This paper evaluates the impact on earnings, pensions, and other labor market outcomes of two parallel educational reforms increasing instructional time in Swedish primary school. The reforms extended the compulsory years of schooling from 6 to 7 years and the annual term length from 34.5/36.5 to 39 weeks per year. Gradually introduced over the 1930-1950 period in more than 2,500 school districts, the extensions generated large exogenous variation in educational attainment at different points in primary school while the overall school system and curricula remained unchanged. The reforms thus constitute an ideal quasi-experimental setting for analyzing the long-run causal impact of compulsory education keeping other school characteristics fixed. With a majority of students receiving only primary schooling, both reforms affected large shares of the population and consequently had large impacts on educational attainment at the compulsory level. We find striking differences in impact between the two reforms, and between males and females. Estimated returns to compulsory schooling are robustly positive only for females, who experience a small increase in early career earnings (~ 2%) when exposed to a 7th year of schooling, and large and persistent increases in earnings (~ 4 – 5%) when exposed to an extended school year. The effects are driven by the extensive margin, in particular increased employment in the public sector.