We investigate horizontal inequality between two conflictive ethnic groups, the Kyrgyz and the Uzbeks in Kyrgyzstan, by employing the Oaxaca-Blinder decomposition. This technique has a long tradition in labour economics but has not been used in the literature that links ethnic inequality and violent conflict. We measure welfare differentials between Kyrgyz and Uzbeks and find that, depending on the welfare indicator used, Uzbeks are either better off or worse off than Kyrgyz. Specifically, we find that Uzbeks are not better off, if welfare is measured in terms of household expenditure. They are, however, clearly more prosperous in terms of the value and the size of their houses - at least in urban areas, where most Uzbeks reside. The picture is mixed when we use ownership of assets as a welfare measure. We conclude that the choice of welfare indicator is essential in studies of horizontal inequality, as it is most likely the more visible aspects of life that drive people's perceptions about other ethnic groups' standard of living. Decomposing welfare differentials between Kyrgyz and Uzbeks shows that the gap in expenditure is due to differences in group characteristics, such as household composition and ownership of livestock, whereas the gap in assets and house values remains unexplained.
measurement of welfare horizontal inequality ethnic conflict Oaxaca-Blinder decomposition Kyrgyzstan