Noncomparable Poverty Comparisons

Abstract

Poverty estimates based on enumeration from a single point in time form the basis for most country-level analysis of poverty. Cross-country comparisons of poverty, and global counts of the poor, implicitly assume that country-level poverty headcounts are comparable. This paper illustrates that the assumption of comparability is potentially invalid when households are interviewed multiple times throughout the year, as opposed to a single-visit interview. An example from Jordan illustrates how the internationally comparable approach of handling data from repeat visits yields a poverty rate that is 26 per cent greater than the rate that is currently reported as the official estimate.

Acknowledgements

This work builds on earlier work done by Johan Mistiaen and Umar Serajuddin. We thank Kathleen Beegle, Ferid Belhaj, Eric Le Borgne, Lidia Ceriani, John Gibson, Talip Kilic, Jin Kim, Gbemisola Oseni, Victor Sulla, Hiroki Uematsu, Dominique van de Walle and Paolo Verme of the World Bank for feedback and comments on the draft. We are grateful to Mukhallad Omari, Zein Soufan and Orouba Al Sabbagh of the Ministry of Planning and International Cooperation (MoPIC), and to Sana Al Momani, Ghaida Khasawneh, Maha Dawas, Ahmad Abu Haider, and Fatmeh Awamreh of the Department of Statistics (DOS) of the Government of Jordan (GoJ) for their cooperation. The data used in this analysis is property of the GoJ, and requests to obtain the data can be made through DOS at [email protected]. The code is available from the authors upon request. The findings, interpretations, and conclusions of this paper are those of the authors and should not be attributed to the World Bank Group or its member countries.

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

1. For example, the global poverty estimates reported in World Bank (2014) are based almost exclusively on cross-sectional household surveys with single-visit enumeration of consumption. Ferreira et al. (Table 2, 2016) document that three fourths of the 125 countries including in the World Bank’s global count of poor use consumption as the measure of wellbeing.

2. Smith, Dupriez, and Troubat (2014) examine household survey data used for poverty estimation from 100 countries and find that 70 per cent of the questionnaires use a recall period for food of two weeks or less.

3. See Carletto, Jolliffe, and Banerjee (2015) for a discussion of the importance of small-scale agriculture to the wellbeing of the vast majority of the poor in sub-Saharan Africa.

4. The analysis is based on independent sub-samples of households implicitly stratified on season to ensure that each sub-sample is nationally representative.

5. They note for example, that in South Africa, 90 per cent of the households living under USD 1 per day spent money on festivals. And, in Udaipur, more than 99 per cent of the extremely poor households spent money on a wedding, a funeral, or a religious festival with the median household spending 10 per cent of its annual budget on festivals.

6. This seasonality is often difficult to uncover since most macroeconomic data are released to the public after it has been adjusted to remove seasonality.

7. Under specific assumptions on the flow of income, Deaton’s model can also imply no smoothing of consumption at all.

8. Parker (2015) examines how households’ consumption behavior changes in response to the receipt of predictable, lump-sum payments totaling USD 25 million in Federal stimulus payments that were distributed randomly across weeks. He finds that households largely do not smooth consumption in response to the transfers.

9. Households are asked about food consumption on a quarterly basis and nonfood consumption on a monthly basis.

10. See World Bank (2014) for an estimated global count of the poor, and Jolliffe et al. (2014) for methodological details on how these estimates are produced and how sensitive they are to methodological differences across countries.

11. See Ravallion (1998) for discussion of CBN and other methods.

12. This line translates to 3.42 US Dollars per day in 2005 PPP terms.

13. The following all use these three measures: Jolliffe, Datt, and Sharma (2004) for Egypt, Datt and Ravallion (1992) cover Brazil and India, Howes and Lanjouw (1998) use examples from Pakistan and Ghana, and Ravallion and Bidani (1994) examine Indonesia.

14. See Jolliffe and Semykina (1999) for details.

15. The estimated per cent difference is more than six standard errors away from zero. The standard error for the difference is estimated as a second-order Taylor approximation. See chapter 6 of Wolter (1985) for discussion, and Jolliffe, Gundersen, Tiehen, and Winicki (2005) for a similar example.

16. As a measure of inequality, the Theil index and the larger family of generalised Theil indices, have many desirable properties that are described in Foster (1983). In particular, Foster shows that an inequality index satisfies the axioms of symmetry, replication invariance, income-scale independence, decomposability, and the principle of transfers only if it is a positive multiple of the Theil index.

17. This difference is statistically significant in the three quarters.

18. When considering the average shortfall of the poor from the poverty line (that is the ratio of the poverty gap to the headcount), we know from above the status quo shortfall is 19 per cent. In contrast, the Rawlsian ever poor shortfall is 25 per cent.

19. The status quo approach relies on a balanced panel and therefore suffers greater levels of attrition relative to the cross-section approach. For the analysis in this paper though, our focus is strictly on the difference in the properties of the two measures and we have therefore chosen to restrict our analysis to the same sample for all measures.