Assessment of the welfare impacts of low-frequency events, such as macroeconomic crises and stabilizations, are often confounded by sampling and nonsampling errors that generate fluctuations in household survey-based welfare indicators; they are also limited by our ability to explain fluctuations in terms of other available data. Basing policy on short-term movements in welfare indicators can thus be hazardous. There was a sharp increase in India's poverty measures in the aftermath of the mid-1991 crisis and the ensuing stabilization reforms. However, only one-tenth of the increase in measured poverty is explicable in terms of the variables one would expect to transmit the shock. Poverty measures soon returned to their pre-reform levels, belying the notion of a reforms-induced structural break.