Measurement of Poverty: a Case Study of Pakistan

1
Measurement of Poverty a Case Study of Pakistan

• Ambar Narayan
• (World Bank)
• Regional Poverty Analysis and Monitoring Workshop
• Islamabad, Pakistan
• March 2002

2
Role of Poverty Lines

• Poverty lines have an element of subjectivity
• E.g. the World Banks 1/day (PPP) poverty line
  useful for comparing across countries
• National poverty lines better suited to
  individual countries
• Consistency in defining poverty lines needed for
  exploring poverty trends
• Poverty lines only useful to measure consumption/
  income poverty
• Non-income dimensions equally important to
  measure capability

3
Outline of Presentation

• Choice of poverty lines
• Important issues different approaches
• Poverty lines used for Bank poverty studies for
  Pakistan
• Why the choice of this line
• Need for consistent poverty lines to examine
  trends
• Important issues in poverty measurement
• E.g. consumption aggregates, price adjustments
• Some results from poverty measurement
• Poverty trends issues of comparability between
  surveys
• Analysis using poverty estimates a few examples

4
Choice of Poverty Lines

• Setting poverty line involves
• Defining a minimum bundle of goods/services
• identifying the cost of a minimum bundle
• Standard approach define household as poor if
  their expenditure is insufficient to consume the
  minimum bundle
• How to define the minimum bundle
• Calorie-based approach minimum calorific
  requirements are converted into minimum
  expenditure requirements
• Basic-needs approach cost of achieving a minimum
  bundle of basic needs -- calorific needs and
  other purchasable needs such as fuel, housing and
  clothing

5
Expenditure-Based Poverty Lines for Pakistan

• Calorie-based approach looking at food
  expenditures only
• Havinga et al (1989), Mahmood et al (1991)
• Cost of basic needs approach incorporating food
  and non-food expenditures
• Ercelawn (1991), Jafri (1999), FBS (2001)
• Fundamental similarity between these approaches
• The poverty line expenditure level corresponds to
  the average or estimated expenditure of
  households which consumes the minimum calorie
  intake

6
Expenditure-Based Poverty Lines (Contd.)

• Why should the cost of minimum bundle depend only
  on calorie intakes ?
• Food is the most important need ?
• Calorie intakes (actual/expected) is a good proxy
  for all basic needs ?
• Non-calorific needs hard to specify ?
• Alternative poverty line is the money value of a
  bundle of commodities that represents a minimally
  acceptable level of living selected a priori
• Gazdar et al (1994), World Bank (1995) modified
  from Ahmad (1993)

7
Poverty Line Used by Bank Studies

• Defining a basic needs bundle (like Ahmad, 93)
  requires intimate knowledge of the country
• Pakistan one of the few developing countries
  where this exercise has been undertaken
• Comprehensive, but somewhat subjective process
• The proposed basic needs package consists of
  food, clothing, housing, health, education,
  transport, social interaction and recreational
  facilities...Discussions were held with
  professional economists in Federal Government,
  Provincial Governments, Research Institutes and
  Universities. A check list thus prepared was
  rechecked with heads of different families. A
  team of economists was constituted to arrive at
  the quantum and value of each componential item
  of various basic needs separately in the rural
  and urban areas. These were rechecked with the
  consumers in different areas (Ahmad, 93 p28)

8
Poverty Line Used by Bank Studies

• Ahmads poverty line for a family of 2 adults 4
  children (91-92 prices)
• Rural Rs. 300 per capita out of which Rs. 150
  allocated to food needs
• Urban Rs. 419 per capita out of which Rs. 212
  allocated to food needs
• Adjustments made by Gazdar et al (1994) using
  HIES (90-91)
• Reducing the urban-rural food price differential,
  using cost-of-living price deflator
• Adjusting housing expenditures by estimating
  expected rural and urban housing expenditure,
  given non-housing expenditure equal to that in
  the minimum required bundle
• With adjustments, rural poverty line falls to Rs
  296 per capita the urban line falls to Rs 334

9
Further Adjustments to Poverty Line

• Poverty lines have to be adjusted for
  inter-temporal price changes, and inter-province
  price differences
• Inter-temporal price changes we have used
  inflation rates from CPIs
• CPIs are only available sources for non-food
  prices
• Food inflation rates using household data over
  the years do not yield substantive changes to
  poverty lines
• Poverty lines not adjusted for inter-province
  price differences
• Household expenditures adjusted for price
  differences instead
• Poverty lines adjusted by equivalence scale
• Into per equivalent adult terms, using Ahmads
  equivalence scale (1 child 0.8 adult)

10
Poverty Lines (Per Equivalent Adult) in Current
Rupees
1990-91 1992-93 1993-94 1996-97 1998-99
Urban 346 424 472 655 767
Rural 307 376 418 581 680
11
More on the Basic Needs Poverty Line

• Comparison with calorie measures
• Regression analysis (for 90-91) reveals that the
  rural and urban poverty lines correspond to per
  capita calorie intake of 2250 (for a family of 2
  adults and 4 children) and 1950 respectively
• Comparison with consumption patterns
• Expenditure shares for food, housing and other
  (90-91 HIES) for households 10 on either side of
  the poverty line, compare well with the shares
  implied by the construction of the poverty line
• Comparison with other poverty lines
• Compares well with Naseem (77), Lanjouw (94)
  10-20 higher than lines by Malik (94), Ercelwan
  (91)

12
Constructing Expenditure Aggregates

• Include those expenditures that correspond to
  basic needs
• Include expenditures on flow of utilities only
• Decisions on what to include and what not may
  involve subjective judgments
• Critical to maintain consistency in methodology
  over time (across different household surveys)

13
Consumption Aggregates in our Study

• Following items are included in household
  consumption aggregates to determine poverty
  status
• Food Expenditure on consumed food items,
  irrespective of whether they were paid for in
  cash or not
• Non-durable goods E.g. Fuel and lighting,
  personal care articles and services, education,
  health, recreation and reading, personal
  transport and traveling, household laundry,
  cleaning, apparel, footwear, housing, maintenance
  and repair charges of household effects
• Important items excluded
• Expenses on house, property, or any other tax,
  fines
• Purchase of durable goods
• Transfers paid out by household members

14
Adjusting Consumption Aggregates for Price
Differences

• Methodology adopted
• Use food prices to construct household specific
  price indices, separately for rural/urban
• Deflate household expenditures by corresponding
  price indices
• Implicit assumption that differences in food
  prices captures cost of living differences to a
  high degree

15
Comparing Poverty Across HIES (1990s) and PIHS
(98-99)

• Factors in favor of comparability of data across
  years
• Consistency in poverty lines and methodology for
  measuring household expenditures
• Consumption module almost unchanged across
  surveys
• Sample sizes and sampling methodology mostly
  similar since 92-93
• Factors that may compromise comparability across
  years
• Change in recall period for some food items in
  PIHS 98-99
• Change in sampling framework for rural areas in
  PIHS 98-99
• Average household sizes higher in PIHS 98-99
• Differences large for lower expenditure deciles
  in rural areas
• Especially large differences for rural Sindh and
  Balochistan

16
Average Rural Household Sizes Across Surveys
Per Capita Exp Deciles HIES 1992-93 HIES 1993-94 HIES 1996-97 PIHS/HIES 1998-99
1 8.4 8.2 8.5 9.1
2 8.0 7.9 7.5 8.5
3 7.8 7.6 7.1 7.9
4 7.3 7.4 7.1 8.0
5 6.9 6.9 6.6 7.6
Total 6.3 6.3 6.1 6.8
17
Our Poverty Estimates(FBS Estimates in
Parentheses)
18
Poverty Trends in Pakistan
19
Analysis with Poverty Estimates Examples

• Sensitivity analysis how poverty estimates
  respond to shifts in poverty line
• Additional insights from looking at distribution
  of population around the poverty line
• Relating poverty to growth and redistribution
• Growth-inequality decomposition of changes in
  poverty rates
• Measuring pro-poor growth
• Non-income/consumption dimensions of poverty
• Find associations between consumption poverty and
  indicators of capability
• Analysis of vulnerability (the prob. of falling
  into poverty)
• Measuring vulnerability finding factors that
  determine vulnerability
• Needs panel data ideally cohort level analysis
  also possible using repeated cross-sections, like
  HIES PIHS

20
Sensitivity of Head Count to Poverty Line - I
21
Sensitivity of Head Count to Poverty Line - II
22
A Snapshot of Vulnerability? Distribution of
Population Around the Poverty Line
23
Growth-Inequality Decomposition of Changes in
Poverty Estimates

• Measures the impact of changes in mean
  consumption and distribution on poverty
• Growth component how much of the change in
  poverty measures is due to variation in mean
  expenditure (per equiv adult) over time, holding
  the distribution constant.
• Redistribution component how much of the change
  in poverty measures is due to a change in the
  distribution of expenditure, holding the mean
  expenditure constant

24
Growth-Inequality Decomposition of Changes in
Poverty Measures between 90-91 and 98-99
25
Final Messages

• Choice of poverty line remains a largely
  subjective judgment
• That said, any approach incorporating basic needs
  (beyond calories) seems preferable in our opinion
• More than the exact line that is chosen, it is
  important to decide on a national poverty line
• Choose a benchmark poverty line and update it
  only for inflation
• Comparability of surveys across years critical
  for examining poverty trends
• Dependent on a number of factors, e.g.
  consistency in sampling framework,
  questionnaires, methodology of field survey
• Poverty estimates useful for a broad range of
  analysis
• Important to relate consumption poverty measures
  to non-income/expenditure dimensions human
  deprivation