Theoretical literature shows that income uncertainty boosts saving, yet empirical work is incomplete. I test for the precautionary motive for saving using panel data. Knowing this motive's size is important for gauging the responsiveness of saving to government programs that reduce uncertainty, and for comparison to other motives, such as bequests. Most empirical studies of precautionary saving use either aggregate time-series or cross-sectional data, which cannot capture the effects of individual income uncertainty. I derive measures of total, permanent, and transitory income uncertainty from panel data—the National Longitudinal Survey—and find a strong precautionary motive. A doubling of uncertainty increases the ratio of wealth to permanent income by 29%.