A Model of On-the-job Learning: Theory and Empirical Evidence

I specify a model in which younger workers learn from more experienced workers. In particular, the ratio of the effective number of experienced workers to the effective number of newer workers determines how much newer workers learn. Workers pay for their learning by reduced wages, and they receive the benefits from their learning by a growth in their wages with experience. In an industry with a higher employment growth rate, experienced workers as teachers are scarce relative to younger workers as students, and consequently younger workers learn less from experienced workers. Therefore I have an important theoretical implication: an industry with a higher employment growth rate has a flatter experience-wage profile. This hypothesis is supported by an empirical evidence using the 1984 and 1986 Korean Occupational Wage Surveys for workers with 15 (or 20) years or less of experience. I also show that a new-born industry's productivity should increase at a decreasing rate over time under a certain stability condition.