The study raises the problem of non-random distribution of unobservable skills across the population of a transition country. At the beginning of the transition, people who were grown up under different political regimes have distinct goals, moral norms, tastes, behavioral patterns, and motivation, which result in the dependence of earnings on cohort. The cohort effects have been separated by two ways. According to the first one, the logarithm of the real wages index was used as a proxy for current period. The second one is based on an assumption on the form of age-earnings profiles in Russia. The results have revealed the strong and robust relationship between earnings and cohort in Russia. Conditioning of the cohort effects on gender has been discovered as well.