Defining chronic poverty: comparing different approaches

Abstract

In this study, we discuss two methods commonly used in the literature to measure chronic poverty, the permanent income approach (where the chronically poor are those whose mean income over time is below the poverty line) and the spells approach (where the chronically poor are those households who are below the poverty line half of the time or more). We check the differences between these two methods considering also several household characteristics. The index we use is that of Foster, Greer and Thorbecke, FGT, with and . When each method identifies the chronically poor (which could be different individuals by each method) and shows the percentage of the chronically poor, while when the methods also take into account the depth of poverty. Our main goal is to show the differences between the two methods, so that policymakers would have different perspectives on the problem of chronic poverty and could make decisions on this basis. The results show that, in the data used, the permanent method and the spells method classify the households into chronically poor and nonchronically poor definitions almost similarly in some cases. Thus, when measuring chronic poverty using the FGT index with , the levels of chronic poverty measured by the two methods are quite the same for the whole population as well as for subgroups of the population. Nevertheless, when using the FGT index with , the permanent method and the spells method give different results, as they take into account the depth of poverty differently.

Keywords:

• chronic poverty
• FGT index

JEL Classification:

• I32
• D31

Notes

¹ We use standardized income, in order to take into account the number of persons in the household. In most studies of poverty, the unit of observation is the household (Jalan and Ravallion, 1998; Haddad and Ahmed, 2003; Foster, 2009). But one could also think of selecting the individual or the family. In fact, the choice of unit of observation should depend on the assumption that is made with respect to the intra-familial allocation of resources. If one supposes that resources are divided equally between the members of the family (or household), one should prefer the family (or household) over the individuals. On the other hand, if one has reasons to believe that resources are not divided equally between the family or household members, the unit of observation should be the individual. In addition, when using household consumption, one can infer things on individual consumption once assumptions are made concerning the issue of private versus public goods. If all goods and services consumed are of a private nature, then individual consumption is simply per capita consumption. If all goods and services consumed are public goods, then individual consumption is household consumption (Danziger and Taussing, 1979).

² Rodgers and Rodgers (1993) have estimated the FGT index for the spells approach for (the HCR), but not for .

³ We have data on 7 years, thus ‘half of the period or more’ means 4 years of poverty or more. We also check for a stricter definition of chronic poverty of 5 and 6 years of poverty (out of 7).

⁴ We again define an individual as chronically poor if his income is below the poverty line half of the periods or more.