MEASURING AND MONITORING POVERTY IN ETHIOPIA

1
MEASURING AND MONITORING POVERTY IN ETHIOPIA

• Tassew Woldehanna
• Department of Economics
• Addis Ababa University
• Ethiopia

2
Content

• Background
• Measuring Poverty
• Monitoring poverty periodic and between surveys
• Mixing Qualitative and quantitative approaches
  to poverty assessment
• Concluding remarks

3
1. Background

• Formulation and evaluation of PRSP is constrained
  by inadequate capacity to carry out surveys,
  quick and timely analysis and monitoring of
  poverty
• Ethiopian PRSP is named as Sustainable
  Development and Poverty Reduction Program - SDPRP
• SDPRP was prepared in August 2002.
• Will be guiding till 2004/05
• It is the main policy document refereed by NGOs,
  regional/state government, parliament, government
  ministries
• This has to be updated continuously
• Hence evaluation is required on the process of
  formulation and monitoring of PRSP

4
Background

• SDPRP objective is to reduce poverty and ensure
  food security via economics growth to be achieved
  via free market.
• SDPRP is built on four pillars
• ADLI and food security (FS)
• Justice systems and Civil service reform
• Decentralization and empowerment
• Capacity building
• The main focus is on the ADLI strategy and food
  security as a means of reducing poverty, but the
  latter three are designed to enhance the
  effectiveness of ADLI and FSS
• SDPRP assumes -labor is abundant, and capital and
  land constrains labor intensive technology,
  land augmenting technologies, capital saving

5
Background

• SDPRP identifies four pro-poor sectors
• Water
• Agriculture and food security
• Education
• Health
• Road
• There is an explicit budget and monitoring plan
  for this four sectors

6
Background consultation

• Consultations were done at Woreda/district/,
  regional/state/, and federal levels
• 117 Woreda and 6000 people participated
• Consultations involved high government officials,
  sectoral bureaus, journalists, religious leaders,
  representative of donor communities, and
  professional associations and the business
  community
• Other forums
• pastoral development forum by state
• Civil organization EEA, FSS
• The consultations was done freely without
  government moderators (Joint IDA-IMF staff
  assessment)

7
(No Transcript)
8
Background consultation

• But consultation was made to react on government
  existed policies instead of crafting new ones.
• conflicting ideas during the consultations e.g.
  b/n business community (pro-urban) and
  Woreda/regional level consultations (pro-rural)
• Is RSP a good generalization of ideas raised
  during the consultations or simply a summary of
  sectoral programs?
• SDPRP has influence the budget allocation in
  2003/04 food security, pro-poor sectors

9
Background institution.

• WMU of the MOFED is the main unit responsible for
  the day to day work related to PRSP monitoring
  and implementation
• WMS established long time before the PRSP(1996)
• Revitalized during SDPRP preparation
• WEM COM poverty oriented sector ministries
  chaired by minister for MOFED
• Provide over all guiding
• WEM TEC Chaired by the Economic Policy and
  Planning Department of MOFED
• supply information technical advise to WMU

10
Background sources information.

• Survey data by CSA
• Administrative data sector ministries
• CSA is the main data collecting authority
• It has its own mid term plan and data collection
  and dissemination regulations
• Recently the CSA has opened a new department to
  handle surveys directly related to poverty
  measurement HICES and WMS, PPA

11
Background sources information.

• Capability of CSA
• Split of population census office
• Good (long-run)
• Bad (short-term)

12
2. Measurement of poverty

• SDPRP adopts WDR(200/01) definition of poverty
• lack of opportunities (material deprivation)
• lack of capabilities (low achievement in
  education and health, malnutrition)
• Vulnerability (low level of security)
• voicelessness (and powerlessness
• While the first and the second are more or less
  well measured, the third dimension was not
  appropriately measured.
• The fourth dimension was not measured at all

13
Measurement of poverty

• Choice of definition of poverty is based on
  experts opinion, not guided by the PRSP process
• Income dimension of poverty is measured based on
  non-welfare approach (based on the expenditure
  required to cover daily calorie requirement (2200
  kcal/day/adult) and essential non-food items
• The poverty line in Ethiopia is 1075 Birr/adult
  in 1995/96 price (??? USD)

14
Measurement of.
15
Measurement of.

• Foster, Greer and Thorbecke (1984) P?-measures of
  additively decomposable poverty measures is used
• Two main data are used for measuring poverty
• Household income and consumption expenditure
  Survey (HICES)
• Welfare Monitoring Survey (WMS)

16
Measurement of.

• HICES provides information on
• consumption of food and non-food item, household
  expenditure,
• payments, receipts and income, and
• household characteristics such as family
  composition, education and occupation.
• used to measure the income dimension of poverty.

17
Measurement of.

• Welfare Monitoring Survey (WMS)
• conducted to assess the non-income dimension of
  poverty such as education, health, malnutrition
  and vulnerability
• It provides extensive information on
• access to education and health facilities
• achievements in education, anthropometric
  measures, and
• underlying asset bases of the poor and
• the opportunity available to households

18
Measurement of.

• Consumption (as opposed to income) is used to
  measure poverty indices
• Income is underestimated in HICE survey data
  less than consumption expenditure for about 70
  of the sample households
• Saving negative (not real) impossible to measure
  saving
• Consumption expenditure is collected based on
• Weekly recall for food, drinks and tobacco and
  six months recall for non-food expenditure
• Non-food expenditure include durable consumption
  goods

19
Measurement of.

• Given households are poor and have limited
  transaction, it is not difficult to remember
• Durable goods are included as expenditure
  (overestimates consumption)
• Fuel wood, water collected from common resources
  are imputed (hence consumption reflect welfare)
  using local prices
• Seasonality of food availability is well
  considered
• HH visited for 8 weeks (4 during January/February
  better food availability) and 4 weeks during
  June/July - less food availability)

20
Measurement of.

• 252 food items and 600 non-food items
• Food items
• pulses oil seeds cereals preparations bread
  and other prepared foods meat fish milk,
  cheese and egg oils and fats vegetables
  fruits spices potatoes and other tubers
  coffee, tea and buck thorn leaves salt, sugar
  and others food taken away from home and milling
  charges

21
Measurement of.

• Non-Food items
• beverages cigarette and tobacco clothing and
  footwear house rent, construction materials,
  water, fuel and power furniture, furnishing,
  household equipment medical care and health
  transport and communication recreation,
  entertainment and education personal care and
  effects and miscellaneous non-food goods

22
Measurement of.

• Real consumption (as opposed to income) is used
  to measure poverty indices
• Income is underestimated in HICE
• temporal price index
• Deflated by regional price index
• Laspeyres price index reason given- simple,
  transparent and easily explained to policy
  makers
• Calculated in terms of adult equivalent (gender
  and age based equivalent scale developed by WHO)

23
Measurement of .price data

• There are two sources of price data in Ethiopia
• Internal price data CSA call it standard prices
• Independent price survey
• Internal price data expenditure/quantity, but
  CSA calculation of consumption expenditure is .
• Reflect the local condition and price faced by
  poor

24
Measurement of .price data

• Independent price survey by CSA
• Conducted in selected markets (a mix of small and
  major towns) using price questionnaire.
• The CSA reports the prices of food and non-food
  items for each zone and major towns in Ethiopia
  both quarterly and every year. The problem with
  this kind of price information is that it is
  difficult to match price from the survey (price
  report) with the expenditure pattern of
  households in the HICE survey

25
Measurement of .price data

• The poverty analysis in the PRSP document used
• internal prices for all food items and few
  non-food items (beverage, alcohol, ..
• external price (independent price survey) when
  internal price data are missing (for few food
  items and for most non-food items) in the HICE
  data

26
Measurement of .price data

• The poverty figures are highly consumed by
  government and non-governmental organizations for
  their projects
• E.g. regional budget subsidy uses a formula for
  allocating regional budget subsidy population
  (55), poverty (10), expenditure need (20) and
  revenue raising effort (15)
• However, it is only the head count index

27
3. Monitoring poverty

• The experience of CSA - not more than 10 years.
• The WM surveys has been conducted in 1995/96,
  1997, 1998 and 1999/2000 and
• two HICES data were collected in 1995/96 and
  1999/2000.
• Hence HICES is a sub-sample of WMS

28
Monitoring poverty

• While WMS represents administrative zones, HICES
  represents rural and urban regions and major
  urban areas.
• In 1999/2000, the WMS collected data from 25,917
  households and HICES collected data from 17,332
  households.
• Both surveys match for about 16,672 households.
• This year both the WMS and HICES will be
  conducted

29
Monitoring poverty

• Apart from these surveys, there are other surveys
  that can be used for poverty monitoring.
• These surveys include
• labor force survey (1999),
• agricultural Sample Census (2002), and
• Household Demographic Survey, HDS, (1999/2000),
• crop survey

30
Monitoring poverty

• Non-of the data collected by CSA are panel
• They have not yet considered to collect such data
• Hence difficult to measure vulnerability
  relatively accurately and and dynamics of
  poverty

31
Selected Welfare Monitoring Data of Ethiopia A.
Nation-wide official data
32
Selected Welfare Monitoring Data of Ethiopia A.
Nation-wide official data
33
Selected Welfare Monitoring Data of Ethiopia
Independent quantitative source
34
Other More Frequent Monitoring Mechanisms
Governmental

• Health Sector HMIS compiles clinic and health
  records from facilities
• Education Sector EMIS compiles educational
  statistics (enrollment, sections, repetitions,
  promotions, calculates drop-out rate, gender
  school to WEO to REB to National/Federal)
• Agricultural Statistics crop production and
  yields (sample), land cultivated, land irrigated,
  some on prices
• National Accounts Ministry of Finance
  accounting
• Annual Progress Report on PRSP
• Regional Plans, Budgets, Programs accounting
  at regional level
• (capacity building still in progress)

35
Rapid Rural Monitoring

• participatory constituency-based monitoring
  (indicators indicate)
• Expenditure Monitoring done by CSO/NGOs (both
  foreign and domestic), usually at lower level
  (Regional, Zonal or Woreda)
• Non-governmental Organizations (NGOs) and
  Community-Based Organizations (CBOs) projects,
  studies, canvass membership on issues
• Academic Institutions on-going work, student
  thesis, think tanks
• Minutes of Local Government Assembly Meetings
• Peasant Association Interviewing (Kebele or
  Village Level) implementation of goods and
  services provision, issues of agricultural support

36
Rapid Rural Monitoring (Cont)

• Service Provider Records and Reports clinics,
  schools, cooperatives
• Customer Service Satisfaction Surveys quality
  of services
• Citizen Report Cards similar to above, but on
  issues of governance and response
• Food Early Warning System (FEWS) keeps track
  of rain, production, pests (produces report at
  least monthly)
• Focus Group Discussions on selected topics

37
4. Qualitative vs quantitative

• Many of the official surveys in Ethiopia are
  quantitative
• Some include qualitative questions wms, ERHS
• Only few surveys qualitative
• Consultation with the poor
• Listening to the poor
• Destitution study
• Young lives
• PPA (preparation underway)

38
Selected Welfare Monitoring Data of Ethiopia
Qualitative data collection and analysis
39
Other qualitative surveys
40
How do they measure poverty?FSS
41
How do they measure poverty?Destitution study
associated with material deprivation
42
Change of poverty overtime

• Qualitative studies indicate
• Poverty is increasing
• Quantitative studies (official)
• Poverty do not change
• Quantitative (independent ERHS)
• Poverty declines (Dercon, 2001)

43
Concluding remark

• SDPRP is an important document both by government
  and other agents
• SDPRP has already influenced policy and budget
  allocation
• More to be done in PRSP consultation process
  include peoples idea instead of experts idea
  only
• Ethiopia is relatively rich in data, but analysis
  of data and making it public takes substantial
  time improvement required
• Improving access of researchers to CSA data,
  otherwise research will continue to depend on
  data that not representative

44
Concluding remark

• Monitoring of poverty is limited by the cross
  section nature of the data CSA should think
  having panel data
• Conduct CWIQ every year to get poverty trend
  using poverty correlates derived from HICE and
  WMS.
• The Decentralization requires local level poverty
  estimates poverty mapping exercise should be
  stated
• Provide attention to qualitative poverty
  assessment help to measure the empowerment
  (voicelessness) dimension of poverty.
• Use depreciation of durable expenditure into
  poverty estimation instead of using the whole
  expenditure on durables