Purpose: The article aims to describe the role of training and examines the impact on the wages of university and college graduates. Design/methodology/approach: We use nationwide administrative data on university and college graduates in 2019. The population includes 1.3 million observations, of which 222,000 ([approximately] 16%) received training after graduation from an educational institution (from July 2019 to 2022). We used OLS and the "difference-in-differences" methods to estimate the returns to training. Estimates obtained using the DID method turned out to be several times smaller because they consider unobserved characteristics (abilities). Findings: We obtained several key findings. First, the participation of graduates in training is high, despite their recent education. Second, undergoing training is conditional on the existence of wage returns. The results show a wage premium of 17.8% (OLS method) and 2.0% (DID method). Third, graduates from nonselective universities (with low state exam score) try to participate more actively in training to acquire missing knowledge and skills. The wage premium for graduates from nonselective universities is 19.1% (OLS method) and 5.1% (DID method). Fourth, there is a high return to training for graduates from socially relevant fields (education and healthcare), where training is regular and mandatory. Originality/value: This paper is one of the first to estimate the involvement and returns to training for graduates using nationwide administrative data in Russia.