Well-Being and Poverty in Kenya. Luc Christiaensen (World Bank), Presentation at the Poverty Assessment Initiation workshop, Mombasa, 19 May 2005

Transcription

1 Well-Being and Poverty in Kenya Luc Christiaensen (World Bank), Presentation at the Poverty Assessment Initiation workshop, Mombasa, 19 May 2005

2 Overarching Questions How well have the Kenyan people fared over the past decade? How could their well-being be improved? Today, we briefly review key topics, the available information base, and appropriate methodologies in addressing these two key questions

3 Information base Nationally representative data National accounts (annual time series, pvt consumption, but aggregate) 1999 Census Data (very disaggregated, geographically referenced, but only for one year and no expenditure data) Demographic and Health Surveys (1989, 1993, 1998, 2003) (repeated cross-sections, highly comparable, but no expenditures and no panel) Expenditure data WMS 1997 (available, but little outdated) KIHBS 2005/6 (comprehensive, but not yet available)

4 Information base (2) Not nationally representative data on specific groups Small holder maize farmers - Tegemeo data very recent, rural panel data with four rounds 1997, 2000, 2002, 2004, but only income data (no good expenditure modules) Urban slum dwellers -African Population and Health Research Center informal urban settlements - panel data on specific groups but no expenditures Pastoralists - Arid Lands Project Participatory Poverty Assessment very recent, can be linked to KIHBS, qualitative in nature, but not nationally representative and not yet available) Others Triangulation is the name of the game!

5 How well have the Kenyan people fared? Dimensions of people s well being,i.e being poor wrt to an absolute (external) standard (poverty) in comparison with others (inequality) wrt one s prospects of future well-being (vulnerability) Indicators of well-being Monetary indicators (e.g. expenditures) Non-monetary indicators (e.g. human capital, empowerment)

6 Evolution of Monetary Well-Being

7 Different approaches Standard practice use of national expenditure surveys to estimate and compare poverty over time (i.e. compare 2005 KIHBS with 1997 WMS) calculate poverty, inequality and vulnerability measures In Kenya, currently not yet available, though even when available, not w/o challenges (snapshots/rainfall; need for appropriate price deflators; comparability issues) Need to complement with alternative approaches Statistical asset index Economic asset index Case studies (Tegemeo and APHRC) Linking national account data to household consumption surveys

8 Statistical Asset Index Sample Means of Household Assets for Use in Asset Index Differences to to to 2003 House floor of low quality (mud, dung, sand) House roof of low quality (thatch) Drinking water - piped or public tap Flush toilet No toilet Electricity connection Owns a radio Owns a TV Owns a refrigerator Owns a bike Head with only primary education Head has secondary education or higher Years of head's educational attainment Sample size 8,561 All variables (except last) are dummy variables (1=yes, 0=no) = "negative" asset

9 Geographical differences in evolution of assets Rural Other Urban Nairobi House floor of low quality (mud, dung, sand) House roof of low quality (thatch) Drinking water - piped or public tap Flush toilet No toilet Electricity connection Owns a radio Owns a TV Owns a refrigerator Owns a bike Head with only primary education Head has secondary education or higher Years of head's educational attainment All variables (except last) are dummy variables (1=yes, 0=no) = "negative" asset Differences (1993 to 2003)

10 Poverty Changes based on Statistical Asset Index Headcount Ratio National Rural Urban without Nairobi Nairobi Decline in national poverty driven by decline in rural poverty and poverty in Nairobi; stagnation of poverty in other urban areas Large decline in poverty in Nairobi? Credible? Relative poverty and not anchored in theory

11 Economic Asset Index Weights are anchored in consumption measures and thus in economic theory It concerns absolute poverty Prediction techniques are similar to those in poverty mapping Estimate model of per capita expenditures using 1997 WMS Explanatory variables common to both WMS & DHS Predict per capita expenditures at household level in the DHS (1993, 1998 and 2003) Calculate poverty & inequality measures

12 Poverty Changes based on Economic Asset Index (Base) National Rural Other Urban Nairobi

13 Economic Asset Index Poverty Poverty Incidenc Rural Other Urban Nairobi

14 Observations Issues with Economic Asset Index Results Rural poverty falling Non-Nairobi urban poverty rising and then stagnating Strong fall in Nairobi s poverty? Insufficient capture of informal economy? Not Issues with Economic Asset Index Methodology Parameter stability time-varying RHS variables (rainfall) Data determining the best fit parameters with wrong sign More theoretical approach larger prediction error? It remains a series of snapshots

15 Price evolution of selected durable goods Radio Cassettes Radios TVs Refrigerators CPI

16 Evolution in poverty among rural maize smallholders Percentage Households Earning Less Than 1 US$/day Zone Coastal Lowlands Eastern Lowlands Western Lowlands Western Transitional High Potential Maize Zone Western Highlands Central Highlands Marginal Rain Shadow Total

17 Linking macro to micro data National accounts data are Annually available Comparable Aggregate These features could be exploited to examine the evolution of poverty by linking them with household consumption surveys Classify households in three groups according to sector of employment of household head as observed in 1997 WMS Apply historical sectoral GDP growth rates Allow for structural transformation based on observed urbanization patterns

18 Per Capita GDP in Kenya DHS 1993 WMS 1997 DHS 1998 DHS Kenyan Shilling (constant 1995)

19 Poverty Change Macro-Micro Linkages DHS 1993 WMS 1997 DHS 1998 DHS Poverty Head Count (%)

20 Work on evolution of monetary poverty is in progress Critical need for further triangulating results regarding poverty and inequality Finalize work on statistical and economic asset indices National Accounts - informal sector properly captured? Recall information from PPA Case studies: Need to finetune results from Tegemeo Urban poverty APHRC Pastoralists Last but not least, analysis of KIHBS once available More work needed on vulnerability?

21 Evolution of Non-Monetary Indicators

22 Primary enrollment rates improved substantially, child malnutrition rates declined, but secondary enrollment still around 1993 levels, and infant mortality rates have risen. Difference National to Enrollment rates Primary (6-13 yrs) Secondary (14-17 yrs) Infant mortality Stunting prevalence

23 Most progress in rural areas and Nairobi, while limited progress or deterioration in other urban Difference Rural primary enrollment (6-13yrs) secondary enrollment (14-17yrs) infant mortality* stunting prevalence (%) Urban (w/out Nairobi) primary enrollment (6-13yrs) secondary enrollment (14-17yrs) infant mortality* stunting prevalence (%) Nairobi primary enrollment (6-13yrs) secondary enrollment (14-17yrs) infant mortality* stunting prevalence (%)

24 Governance Indicators

25 Governance indicators (2)

26 The husband s beating stick is like butter % women agreeing with justification of husband beating wife Burns the food 16.3 Argues with him 45.9 Goes out without telling him 39.4 Neglects the children 55.1 Refuses sexual relations 29.4 Agrees with at least one specified reason Source: Demographic and Health Survey

27 How could the Well-Being of Kenyan People be improved? Determinants of monetary and non-monetary wellbeing marginal benefits of different interventions (path to reach MDGs) Evaluation of the poverty reducing, distributional and vulnerability reducing effect of particular policies (e.g. NPCB???) Evaluation of the effectiveness of public expenditures (investment (roads vs malaria vs extension), recurrent costs teacher salaries) in reducing poverty, inequality and vulnerability

28 Concluding Remarks Critical requirements for assessing evolution of poverty are: Comparability of data, survey and questionnaire design Regular availability of data, b/c annual variations No single source meets all these criteria innovative approaches (economic asset index) and triangulation necessary Work has started, but need for further triangulation using case studies, ppa and KIHBS Work on vulnerability so far limited

29 Concluding Remarks (2) Evolution of Non-Monetary poverty More robust information base (3 rounds of DHS) Human assets Overall progress in enrollments (especially primary), child malnutrition and stagnation in secondary enrollment and deterioration in infant mortality Progress mainly driven by rural areas and Nairobi, and much less so by other urban areas where child malnutrition actually increased Empowerment issues More work needed on gender issues DHS provides good opportunities to explore determinants of attitudes and incidence of gender based violence Governance issues

30 Concluding Remarks (3) Determinants of Well-Being regression analysis of DHS data to explore determinants of human assets (work is ongoing) link with MDGs Determinants of gender based violence? Other sections will discuss Evaluation of the poverty reducing, distributional and vulnerability reducing effect of particular policies (e.g. NCPB) Evaluation of the effectiveness of public expenditures (investment (roads vs malaria vs extension), recurrent costs teacher salaries) in reducing poverty, inequality and vulnerability