Observational equivalence of the overlapping generations and the discounted dynamic programming frameworks for one-sector growth (Q1068681)

This paper considers ''observational equivalence'' of two alternative one- sector growth models with neoclassical technology. The first is a two- period overlapping generations (OLG) model with strictly selfish agents. The second is the dynamic programming model with infinitely-lived agents who maximize the discounted sum of the stream of one-period utilities. The paper shows that under certain conditions, the two models are observationally equivalent in the sense that they lead to identical paths for aggregate capital, output, consumption, savings, real wages and interest rate. One of the main conditions is that the steady-state rate of interest in the OLG model must be strictly greater than the growth rate of labor force. In other words, the steady-state competitive equilibrium must be Pareto efficient. The equivalence proposition is shown for both the deterministic case and the case where the OLG model is characterized by a random utility function. The major contribution of this paper is that it justifies the use of OLG models for short-run business-cycle studies.