Measuring income poverty in Bangladesh

Measuring income poverty in
Bangladesh
Hassan Zaman
Senior Economist
World Bank
Presentation structure
o World Bank analytical work and TA on
poverty measurement
o Approaches to poverty measurement
o Poverty estimation with HIES 2005
TA for BBS for Poverty Monitoring
o Longstanding relationship with
poverty section of BBS since early
nineties
o Grant financed HIES 2005 and will also
finance a safety net survey
o World Bank staff conducted refresher
training on poverty analysis in Nov
2005 at BBS
Poverty Assessment Report
o One important part of this on-going
poverty work is a Poverty Assessment
report.
o The last Poverty Assessment Report
2003 for Bangladesh, in conjunction
with ADB, was based on HIES 2000
Using HIES 2005 for policy analysis
Very preliminary issues that next Poverty
Assessment could examine include:
o Targeting of public expenditures (incidence
analysis)
o Relationship between growth, inequality
and poverty
o Remittances
o Poverty and social impact of specific
interventions (e.g. oil price hike)
Different Approaches to Poverty Measurement
o Direct Calorie Intake (DCI)
o Household is poor if its per capita calorie intake is
less than the standard per capita nutritional
requirement - 2,122 kcal per day.
o Best used to measure under-nourishment as it
doesn’t include non-food items
o Food Energy Intake (FEI)
o Food poverty line is the monetary value of the
food expenditure that allows households to just
meet the stipulated calorie requirement.
o The FEI is normally derived through regression of
the relationship between calorie intake and
expenditure.
o However, rural households are more willing to
consume food that is cheaper per calorie. This
could result in urban households appearing to be
poorer than rural households even if in fact they
are better off.
Different Approaches to Poverty Measurement
o Cost of Basic Needs (CBN)
1. a basic food basket is identified from the
data, consistent with consumption patterns
2. the quantities in the basket are scaled
accordingly to correspond to the nutritional
requirement
3. the cost of acquiring the basket is
calculated.
This results in the food poverty line.
o A non-food poverty line is calculated by
estimating the cost of consuming a basic set of
non-food goods for (i) extreme poor
households whose total expenditures equals
the food pov line (ii) moderate poor hhs whose
food expenditure is at food pov line.
Contd.
Poverty trends 1991-2000
o World Bank/BBS estimates are based on the Cost of Basic Needs
(CBN) method.
o The base year for the CBN poverty line was 1991-92, which was
then updated for 1995-96 and 2000 for changes in the cost of
living using a price index.
o Difference with Sen/Mujeri estimates is the use of unit record data
vs grouped data and different non-food poverty line as base
year.
Bangladesh: Poverty Headcount Rate
Upper Poverty Line
1991-92
1995-96
Lower Poverty Line
2000
1991-92
1995-96
2000
National
58.8
51.0
49.8
42.7
34.4
33.7
Urban
44.9
29.4
36.6
23.3
13.7
19.1
Rural
61.2
55.2
53.0
46.0
38.5
37.4
Updating vs. Re-Estimating the Poverty Line using HIES 2005
o Merits of simple updating of the poverty line
o The PL can be updated using inter-temporal
price indices
o Updating the PL allows for easy comparison
across years since the same reference
basket is being used.
o Merits of re-estimating the poverty line
o Consumption patterns change over time and
commodities may need to be added to or
dropped from the reference bundle
o Relative prices change over time and the
original reference bundle may not reflect the
optimum consumption choice of those near
the poverty line
Updating vs. Re-Estimating the Poverty Line using HIES 2005
o There are some problems with simply updating
the PL in Bangladesh:
o Sampling frame differences make inter-
temporal price indices harder to estimate.
o 2005 food bundle for the poor could well
be different from current food bundle.
o Main problem with re-estimating the PL is that
MDG and PRSP targets are based on data
corresponding to the old PL.
o This problem can be addressed by estimating
poverty measures with both the “old” and “new”
poverty lines.
Contd.
Next steps
o When full data set is available we will check
how existing food bundle corresponds with
2005 bundle for the poor
o We will check how inter-temporal price
indices can be constructed despite new
sampling frame
o We will hold discussions with professionals in
this field to seek advice and build
consensus