But to arrive at the magical figure of 19 per cent, though not deliberately, the Planning Commission adjusted the NSS data with the much higher income data of the Central Statistical Organisation. But, says Abusaleh Shariff, head, HRD Programme Area, NCAER: "The problem of using state domestic product data is that it includes productivity of machines too and is consequently much higher. Its not survey data. This is done on the plea that the poor tend to understate their consumption. Thats wrong, only the very rich do that. Secondly, the last Commission estimate came out of a much smaller sample of NSS (50th round), compared to the five-yearly samples of 1.5 lakh people. As a result, the disparity between the actual observed expenditure and state incomes is growing. And poverty is falling dramatically." NCAERs own figure for rural poverty is 39 per cent in 1993-94 (poverty line equals Rs 2,944 per year), a little higher than the new estimates. It has found that poverty has increased in intensitythe Amartya Sen index for India is only 0.18. For instance, 16 per cent of the poor have a daily income of Rs 3 and another 18 per cent get Rs 5.5.