By Roger Farmer, Carine Nourry and Alain Venditti
Existing literature continues to be unable to offer a convincing explanation for the volatility of the stochastic discount factor in real world data. Our work provides such an explanation. We do not rely on frictions, market incompleteness or transactions costs of any kind. Instead, we modify a simple stochastic representative agent model by allowing for birth and death and by allowing for heterogeneity in agents’ discount factors. We show that these two minor and realistic changes to the timeless Arrow-Debreu paradigm are sufficient to invalidate the implication that competitive financial markets efficiently allocate risk. Our work demonstrates that financial markets, by their very nature, cannot be Pareto efficient, except by chance. Although individuals in our model are rational; markets are not.
That birth and death matters should be no surprise. As in overlapping generation models, the fact that the yet-to-be-born cannot trade with current generations leads to inefficiencies. That discount factor heterogeneity matters here is more of a surprise, at least to me. I would have expected this to simply to two classes of agents, one borrowing to the limit, the other accumulating to the other limit and that each category would otherwise enjoy efficient allocations, just with a higher discount rate due to the risk pf death. Apparently this is more complex than I thought.