Returns to Education

Education is the canonical application of the treatment-effects machinery from the previous weeks. Mincer (1958, 1974) gave labour economics its workhorse regression of log earnings on years of schooling; sixty years on, we have credible IV evidence that schooling raises earnings, but the level of the return still depends on what we assume about earnings dynamics, employment risk, and the consumption-smoothing technology agents use. Liu, Mogstad, and Salvanes (2025) sit on the interface between IV identification and a structural life-cycle model — exactly the joint we will pull apart in class.