Measuring the Cumulative Effects of Embeddedness on Earnings Trajectories in North America in Psid and Slid

According to leading sociological theories, economic action is embedded within social relations and customs. This “embeddedness” should give rise to path dependent and cumulative dynamics in earnings trajectories when large and/or, especially, unexpected changes in earnings move people into new social positions to which they have not yet adapted. Over the life course, these effects will induce negative autoregressive biases in earnings trajectories which are not accounted for by the kinds of earnings models ordinarily used by economists. In this paper, I model and evaluate this hypothesis using dynamic panel data methods and estimation, and longitudinal earnings data in PSID and SLID. I find robust evidence in support of the sociological concern with embeddedness: large changes in earnings, positive or negative, do in fact induce a long run negative effect on earnings over time. Moreover, these effects are more pronounced among the poor.