Food Security Assessment in Bangladesh

1
Food Security Assessment in Bangladesh
Issues and Implications for Mapping Food
Insecurity and Vulnerability
2
Programme

1. Geographic targeting
2. Data sources for food security analysis and
   mapping
3. VAMs strategy for a new set of targeting maps
4. Discussion

3
Geographic targeting

• WFP resources to be concentrated in areas
  identified through VAM assessments of food
  insecurity and vulnerability.
• Considerable variation in food consumption
  between districts. (E.g. range from 2470 k.Cal in
  Dinajpur to 1819 K.Cal. in Bagerhat)
• Explained by regional difference in poverty
• Poor areas are not just poor because poor
  households concentrate in these areas (WB)
• Geographical indicators can provide indication
  where to find a concentration of higher levels of
  poverty. (land productivity, electricity,
  infrastructure, distance to major markets etc.)

4

• One of the most widely used maps for targeting
  of development programmes
• Four relative categories of food insecurity
  very high, high, moderate, low
• Composite index of 7 proxy indicators
• incidence of natural disasters
• Food grain deficit/surplus
• Agricultural wage rate
• Households not owning land
• Unemployed persons
• Female-headed households
• Literate women
• Scope for improvement
• Type of indicators used
• Double counting
• Availability of new data

5

• Based on HES 1995/96 data
• 7 indicators
• Income and assets
• per capita income
• per capita income from own agricultural
  production
• Land owned
• Food intake
• Daily calorie intake
• Percentage of calories from cereal (Q)
• Expenditure
• Per capita expenditure
• Share of expenditure on food
• Limitations
• Additional outcome indicators
• rational for targeting or not targeting certain
  areas is not clear
• level of targeting at cluster level

6
Related mapping efforts

• World Banks Poverty targeting exercise (HES
  1995-96)
• Bangladesh human development report
• Unicefs child risk measure
• IRRIs study into the geographical distribution
  of poverty and food security

• NW - Kurigram, Lalmonirhat, Nilphamari,
  Panchagarh
• North Jamalpur, Mymensingh, Sherpur, Tangail.
• South of Dhaka Faridpur, Gopalganj, Madaripur,
  Rajbari, Shariatpur
• Barisal Division

• Jamalpur, Sherpur, Kusthia

• Sunamganj, Bhola

7
(No Transcript)
8
Data sources for food security analysis and
mapping

• Household Income and Expenditure Survey, 2000
• Poverty Monitoring Survey, May 1999
• Child Nutrition Survey, 2000
• Demographic and Health survey, 1999-2000
• Multiple Indicator Cluster Survey, 2000
• Nutrition Surveillance Project, HKI/IPHN
• Census of population, 2001
• Census of slum areas and floating populations,
  1997
• Census of Agricultural, 1996

9
VAMs strategy for a new set of targeting maps.
Considerations

• Level of targeting
• Incidence of chronic food insecurity, but
• Nature of the development programme
• Accessibility and cost-effectiveness
• Presence of partner capabilities
• Political and security considerations
• Geographic targeting vs beneficiary targeting
  (access to land, gender, source of income)

10
(A) Targeting maps for VGD IFS
Approach 1 Mapping of food security indicators

• 1st tier - Based on indicators of extreme
  poverty (food intake, income, expenditure, assets
  and nutrition)
• 2nd tier Based on household (e.g. quality of
  housing, source of income, landownership,
  literacy, etc.) and area characteristics (e.g.
  rural electricity, land-use, infrastructure etc.)
• Ranking of areas based on a combination of
  methods (scoring, indices, principle component)

• Poverty/lack of sufficient purchasing power is
  main obstacle to achieving food security
• 44 percent falls below the absolute poverty line
  (2,122 K.Cal/day)
• 20 percent (25 million people) consume less then
  1,805 k.Cal/day

11
Rural Poverty by greater district
12
(No Transcript)
13
Approach 2 Small area estimation
Steps
Survey data Census data
Calorie intake Literacy
Literacy Landownership
Stunting/underweight Sanitation facilities
Landownership Employment
Sanitation facilities Household size
Housing material Housing material
Enrolment
Expenditure on food

1. Regression model for each region (urban rural)
   using survey data
2. Estimate small area indicators of food insecurity
   by substituting survey data for census data
3. Ranking of areas based on level of food insecurity

14
Analysis of the differences between approach 1
and 2

• Compare and validate different approaches and
  outcomes
• Identify the underlying factors that were most
  significant in explaining the incidence of food
  insecurity
• Overlay food insecurity maps with area resource
  maps (e.g. differences in operational costs and
  accessibility, partners presence and capacity)
  for additional insight into where to target WFPs
  development programmes.

15
(B) Targeting maps School Feeding Programme

• Urban and rural targeting maps based on
• Food Insecurity and Nutrition (food consumption,
  anthropometric data)
• Primary education (no. schools, enrolment,
  attendance, teacher/pupil ratio, drop-out.
• Local partners capabilities

16
Rural targeting
Source Multiple Indicators Cluster Survey, 2000
17
Urban targeting

• HIES 2000, poverty in urban areas increased with
  7.2 percent
• Nutritional status of children living in urban
  slum areas is worse than in rural areas (stunting
  (54.1 52.9, Underweight (72.6 59.3), CNS
  1995/96
• Literacy rate in urban slums (14.7) is more than
  half the national rate found in the 1991
  population census (32.4)
• HKI. Wide differences in living conditions
  between slums within a city. Percentage of
  underweight mothers ranging from 10 percent in
  one slum to 45 percent in another in August 2000.

• Map based on proxy indicators obtained from
  Census of slum areas, 1997 (slum characteristics,
  literacy, employment, housing, toilet facilities,
  source of drinking water, land ownership,
  electricity).
• Mapping of partner presence and capabilities

18
(C) Vulnerability maps for emergency operations

• Vulnerability map based on
• Identification of the most physically affected
  thanas based on historical disaster data (DMB
  CAREOFDA/CRED)
• Coping potential as indicated by level of poverty
• Existing WFP and partner programmes in these
  areas
• Scenario maps, estimate likely impact on
  household access to food.

• OFDA/CRED International disaster database
  registered 170 events during the period 1975-2000
• Vulnerability F(fs, exposure to shocks,
  ability to cope)

19
Thank you!