There have always been big differences in estimates of the number of poor people India has "" Surjit Bhalla's estimates, for instance, are significantly lower than the official NSS-based ones as he adds back the consumption that NSS estimates don't capture, while retaining the NSS sample's distributional properties. Another major issue that changes poverty estimates is that of the prices used to calculate poverty levels. Angus Deaton has argued, in a recent article in the Economic & Political Weekly (Feb 9, 2008), that the rate of inflation in food prices between 1999-00 and 2004-05 was around 70 per cent higher than that showed in the food component on the CPIAL (Consumer Price Index for Agriculture Labour), thanks to the excessive weight given to cereals. This, and the higher weight given to food (nine percentage points too much), according to Deaton, has resulted in the CPIAL being underestimated by a little more than 40 per cent. When this is corrected for, Deaton finds the rural poverty numbers in 2004-05 rise from the official 28.3 per cent of population to around 31.1 per cent. With some further changes in assumptions on the share of food in poverty-line households, this rises to 32.1 per cent. So, apart from resolving the issue of the consumption that the NSS misses out, the National Statistical Commission also needs to deal with this issue of price lines.
