This paper presents a methodology for adjusting measures of income and poverty for the risk faced by a household. The approach draws on the standard economic concept of risk aversion, and it is based on the intuition that households will prefer a steady stream of income to a variable one with the same mean. Relying on a Constant Relative Risk Aversion utility function, we use panel data for Argentina to compute risk-adjusted income and poverty measures. At the aggregate level, we find that taking risk into account substantially increases the poverty headcount. Moreover, a regression analysis suggests that many household characteristics are correlated not only with the average income of the household over time, but also with its variability.