Yemtsov, Ruslan (2008). Through the looking-glass: what is behind official data on inequality in Russia over 1992-2003?.
Russia represents an interesting case in which officially published data on income inequality are not directly derived from any specific survey. Rather they come from combining four elements: household survey of expenditures, household balances of income and expenditures, macroeconomic statistics tracing changes in the aggregate household money incomes , and a parametric model (lognormal) of the distribution, calibrated using data from these sources. This method is called to deal with problems which are not unique to Russia: low response rate in household surveys and under-reporting of incomes by survey respondents. The paper describes the evolution of the method over the period under study. It finds several breakpoints during 1992-2006, affecting comparability of inequality statistics over time. The paper shows that the modeling assumptions influence the level and trends of published inequality indices, and that these assumptions can be challenged. The raw survey data are adjusted by the Russian statistical office in line with these assumptions. Based on documentation of the official survey the paper proposes ways for restoring crude survey weights, which is necessary for comparisons to other survey-based inequality statistics. Relying on raw data from three surveys for the same year 2003: the official expenditure survey (HBS), large alternative survey (NOBUS) and a well-known panel survey (RLMS), the paper shows that the choices of welfare indices have at least as large an influence on measured inequality as modeling assumptions . Hence, improvement in the consistency of welfare indices can help to improve the accuracy of inequality data in Russia. Initiative of Rosstat to open access to its survey data will help to make a significant progress in inequality analysis, and will motivate improved data quality. The paper advocates for better sampling techniques, expanded definitions of welfare aggregates, expanded open access to survey data, and the use of other data sources (e.g. tax data) to provide new, more reliable evidence on the levels and changes of inequality in Russia.