8A Gisele KAMANOU Chapter 5

Views:
 
Category: Entertainment
     
 

Presentation Description

No description available.

Comments

Presentation Transcript

Expert Group Meeting on Handbook on Poverty Statistics United Nations Statistics Division United Nation, New York 28-30 June 2005: 

Expert Group Meeting on Handbook on Poverty Statistics United Nations Statistics Division United Nation, New York 28-30 June 2005

Chapter 5 Statistical Issues in Measuring Poverty from non-survey sources (Havinga, Kamanou, Ward) : 

Chapter 5 Statistical Issues in Measuring Poverty from non-survey sources (Havinga, Kamanou, Ward)

Introduction (Ward) 5.1 Prospects for meeting broader data requirement and quality issues in poverty assessments (Kamanou) 5.2 Capturing the Multidimensionality of poverty (Ward) 5.3 National Accounts (Havinga) : 

Introduction (Ward) 5.1 Prospects for meeting broader data requirement and quality issues in poverty assessments (Kamanou) 5.2 Capturing the Multidimensionality of poverty (Ward) 5.3 National Accounts (Havinga)

Prospects for meeting broad data requirements and quality issues in poverty assessments: 

Prospects for meeting broad data requirements and quality issues in poverty assessments Gisele Kamanou

Rationale: 

Rationale Income and Non-income (non-material needs) are essential Household income/expenditure do not capture even some important material needs, e.g. those provided by the government) Need for adequate data for addressing poverty in its multidimensional aspects

Rationale (cont): 

Rationale (cont) A wealth of data exists - BUT Disparate source Cross validation and linkage difficult Varying quality of “other sources” (e.g. # lines ministries) compound the problem Issues of comparability/consistency of data collected with different techniques

Rational (cont): 

Rational (cont) Focus of the section: Identify the more pressing outstanding issues where further improvement is essential to provide a sounder basis for PA Complement Ch3 with a review of the literature for a broad assessment of poverty data requirements Weak aspect of the section: Practical experience in non-surveys based analysis of poverty is severely lacking Yet – few cases studies are cited to exemplify the need to complement HHS data with non survey data.

Slide8: 

5.1.1 Conventional Poverty Assessment techniques and data requirements. 5.1.2 Practical avenues for Strengthening Household Surveys based Poverty Assessments Revisiting the practice of multi-topic household surveys Qualitative Assessment and participatory techniques Use of population census data and administrative record

5.1.1 Conventional Poverty Assessment techniques and data requirements (cont.): 

5.1.1 Conventional Poverty Assessment techniques and data requirements (cont.) Main purpose of HHS not often matches the needs for specific poverty data Financial constraints might lead to changes in the primary (poverty) focus of the HHS Other side of the coin: Countries are reluctant to making changes in their (customary) survey design Fear of non continuity in time series data Lack of openness to learn new techniques Lack of resources to train staffs on the new approaches

5.1.1 Conventional Poverty Assessment techniques and data requirements.: 

Conclusion: Great need to revisit the main objectives and priority of the surveys. However : Major challenges in defining the objectives of (poverty focus) surveys The section focuses on adequacy of poverty data 5.1.1 Conventional Poverty Assessment techniques and data requirements.

Slide11: 

Defining data adequacy is also a problem Grosh and Glewwe (1996) – Good review and some solutions are proposed – BUT they mainly look at HHS surveys and those funded by international aids. From the review of other literature, one main observation with important practical implication for NSO: Data requirements are for the majority conflicting in terms of survey design and cost. (e.g. a well known problem is the data requirements for specific VS consistent poverty lines)

Slide12: 

Few crucial data requirement to be met (which have been overlooked by data practitioners) 1. Data for monitoring trends in poverty fundamentally different (in terms of statistical properties) than those for understanding its manifestations (See Ch 4 and 6) 2. Increased concerns on the multidimensionality of poverty > greater demand for data to inform a wide range of policy issues YET – at the same time, monitoring poverty (e.g MDGs) put more emphasis on longitudinal data > countries limit topic and geographical coverage to prioterize frequency Undesirable result : monitoring data become of limited usefulness for policy formulations

Slide13: 

3. Micro vs macro determinants of poverty : Fundamental analytical dimensions with opposing data requirements: Interplay between micro and macro determinants of poverty not well understood. In particular, need data to assess the direct impact of macro-economic policy on poverty alleviation ( e.g. through retail prices) Section 5.2 on the use of supplementary financial data to complement price data A well researched topic with micro/macro policy implication is the role of investment in human capital on poverty alleviation. Both micro and macro policy need to be assessed jointly when looking at – say – the role of social spending on human development as the causation between the poverty and education outcomes are to straight forward (Pyatt and Ward)

Slide14: 

More generally, empirical work on the cause and effect relationship between economic development and social outcome are not conclusive (e.g Raminez et al. vs Lorenz King) and one of the main reasons is the lack of adequate data

Slide15: 

Aggregate data (life expectancy, infant mortality, GDP are often used to model the impact of growth of level of poverty. Weak results due to: Non-reconcilability of household level data with those available only at higher level of aggregation (life expectancy) Large degree of variation among (often too few) countries used in these analyses Pooled time series of these sorts are likely to suffer from heteroschedasticity (across the countries and over time period of the studies)

Slide16: 

However: Two important lessons with respect to the data implications of the (preliminary) empirical findings. 1. With respect to the problem of heteroschedasticity: Limit the scope of the studies to regions and /or periods for which the model assumptions are more reasonable. (King. 1998) study is illustrative in this respect, with the unique future that it enable to account for time specific stochastic effects, an improvement in comparison to other studies that mainly used an OLS.

Slide17: 

2. Relative importance of high growth to government expenditures in reducing poverty cannot be assessed with cross section aggregated economic and welfare data alone. > Great need for country-specific data at both micro and macro levels. Few case studies and illustrative: Ravallion and Anand (1993) on Sri Lanka, Hanner and al (2003) who estimated 420 000 equations to challenge the claim that health expenditure are ineffective in reducing child mortality.

Slide18: 

3. Other cross cutting issues with great data needs: Space dimension is as important as the time effect in these cross section, but always neglected in analysis of cross section aggregated data. (More in other chapters) Feminization of poverty has been hard to document empirically (patterns and trend wise) (Sex of head of household very weak and noisy variable) Need more data on intra-household resources allocation including and time use.

authorStream Live Help