Open defecation facilities


8.0 HEALTH PRACTICES AND ILLNESS

8.1 Hand washing

Overall, the majority of respondents wash their hands before eating but less than half do so before

preparing food and only one-third wash their hands before feeding children. The majority of

respondents wash their hands after defecation but only one-third of respondents do so after cleaning

a baby‟s bottom. Qualitative data shows that hand washing with soap and ash after defecation is

uncommon.

When comparing across districts, hand-washing behavior before food preparation is significantly

higher in Netrokona than in the other districts. Hand washing after cleaning baby‟s bottoms and before

feeding children is significantly lower in Netrokona. There are no significant differences across Haor

types.

Table 33: Hand-washing behaviors among the FSUP baseline study households, by District and Haor

type (1)

When are hands washed… District Haor Type Total


N 628 634 630 947 945 1892

Before food preparation 49.7 59.0c 40.6 49.0 50.6 49.8

Before eating 91.9 87.4 94.4 91.0 91.4 91.2

Before feeding children 38.2 25.1 c 40.6 35.7 33.5 34.6

After defecation 96.5 88.0 98.4 94.6 94.0 94.3

After cleaning babies bottoms 43.3 22.1 c 38.9 35.1 34.4 34.7

Other 13.7 c 5.7 7.9 9.7 8.5 9.1

Do not wash hands 0.0 0.0 0.0 0.0 0.0 0.0


The use of ash or clay for hand washing is most common followed by use of only water. The use of

soap is least common, which is confirmed by qualitative data. When comparing across districts, the

use of water only is significantly higher in Sunamganj, the use of ash or clay is significantly higher in

Netrokona, and the use of soap is significantly higher in Kishoreganj. When comparing across Haor

types, the use of water only is significantly higher in moderate Haor, the use of ash or clay is

significantly higher in deep Haor. There are no significant differences in use of soap between deep

and moderate Haor types.

Table 34: Hand-washing behaviors among the FSUP baseline study households, by District and Haor

type (2)

Hands normally washed

with…



N 628 634 630 947 945 1892

Water only 33.9

32.6 42.9 c 34.0 38.9

a 36.5

Ashes or clay 48.1 58.7 c 44.0 53.7

c 46.8 50.3

Soap 18.0 c 8.7

b 13.2

b 12.2 14.3 13.3



8.2 Illness among adults and health-seeking behavior

The average number of illnesses cited per household was 2.4: 2.4 for Kishoreganj, 2.5 for Netrokona

and 2.2 for Sunamganj. The average number of illnesses reported by households in deep and

moderate Haors was 2.4 for both Haor types. Only 2.6% (49 households) experienced no illnesses at

all in the last 12 months. The most common illness experienced by adults during the previous 12

months is a cold attack, followed by gastric illness and diarrhea. There are no significant differences

among districts and between Haor types.

Table 35: Top ten illnesses experienced by adults in households during the previous 12 months, by


Illness

(multiple response)



Number of Responses 1521 1611 1364 2265 2231 4496

Cold attack 77.9 90.4 61.7 76.9 76.5 76.7

Gastric illness 46.2 44.0 41.6 43.4 44.4 43.9

Diarrhea 21.3 30.4 17.0 23.8 22.1 22.9

Dysentery 20.1 14.2 16.5 19.4 14.4 16.9

Anemia 13.9 19.1 17.8 17.6 16.2 16.9

Rheumatic fever 8.1 9.5 15.6 11.4 10.7 11.0

High/low blood pressure 4.6 9.6 6.3 5.4 8.4 6.9

Typhoid fever 9.9 8.7 2.2 6.8

7.1 6.9

Skin diseases 8.9 6.2 4.4 7.1 5.9 6.5

Asthma 4.9 6.3 4.3 5.0 5.4 5.2

Other 11.6 5.7 10.2 8.4 9.8 9.1

„Other‟ illnesses, cannot be disaggregated into individual illnesses, and therefore are not included in the top ten

illnesses.

Table 36 shows that medicine shops and village doctors are the most common treatment sources for

household members. Other treatment sources not reflected in table 36 include private paramedics/

LMA, Union Health Center, District Hospital, Homeopath, Kabiraj, untrained doctor, and Ojha/Jhar

Fuk; each accounted for less than 2%.

Table 36: Usual treatment source for household members, by District and Haor type

Treatment sources



628 634 630 947 945 1892

Medicine shop 23.1 c 40.7 41.7 31.4 39.0

c 35.2

Village doctor 32.3 33.0 35.1 35.9 b

31.0 33.5

Upazila Health Center 14.2 14.0 5.4 c 11.7 10.7 11.2

Private MBBS 14.2 c 4.4 6.2 9.6

b 6.9 8.2

Private clinic 3.5 2.1 3.2 2.9 3.0 2.9



Picture 16: Village medicine shop

Picture 17: Union Health Center


9.0 PARTICIPATION AND ACCESS

9.1 Participation in development

Participation in the development process is overall low at 4.5% of all households, which was too low

for meaningful analysis. In Kishoreganj it is 4.9 percent, in Netrokona 7.3 percent, and in Sunamganj it

is 1.1 percent. Participation in Sunamganj is significantly lower than the other two Districts (p=.000),

and in Kishoreganj it is significantly lower than in Netrokona (p=.044). Participation averages at 4.4

percent in both Deep and Moderate Haor.

Respondents were also asked who in the household participated but only 186 responses were

received from 174 households (a few households identified more than one person participating) -

9.2% of all respondents. Among the 186 responses, household head was mentioned as the most

common household member involved in development processes. Females (spouses plus female

heads of household) accounted for 15.1% of the 9.2%, or about 1.6% of the overall population.

Overall, the responses are too low for meaningful analysis.

More than 30% of responses were for the category other, which may have been used to record the

option “all household members‟, instead of checking all multiple response boxes.

Table 37: Household members involved in development processes

Household member involved

(multiple response)

Responses

N Percent

Household head

91 48.9%

Spouse 20 10.8%

Son/daughter 12 6.5%

Father/mother 4 2.2%

Daughter/son-in-law 1 0.5%

Grandson/granddaughter 1 0.5%

Other 57 30.6%

TOTAL 186 100.0%

Note: Female participation included 8 female heads of household plus spouses. There could

also be other females participating (e.g., mothers) but the data does not allow for

disaggregation at this level.

When asked about the type of development institution that the household member was

involved/engaged in, while participating in local development processes in the last 12 months, only

162 responses were collected. Again, this cannot be meaningfully disaggregated by district and Haor

type. Nonetheless, statistical testing among Haor types showed no significant difference.

Among the 162 responses, the Masjeed or religious committee was the most common response

(24%) followed by participation in NGOs (19%) as village group members, which is often a

prerequisite to receiving microcredit. Almost 22 percent of respondents who answered this question

stated that they did not know the type of development institution the household was involved with.

Again this may indicate that there were multiple institutions involved and respondents found it difficult

to recollect which ones exactly.


Table 38: Type of development institution/person that HH members were involved with

Type of development institution involved

(multiple response) Responses

N Percent

Union Parishad Chairman/Counselor

3 1.9%

Union Parishad Standing Committee

4 2.5%

Bazar Committee 12 7.4%

Masjeed or Religious Committee

39 24.1%

School/Madrasa Management Committee

12 7.4%

PTA 2 1.2%

Village Court/Salish 5 3.1%

NGO 31 19.1%

CBO 2 1.2%

Other 17 10.5%

DNK 35 21.6%

TOTAL 162 100.0%

When asked about the nature of household member‟s involvement/engagement with the development

institutions/persons stated in table 38, half (49.3%) of respondents reported that they had received

services, which may support the idea that the 30% „Other‟ in table 37 was used to indicate

participation by the entire household. Other types of participation were: volunteer (27.9%); committee

member (19.3%); participant in activities (19.3%); and recipient of training (1.9%). Only 2 households

received training: 1 in awareness on social issues, the other in awareness on H/N issues.

Only 5% of households had experience with collective action in last 12 months. In Kishoreganj it was

6.5 percent, in Netrokona 6.6 percent, and in Sunamganj it was significantly lower at 3.5 percent

(p=.019). Participation averaged 4.6 percent in Deep and 6.4 in Moderate Haor, but these values are

not significantly different. Among the 5% of households that had experience with collective action, the

main types of action were road construction/repair and Mosque construction/repair. Again, this cannot

be meaningfully disaggregated by district and Haor type.

Table 39: Type of collective action that households have participated in, by District and Haor type

Collective Actions

(multiple response)



ganj Deep Moderate

N 42 42 22 44 62 106

Road construction/repair 59.5 35.7 54.5 27.3 64.5 49.1

Canal/pond digging 0.0 4.8 0.0 0.0 3.2 1.9

Bamboo bridge construction 2.4 9.5 0.0 11.4 0.0 4.7

Embankment construction/repair 7.1 9.5 4.5 13.6 3.2 7.5

Graveyard construction/repair 4.8 9.5 0.0 6.8 4.8 5.7

Mosque construction/repair 31.0 61.9 31.8 59.1 32.3 43.4

Homestead raising/protection 4.8 7.1 9.1 4.5 4.8 4.7

School construction/repair 4.8 4.8 4.5 6.8 4.8 5.7


Picture 18: Community collective action to improve road infrastructure

9.2 Access to GoB services

Over two-thirds of households (68.7%) had accessed one or more GoB service providers in the

previous year. Table 40 shows the types of GoB service providers used by households in the last 12

months. The most common service providers used were Union Parishad and Government

Immunization Services, followed by Government Family Planning, Upazilla Health Services and Union

Health Services. All other service providers listed in table 40 were 0.5% or less. Department of

Fisheries, Department of livestock, Department of Cooperatives, and Government Vocational/

Educational Training all recorded zero responses.


Table 40: Proportion of households using various types of Government service providers, by District

and Haor type

Service Provider District Haor Type Total


N 446 433 421 681 619 1300

Dept. of Agr. Extension (DAE) 0.2 0.2 0.7 0.4 0.3 0.4

Government Land Office 0.0 0.2 0.0 0.1 0.0 0.1

Dept. of Youth Development 0.0 0.5 0.0 0.3 0.1 0.2

Dept. of Women‟s‟ Affairs 0.0 0.9 0.5 0.6 0.4 0.5

Government Family Planning 23.5 b

15.2 15.2 18.2 17.9 18.1

Govt. Immunization Services 41.3 40.6 55.1 c 39.6 52.0

c 45.5

Union Parishad 47.8 67.4 c 57.5

b 61.5 53.0

a 57.5

BADC Seed Department 0.4 0.0 0.0 0.0 0.2 0.2

Union Health Services 16.2 c 1.4 5.0 10.6

c 3.7 7.3

Upazila Health Services 24.2 20.6 7.1 c 18.5 16.3 17.5


Figures 13 and 14 show the types of GoB service providers accessed by District and Haor type. When

comparing across districts, Government Family Planning and Union Health Services was significantly

higher in Kishoreganj than in the other two districts. Union Parishad was significantly higher in

Netrokona. In Sunamganj, Government Immunization Services was significantly higher and Upazilla

Health Services was significantly lower than in other districts.

Figure 13: Types of service providers accessed, by District

(N=1300)

When comparing across Haor types, Government Immunization Services was significantly higher in

the moderate Haor, and Union Parishad and Union Health Services was significantly higher in the

deep Haor.


Figure 14: Types of service providers accessed, by Haor type

(N=1300)

Table 41 shows the types of services received by GoB service provider, by District and Haor type. For

Union and Upazila Health Services, the most common service received is medication followed by

suggestions. For Family Planning, suggestions, medicines and vaccinations are the main services

received. For Government Immunization Services, vaccinations are the main services received, as

was to be expected.

Overall, training provided by GoB service providers is very low. Union Parishad was not included in

the table because 95% of services provided were reported as „Other‟. „Other‟ could refer to safety nets

such as the government programs for supporting vulnerable populations: the Vulnerable Group

Feeding (VGF) program, which provides food to low income and other vulnerable groups who cannot

meet basic needs for survival due to natural disasters or socio-economic circumstances, such as age,

illness or disease; and the Vulnerable Group Development (VGD) program, which aims to enable the

poorest rural women and their family members to overcome food insecurity and their low social and

economic status. These kinds of safety net programs were indicated in qualitative data collection as

the main service that Union Parishads are known for. It is important to note that although these

programs were highly valued, community members had concerns about the transparency and equity

in recipient targeting.

It is apparent that the majority of services provided are health related. Qualitative data confirmed the

lack of services on economic activities. Community members particularly mentioned the need for

more and better technical assistance in the areas of livestock rearing, fishery and agriculture.


Table 41: Types of services received by GoB service providers, by District and Haor type

Service Provider District Haor Type Total


Union Health Services N 95 7 38 100 40 280

Suggestions 26.5 33.3 54.5 26.4 54.2 33.3

Medicines 77.9 83.3 77.3 79.2 75.0 78.1

Vaccinations 32.4 0.0 31.8 30.6 29.2 30.2

Other 2.9 0.0 9.1 2.8 8.3 4.2

Upazila Health Services N 167 141 51 191 168 718

Suggestions 52.3 43.5 48.4 44.9 52.4 48.3

Medicines 86.2 77.2 58.8 90.6 76.2 84.1

Vaccinations 13.8 22.8 19.4 11.8 25.7 18.1

Other 0.9 9.8 0.0 3.1 5.7 4.3

GoB Family Planning

N

188 120 167 240 235 475

Formal training 0.0 0.0 1.6 0.0 0.9 0.4

Suggestions 73.1 67.2 89.1 75.2 76.6 75.9

Medicines 54.8 75.0 95.3 72.7 70.3 71.6

Vaccinations 51.0

1.9

43.8 75.0 49.6 62.2 55.6

Other 1.9 1.6 1.6 0.8 1.8 1.3

GoB Immunization Services

N

262 259 393 421 493 914

Informal training 0.5 0.0 0.0 0.0 0.3 0.2

Suggestions 29.6 19.7 22.6 26.1 22.1 23.9

Medicines 15.6 25.8 47.9 27.6 34.4 31.3

Vaccinations 93.0 97.8 93.2 96.0 93.3 94.5

Other 2.2 2.2 4.3 5.1 1.2 3.0

Figure 15 shows the level of satisfaction with the services received through the various GoB service

providers. For all service providers, the majority of respondents indicated they were satisfied or highly

satisfied.

Figure 15: Level of satisfaction with selected GOB services


9.3 Access to other services

Respondent were also asked about services that they had received from non-government service

providers. Overall 67% of households reported not receiving any services from other non-government

service providers. When disaggregated by District, the number of households receiving no services

from non-government service providers was significantly higher (p=.020) in Netrokona (78%) than in

Kishoreganj (65%) and Sunamganj (58%). When disaggregated by Haor type, the number of

households receiving no services from non-government service providers was also significantly higher

(p=.000) in moderate Haor areas (74%) than in Deep Haor (60%).

Overall, the three most common non-government service providers (for individuals who reported

receiving services) were NGOs (76%), Grameen Bank (16%), and Local Service Providers (18%).

Less than 1% of households reported receiving services from Commercial Banks, CBOs, input

retailers/dealers and non-Government Vocational Education/Training, respectively. The most common

services received from NGOs were credit (68%), suggestions (16%), and relief/aid (4%). The most

common services received from Grameen Bank were credit (99%) and suggestions (13%). The most

common services received from Local Service Providers were suggestions (75%), credit (65%),

suggestions (16%), medicines (71%) and relief/aid (12%).

Qualitative data collection showed that community expectations for economic and development

activities primarily revolve around facilitating access to Khas water bodies, access to credit, and

capacity development. Men and women share expectations around external support for increased

participation in community decision making, improved flood protection and improved children‟s

education. While men‟s expectations focus mainly on economic opportunities and strengthened links

to livestock and agriculture services; women expressed expectations around increased opportunities

to have a voice in community affairs, improved health services and access to life skills training.

Picture 19: Women engaged in alternative livelihood activities


At present, the main sources of knowledge and skills for economic/livelihood activities are knowledge

transfer from previous generations, and from relatives and neighbors. The little external assistance

that ultra-poor households do receive comes primarily from NGOs. Examples mentioned by

respondents included livestock and poultry rearing, credit groups, increased crop production, market

development, homestead gardening. Women appear to be mainly involved in micro credit and men

are also involved in earthwork, flood protection and infrastructure projects. Among the limited number

of women who participate in income-related activities in deep haor regions, it was mentioned that a

large number are widows and that access by married women with families is more difficult. NGOs also

facilitate community participation in development/economic activities. This is valued by community

members as the limited participation of the ultra poor in local committees is an apparent concern.

Knowledge of economic and development opportunities appears to be low. The main sources of

information about external assistance for economic and development activities are NGO workers. This

information is often channeled to community members by village leaders and prominent community

members such as school teachers, health workers and local elites; and does not reach everyone

equally. Union Parishad officials are the main source of information for GoB development activities. It

was noted that Union Parishad officials do not pass information to everyone; they prefer to share

information with their supporters only.

Similar to participation in economic activities, decision making around participation in NGO and GoB

development activities is heavily influenced by the rich and politically powerful, as well as by kinship

ties. Although the poor and ultra poor do participate in the development process, they have little voice

regarding types and recipients of benefits, resource allocation and arbitration. Community members

recognize the purposive targeting of women by NGOs and are very supportive of this. However, it is

important to note that within the family the nature of participation by women is often still determined by

male household members.

Community members stated that the community benefits generated from participation in development

activities are more important than monetary benefits. These benefits include reduced cost of travel,

market and health care connectivity, improved drinking water, flood protection, improved access to

other service providers, and improved school attendance/ reduced dropout rates. Non-monetary

benefits generated at the individual and household level include improved social dignity for the ultra

poor, increased women‟s participation and more joint decision-making between men and women.

There are serious concerns about how benefits are distributed; with benefits going more to those with

kinship relations and the economic means to bribe officials or invest in development activities. An

example of the latter is the installation of village tube wells; the more well-off community members

usually pay the security deposits for the wells and then end up controlling irrigation to the benefit of

their own crops and those of their kin. Community members also mentioned the high fees provided to

local experts hired for training, which in some cases come to 25% of the total available funds for the

local project.

There appear to be significant costs involved with participation in development activities. Similar to

participation in economic activities, there are non-monetary costs associated with time away from

home by women – such as reduced care for children, and inability to do household chores, which

causes stresses between husband and wife. At the same time there are monetary costs (primarily for

males) as participation in development activities takes time away from work. There are also reports of

food being stolen when males are away.


There are also direct monetary costs involved in participating in development activities. To get

VGD/VGF cards, community members commonly must take high interest loans to bribe Union

Parishad members. Convincing Union Parishad members to allow their participation also requires a

significant time investment and in some cases community members must also provide physical labor

to help convince them.

Power relations due to kinship and politics are considered the main barriers to access to development

opportunities by the ultra poor. In addition, women face additional barriers due to their lack of access

to information, their lack of confidence in speaking publicly and the ongoing discouragement by men

(and some women as well) that prevents them from participating in community dialogues and

meetings. To overcome these barriers, community members stated that there is an urgent need for

capacity development on rights issues.

9.4 Access to common property

Table 42 shows the proportion of households that have various types of property available in their

area, disaggregated by District and Haor type. Overall, Beel/Haor and canal/river are the most

common land property types, followed by Khas land, roadside sloping and Khas ponds. Khas pond

and Khas land are most common in Kishoreganj, and most common in Deep Haor. Roadside sloping

and grazing land are both more common in Netrokona.

Table 42: Proportion of households that have various property types available in their household

area, by District and Haor type

Types of property District Haor Type Total


N 628 634 630 947 945 1892

Khas pond 11.5 2.4 4.1 7.4 4.6 6.0

Khas land 32.8 18.8 9.4 32.5 8.0 20.3

Roadside sloping 3.3 18.3 0.2 6.2 8.4 7.3

Embankments 7.3 0.6 1.7 3.6 2.9 3.2

Railway grounds 0.5 0.2 0.2 0.4 0.1 0.3

Beel/Haor 74.7 92.7 74.4 84.3 77.0 80.7

River/Canal 78.0 88.8 88.4 91.7 78.5 85.1

CBO water body 5.7 1.3 1.1 6.3 5.7 6.0

Grazing land 2.9 10.9 4.1 6.0 5.9 6.0

Forest 1.3 0.2 0.2 0.1 1.0 0.5

Hills 0.3 1.1 0.2 0.1 1.0 0.5


Table 43 shows the proportion of available property that is accessible by households, which means

they can use the resources for household or livelihood purposes. Railway, forest and hill land were

excluded from the table due to an inadequate number of responses for meaningful analysis. Overall,

the highest proportion of households has access to river/canals, followed by roadside sloping and

beels/haors. Access to Khas land is lowest. When comparing across districts, access to Khas pond,

road side sloping and river/canals is highest in Sunamganj – with access to roadside sloping reported


as 100%. Access to embankments and CBO water bodies is highest in Kishoreganj. Access to Khas

land in Kishoreganj, and embankments in Netrokona and Sunamganj was reported as 0%.

Table 43: Proportion of available property that is accessible by households, by District and Haor type

Types of property District Haor Type Total

N Kishoreganj Netrokona Sunamganj Deep Moderate

Khas pond 116 12.5 18.8 64.3 7.0 55.6 25.9

Khas land 385 0.0 15.1 23.3 8.1 9.2 8.3

Roadside sloping 138 23.8 46.6 100.0 50.8 30.8 43.5

Embankments 64 40.8 0.0 0.0 11.8 53.3 31.3

Beel/Haor 1526 31.8 50.0 46.3 45.4 40.9 43.3

River/Canal 1613 68.2 66.6 79.2 76.5 65.6 71.5

CBO water body 115 52.8 37.5 15.5 39.3 16.7 28.7

Grazing land 113 33.3 21.7 19.2 17.5 28.6 23.0


Respondents were also asked what kind of activities household members were allowed to do on the

properties they had access to. However, answers to common property uses allowed were quite

varied, suggesting a need for clarification and awareness-building in this area.

Qualitative data showed that community members specifically stated the restricted access to open

Khas water for fishing purposes and restricted land access for rice cultivation as main barriers to

economic development. Community members emphasized the need for increased advocacy by NGOs

and other stakeholders for increased access to Khas land and water to increase participation of the

ultra poor in economic activities.

Picture 20: Government-owned Khas land


10 DISASTERS AND CRISES

10.1 Natural disasters: effects and coping strategies

Overall, 78% of households reported that they did not experience a natural disaster in the previous

year. In both Netrokona and Sunamganj, 81% experienced no disaster, and in Kishoreganj

significantly fewer (71%) experienced no disaster. In deep Haor, 75% of households did not

experience a disaster, while in moderate Haor the proportion was significantly higher (p=.022) at 80%.

Table 44 provides data for those households that did experience a natural disaster. The highest

proportion of disasters experienced in the last 12 months were wind damage, floods, excessive rain

and storms. Wind damage is locally called „Aphal‟; strong winds that damage standing crops, cause

soil erosion and uproot trees.

Table 44: Disasters experienced by households in the last 12 months, by District and Haor type

Type of natural disaster District Haor Type Total


N 181 122 121 233 191 424

Flood (flash/monsoon) 37.6 29.5 28.1 38.6 25.1 32.5

Drought 2.2 0.8 3.3 1.7 2.6 2.1

Storm 18.1 18.0 26.4 17.6 24.1 20.5

River erosion 0.0 1.6 5.0 3.4 0.0 1.9

Excessive rain 44.2 3.3 22.3 25.3 27.2 26.2

Water logging 12.7 15.6 8.3 14.6 9.4 12.3

Land slide 0.6 0.0 0.0 0.4 0.0 0.2

Wind damage 50.8 45.1 17.4 46.8 40.9 39.6

Soil erosion 1.1 5.7 0.8 3.4 1.0 2.4

Picture 21: Damage to buildings as a result of natural disasters


Respondents, who reported experiencing a natural disaster in the last 12 months, were subsequently

asked what the effect of that particular disaster was on their household. The highest proportion of

households experienced partial damage to their house, followed at a distance by loss of working days

and full damage to their house.

Table 45: Proportion of households experiencing various consequences of a natural disaster in the last

12 months, by District and Haor type

Effect of natural disaster District Haor Type Total


N 184 124 123 238 193 431

Loss of working days 54.3 c 7.3 4.1 31.1

c 20.7 26.5

Damaged house fully 6.5 c 16.1 14.6 10.9 12.4 11.6

Damaged house partially 66.8 a 72.8 74.0 77.7

b 61.1 70.3

Damaged poultry and livestock 8.2 7.3 0.0 c 8.0

c 2.6 5.6

Loss of productive assets 1.1 0.0 0.0 0.8 0.0 0.5

Crop loss 5.4 1.6 c 8.1 2.9 7.8

b 5.1

Loss of HH goods 3.3 8.1 0.0 c 4.2 3.1 3.7

Loss of trees 13.6 b 8.1 1.6 6.7 10.9 8.6

Tube well damage 0.5 0.0 0.0 0.0 0.5 0.2

Latrine damage 9.8 5.6 0.8 c 8.0

a 3.6 6.0

Other 1.1 5.6 0.0 2.5 1.6 2.1

Figure 16 shows the mean asset/income loss reported by households who experienced a disaster in

the last 12 months, disaggregated by district and Haor type. The mean asset loss per disaster was

reported at around Taka 3,017. There are no significant differences among districts but the asset loss

in moderate Haor is significantly higher than in deep Haor.

Figure 16: Mean asset loss from households experiencing asset loss in a natural disaster

in the last 12 months, by District and Haor type


Figure 17 shows the mean number of working days lost reported by the 26.5% of households who

indicated this effect in table 45, disaggregated by district and Haor type. The mean number of working

days lost is 10. When comparing across districts, the mean number of working days lost is

significantly higher in Kishoreganj than the other two districts. There is no significant difference across

Haor types.

Figure 17: Mean number of working days lost from households experiencing a natural

disaster in the last 12 months, by District and Haor type

The most common coping strategies used by respondents to recover from a natural disaster were:

taking out a loan from friend/neighbor (41%), taking loan from a moneylender (31%), adjusting meals

(25%), using savings (25%), accepting help from others: (24%), purchasing on credit (21%) and

taking a loan from NGO (11%).

10.2 Household crises: effects and coping strategies

Respondents were also asked the same range of effect and coping strategy questions for a range of

household crises, not caused by natural disasters. Only 16.7% reported the occurrence of such crises

in the last 12 months – ranging from 16-19% among Districts and across Haor type, with no significant

differences. The most common types of household crises reported were illness of income earners

(57.2% of cases where a household crisis was reported) and illness of other household members

(32% of cases where a household crisis was reported). All other responses were less than 5%.

The main effects of the household crises were asset/income loss and work days lost. Figure 18 shows

the mean loss of assets, disaggregated by district and Haor type. The mean loss of assets was just

under Taka 5,000. Comparison among districts shows that asset loss was significantly higher in

Netrokona than in Sunamganj. There is no significant difference across Haor types.


Figure 18: Loss of assets among households experiencing household crises

in the last 12 months, by District and Haor type

(N = 306)

Figure 19 shows the mean number of working days lost, disaggregated by district and Haor type. The

mean number of working days lost was 36. Comparison among districts and across Haor types

showed no significant differences.

Figure 19: Loss of work days among households experiencing household

crises in the last 12 months, by District and Haor type

(N = 306)

Figure 20 shows the mean number of working days lost as a result of illness of either income earner

or other household members, disaggregated by district and Haor type. The mean number of working

days lost due to illness was 36 (mode=15). Comparison among districts shows that the number of

working days lost due to illness is significantly higher in Netrokona than in Sunamganj. There are no

significant differences across Haor types.


Figure 20: Average number of days lost due to illness for those households with an ill member designated as a household crises in the last 12 months, by District and Haor type

(N=305)

The most common coping strategies used by respondents to cope with household crises were

(n=316): took out a loan from friend/relative (42.7%), took out a loan from moneylender: (36.0%),

made adjustment to meals (27.8%), accepted help from others (20.6%), purchased goods on credit

(18.4%), used savings (11.7%), took out a loan from an NGO (11.1%), took a grain loan (10.4%), ate

famine foods (8.2%), and accepted aid (5.4%)

10.3 Climate change

Qualitative data collection included some exploratory questions around climate change in qualitative

data collection, the findings of which are by no means robust. Community members reported

increased temperatures, more extreme storms, irregular flash floods and irregular/infrequent rainfall.

They inferred multiple linkages between these changing weather-related characteristics and livelihood

impacts such as reduced crop and fishing yields, less migratory birds, increased insect infestation and

crop disease, and reduced soil fertility. On e common example mentioned was that climate variability

has reduced ability to predict flash flood; previously crops could be harvested prior to flash flood

It was noted that in the last ten years the water levels of the Haor have been reduced significantly and

sedimentation has increased; beels and marshlands were filling up, which negatively affected fishery

and agricultural practices due to lack of water in the dry season. In addition, the increased irrigation

required as a result of the reduced rainfall and lower water levels make agricultural practices more

costly; reducing profits derived from agriculture.

As a result of these changes, many poor and ultra poor households can no longer rely on daily fishing

labor as the main source of income but must now do agricultural day labor and poultry rearing to

make a living. Overall, community members noted limited capacity to adapt to these changes,

particularly for the ultra poor.


In addition, community members highlighted man-made problems that compounded the problems

considered to be caused by climate variability. For example, the use of insecticides on crop land

reduced fishing yields and also decreased day labor opportunities for pulling weeds.

To address the impacts of climate change, community members stated the need to organize

development activities that focus on river dredging, promotion of more resilient crops and agricultural

practices, and provision of training on climate change and how to adapt.

Picture 22: Social mobilization around community issues


11 FAMILY AUTHORITY AND DECISION MAKING

11.1 Household decision making

Tables 46 to 49 show the types of decision making for 12 different types of household decisions,

disaggregated by Haor type and district. All questions were answered by a female household

member. The highest proportion of decisions is made by the husband after discussion with the female

household member. It is also apparent that women have greater involvement in household decisions

such as minor household purchases, children‟s clothing and education, medical expenses and in

spending money that they have directly earned. Women have less involvement in expenditures that

relate to livelihoods, higher value assets, loans/savings and events such as weddings and ceremonies

and shelter in case of disasters. The proportion of decisions made without any involvement by the

female is low for almost all decision types, except salish decision making.

Tables 46 and 47 show that when the data is disaggregated by Haor type, the proportion of decisions

made by the husband after discussion with the female household member is significantly higher in

deep Haor than moderate Haor areas for many of the decisions. For several decisions, the proportion

of women not involved in decision making is significantly higher in moderate than in deep Haor. It is

important to note that a relatively high number of women answered not applicable (not listed in tables

below) to the various decision types, which could be interpreted that they were uncomfortable

responding.

Table 46: Household decision making, by Haor type (1)

Decision

Haor Type

Overall Deep Moderate

Buying small food items, groceries, toiletries

Can decide alone 20.3 20.2 20.3

Decide w/ husband or other adult male 14.2 14.7 14.4

Husband decides after discussion 58.1 a 53.4 55.8

Not involved in decision 7.4 11.7 c 9.5

Buying clothing for yourself and your children



Husband decides after discussion 65.8 b 59.3 62.5

Not involved in decision 6.6 12.5 c 9.6

Spending money that you yourself have earned



Husband decides after discussion 63.5 a 58.6 61.1

Not involved in decision 5.0 6.5 5.7

Buying or selling major household assets (land, livestock, crops)



Husband decides after discussion 65.2 b 60.7 63.0


Buying or selling jewelry



Husband decides after discussion 72.1 71.6 71.8



Use of loans or savings




Not involved in decision 4.8 7.5 b 6.2

Table 47: Household decision making, by Haor type (2)

Decision

Haor Type

Overall Deep Moderate

Expenses for your children’s education





Expenses for your children’s marriage


Decide w/ husband or other adult male 17.9 26.7 a 22.1

Husband decides after discussion 69.7 c 58.1 64.2

Not involved in decision 2.6 5.6 a 4.0

Medical expenses for yourself or your children

Can decide alone 13.6 16.9 b 15.3




Expenses for family planning (contraceptives)





To move to shelter during time of disaster



Husband decides after discussion 54.5 c 48.4 51.5

Not involved in decision 9.4 15.5 b 12.4

Actively participate and involved in salish decision making

Can decide alone 4.9 8.3 c 6.8

Decide w/ husband or other adult male 7.4 11.8 c 9.8

Husband decides after discussion 19.0 23.3 b 21.4

Not involved in decision 68.8 c 56.

6 62.0

Tables 48 and 49 show that when the data is disaggregated by district, the proportion of decisions

made by the husband after discussion with the female is significantly lower in Netrokona than in the

other two districts. Correspondingly, the proportion of decisions wherein the female is not involved at

all is significantly higher in Netrokona for many decision types. However, it is interesting to note that

the proportion of decisions that females can make on their own is also significantly higher in

Netrokona than in the other two districts for several of the decision types.


Table 48: Household decision making, by District (1)

Decision

District


Buying small food items, groceries, toiletries

Can decide alone 16.1b 23.1 21.5

Decide w/ husband or other adult male 12.9 21.5b 8.5

Husband decides after discussion 68.2 39.5c 60.4

Not involved in decision 2.8 15.9c 9.6

Buying clothing for yourself and your children

Can decide alone 12.7 19.0c 10.8

Decide w/ husband or other adult male 11.4 21.9c 7.8



Spending money that you yourself have earned


Decide w/ husband or other adult male 7.2 11.8a 5.7


Not involved in decision 4.1 1.9b 11.4

Buying or selling major household assets (land, livestock, crops)

Can decide alone 12.1 15.6 6.7b

Decide w/ husband or other adult male 12.8a 22.4 19.0

Husband decides after discussion 69.1 51.5b 66.7


Buying or selling jewelry



Husband decides after discussion 75.0 52.4b 80.2


Use of loans or savings

Can decide alone 9.9 13.1a 8.0





Table 49: Household decision making, by District (2)

Decision

District


Expenses for your children’s education




Not involved in decision 1.8c 4.8 6.6

Expenses for your children’s marriage

Can decide alone 10.2 17.5 5.2b

Decide w/ husband or other adult male 17.1b 22.1 26.8



Medical expenses for yourself or your children





Expenses for family planning (contraceptives)




Not involved in decision 1.0b 5.5 8.8

To move to shelter during time of disaster


Decide w/ husband or other adult male 20.0b 28.0 26.1



Actively participate and involved in salish decision making

Can decide alone 10.8 7.9 1.9c


Husband decides after discussion 20.1 14.9 30.7c


Qualitative data supports the quantitative findings around household decision making presented in the

tables above. Qualitative data also shows a trend of an increasing role of women in household

decision making. For example, women‟s participation in economic activities is increasingly jointly

discussed, although it is important to note that men still make the final decision and many women

believe that the men often know best. Profit and loss decisions vary by household. There is also

indication of increased independence in household processes. For example, women‟s freedom of

movement appears to be expanding. Previously women shopkeepers relied on their husbands to

acquire shop stocks, now more women are able to acquire the goods themselves. External contact

(like with NGOs) appears to have been an important factor in these trends.


11.2 Family life attitudes

Tables 50 to 51 show the attitude about family life, disaggregated by Haor type and district. All

questions were answered by a female household member. Overall, a higher proportion of women

agree that the husband should help with household chores if the female is working; and that they have

the right to express their opinion, even when they disagree with their husband. The proportion of

women overall who disagree with the statement that it is better to send a son to school than a

daughter is also significantly higher. However, it is interesting to note that despite the more liberal

attitudes about family life expressed by women, the proportion of women who agree that a wife should

tolerate being beaten is significantly higher that the proportion who disagrees.

When disaggregated by Haor type, Table 50 shows that a significantly higher proportion of women in

deep Haor agree that the husband should help with chores if the wife is working and married women

should be allowed to work outside the home. However, a significantly higher proportion of women in

deep Haor also agree that a wife should tolerate being beaten and that it is better to send boys to

school instead of girls.

Table 50: Attitudes about family life, by Haor type

Attitudes about family life

Haor Type

Total Deep Moderate

N= 947 945 1892

The important decisions in the

family should be made only by men

Agree 49.2 45.8 47.5

Disagree 47.7 52.2 49.9

DNK 3.1 2.0 2.5

If the wife is working outside the

home, then the husband should help

her with household chores

Agree 66.1b 60.5 63.3

Disagree 27.8 34.5 31.1

DNK 6.1 5.0 5.5

Married women should be allowed

to work outside the home

Agree 53.1b 46.3 49.7

Disagree

42.2 49.0

45.6

DNK 4.6 4.7 4.7

The wife has a right to express her

opinion even when she disagrees

with her husband

Agree 66.8 66.0 66.4

Disagree 28.7 29.1 28.9

DNK 4.4 4.9 4.7

A wife should tolerate being beaten

by her husband in order to keep the

family together

Agree 83.5c 74.3 78.9

Disagree 14.6 23.3 18.9

DNK 1.9 2.4 2.2

It is better to send a son to school

than a daughter

Agree 27.1b 21.1 24.1

Disagree 66.8 72.4 69.6

DNK 6.0 6.6 6.3

When disaggregated by district, Table 51 shows mixed results. A significantly higher proportion of

women in Netrokona agree that the important decisions in the family should only be made by men.

There are significant differences among the three districts in the proportion of women who agree with

the statements that the husband should help with chores if the wife is working and married women

should be allowed to work outside the home; whereby the highest proportion of women who agrees is

in Kishoreganj and the lowest proportion is in Sunamganj. The proportion of women who agree that

the wife has a right to express her opinion even when she disagrees with her husband is also

significantly lower in Sunamganj.


Although findings indicate that Sunamganj is the most conservative of the three districts, it is

important to note that the proportion of women who agrees that it is better to send a son to school

than a daughter is significantly lower in Sunamganj than in the other two districts; whereby

Kishoreganj shows the highest proportion of women who agrees. In contrast, Kishoreganj shows a

significantly lower proportion of women who agree that a wife should tolerate being beaten by her

husband in order to keep the family together

Table 51: Attitudes about family life, by District

Attitudes about family life

District


N=628 N=634 N=630

The important decisions in the

family should be made only by

men

Agree 41.6 58.2 c 42.7

Disagree 55.3 40.4 54.3

DNK 3.2 3.2 3.0

If the wife is working outside the

home, then the husband should

help her with household chores

Agree 76.3 c 66.6

c 47.1

c

Disagree 18.6 30.6 44.1

DNK 5.1 2.8 8.7

Married women should be allowed

to work outside the home

Agree 62.1 c 47.5

c 39.7

c

Disagree 31.8 50.3 54.6

DNK 6.1 2.2 5.7

The wife has a right to express her

opinion even when she disagrees

with her husband

Agree 68.9 70.3 60.0 c

Disagree 24.7 26.7 35.4

DNK 6.4 3.0 4.6

A wife should tolerate being

beaten by her husband in order to

keep the family together

Agree 75.6 b 80.3 80.8

Disagree 21.5 17.8 17.5

DNK 2.9 1.9 1.7

It is better to send a son to school

than a daughter

Agree 33.3 22.9b 16.2

Disagree 62.9 68.8 77.1

DNK 3.8 8.4 6.7

11.3 Daily time patterns of men and women Qualitative data collection was organized to gain a better understanding of daily time spending of

men, women involved in work and women who stay at home. Graphic representations are added in

annex 3. Figure 20 below provides one example from a moderate Haor village in Kishoreganj. It is

important to note that the data collected is specific to the month of February around the time when

agricultural day labor, primarily in rice fields, is coming to an end. Patterns will likely be different in

other seasons.

Men typically start their day between 5-6am. The first thing they do is pray, put the cow to field and

clean the shed. Men then usually put in 1-2 hours of work in the rice fields such as transplantation,

weeding, fertilizing and irrigation before taking breakfast between 8-9am; followed by more work in the

rice fields until taking a 1-2 hour lunch break around 1pm. They return from the fields between 5-6pm,

feed the cow and put it to shed. The time between 6-8pm is commonly spent resting or gossiping,

wandering around or going to the market. Dinner is taken around 8-9pm after which men go to sleep.

Income derived from a typical day describe above ranges between 150-200 Taka per day. In some

cases men get 200 Taka, excluding meals or 150 Taka including 3 meals, although 200 Taka

including meals is also reported.


Women who stay at home get up at the same time as their husband, between 5-6am. They start the

day by fetching water, sweeping the house and the courtyard, cleaning the cooking pot, helping the

husband put the cow out, and cooking breakfast. If there are school-going children in the household,

they are sent to school after breakfast. After breakfast, the women start a range of household chores,

including: coiling dung for fuel, collecting vegetables and fish, collecting firewood, washing clothes,

cleaning pots and cooking meals. Women usually bath before preparing lunch and take two rest

periods. One in late morning and one after serving lunch, during which time they gossip and stitch

kantha. Children are washed between 5-6pm and preparations are made for dinner. Before dinner,

the women spend about one hour with the children to help them learn. After dinner, the women stay

up longer to clean the cooking pot and house, and to pray before going to bed around 10pm.

Women who are involved in income-generating activities get up around 4am to give them enough time

to complete the first chores of the day. If breakfast and dinner are provided by the employer, they then

work from 5am to around 7pm. If not, they work from 8am to 5pm. Lunch is almost always included.

After work they must fetch water, clean the cooking pot, feed the children and cook dinner for their

husband. After dinner they prepare the bed for their husband and do some small household chores

before going to sleep at around 10pm. Women commonly spend around 12-13 hours working and

about 2-3 hours doing household chores. Payment for women varies and can range from 50-100 Taka

for uprooting rice seedlings without meals to proportions of the harvest.

Figure 21: Daily time use of men and women


12 CHILD NUTRITION, ANTENATAL CARE AND FAMILY PLANNING

12.1 MCHN characteristics

All questions in this section relate to < 2 children. The respondent is always the child‟s mother. Of the

total number of respondents, 70 percent did not have any children < 2 years of age. Of those that did,

29 percent had one and 1 percent had two < 2 children. Table 52 provides an overview of Maternal

Child Health and Nutrition characteristics, disaggregated by District and Haor type. Note that there

were no significant differences between Haor types.

Virtually every mother has breastfed her child (99.5%) and 45% of overall mothers initiated

breastfeeding with the first hour of birth. The percentage of women who initiated breastfeeding after

the first hour was significantly lower in Netrokona than in the other districts.

The average age for introducing solid/semi-solid foods (weaning) was just over 5 months age,

whereby the average age in Sunamganj was significantly lower at 4.6 months. Table 52 shows the

proportion of women introducing solid/semi-solid foods for 0-3, 4-6, 7-9 and 10+ months. This data

can be used to track change over time in the introduction of solid/semi-solid foods, based on project

recommendations on proper weaning practices.

Qualitative data showed that most women know the value of colostrums to newborn health. However,

in some villages traditional practices and superstitions prevent mothers from providing colostrums to

newborns. Nutritional information for newborns was usually obtained from the village doctor or health

worker. Exposure to media is an additional source of nutritional info.

Overall, 35.5% of mother‟s took iron or folic acid supplements. The proportion of mother‟s who took

these supplements was significantly higher in Kishoreganj. Taking iron and folic acid during

pregnancy varied by village; where health workers were present, use was common. Lack of

knowledge and mother-in-laws who discourage use of both supplements presented barriers in other

villages.

The majority of mother‟s did not change the amount of food that they consumed during their last

pregnancy; 16% increased their food intake and 33% decreased their food intake. Qualitative data

showed that some mothers reduce food intake to two meals a day during pregnancy to keep the size

of the baby smaller. A large baby during pregnancy makes it hard to work and the want to avoid the

complicated delivery of a large baby as the hospital is far away and costly to access. Mothers do have

some knowledge on food supplements during pregnancy – primarily from health workers, media, and

village elders. However, limited financial resources often prohibit taking more nutritious foods. Dietary

diversity while breast feeding is also poor.

The majority of women also did not change the amount of rest they took after the last birth. Only 23%

took more rest than usual. Qualitative data indicated that the gender of the baby determines the

amount of rest (this means a period of light work) for new mothers: 9 days rest for a male child; 7 days

for female. After the 7-9 day period, women resume regular work and only rest while breast feeding.

Qualitative data also shows that most women do not take rest during their pregnancy either. They

continue to complete their daily activities, taking only a little rest after chores are done. Household and

community members generally help take on some of the household chores if the pregnant mother

becomes sick.


Picture 23: Balanced meal taken by a pregnant woman

Overall, mothers attended on average 1 ANC session. Qualitative data shows that this is usually after

3 months of pregnancy. The proportion of mothers attending ANC was significantly higher in

Kishoreganj than in the other Districts. Qualitative data shows that there are very few periodic medical

checkups during pregnancy due to lack of knowledge, lack of money, and difficulties in

communicating with the medical centers. Medical check-ups are only used for serious complications.

Additional barriers to obtaining pre-natal care are: mother in law and some elders do not approve; and

when the husband is unable to accompany wife to doctor, social norms prevent a woman from

traveling alone.


Table 52: MCHN characteristics by District and Haor type

Hand-washing Behaviors District Haor Type Total


N= 196 166 202 271 293 564

Average age solid/semi-solid food introduced (months)

5.4 5.3 4.6 b 5.1 5.1 5.1

When breastfeeding

initiated

Within 1st

hour (%) 45.4 55.4 37.1 46.5 44.4 45.4

After 1st

hour (%) 54.6

b 44.6

c 62.9 53.5 55.6 54.6

Age of introducing solid foods

0-3 months 15.8 21.4 43.9 c 23.8 29.4 26.9

4-6 months 63.7 57.2 25.9 c 51.9

b 45.1 48.1

7-9 months 18.4 17.9 25.8 b

22.1 20.7 21.4

10+ months 2.1 3.5 4.3 2.2 4.8 3.6

Took iron or folic acid supplements (%)

47.4 c 28.9 29.2 34.3 36.5 35.5

Changes in amount of food

consumed

More (%) 15.2 16.9 15.8 17.2 14.7 15.9

Same (%) 47.7 50.6 55.6 50.2 52.6 51.4

Less (%) 37.1 32.5 28.6 32.6 32.8 32.7

Number of ANC sessions attended

1.2 c 0.7 0.8 0.9 0.9 0.9

Amount of rest

after last birth

More than usual 25.4 18.7 23.2 22.3 22.9 22.6

About the same 62.9 72.3 65.0 67.8 65.2 66.4

Less than usual 11.7 9.0 11.8 9.9 11.9 11.0



Table 53 shows the different types of weaning foods used by mothers for their most recently born

child, disaggregated by District and Haor type. On average, mothers in Kishoreganj used 2.2 different

weaning foods, mothers in Netrokona used 1.8 weaning foods, and mothers in Sunamganj used 1.6

weaning foods.

When comparing across Districts, the use of kichori, Soji/Sagu/Barli, and cow/goat milk is significantly

lower in Sunamganj. The use of rice powder/soup is significantly lower in Kishoreganj; and the use of

potato and egg is significantly lower in Netrokona.

When comparing across Haor types, the use of Soji/Sagu/Barli is significantly higher in deep Haor;

and the use of cow/goat milk, rice powder/soup and fruits/juices is significantly higher in moderate

Haor. Qualitative data shows that nutrition-related information for newborns is obtained from the

village doctor and health assistant.


Table 53: Weaning foods used, by District and Haor type

Weaning Foods District Haor Type Total


N 196 166 202 271 293 564

Baby formula/Cerelac 3.4 3.1 2.1 4.3 1.7 2.9

Khichori 35.6 27.5 16.4 b

28.6 25.0 26.6

Soji/Sagu/Barli 37.7 39.7 25.7 b

44.9 b

25.9 34.3

Cow/goat milk 21.2 17.6 2.9 c 9.8 17.2

b 13.9

Rice powder/soup 58.9 a 69.5 77.1 64.3 71.6

c 68.3

Potato 21.9 4.6 c 17.1 15.7 14.2 14.9

Egg 8.9 0.8 c 3.6

b 3.2 5.6 4.6

Banana/other fruits and juices 28.8 22.1 17.9 19.5 25.9 b

23.0

Letters denote significant differences among Districts or between Haor types for a given weaning food. Significance levels

for comparisons: a = .10; b = .05; c = .00

Table 54 shows birth attendance during the last delivery, disaggregated by District and Haor type.

Overall, the majority of births were attended by Traditional Birth Assistants. Less than 1% of births

were attended by a doctor. It is important to note that in the qualitative data collection most women

reported that they had experienced a newborn die, many before the newborn was six months old.

Qualitative data further shows that local doctors and health workers are used for small complications;

hospital for serious complications. Most women have home births with traditional birth attendants or

family assisting, even though they recognize these individuals are often untrained. Few can save

money for emergency delivery. In the case of emergencies, community members will help to hire

transport and cover costs; neighbors will help with childcare. Some women reported being reluctant

to go to the hospital because doctors frequently will not attend to the ultra-poor. Many birth attendants

are untrained. Hygienic practices during and after delivery are uncommon in most areas.

Table 54: Who attended last delivery, by District and Haor type

Birth Attendee District Haor Type Total


N 197 166 203 273 293 564

Friend/relative 11.7 9.0 2.5 8.4 6.8 7.6

TBA 67.0 47.0 76.8 62.3 66.9 64.7

TTB 19.3 40.4 16.7 25.6 23.5 24.6

Doctor 0.5 0.6 1.5 0.7 1.0 0.9

FWV (nurse/paramedic/FWV) 1.5 3.0 1.0 2.2 1.4 1.8

Other 0.0 0.0 1.5 0.7 0.3 0.5

For households currently with a child 2 years of age or under, 81.8 % of the oldest child in this age

group has received at least one immunization. The proportions by District do not vary statistically

(p=.136) nor do they vary by Haor type (p=.484). For those children who did receive immunizations,

72.9% have immunization cards. These cards are significantly more common in moderate Haor but do

not vary significantly by District (p=.257). Of children 9 months and older, just over 50% were fully

immunized, as verified through their immunization cards.


Qualitative data shows that when health care workers come to villages, virtually all babies are

immunized and receive their full dose of vaccines, although most mothers do not know about the

different types of immunization or the benefits/risks of immunization.

For those children who needed antihelmintics, 47% received them. The proportion of children

receiving antihelmintics was significantly higher in Kishoreganj. Just over 47% percent of children

received vitamin A supplements in the last 6 months, whereby the highest proportion of children

received vitamin A supplements in Kishoreganj.

Table 55: Child health and immunization, by District and Haor type

Child health and

immunization



N 196 166 203 220 242 565

Oldest under 2 receiving at least

1 vaccination (%) 86.2 77.7 80.8 80.6 82.9 81.8

Proportion of those immunized

with immunization cards 72.8 68.2 76.8 79.8 c 65.5 72.8

Proportion of children 9 months

and older fully immunized 49.5 51.1 51.5 47.9 53.3 50.7

Proportion receiving

antihelmintics, if needed 39.0

c 23.4

b 31.5

b 30.9 32.4 31.7

Proportion receiving vitamin A

supplements 63.4

b 47.5 41.3 46.4 48.1 47.3

Letters denote significant differences among Districts or between Haor types for a given variable. Significance levels for

comparisons: a = .10; b = .05; c = .00

Table 56 describes the health issues of mothers with children under 2, disaggregated by District and

Haor type. Overall, only 7.4% of mothers reported suffering no illnesses in the last 12 months. The

highest proportion of women suffered from cold attacks, followed by gastric complications and

anemia. The lowest proportion of women suffered from Typhoid. When comparing across districts, the

proportion of women suffering from cold attacks, gastric complications and diarrhea is significantly

higher in Kishoreganj than in the other districts. The proportion of women suffering from anemia is

significantly higher in Netrokona; and the proportion of women suffering from cold attacks and

dysentery is significantly lower in Sunamganj than in the other districts.

When comparing across Haor types, the proportion of women suffering from four out of the seven

listed illnesses is significantly higher in deep Haor than in moderate Haor. In turn, the proportion of

women suffering no illnesses in the last 12 months is significantly lower in the moderate Haor.

Qualitative data showed that physical weakness, anemia, abdominal pain, back pain, fever, bleeding

and uterus complications are common for lactating mothers. Health care for complications are

commonly addressed locally. Few have the resources to seek treatment at the hospital.

Relatives/community members assist with household chores and childcare when mothers suffer

complications. Information about complications is commonly obtained from NGO and GoB health

workers, village doctor, TBA and village elders. Post-natal check-ups are rare due to limited

knowledge, resources, and communication challenges. In some cases, mothers report that their


mother-in-laws do not permit them to seek medical care. Use of Vitamin A supplements after giving

birth varies. In some villages it is not taken even when distributed by health workers. Ultra poor and

poor women do not take vitamin A supplements. There appears to be limited knowledge of the

benefits of these supplements.

Table 56: Health issues of mothers with children under 2, by District and Haor type

Illnesses District Haor Type Total


N 197 166 203 273 293 566

Proportion of mothers suffering

no illnesses 4.1 1.2 15.7

c 5.1 9.6

b 7.4

Cold attack 80.7 b

90.4 60.1 c 79.9

b 72.7 76.1

Gastric complications 32.5 b

21.7 24.6 26.4 26.6 26.5

Anemia 22.8 35.5 b

21.7 26.4 25.9 26.1

Diarrhea 24.4 b

16.3 13.8 22.0 b

14.7 18.2

Dysentery 20.3 17.5 11.8 b

19.8 b

13.3 16.4

Rheumatic fever 3.0 3.6 10.8 c 3.7 8.2

b 6.0

Typhoid 6.6 b

4.2 1.0 5.1 a 2.7 3.9

Letters denote significant differences among Districts or between Haor types for a given variable. Significance levels for

comparisons: a = .10; b = .05; c = .00

Table 57 describes the health issues of children under 2, disaggregated by District and Haor type.

Overall, only 5.8% of children did not suffer any illnesses in the last 12 months. The highest proportion

of children suffered from cold attacks, diarrhea and pneumonia. The lowest proportion of children

suffered from skin diseases and other illnesses. There are no significant differences between Haor

types. When comparing across districts, the proportion of children suffering from cold attacks is

significantly lower in Sunamganj than in the other districts; and the proportion of children suffering

from diarrhea and skin diseases is significantly higher in Netrokona.

Qualitative data shows that Children of two years commonly experience pneumonia, influenza,

typhoid, hepatitis, pneumonia, cold, fever, chicken pox, measles and diarrhea. TBA, village elders and

health workers provide advice to mothers about childhood disease and will lend money for children‟s

emergency treatment. In most cases there is no gender disparity with respect to health care for

children, although in some villages it is common for male children to receive foods of higher nutritional

quality and more of them.


Table 57: Health issues of children under 2, by District and Haor type

Illnesses District Haor Type Total


N 197 166 203 273 293 566

Proportion of children suffering

no illnesses 5.6 3.0 8.4

b 7.0 4.8 5.8

Cold attack 86.2 92.8 80.8 b

86.4 86.0 86.2

Diarrhea 34.2 48.2 b

32.0 36.3 38.7 37.5

Pneumonia 25.0 22.9 23.2 23.1 24.3 23.7

Dysentery 13.8 16.9 12.3 15.8 12.7 14.2

Skin diseases 7.7 13.3 a 4.4 8.1 8.2 8.1

Other 0.5 2.4 3.4 1.8 2.4 2.1

Letters denote significant differences among Districts or between Haor types for a given illness. Significance levels for

comparisons: a = .10; b = .05; c = .00

12.2 Anthropometric measurements

Table 58 shows the result of anthropometric measurements carried out with 398 children aged 6-23

months: 54% boys and 46% girls. The average age was 14.7 months, with no significant differences

among districts, Haor types or gender.

Weight-for-age (underweight) is a composite index of height-for-age and weight-for-height. A child can

be underweight for his/her age because s/he is stunted, wasted or both. Weight-for-age is a useful

tool in clinical settings for continuous assessment of nutritional progress and growth. Children whose

weight-for-age is below minus two standard deviations from the median of the reference population

are classified as underweight. Table 58 shows that 57.7% of < 2 children are underweight: 39.4% are

moderately underweight and 18.3% are severely underweight. A prevalence of > 30% is considered to

be „very high‟.

There is limited national data available for < 2 children; the majority of published data available for

comparison reflects anthropometric scores for < 5. However, the 2004 BDHS sample for < 2 children

in rural Bangladesh (< -2 SD) estimated stunting at 42.5%, wasting at 19.4% and underweight at

50.2%. Compared to such national data sets, the anthropometric scores for stunting and underweight

in the Haor region appear high.

When disaggregated by district, the proportion of children with moderate stunting is significantly lower

in Netrokona than in Kishoreganj. The proportion of children with moderate wasting is significantly

higher in Netrokona than in Kishoreganj, and significantly lower in Sunamganj than in Netrokona. The

proportion of children with moderate underweight is significantly higher in Sunamganj than in

Netrokona. When comparing across sex, the proportion of girls with severe stunting is significantly

lower than boys. There are no significant differences across Haor type.


Table 58: Anthropometric measurements

HAZ

stunting

WHZ

wasting

WAZ

underweight

Moderate Severe Moderate Severe Moderate Severe N

District

Kishoreganj 41.3 20.3 8.4 0.7 40.6 18.9 143

Netrokona 26.81 22.0 14.6

4 0.0 31.7 19.5 123

Sunamganj 34.8 25.0 6.83 0.0 45.5

3 16.7 132

Haor Type

Deep 34.7 22.5 9.2 0.0 36.4 19.7 173

Moderate 34.7 22.2 10.2 0.4 41.8 17.3 225

Sex of child

Male 32.1 26.5 9.8 0.5 38.6 19.5 215

Female 37.7 17.57 9.8 0.0 40.4 16.9 183

Total Sample 34.7 22.4 9.8 0.3 39.4 18.3 398 Notes: Moderate (-2.01 to -3.00 SD) Severe (< -3.00 SD) 1Netrokona different from Kishoreganj at 0.05 significance level

3Sunamganj different from Netrokona at 0.05 significance level

4Netrokona different from Kishoreganj at 0.10 significance level

7Female different from Male at 0.05 significance level

HAZ=Height-for-age z-score WHZ=Weight-for-height z-score WAZ=Weight-for-age z-score

The estimates in FSUP-H appeared reliable when compared with corresponding baseline estimates

from the SHOUHARDO anthropometric surveys for < 2 in the Haor area, assuming no overlap in

beneficiaries. Figure 21 shows that the 2006 pre-intervention values for the Haor region are very

similar to the FSUP-H pre-intervention findings for the same area. A comparison between

SHOUHARDO baseline and endline findings also shows the impact that effective interventions can

have on child malnutrition.

Figure 21: Comparison between SHOUHARDO and FSUP-H malnutrition levels


13 STATUS OF FEMALE-HEADED HOUSEHOLDS Table 60 provides an overview of key variables for female-headed households that provide a good

overview of their food security and livelihood status, as compared to male-headed households. Almost

15% of the households sampled had female heads of household. When comparing female- and male-

headed households across Districts and Haor type, the following observations can be made:

- female-headed households have significantly lower per capita monthly income levels than

male-headed households

- there are no significant differences in per capita expenditures between female- and male-

headed households, except in deep Haor areas where expenditures in female-headed

households are significantly lower

- female-headed households have significantly lower food consumption scores than male-

headed households

- female-headed households have a significantly higher coping strategy index score in

Kishoreganj and Sunamganj, and in deep Haor areas

Table 59: Key variables for female-headed households, by district and Haor type

Collective Actions


Kisho-reganj

Netro-kona

Sunam-ganj

Deep Moderate

N 628 634 630 947 945 1892

Female-headed HHs (%) 14.5 17.4 12.1 14.4 14.9 14.6

Monthly PC Income (Taka)– Male HHH 866 911 c 676 843

c 788 816

Monthly PC Income (Taka) – Female HHH 718 b

732 660 692 723 707 c

Monthly PC Expenditures (Taka) – Male HHH 1242 1426 1488 1361 a

1411 1327

Monthly PC Expenditures (Taka) – Female HHH 1158 1344 1096 1123 1303 1395

Food consumption score – Male HHH 9.2 c 8.3

b 7.6 8.6

c 8.1 8.3

c

Food consumption score – Female HHH 8.4 7.7 7.2 7.7 7.9 7.8

Coping strategy index – Male HHH 24.2 22.6 24.3 23.9 c 23.5 24.2

Coping strategy index – Female HHH 26.3 c 22.4 25.8

c 25.7 23.6 23.5

Letters denote significant differences between gender of head of Household. Significance levels for comparisons: a = .10; b = .05; c = .00


14 CONCLUSION AND RECOMMENDATIONS The baseline study findings provide important information, which can be used by FSUP-H partners to

measure impact-level changes in food security and livelihood trends over time. An overview of the

relevant findings for the FSUP-H baseline logframe indicators that state baseline and endline surveys

as means of verification is provided in table 60 below, including recommendations to improve the

indicators.

Table 60: Baseline values and recommendations for FSUP-H logframe indicators

Objective indicators Relevant baseline finding Recommendation

Overall Objective: At least 40%

of the targeted 55,000 ultra-

poor women have graduated

out of extreme poverty

/ Index to be developed by CARE

Bangladesh using the livelihoods

and economic security baseline

values

Specific Objective: Prevalence

of chronic malnutrition among

women has decreased by <to

be determined by baseline> %

by 2013

Anthropometric findings for < 2

Stunting: moderate 34.7 / severe 22.4

Wasting: moderate 9.8 / severe 0.3

Underweight: moderate 39.4 / severe 18.3

There was no malnutrition

measurement for adult women.

Revise this indicator to < 2.

Specific Objective: At least 70%

of households reported at least

3 meals/day, including during

lean periods

The mean value for households that take 3

meals/day „most of the time‟ in the last 12

months is 14% („most of the time‟ and often

combined is 56.3 %)

If CARE takes the 14% as the

baseline value, then the 70%

target is likely too high.

Specific Objective: Reduced

asset loss due to improved

resilience to natural disasters

and shocks

The mean asset loss per disaster, among

households who experienced a disaster in the

last 12 months, was reported at around Taka

3,017.

CARE needs to determine a

realistic target based on intensity

of coverage by project activities

Indicator Result 1 Relevant baseline finding Recommendation

At least 70% of individuals are

able to negotiate access to

services with local government,

service providers and local

leaders (in the areas of health,

livestock, agriculture, fisheries,

social protection)

68.7% of households had accessed one or

more GoB service providers in the previous

year

This indicator should be

reformulated around actual

levels of access, as ability to

negotiate is hard to measure and

is a lower-level indicator. To

make the indicator more

meaningful, CARE could

consider reformulating the

indicator around particular

services (see table 40), which

are expected to be the focus of

project interventions.

At least 30% of women

participate in any of the

following: UP standing

committees, SMC, PTA and

local arbitration in project areas

Participation in the development process was

low at 4.5% of all households. The response

rate for this variable was too low for meaningful

analysis. Regarding participation of women: a

total of 186 responses were given from 174

households – 9.2% of respondents. Females

(spouses plus female heads of household)

accounted for 15.1% of 9.2%, which is about

1.6% of the overall population.

If CARE wants to keep this

indicator, we would suggest

stating the baseline value as 1%.


Indicators Result 2 Relevant baseline finding Recommendation

% of women have increased

income, particularly through

rural sales networks and

assemble markets

Income data was collected at the household

level. Per capita monthly income of female-

headed households is 707 Taka.

Reformulate this indicator to

capture per capita monthly

income of female-headed

households

At least 40% of women from

ultra-poor households reduced

debt from unsustainable

sources (particularly

moneylenders)

Almost all women (98%) had taken a loan from

the Grameen Bank, which reflects the

Grameen Bank‟s policy of lending to women.

The proportion of women who took a loan from

NGOs is also high (88%), for similar reasons.

The proportion of women taking loans from

moneylenders is the lowest among all loan

sources.

Perhaps this indicator could

better be formulated around

(female-headed) household debt

burden. If the indicator is not

changed, then CARE should

consider which loan sources

qualify as unsustainable.

% in productive utilization of

income (% of expenditure on

assets, % of expenditure on

food)

Overall, daily expenditure on food is 113 Taka.

Less than 1 in 10 households own productive

assets: an overview of household productive

asset ownership is provided in table 18.

It is recommended to split this

indicator into two separate ones:

productive asset ownership and

daily food expenditure.


No indicators measured by

baseline, as per FSUP-H

logframe

Based on the baseline findings, the following indicators are suggested for tracking

as proxies for household resilience to natural disasters and household crises:

Mean value of asset loss (baseline = 3,017 Taka)

Mean number of working days lost (baseline = 10 days)

Combination of coping strategies applied by households (baseline values =

see sections 10.1 and 10.2)


% of households reporting

increased food consumption

and improved dietary diversity

This survey utilized the Food Consumption

Score (FCS) to measure food consumption and

dietary diversity: 16.2% of households had

poor FCS, 31.5% had borderline FCS and

52.3% had acceptable FCS.

CARE needs to determine a

realistic target based on intensity

of coverage by project activities

Improved infant and young child

feeding practices in 80% of

VDCs (including exclusive

breastfeeding, early initiation of

breastfeeding, weaning)

Overall, 45.4% of mothers initiated

breastfeeding with the 1st hour.

26.9% of mothers started weaning between 0-3

months, 48.1% between 4-6 months, 21.4%

between 7-9 months and 3.6% after 10

months.

It is recommended to

reformulate this indicator as two

separate indicators around

breastfeeding and weaning.

% of household reporting

reduced prevalence of diarrhea

22.9% of household reported diarrhea in the

last 12 months

/

At least 80% of pregnant

women from ultra-poor

households received

appropriate supplements (i.e.

folic acid, iron and vitamin A)

from government health

services

Overall, 35.5% of mother‟s took iron or folic

acid supplements. No data was collected on

vitamin A intake of pregnant women, only for

children.

It is recommended to

reformulate the indicator around

iron and folic acid supplements.

_______

FSUP Baseline Report Final 23 June

Documents

study methods

fsuph project

quantitative study design

qualitative study design

fsuph site

food security

x executive summary

x glossary of english