Official data has always been questioned but by not releasing jobs survey, Modi made it appear he was manipulating the system.
You can question the timing of the 108 economists and social scientists who have put out a note doubting the credibility of India’s official data, and alleged that the data has been influenced by the government, but most will accept that the government has hurt its own case. By refusing to release the latest jobs survey especially, the government has given the impression that it will not release any data that goes against its narrative of a booming economy with jobs aplenty. Indeed, while many questions have been raised about the quality of official data in the past as well – even the economists/social scientists acknowledge this – the government has needlessly politicized the process.
So, when the GDP back-series data was presented to the public – this is when the NDA’s growth numbers first rose above those of both UPA-1 and UPA-2 – this was done by Niti Aayog instead of the official statisticians. To quote the press release of the 108 academics, Niti is “an advisory body which had hitherto no expertise in statistical data collection”. It didn’t help that, while the Sudipto Mundle-panel’s GDP back-series bumped up the UPA-2 average growth to 8% per year, this got lowered to 6.7% in the revised back-series versus the Modi government’s average of 7.35%.
If the official statisticians had been handling various briefings instead of Niti Aayog, the public perception could possibly have been different and the nub of the allegation, that of political manipulation, could probably have nipped in the bud since the rebasing exercise which led to the change in GDP growth rates – the base year was changed from 2004-05 to 2011-12 – was set in motion by the UPA. The methodology for this was also approved by the UPA; that the numbers came out when Modi was in power is just a coincidence.
It is certainly difficult to understand how GDP growth of 9.3% in FY06, for instance, became 7.9% in the rebasing exercise; in overall terms, the rebasing lowered UPA-1’s GDP growth from 8.1% to 6.7%. But it can surely be argued that this exercise also significantly raised GDP growth in FY13 and FY14 – from an average of 4.9% to 6% – which were UPA years. And while the academics argue the revisions “did not square with related macro-aggregates”, it is possible to argue that the fall in WPI in these years boosted growth numbers. And while the much lower credit growth in the NDA years as compared to the UPA’s is seen as clinching the argument that the new GDP numbers are incorrect – if credit growth slows, how can GDP grow faster? – the argument is less convincing once you factor in the much lower inflation in the NDA years. As compared to 6.9% in UPA-2, WPI grew a mere 0.6% in the NDA period.
A similar argument is made by the academics about FY17 – the demonetization year – where GDP growth estimates were raised from 7.1% to 8.2% between January 2018 and January 2019. This does seem like the statistical system being manipulated since a higher growth – the highest since FY11 – would suggest demonetization didn’t hit economic growth. But keep in mind that this is also a year in which passenger vehicle sales grew at 30.5%, and two-wheelers at 6.9% versus 27.9% and 3% in the previous year; also, demonetization took place only in November, so the major part of the year was over before its impact could be felt. Indeed, FY18 GDP growth fell to 7.2%, which is consistent with the lagged impact of demonetization.
An example of how getting statisticians to present the data works better is former chief statistician TCA Anant’s defense of the back-series. Anant argued that one reason for higher growth in more recent years was that corporate data from the MCA database – this boosted estimates of GDP – was not available for earlier years; in other words, when better-quality data was available, the government had to use it to improve the quality of GDP estimates, never mind if the same series was not available for the past.
But what really cemented the view that the government wouldn’t release data that didn’t suit it was the refusal to make public the latest jobs survey; this, in fact, led to the resignation of the head of the National Statistical Commission along with one other member. The increase in unemployment that the survey showed – according to a leaked version of it in Business Standard – looked plausible since, over time, India’s employment elasticity has been falling, and, in any case, not much of recent GDP growth has come from employment-intensive sectors like readymade garments; agriculture growth has also slowed under Modi.
This is embarrassing for a government going to polls, and it also runs contrary to the narrative – created by Ghosh & Ghosh using EPFO data, primarily – of healthy jobs creation, but this doesn’t justify not releasing the latest jobs survey. But, more important, as Avik Sarkar, Niti Aayog’s data analytics’ head wrote in this newspaper (goo.gl/Bv3rTi), the latest survey was not comparable with the earlier ones on jobs that the NSS carried out since it canvassed more educated people than in the past; while less than a fifth of Indians have studied beyond Class 12, over 75% of those canvassed in the jobs survey were those who had studied beyond Class 10. As a result, Sarkar argued, it gave lower estimates of jobs and labour force participation as compared to earlier NSS surveys. Instead of stressing this vital difference, by refusing to release the survey, Modi’s advisors have let him down, and even though there is no evidence of it, opened him to the charges of manipulating data.