Rare Events, Financial Crises, and the Cross-Section of Asset Returns

This paper shows that rare events are important in explaining the cross section of asset returns because of their role in shaping agents' expectations. I reconsider the "bad beta, good beta" ICAPM proposed by Campbell and Vuolteenaho and I point out that the explanatory power of the model relies on including the stock market crash that opened the Great Depression. When using a Markov-switching VAR, a '30s regime is identified. This regime receives a large weight when forming expectations consistent with the ICAPM, suggesting that the way agents think about financial markets is shaped by what happens during extreme circumstances. From a technical point of view, the paper extends the present value decomposition of Campbell and Shiller to allow for Markov-switching dynamics in the law of motion of the state variables. This approach could shed new light on the sensitivity of the present value decomposition methodology to the sample choice.