Wednesday, January 13, 2010

Missing state by Manas chakravarhty , Mint


India's GDP (gross domestic product) data do not add up. ndia's GDP (gross domestic product) data do not add up.
An entire state about the size of Uttar Pradesh (UP) appears to have gone missing. This Lost Pradesh, or Errors and Omissions Anchal or Discrepancy Nadu, call it what you will, fluctuates in size from year to year, but seems to be mysteriously growing larger and larger every year.

Why do I say there's a missing state? The answer is simple: if you add up the gross state domestic products at constant prices put out by the various directorates of economics and statistics in the states, the total is lower than the GDP of the country, as computed by the Central Statistical Organisation (CSO).

If you go to the CSO website and look at the state domestic product at constant prices, 1999-2000 series, you'll find a list of all states and Union territories (UTs) with their GDPs. At the bottom of the list they also have a line for all-India GDP 1999-2000 base, which seems to indicate that it's the sum of the GDP numbers of all the states and UTs. But it isn't so.

What is remarkable is that the totals of the states and UTs do not agree with the all-India number. No wonder CSO is at pains to point out, at the bottom of the table, that the state GDPs have been supplied by the states directly, which implies they aren't responsible for it. (This is GDP at constant prices, 1999-2000, at factor cost. The all-India figure for GDP at market prices, base 1999-2000, is even higher). At the very least, they should have had a row containing the states' totals and added a note about the discrepancies.

I would have expected that, left to themselves, the states would have been interested in showing a higher state domestic product. After all, West Bengal's growth in GDP in the 1990s had attracted plenty of scepticism, because there was little evidence on the ground to show that the growth was indeed taking place. The implication--the state was fudging its GDP data.

But if all the states inflated their GDPs, then their totals should have exceeded the all-India figure. But the reverse seems to be true. All-India GDP is much higher than the combined GDP of all the states and UTs. And not by a small amount either; in 2006-07, for instance, the total of the states' GDP was Rs26,17,531 crore, compared with Rs28,71,118 crore for the all-India figure. That's a difference of Rs2,53,587 crore, or more than the GDP of a large state like UP, whose GDP in that year was Rs2,37,420 crore. The discrepancy is as high as 8.8% of the all-India GDP for 2006-07.

But why should the states underestimate the size of their economies? That's why it looks like we have a very large well-hidden state that appears in the national GDP data, but doesn't figure in the states' list.

The first chart gives you a picture of the discrepancies.
Since the discrepancy is not constant, there's also a difference in the rates of growth of the economy if we take GDP according to the all-India figure computed by CSO, or if we go by the totals of the states. The second chart shows the difference between the growth rates. The difference in growth percentages is not much, except for 2005-06, when it was as much as one percentage point.

This is, of course, not the only discrepancy in the computation of GDP. When calculating GDP at market prices, for example, discrepancies add up to a substantial amount. For instance, CSO says that the sum of the various components of expenditure in the second quarter was lower than GDP at market prices by 2.1% in the second quarter of 2009-10. In the second quarter of 2008-09, this difference was as high as 7.6%.

Interestingly, the difference between the states' total and the CSO all-India figure has been increasing, as is seen from the chart. It was 8.2% of all-India GDP in 2001-02 and 9.4% in 2007-08. The 2008-09 states' total, however, should be larger, because GDP figures for Nagaland, Tripura and the Andamans are not available. Once these numbers are in, the discrepancy for the year should be slightly lower.

The data for 2008-09 haven't been taken into account as they are incomplete, with only 18 states having declared their GDP, according to the numbers made available by CSO.

No comments: