This paper examines a mechanism of liquidity-preference fluctuations caused by changes in people's belief about a random liquidity shock. When observing the shock, they rationally update their belief so that the shock probability is higher; consequently they raise liquidity preference and reduce consumption. As the period without the shock lasts, they become more optimistic so that they gradually lower liquidity preference and increase consumption. The recovery pattern depends on the realized frequency of the shock: when the shock occurs many times in succession, the consumption recovery is first slow, gradually accelerates and eventually slows down, tracing an 'S'-shaped curve.
Bayesian Learning Liquidity Preference Precautionary Motive Business Cycle Markov Switching