Column: Disturbing datasets

Written by Madan Sabnavis | Updated: Feb 8 2014, 08:14am hrs
Two recent economic releases from the government are both encouraging and disturbing. The revision in GDP growth number for FY13 to 4.5% from 5% provides a lower base for FY14 which will provide the statistical impetus in an otherwise disappointing environment. The fact that our fiscal deficit is 95% of the targeted amount is positive for achieving the redline figure of 4.8% of the GDP for this year. But where, then, is the cause for concern

Revisions in GDP data have become a habit in the last four years. There are significant divergences from the initial number released in May and the revised number a year later in January and the final number. In FY09, the CSO got the number right all through, at 6.7%. In case of FY10, the initial number of 7.4% went up to 8% and then 8.6%. For FY12, it moved from 8.5% to 8.4% to 8.9% while the FY12 number was scaled down from 6.5% to 6.2% but finally reverted to 6.5%. The issue is that with such revisions, the sanctity of the data gets questioned. In fact, policy decisions taken based on such numbers, in retrospect, would sound inappropriate.

There are two problems here. The first is that data collation is a challenge considering that ours is a complex economy where the unorganised sector dominates real estate, retail trade, business services, eateries, etc, cannot be tracked and their output would be subject to imputations. The second issue is conceptual. Ideally, one should not be changing weights usedas has been done this time, by using ASI data instead of IIP. While the CSO release says that the changes are marginal, it still does cause confusion in interpretation of the data. Ideally, any change in methodology should be done when the basic index or variable is being benchmarked with a new base year.

Carrying this problem further, we have seen substantial revisions in both, the WPI and the IIP, this year which poses a problem for RBI. When Governor Raghuram Rajan presented his first policy, he would have used the August WPI inflation figure of 6.1%which actually is 7%while, in the October review, he could have gotten more aggressive if he knew that the September inflation figure was not 6.5% but 7.1%. In fact, for the year so far, the standard deviation for the differences between initial and final inflation figures was 0.37 which works out to annualised monthly volatility of 1.29%. The same story holds for the IIP which had a standard deviation of 0.44.

The suggestion is that it is preferable to wait before releasing data especially when there are problems with the system in terms of information flow not being robust. Therefore, instead of having a lag of 14 days for the release of WPI numbers, it should ideally be deferred to the point that we are sure of the data. The same should hold for IIP, where we should ideally present only cumulative data as monthly numbers are misleading as several industries go through phases of bunching of productionwhich holds for both consumer and capital goods. In fact, given the impact of inventories, the CSO should provide separate data on saleswhich can be procured from the income statements of companieswhich will better capture manufacturing activity. In the case of the GDP, quite clearly, the processes have to be cleansed, else the credibility of data would be at stake as critics would disparage the entire growth story based on such fickle data.

The fiscal deficit data presented, which shows that 95% of the target has been attained, when combined with the news that the government borrowing programme would be lower than budgeted should normally generate enthusiasm. There are reservations here as while it would statistically be achieved; the quality of the budget would necessarily have to take a hit. On the revenue side, there are moves to get the PSUs to pay higher dividend as tax revenue would be lower due to lower growth in GDP with excise and customs unlikely to pick up in the last quarter given that industrial growth, which is negative so far, can just about move to the positive zone. The Budget also has a huge sum of R40,000 crore as non-tax income from communications and the spectrum sale, which would be critical here, along with the disinvestment amount of R54,000 crore, likely through the SUUTI sale as well as cross-holdings by other PSUs. Quite clearly, any success on the revenue side will be on the account of the PSUs providing support rather than fiscal efficiency.

It may still be argued that the government would be within its rights to garner revenue through these routes. But, on the expenditure side, the slippages are more compelling. If the food, fertiliser and fuel subsidies are summed up, so far around R2 lakh crore out of budgeted R2.2 lakh crore has been consumed. With three months to go, logically, on a pro-rata basis, there should be a spillage of at least R40,000 crore (at the rate of R22,000 crore a month). If the redline is to be met, it has to be through either rolling over this amount to next year by not paying it (budgets are based on cash being paid and not on accruals) or by cuts in capex. Last year, the government had cut down on capex from R87,500 crore to R73,500 crore. This year, capex is likely to be R99,000 crore. While assuming the role of a welfare state can be justified in an economy like ours where poverty ratios could be anywhere between 25-30%, cutting down on this expenditure may not be the best way out to attain the 4.8% fiscal deficit number. More importantly it will affect GDP growth in FY15 too, and spoil the party.

The author is chief economist, CARE Ratings. Views are personal