Advance Information and Consumption Insurance : Evidence from Panel Data

We investigate whether US households possess advance information about their future income and what this means for consumption insurance. Based on insights from a theoretical model, we propose a new test to detect advance information, which requires only panel data on consumption and income. Using the Panel Study of Income Dynamics, we find---in contrast to the existing literature---strong support for the existence of advance information. We use this evidence to estimate a standard incomplete markets model and find that advance information reduces households' income forecast errors by 15%. Our estimation results imply that 27% of all unexpected income changes are passed through to consumption. Ignoring advance information leads to a significant overestimation of consumption insurance and even more so at the bottom of the wealth distribution