Martin Ravallion IN an article in the January 10, 1998 issue questioned whether the available data on poverty and food prices in India have been interpreted correctly in some recent debates on policy reform,1 In the April 4 issue of EPW, Mohan Rao disputes some of my argument,2 He presents no new data,3 but he does appear to question aspects of my methods of analysing the data that are available. The issues at stake here are too important to ignore, so I would like to take this opportunity to respond. I will confine my attention to Rao's main substantive concerns.4 My article began by confirming one claim often made by critics of policy reform in India, namely, that there is a strong positive correlation over time between the poverty rate, as conventionally measured, and the relative price of food (a correlation coefficient of 0.76 using 24 observations from the National Sample Surveys)' The article then tried to test various explanations for that correlation. It argued that there are plausible explanations which have little or nothing to do with the interpretation that critics of policy reform have given.the correlation.