This paper studies the cyclical nature of individual income risk using a confidential dataset from the U.S. Social Security Administration, which contains (uncapped) earnings histories for millions of individuals. The base sample is a nationally representative panel containing 10 percent of all U.S. males from 1978 to 2010. We use these data to decompose individual income growth during recessions into "between-group" and "within-group" components. To study the former, we group individuals along several observable characteristics at the time a recession hits. We find two variables to be excellent predictors of fortunes during a recession. First, prime-age workers that enter a recession with high average earnings suffer substantially less compared with those who enter with low average earnings. Second, we estimate "individual betas" (analogous to "stock betas" in finance) and examine their out-of-sample predictive power. We find that the earnings of high-beta individuals (those that exhibited higher sensitivity to prior recessions and expansions) fall significantly more during subsequent recessions. Next, we turn to within-group differences. Contrary to past research, we do not find the variance of idiosyncratic income shocks to be countercyclical. Instead, it is the left-skewness of shocks that is strongly countercyclical. That is, during recessions, the upper end of the shock distribution collapses--large upward income movements become less likely--whereas the bottom end expands--large drops in income become more likely. Thus, while the dispersion of shocks does not increase, shocks become more left skewed and, hence, risky during recessions. Finally, we find that the cyclical nature of income risk is dramatically different for the top 1 percent compared with all other individuals--even relative to those in the top 2 to 5 percent.