“…While earning an associate's degree and then transferring to a bachelor's program with a year or two worth of existing credit is possible in principle, actually earning a bachelor's degree on this path is rather rare (see discussion inOdle and Russell (2023)), so we do not incorporate this in the model.18 At T , agents receive 1 1−β times the utility of the average annual income earned by individuals with their particular degree attainment status.19 This is smaller than the full sample used in Section 3, but larger than estimation samples that require other covariates and outcome variables (Tables3-5), particularly because the state variables used to construct…”