Predictability in Consumption Growth and Equity Returns: A Bayesian Investigation
使用贝叶斯方法估计一个包含长期风险的消费资产定价模型，发现模型虽与消费和股息增长数据一致，但消费增长的条件均值不够持久，且需要较高的跨期替代弹性，导致只能解释资产收益总变动的50%。
Abstract We use a Bayesian method to estimate a consumption‐based asset pricing model featuring long‐run risks. Although the model is generally consistent with consumption and dividend growth moments in annual data, the conditional mean of consumption growth (a latent process) is not persistent enough to satisfy the restriction that the price‐dividend ratio be an affine function of the latent process. The model also requires relatively high intertemporal elasticity of substitution to match the low volatility of the risk‐free return. These two restrictions lead to the equity volatility puzzle. The model accounts for only 50% of the total variation in asset returns.