Financial Instability via Adaptive Learning

This paper develops a simple model in which adaptive learning by investors leads to recurrent booms and busts in asset prices. The model captures aspects of Minsky’s “financial instability hypothesis” in which periods of tranquility lead investors to increase their estimates of expected returns and reduce their estimates of return volatility. The changes of beliefs drive up asset prices and hence realized returns. However once agents invest a significant fraction of their wealth in stocks, the economy becomes fragile and so small negative shocks can lead to large declines in prices. I show how this process recurs over time, and discuss the features of the model which drive the boom-bust cycles in asset prices.