Measurement of population income: Variants of estimating biases

Abstract

Income is one of the most obvious and frequently used indicators of economic status and living standards. Surveys of households and individuals are the main sources of income data for sociologists and economists. Administrative data is added to them on a growing scale. Comparison of data obtained from different sources or surveys using different methods allows us to estimate biases, sources of errors, and demonstrates the absence of “ideal” income data in general. The review of foreign studies on this problem is supplemented by an example of calculations on data from the The Russia Longitudinal Monitoring Survey — Higher School of Economics (RLMS—HSE): we compare the compositional individual income, calculated as the sum of types of income, and the total personal income reported by respondents. The first measurement of individual incomes has turned out to be more consistent and definite, less prone to measurement error, but gives lower values of individual incomes. The differences of the total personal income reported by respondents and compositional individual income are due not so much to the inaccuracy of the summation and rounding as to “conceptual” features of understanding of personal income by some respondents. Such comparisons are necessary in order to understand the limitations of various measurements of income, grounded and reflexive choice of its specific indicators.