Measuring Inequalities: Do The Surveys Give The Real Picture? Study Of Two Surveys In Cote D’Ivoire And Madagascar

Measurements of standards of living and its distribution are affected by methodological choices made before the consumption and income aggregates are calculated and by failure to correct the primary databases, but several sources of bias can also have an impact. This study, based on surveys held in Madagascar and Côte d’Ivoire, aimed to detect these biases by applying several scenarios for calculating living standards aggregates, by analysing the internal coherency of the surveys and by confronting the survey data with other sources of data, namely the National Accounts and the Balance of Payments. Methodology was found to have little impact, except for the question of whether or not regional prices were taken into account. Although there was significant bias due to under-declaration, this was not easy to correct, notably with the multiple imputation method. However, the results show that average income levels appear to be underestimated by 15 to 50% in the two surveys in question. The different corrections bring inequality levels in both countries nearer to levels in the most inegalitarian countries such as Brazil.