When computing measures of poverty and inequality, most studies treat households as
homogenous groups, wherein all members benefit from the same access to resources.
Household surveys are based on nationally representative samples and allow a precise
measurement of household consumption and provide with reliable indicators of the level of
and changes in inequality and household poverty. However, these measurements lack of
precision in particular because consumption is always measured at the household level and
the individual well-being is measured by household per capita or per adult equivalent
consumption or income. Indexes are computed on this basis.
Independently, a large literature on intrahousehold allocation of resources now exists, that
aims at understanding the mechanisms at play within household. This literature suggests that
inequality within household could exist, depending on the bargaining power of each member.
Very recently, a paper by Dunbar et al. (2013) tries to put numbers on the implication of
potential within household inequalities on measures of poverty in a poor African country.
Given they only have data on household consumption (as is generally the case), they do it
based on the standard collective household model and using heavy structural hypotheses to
estimate the actual sharing of resources. Their attempt is not completely satisfactory for a
number of reasons, starting with the fact that it requires admitting the Pareto efficiency
hypothesis which in Africa is fairly unlikely. Nevertheless, it is the first attempt to take
seriously the fact that within household inequalities might hide the extent of poverty, and in
particular for certain type of household members (they focus on children).
In this paper we go one step in the direction of the true measurement of individual welfare.
Our data are taken from the Pauvreté et Structure Familial Survey (Poverty and Family
Structure - henceforth PSF, according to its French acronym) that has been conducted in
Senegal in 2006-2007. This survey has been specifically designed to obtain a measurement of
consumption at the "cell" level - a clearly identified budgetary decision unit within Senegalese
households - that allows coming closer to true individual consumption than household
consumption per capita. We evaluate inequality and poverty using household and cell
measurements of consumption.
In section 1 of this paper we present the principal features of the PFSS, give the precise
definition of cells within households and explain why this level of observation is pertinent to
analyse the extent and determinants of poverty.