Training for poverty economists
The materials on this site aim to help enhance the quality of advice and integrated knowledge on sustaining past welfare gains in the face of slower global growth. These resources are taken from the workshop, which provides a comprehensive technical review of methods, using actual household survey data. The training covers all steps of poverty measurement, from household survey data collection (including basic knowledge of sampling), to the creation of consumption aggregates, price and household composition adjustments, the construction of poverty lines, different measures of poverty, and sensitivity analysis.
| Time | 15-Jul | 16-Jul | 17-Jul | 18-Jul | 19-Jul |
|---|---|---|---|---|---|
| 8:30AM | Breakfast | ||||
| 9:00AM | Survey Design (Kristen) | Introduction to Poverty Measurement (Kristen), Consumption Aggregate Overview (Dhiraj) | Consumption Aggregate: Non-Food (Kristen), Consumption Aggregate: Durables (Andres) | Hands-on Exercise using Stata (Andres - Housing) | Poverty Indices & Vulnerability (Emmanuel) |
| 10:45AM | Coffee Break | ||||
| 11:00AM | Sampling (Kristen) | Consumption Aggregate Food (Dhiraj) | Consumption Aggregate: Housing (Andres) | Hands-on Exercise using Stata (Judy - Prices & Household Composition) | Sensitivity Analysis (Andres) |
| 12:30PM | Lunch Case Study #1 | Lunch Case Study #2 | Lunch Case Study #3 | Lunch Case Study #4 | Lunch Case Study #5 |
| 1:30PM | Hands-on Exercise using Stata (Kristen - Sampling) | Adjustments: Prices & Household Composition (Carolina) | Hands-on Exercise using Stata (Judy - Food & Non-Food) | Poverty Lines (Gabriel) | Hands-on Exercise using Stata (Andres) |
| 3:15PM | Coffee Break | ||||
| 3:30PM | Hands-on Exercise using Stata (Judy - Data Cleaning) | Hands-on Exercise using Stata (Judy - Food & Non-Food) | Hands-on Exercise using Stata (Andres - Durables) | Hands-on Exercise using Stata (Judy - Poverty Lines) | Hands-on Exercise using Stata (Andres), Discussion & Wrap-Up (Kristen) |
| 5:00PM | End of Training | ||||
There would not be any poverty measurement without data to analyze. In order to have a precise and effective poverty measurement, practitioners and researchers must rely on quality datasets. This session gives an overview on how sampling errors affect the quality of our estimates. Predicting the sampling errors of poverty surveys require knowledge of different sampling techniques. We will learn these techniques, and how to implement them in Stata. The hands-on training session gives an overview of the statistical foundations of sampling, and discusses the different possible random sampling techniques and how to calculate sampling errors and weights.
The Poverty and Shared Prosperity Report will henceforth be published on a two-year cycle. The next issue will come out in the Fall 2018. This report will be focused on the measurement of global poverty and shared prosperity, and introduce some of the complementary indicators recommended by the Atkinson Commission on Global Poverty, which are intended to broaden the World Bank’s conception of poverty. This presentation shows preliminary findings from the empirical part of the report.
There would not be any poverty measurement without data to analyze. In order to have a precise and effective poverty measurement, practitioners and researchers must rely on quality datasets. This session gives an overview on how non-sampling errors affect the quality of our estimates. We will explore the options of decentralized data collection (CAFÉ) and computerized personal interviews (CAPI), and explain the different phases of the data collection process, from the preparation of the questionnaire, to data entry programs and training/supervision for enumerators. PPT
Using consumption as a welfare aggregate for poverty measurement finds its theoretical justification in Ravallion (1994)’s money-metric utility theory. The consumption aggregate is the foundation over which all poverty analysis rests. Its components can be aggregated into four main classes: (i) food items, (ii) non-food items, (iii) consumer durables, and (iv) housing. This session will present and discuss Ravallion’s conceptual framework. and it will concentrate on the first two components of the consumption aggregate, as well as on the treatment of outliers. An empirical application will provide practical guidance on the implementation of the theoretical setting.
Housing is often on of the most problematic components of the household consumption aggregate. As with durable goods, the appropriate measure of housing consumption is a measure in monetary terms of the flow of services that the household receives from occupying its dwelling. If rental markets are perfect and all households rent their dwellings, the rent paid is the obvious choice to include in the consumption aggregate. But this is not often the case, either because rents are subsidized, or because households own their dwellings. Even if the dwelling is owned by the household or received free of charge or with a subsidized rent, an estimate of the annual rental equivalent must be included, when possible, in the consumption aggregate. After presenting motivations and a brief literature review on the problem of rent imputation, this session will focus on the different methods to estimate rent for home-owners. It ends with examples of empirical applications of the different methodologies and their impacts on poverty measures.
Durable goods are relevant components of the welfare aggregate. One of their characteristics is that they last for several years and it is clearly not their expenditure on purchases the appropriate measure of their consumption. Instead, it is the value of services that the household receives from all durables goods in its possession over the relevant time period. This session presents first a brief discussion of the theoretical framework. It follows with an introduction to several methods of estimating the value of services from durable goods conditioned on the information available in the household survey. It ends with examples of empirical applications of the different methodologies and their impacts on poverty measures.
Expenditure is utility consistent, and therefore a good indicator for individuals’ welfare, only if a set of hypothesis is satisfied. One of the hypotheses is that individuals must face identical prices. This is often not the case. Prices usually vary both over time and across space: for instance, the same good can cost more in urban as opposed to rural areas, or more at the end of the year than the beginning of the year. In order to have meaningful utility comparisons, prices need to be adjusted to take into account these differences. After discussing the problem of price variation for poverty measurement, this session present three different price indices that can be used to approximate the first best scenario of equal prices for all individuals. A hands-on problem set concludes the session.
Household composition is a key input to the unit of analysis in poverty measurement. The conventional and most commonly used approach is to divide total household expenditure by household size to obtain per capita expenditure. However, some analysts use consumption per adult equivalent, in order to capture differences in need by age and gender, and economies of scale in consumption (Haughton & Khandker, 2009). While the use of an adult equivalent scale is compelling in theory to ensure utility consistency of the welfare indicator, a practical difficulty remains in practice as to what is the appropriate scale. This session introduces some of the most widely used scales and discusses implications of using such scales on the resulting poverty estimates.
A poverty line can be defined as the minimum expenditure required by an individual to fulfill his or her basic food and nonfood needs (Haughton & Khandker, 2009) so that those who live below this line is considered poor. Setting poverty lines is the last of the three major steps in poverty measurement, and often considered the most difficult and contentious process in poverty measurement (Ravallion, 1998). There are three types of poverty lines; absolute, relative, and subjective. The session focuses on absolute poverty lines using the cost of basic need (CBN) approach.
Once welfare aggregate and poverty lines have been computed, all information needs to be summarized in a single measure to quantify the level of poverty in a country, and to make comparisons with other countries or across time. Among the different poverty indices that have been developed in the literature, this session focuses on a specific family, the Foster Greer and Thorbecke (1984) poverty indices. The session is organized as follows. After a short introduction to several poverty indices, the FGT class is defined. For each index of the FGT class, pros and cons will be discussed and explained with several examples. Finally, a specific and desirable feature of the FGT index, decomposability, will be introduced and discussed.
Many arbitrary choices are made in the process of measuring poverty: for instance, the choice of the poverty index, the setting of the poverty line, and the definition of the equivalence of scale. In order to draw robust poverty comparison, it is important to perform some sensitivity checks to the alternative assumptions. This session will cover the main sensitivity analysis tools, with a particular focus on poverty dominance analysis.