Deconstructing the controversy over GDP growth: Modi’s ‘army’, Modi’s  ‘statistics’ ?

There are, it is said, two kinds of statistics, one that you look up and the other that you make up. Statistics are pliable and that is why it is also said that there are lies, damned lies and statistics. But the Modi government’s think tank, Niti Aayog, and the Central Statistical Office (CSO) now have the dubious distinction of showing that statistics can also be ‘political’. Just as Prime Minister Narendra Modi is fond of saying, ‘my’ Army, ‘my’ soldiers, he can now lay claim to ‘my’ statistics!

Indeed, Bloomberg’s Andy Mukherjee was quoted as asking whether India’s statistics will now depend on “which government is in power”.

Bill Gates (Microsoft and Gates Foundation) is said to have wondered at the statistical probability of 80 per cent of the Chinese approving their unelected government and only 20 per cent of the Americans approving their elected one. What he didn’t say was the old saying that statistics are often used as drunks use the lamp post, for support and not for illumination.

The CSO has been calculating the national income every quarter. But last week for the first time, the CSO took a back seat as Niti Aayog’s vice-chairman Rajeev Kumar took centerstage and made a power point presentation to announce the revised ‘back series’ data. Former Chief Statistician Pronab Sen found the role of Niti Aayog problematic and even a pro-Modi economist like Surjit Bhalla, a member of the Prime Minister’s Economic Advisory Council on the economy, reluctantly declared, “I, along with others, also found it inappropriate for NITI Aayog to be directly involved in the presentation of statistical data by the CSO.”

A panel, headed by economist Sudipto Mundle, had already done the exercise earlier this year and shown that the economy had grown at an average 8 per cent per annum during 10 years of the UPA between 2004 and 2014. This was higher than the rate calculated by the CSO earlier, which was 7.5 per cent.

The Mundle panel’s data were, however, withdrawn and the CSO got down to re-calculate the back series. It has now concluded that the growth during the UPA years was actually lower at an average of 6.9 per cent. This is also lower than the average growth of 7.4 per cent during the Modi years, between 2015 and 2018.

While former Union Minister P Chidambaram angrily described it as a ‘bad joke’, statisticians quibbled over the integrity, consistency and the continuity of data. Former RBI Governor YV Reddy wryly commented that while the future looks uncertain in most parts of the world, in India even the past is uncertain.

Economists have taken pains to point out that there may be nothing wrong in the methodology used by the CSO, that the methodology may even be superior to what was done earlier. But they also maintained that the CSO has not been fully transparent about the methodology, and it should allow an independent panel to examine it.

The problem with the back series is that the figures are not consistent with economic indicators. The gap between the ‘real’ economy, measured by real people on the ground as opposed to statisticians poring over data, cannot be so wide so as to defy reason. In an editorial, financial daily Business Standard called upon the government to withdraw the announcement. “[The government] should withdraw the announcement regarding the back series of the GDP data…the data does not align with that from the real economy... this reflects poorly on the ability of the back series to accurately reflect what happened during these 10 years.”

Economist Vivek Kaul also joined issues and said that ‘reality’ appeared very different from the back series data while looking at indicators like domestic car sales, domestic tractor sales, coal despatches etc. The data in short have not passed these basic ‘smell tests’, prompting demands for a clearer explanation of the methodology that went into calculating the new numbers.

In a Television interview, Mundle explained that while the economy changes over the years, the changes rarely take place overnight. It is an incremental change over several years that statisticians try and capture while re-calculating national income and growth of the past years. His panel, Mundle explained, had distributed the changes over the years, making sure that the changes in the data are also small and incremental.

Since then, other commentators and economists too have weighed in and the controversy shows little sign of abating. Clearly outraged, Ashok Desai suggested that the government may have invented a ‘new economics’ and added, “ Anything is possible in a country that manufactured airplanes three millenia ago.” Pointing to one of the absurdities relied upon by the CSO, Desai wrote that the CSO had treated the banking department of the Reserve Bank of India as a market enterprise.

“I cannot imagine what this business produces. I can understand the inclusion of stockbroking in GDP, but I still cannot see what stock exchanges produce: they are just markets, like any other,” was Desai’s acerbic comment. What’s more, a GDP growth of 6-7 per cent, he explained, would mean an increase in both productivity and income of the ‘average man’ by 4-5 per cent. “The common man must wonder where this lucky average man is to be found,” he concluded. In any case, figures for past years have never been recalculated with newer and more recent weights.

Theoretically, the national income of 1818 or 1918 can also be calculated using weights of 2018-19. But as economists have pointed out, the exercise would be futile because many of the services and goods in 2018 did not exist 200 years ago.

Raghav Bahl in The Quint explained it by citing the example of the mobile phone. In 2005, a mobile phone user paid much more for voice and data. Because of higher rates, the usage of voice and data was much less then than in 2018. But if the service were to be measured in terms of minutes and hours of data consumed in 2005, it would provide an unrealistic picture because there was no Facebook, YouTube videos or iPhones.

Similarly, the CSO and Niti Ayog’s defence that they have used a much larger data set to determine growth (five lakh companies as opposed to five thousand) has also been questioned. A larger data does not necessarily lead to better estimates, say statisticians. Using data which did not exist in 2005 to re-calculate national income would also be a statistical impropriety because different sets of data cannot be used to make comparisons.

This article was first published in National Herald on Sunday.