GDP calculation: Is the CSO right in its approach?

Academic papers have their own charm, as in economics, and once the assumptions are stated and the models run, there is nothing amiss if all the tests are performed and are shown to work.

Are our GDP growth numbers right? One is not sure, for certain, because the 6.8% number for FY19 does not seem to rhyme with other indicators such as industrial growth or exports or even the state of the banking sector, employment and so on. In fact, the growth number of 8.2% in GDP in 2016-17 when the economy came to a standstill for five months is still hard to reconcile. But the methodology is of global standard and hence there can be no debate on the approach used by the Central Statistics Office (CSO).

Now the latest controversy has been raised by the former Chief Economic Advisor’s academic paper, which shows that growth has been overstated in the last few years, starting 2011-12, by 2.5 percentage points. Academic papers have their own charm, as in economics, and once the assumptions are stated and the models run, there is nothing amiss if all the tests are performed and are shown to work. But that is academia, which has an important position on the library shelf or the hard disk.

Can the GDP of any country be calculated based on a set of variables, which here are around 17? The curious part of this set, which has been used, is that almost all of them are already a part of the GDP calculation. Also, a statement made that the variables chosen have the advantage of not coming from the CSO is curious because the IIP growth numbers for manufacturing and industry are actually brought out by the CSO!

Read Also| Jet Airways stock tanks 41%, hits all-time low

While this approach has been validated across countries, it is not pursued anywhere. No country uses regression models while calculating GDP growth, as it would be too facile an approach, fraught with dangers of oversimplification. This is relevant even more for India where there is a large unorganised sector and data is opaque, which requires the use of proxies for better representation.

In short, the methodology used is that a set of variables are used to see how the economy grew in the past, which validated the GDP growth numbers. This includes auto sales, power consumption, IIP, petroleum consumption, cement, steel, overall real credit, real credit to industry, exports of goods and services, etc. There are a host of jargon and complicated equations and sophisticated diagrams that are presented to prove the final GDP growth numbers. The approach shows that post 2011-12 these variables showed growth, which was around 2.5% lower than CSO’s estimates, but which was not the case when other countries were tested.

The ultimate justification for this thought is that other problems in the economy, which are in the form of unemployment or banking, are well-aligned with the new number and not 7% growth that we have been speaking of. As a corollary, all the policies that have been pursued so far were incorrect because these went on the basis of wrong numbers of GDP, which is serious as results would have been different if RBI had cut interest rates sharply at a time when it was increasing the repo rate. Therefore, there is a need to kick-start the economy today.

Read Also| HDFC AMC shares drop nearly 7% to Rs 1,807

The methodology can be examined a bit closely. The 17 variables chosen are all from the organised sector and leave out the unorganised, which is 40-50% of the economy. There is no mention of agriculture in the model, which makes it weak as 60% of the workforce in the country resides in this segment. The government is not a part and taxes are excluded as they are not reliable. Such a model is bound to be questioned on the assumptions made. Also, if it were so easy to arrive at the GDP number based on these independent data series, it would be a plain desk job that would not involve any surveys or ground studies. This is not convincing, more so because it appears that the 17 variables that have been pulled out from the existing GDP are used separately to reckon GDP as they were well-correlated during the earlier time period.

The exercise would have normally not gone beyond academic discussions because there have been other experts who have put forward their versions of GDP growth in the demonetisation year, which was negative. Yet it has not made the headlines. The fact that the current revelation comes from an ex-CEA stands out as he was from the establishment. All through those years, the Economic Survey, which was authored by him, made the point that India was in a sweet spot, with a touch of hubris. This could not have been said unless there was conviction that growth was blooming everywhere. The fact that the new exercise shows that the economy was abysmal is a turnaround. The audience would ask: Which version was right? Also, the former CEA had defended demonetisation when it was on, but took a different view when out of the government. This raises a broader issue at the government level, because an ex-employee speaking differently once out of the system questions several foundations of the regime.

The issue of credibility, hence, comes in. There has already been a storm over the GDP series, both the back series as well as the revised numbers, as it became a tool for political gamesmanship. With the former CEA now adding a new dimension, data from the government would tend to be questioned. The situation is not very different from, say, a CMD of a bank saying after retirement that all the loans that turned NPAs were given based on impressionistic views and not a formal credit score! The government should probably really consider having confidentiality-like clause in place, which prevents those occupying ‘systemically important positions’ from talking on related subjects post tenure as there is a reputation issue involved.

Also, academic institutes that sponsor such research should be discreet when allowing for such paper releases, given the background of the scholars, as this can also question the reputation of the concerned institute. The processes are surely rigorous, with peer review being mandatory, but when the results potentially can be controversial and struck down as being incorrect, serious questions can be posed on the release of such research.

Should we be concerned that we pursued incorrect policies in the last seven years? The answer is ‘no’, as the ex-CEA’s views are a personal academic venture, which need not be accepted howsoever dramatic the results may be. The conclusion would not have been alluring but for the fact that the release of GDP series of the CSO has been fraught with controversy for different reasons. While we may definitely have to take a harder look at the calculations and proxies being used, deviating from the approach being pursued is definitely not called for. The CSO will have to continue working on ways to make the data more robust, but a change of approach is definitely not on.

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