Pres Feulefack Zeller

How Accurate is Participatory Wealth Ranking (PWR) in Targeting the Poor? A Case Study from Bangladesh: 

How Accurate is Participatory Wealth Ranking (PWR) in Targeting the Poor? A Case Study from Bangladesh Joseph Florent Feulefack and Prof. Dr. Manfred Zeller October 12, 2005 Georg-August-Universität Göttingen

Introduction: 

Introduction Global Consensus towards Poverty eradication 1.2 billion live on less than $1 a day worldwide (World Bank, 2002) Unfortunately, lack of accurate and operational tools to identify the poor

Questions and Objective: 

Questions and Objective Can PWR target extreme poverty? Can PWR be used in ex-post assessment? How does it compare with other methods? Questions Evaluate the accuracy of PWR as a poverty targeting and assessment tool, and compare it with alternative tools: - Ladder of Life (LL) - Visual Impression Interviewer (VI) - Poverty Assessment Tool (PAT) Objective

Methodology: 

Methodology Research Design Sample =320 households CGAP PAT Consultative Group to Assist the Poor Poverty Assessment Tool Conceived by IFPRI (Henry et al., 2003) Index derived by Principal Component Analysis (PCA) model Diverse poverty indicators -Secondary Data from Bangladesh analysed using SPSS Benchmark=“Truth” - Followed World Bank’s LSMS guidelines - Per capita daily expenditure survey: Assets, Food, Education, Housing, Transport, Savings, Social Capital, etc Poverty Line = $1 US /Day (=23.18 Taka in March 2004 in Bangladesh)

PWR Field Process: Sorting: 

PWR Field Process: Sorting Methodology Source: DATA, Survey Firm, Bangladesh Reference groups made only of women Follows rules prescribed by Gibbons and Simanowitz (1999) Conducted by DATA, survey firm, Bangladesh (IRIS, 2004)

Methodology: 

Methodology Field Process of Tested Tools Ladder of Life (LL) “On which step of this ladder are you located today?” Visual Impression (VI) self-assessment by the enumerator “Based on the information you have received so far, how would you rate the standard of living of the household?” Benchmark: LSMS-type per–capita Daily Expenditure Ranking Step 1= Poorest Step 10 = Richest

Methodology: 

Methodology Accuracy test of tools that measure relative poverty BEST Score = Score of preferred highest accuracy (Zeller et al., 2004) Participatory Wealth Ranking (PWR) Benchmark `Truth´ Total Accuracy (TA)= [(25 + 45)/100] x 100 = 70% Poverty Accuracy = [25/(25+20)] x 100 = 56%

Main Result 1 : 

Main Result 1 PWR’s geographic poverty comparisons TA is lowest at the largest geographical area and highest at the narrowest

Main Result 2: 

Main Result 2 PAT has the best BPAC at all the levels At community level, PWR is second to the last whereas the LL performs worst Comparison of Balanced Poverty Accuracy Criterion (BPAC) of Tools BPAC = Balanced Poverty Accuracy Performance Criterion

Main Result 3: 

Main Result 3 Comparison of Total Accuracy of Tools PWR has been outperformed by other tools at all level Results specific to Bangladesh: Externally judged criteria produced less accurate results (Gibbon et al., 1999) PAT is far the best of all at the nation level

Main Result 4: 

Main Result 4 Comparison of PIE across Tools PAT most robust to misclassification errors across geographic levels PWR is the subjective tool the least prone to misspecifications at the district and at the “nation”

Conclusions: 

Conclusions Poverty Assessment Tool (PAT) Suitable tool for ex-post assessment Participatory Wealth Ranking (PWR) Targeting method for clients of development programmes or projects.

Slide13: 

Thank you for your attention! Germany Cameroon