How does the relationship between earnings and schooling change with the introduction of comprehensive economic reform? This Paper uses a unique dataset (covering about 3 million Hungarian wage earners, from 1986 to 1998) and a novel procedure to correct sample selection bias (based on DiNardo,...