Preference Parameter Changes in Life-cycle Consumption Models: the Measurement-error-robust Approach
This study investigates preference parameter changes after a natural disaster by using observational data. The heterogeneous preference parameters in a structural model is estimated by an existing measurement error robust approach, where assumptions set on measurement errors are tested by a novel moment inequality approach proposed in this study. We find that the disaster affects households' preference parameters heterogeneously, depending on households' future risks of facing a similar disaster-households facing a high risk become more risk averse compared with those facing a low risk. Our simulation results suggest that consumption predictions can be improved by using updated preference parameters.
*The latest version of the discussion paper reflects readers’ comments. First, the assumption related to stationarity is revised. Second, the mechanism that disasters affect consumption and investment is reconsidered. Third, we change our presentation to focus on the empirical results of the study.