We study a reform that increased compulsory schooling in Britain. Using a regression discontinuity design, we find that, in line with prior studies, compulsory schooling caused an increase in income. However, it also caused a decrease in subjective well-being, or happiness. Using an extensive survey, we provide evidence that education affects both reference points and outcomes, but the former more than the latter. That is, while outcomes improve for those subject to compulsory education, they improve less than the individuals themeselves expect they will.