Statisticians Vindicated: The GDP Numbers Were Not Fudged After All

The clear winners in this GDP versus that GDP debate are the Indian statisticians. Beyond the important question of whether the slowdown in 2016-17 predates or follows demonetisation, it is nice to know that the avalanche of articles (here and here) that portrayed the Indian statistical measurement system as highly suspect, have hit a bulwark.

The bulwark is the compelling evidence the Central Statistical Organisation (CSO) has provided that it has not fudged the gross domestic product (GDP) data. It has laid low a highly vituperative campaign that had begun to harshly equate the quality of Indian statistics with the long held justifiable suspicion about China’s official statistics. The crescendo affected foreign commentators writing about India too.

“China has long been suspected of manipulating its GDP numbers to suit policy purposes. India's figures are also the focus of skepticism. It therefore may be wise to stop treating GDP as the principal indicator of such a country's economic performance,” the Nikkei Asian Review said on 20 December 2016. There were plenty of such examples especially after February this year, trashing the Indian GDP estimates.

It was, therefore, amusing – even touching – to see the avid faith with which the same CSO’s provisional estimates of GDP for 2016-17 released on 31 May, being accepted as true.

This is what makes any analysis of the Indian economy by a large segment of commentators suspect. If the data released in February was suspect, how can data released on Wednesday by the same organisation suddenly be trustworthy? There is no report that the government had, in the interregnum, purged the entire army of enumerators and appointed a fresh set including the chief statistician to measure the economy afresh. There is also no evidence that the changes which they were bringing about in the system of measurement were therefore a Houdini trick to make the economy look better than what it was actually capable of. And now worse than what it actually is.

As the Ministry of Statistics and Programme Implementation had repeatedly told Parliament through these months, the new series of national accounts with base year 2011-12 was released in January 2015 “as per internationally accepted practice, to capture the changing structure of the economy. This ensures capturing latest information and hence accurately reflects the current economic situation in the country. GDP based on old series would not reflect the current economic situation correctly”.

It also noted that the new series was compliant with the latest United Nations guidelines in System of National Accounts, 2008. “The new series represents a structural break from the old series… (and) therefore, is not comparable with the old series with base year 2004-05.”

The revision exercise for the Indian GDP had begun in 2012. It had naturally taken time to bring it to implementation. Yet writing in Harvard Business Review Prof Bhaskar Chakravarti, of Fletcher School at Tufts University was clear: “The official economy-wide data struggled to reflect the reality on the ground precisely because cash transactions are fragmented and defy accurate data capture”.

If that is so, how has it captured the weakness now? Rather it has escaped attention of many that the slowdown may be more pervasive than initially estimated since the estimates of GDP for say 2015, were based on a wholesale price indicator (WPI) and so may have overestimated growth. This is quite possible as the WPI had swung from positive to negative sharply by the middle of 2015. Since more elements of the Indian GDP is deflated on the basis of WPI, this is quite possible. No attention was paid in the literature to this aspect which would have showed that the Indian economy had begun to cool off way before demonetisation.

Has it got something to do with the fact that few commentators bother to learn the intricacies of measuring GDP as it is actually done by the statisticians? Else they could have productively read this article by Deep Mukherjee, visiting faculty at IIM Calcutta, to figure out the problems.

In any case, it is quite reminiscent of another kerfuffle some years back that surrounded the estimate of so called Rs 32 per day poverty line that was ascribed to Prof Suresh Tendulkar. Try as he might, there were few ready to take the pains to understand what the distinguished scholar was driving at. Criticism was not in short supply for him.

Finally, right now there is work going on to refine India’s real sector statistics by a committee set up by the National Statistical Commission. The committee is expected to cover inter alia the requirement of statistics for estimation of GDP, and data governance for quality, timeliness and credibility. Basically, they will be expected to suggest putting in more tech “for collection of granular data from primary sources”. But without examining the details once the report becomes due for release, expect more fun and games about Indian statistics.