Low safe interest rates: A case for dynamic inefficiency?

We reexamine the tests for dynamic inefficiency in productive overlapping-generations economies with stochastic growth. Contrary to certain claims in the recent literature, we argue that the size of real safe interest rates relative to average GDP growth is an inconclusive test for dynamic inefficiency. A more accurate test should take into account the correlation between growth and the marginal utility of wealth. We provide an exhaustive criterion based on the growth-adjusted dominant root of the stochastic discount factor emerging at the competitive equilibrium. Surprisingly, a preliminary rough empirical application of this criterion uncovers dynamic inefficiency of the US economy for any reasonable degree of risk aversion. We also distinguish capital overaccumulation from an inefficient distribution of consumption risk. The refined test for capital overaccumulation is rather stringent: Capital is not overaccumulated if the net dividend remains positive with some probability, as opposed to always, as in the original Abel et al. (1989)'s formulation.