Sunil Jain: Will India measure up?

RATIONAL EXPECTATIONS

India may grow at 8 per cent if all goes well, or it may grow at seven if they don't. But the real tragedy is we'll never know for sure, given just how poor the quality of measuring data is in the country. So, in the new year let's hope that the newly constituted National Statistical Commission (NSC) is able to make sense of India's data jungle. Apart from the fact that having the correct data is important for those investing billions on the basis of this, what makes this all the more important is that the UPA government is embarking upon so many new policy measures for the so-called dispossessed, based on precisely this dodgy data.

The most well-known example of this, of course, is the difference between the official poverty numbers and those arrived at by Surjit Bhalla, who's now a member of the NSC. When the official poverty numbers were around 26 per cent, Bhalla said the actual numbers were more likely to be 13-14 per cent. He argued that the National Sample Survey (NSS) data used to calculate poverty reflected just around half the total consumption that the National Accounts Statistics (the NAS gives the GDP numbers) recorded""so, the NSS consumption figures had to be blown up, hence the lower poverty numbers.

If this isn't bad enough, consider the problems that arise when the same survey gives you two estimates of poverty! That's right, the NSS 2004-05 has two estimates of poverty, one is 22 per cent and the other 28 per cent. Those in the know will tell you the 28 per cent figure is comparable with the 36 per cent poverty figure of the NSS survey of 1993-94, while the 22 per cent figure is comparable with the 26 per cent poverty estimate of the 1999-00 NSS round. That is, in both cases, the proportion of the poor declined by between 0.7 and 0.8 percentage point each year.

But what's the difference between the 1993-94, 1999-00 and 2004-05 NSS rounds? Apparently it's got to do with the manner in which the question was posed about a household's consumption""what did you consume in the last week versus what did you consume in the last one month? You'd presume both would give the same result in terms of the number of the poor since consumption levels determine poverty status, but for some reason they don't. That's what this big debate was about between economists when the 1999-00 results came out showing poverty levels of 26 per cent""it was argued that while the 1993-94 NSS asked one question, the 1999-00 NSS asked the other one. So, the 2004-05 NSS asked both questions, but to two different sets of respondents. Hence the two answers. Imagine the problem this now poses. The next time you want to know what India's poverty is, the answer will depend upon the way the question is framed!

This is not the only set of wonky conclusions the NSS throws up. The latest NSS shows, for instance, that employment growth in the country is about the highest you've seen in the last two decades. While this vindicates those who argue that economic reforms lead to more jobs and not less, the data on wages from the same NSS survey give the opposite picture. To cite one example, wages for men in urban areas who'd passed at least secondary school grew by 12.3 per cent per annum between 1993-94 and 1999-00, a period in which, the NSS says, employment growth was a mere 1.2 per cent per annum. But, in the 1999-00 and 2004-05 periods, when employment growth rose to 2.7 per cent per annum, the NSS says wage growth for the same lot of persons fell to a mere 1.6 per cent! In real terms, that means wages fell when employment was growing faster.

Or take the figures that the 2004-05 NSS throws up on Other Backward Classes, or OBCs. While the NSS 2004-05 puts the share of OBCs in the country's population at around 40 per cent, the 1999-00 NSS put the proportion at around 36 per cent, which means OBCs would have had to be breeding faster than rabbits in the last five years to get such a massive hike in population share. The other, and more reasonable explanation, of course, is that each state has probably increased its list of castes covered under the OBC umbrella. While this may be true, how do you then account for the fact that, as PS Krishnan pointed out in The Indian Express last month, the NSS showed the share of OBCs among the Muslims in Karnataka has gone down between 1999-00 and 2004-05 when the entire state's Muslim population has been included as a sub-category of backward castes? It showed SCs and STs in states where there were none, and so on. Krishnan, who's an advisor to the human resources ministry and a former member-secretary of the National Commission for Backward Castes, presumably should know these things.

All of this presumes there is no problem when it comes to estimating economic growth""Shankar Acharya, who was the chief economic advisor when Manmohan Singh was FM, has however, made a pretty devastating critique of this as well (see "If growth were slower ...," Business Standard, Dec 28). So, when Prof. Suresh Tendulkar, who's the Chairman of the NSC, comes back from his holiday, he's going to have his hands more than full. Happy New Year!