Given the increasing importance of markets (and hence information) in India?s economy, the need for an accurate reflection of economic reality, in order to calibrate policy and commercial decision making, is becoming acute. Capex plans of corporates are increasingly data driven. Data integrity is particularly important for financial markets, which intermediate funds for investment, having to position their high frequency data-driven decisions within a stable framework of macro-economic data.
Unfortunately, there is an increasing apprehension about the quality of national accounts and industrial output data, given the magnitude of revisions that we have been witnessing. Specifically, for the GDP data, the proximate cause of these revisions is stated to have been the recent re-weighting of delineated baskets of commodities in the WPI, but there seems to be much more to the story than this.
Earlier this week, for instance, India?s 2009-10 GDP estimates have been significantly revised upwards, with growth up from earlier estimates of 7.4% to 8%. This is significantly higher than the extent of revisions in FY09 and FY08, where the revisions were about 0.1-0.2 percentage points.
A portion of these revisions in real GDP is due to the change in the GDP deflators, due to the recent change in the composition of the WPI. However, there is significant ambiguity in interpreting the revisions, given the large gaps in the data. While the revision from the production side (all in constant 2004-05 prices) is Rs 29,662 crore, that from the expenditure side is Rs 62,095 crore. This is the gap in just the revisions, mind you, not the actual GDP figures. The difference, conceptually, between the production and expenditure sides is ?net indirect taxes?, since the production side is measured at Factor Cost and the expenditure side at Market Prices. It is difficult to validate the gap in revisions, particularly the composition of the subsidies.
What is peculiar however, is the composition of the revision of the expenditure side of GDP. While the overall revision is Rs 62,095 crore, as noted earlier, that of the consumption and investment components are Rs 86,300 crore and Rs 1,20,600 crore, respectively, implying a negative revision of Rs 1,44,700 crore in the two other components that make up the expenditure side of the GDP. This residual, whose breakups are not given in the GDP data, comprises net exports (i.e., exports minus imports) and ?discrepancies?. The total net trade (goods and services) deficit in 2009-10, to put the residual revision in perspective, was Rs 1,80,700 crore. If you consider a revision of even as large a proportion of 20% in this number, that?s Rs 37,000 crore, leaving Rs 1,08,000 crore allocated to the ?discrepancies? line item. If you were to hypothetically allocate this gap to either consumption or investment (or to manufacturing or services growth on the production side), this would introduce significant uncertainty about the revised growth numbers.
From a sector perspective, the revision in FY10 GDP estimates is due to higher agriculture and services on the production side and rising inventory accumulation (?changes in stocks? in GDP parlance) on the expenditure side. Increase in investment growth has been mostly due to rising inventories and holdings of valuables, even after a downward revision in fixed capital formation (about 90% of total capital formation) both in FY09 and FY10.
That was the inflation adjusted side. To understand the impact of the GDP deflators, the ratio of the old 2009-10 GDP at current and constant prices was 1.31 but 1.35 for the revised estimates. That?s a huge revision, given the magnitude of nominal GDP (Rs 66 lakh crore). This magnitude of revision, however, is not seen for the GDP deflators of the two previous years, which remain virtually unchanged at 1.18 and 1.27, respectively. This was the result of the revisions of the basket of weights in the WPI, which was revised in September 2010 (which will be a separate story altogether).
It is certain that the current magnitude of revisions is transient and the changes in data capture processes and the design of the new indices are making the data apparatus of the government more robust. The endeavours of the former and current heads of the statistical machinery in modernising India?s data processing and dissemination give some cause for reassurance. Unfortunately, however, there is a thick pipeline of impending changes in various measures of prices and output, with a series of revisions in the Wholesale Price Index and the Consumer Price Index, which will keep various deflators in a state of flux, although the end result of the exercise is likely to be a more representative and robust set of numbers. During the transition, it would be helpful to users of the data if a mapping was provided of the changes in the prices indexes (which are well documented) to the construction of the GDP price deflators, to reduce the ambiguity of the effects on the changing deflators and to enable analysts to distinguish these effects from other structural changes.
The author is senior vice-president, business & economic research, Axis Bank. These are his personal views