Pronab Sen interview: ‘Data release by anybody other than the CSO sends out wrong message’https://indianexpress.com/article/business/economy/pronab-sen-interview-data-release-by-anybody-other-than-the-cso-sends-out-wrong-message-gdp-back-series-data-5475392/

Pronab Sen interview: ‘Data release by anybody other than the CSO sends out wrong message’

The NITI Aayog’s like the old Planning Commission is a political one. It is supposed to be an extension of the Prime Minister's Office and the moment the NITI Aayog comes into the picture, it tends to give a political colour.

Interview with Pronab Sen: There needs to be some distance between govt and statistical system
Pronab Sen

Pronab Sen, the Country Director for the India Programme of the International Growth Centre and former chief statistician of India recently spoke to The Indian Express on the release of the GDP back series data. Edited excerpts:

You have said that downward revision for tertiary sector’s growth was still expected but not for secondary sector.

That’s correct. I was expecting a downward revision for tertiary sector’s but secondary sector downward revision comes as a big surprise.

What’s the reason for it? Are the proxies used starkly different from the current 2011-12 series?

Advertising

There’s a big difference, that’s why the back series was so difficult to do. The current series is based on company data, which is MCA-21, which is the balance sheet data, whereas the back series is based on volume data. So, in the case of secondary sector, it was based essentially upon the Annual Survey of Industries (ASI). We do know that ASI was underestimating the growth in manufacturing, so I had expected that to be revised upwards. But that has been revised downwards as well.

The data now being in sync with UN System of National Accounts, that’s been cited by NITI Aayog. How pertinent is that argument?

The System of National Accounts prefers to go with volume indices, so if the new series had been volume index based then that would have been OK. Unfortunately, the new series is not. And the big difference between the volume index approach and the financial data approach is that the financial data captures changes in quality which the volume approach does not. So if a substantial part of the growth has been coming from quality, then if you take the volume approach, you would underestimate the growth. That’s the big difference.

Another argument has been that credit growth was much higher post-2008 crisis or the company earnings not getting reflected.

Again the same issue. As far as financial services are concerned, there are two arguments to it. One element is financial services as captured by the data on organised financial sector, which are banks, insurance companies and so on. Even earlier, as far as I remember, for lot of these companies because they were organised, the RBI used to put together that data and that was done from balance sheets. So there should have been no change in it.

You have questioned the role of NITI Aayog in the release of back series.

Well, that is what came as a shock. This is a technical exercise and this sort of data to be released by anybody other than the CSO itself, to my mind sends out a wrong message. The NITI Aayog’s like the old Planning Commission is a political one. It is supposed to be an extension of the Prime Minister’s Office and the moment the NITI Aayog comes into the picture, it tends to give a political colour. This may not actually be true but that is the impression that comes out. Otherwise, they have absolutely no business in convening the data release.

On the choice of proxies used, such as minutes of usage instead of Telecom subscribers, are those choices subjective in some sense?

Not really. You see the number of subscribers was a convenient way of doing at one time that really is not the best pick. In the new series, we do use the amount of time use of the telecom services. That’s a better way of using it. There’s nothing implicitly wrong in that.

So statistically you are OK with the back series?

Well, I don’t know. Again, one would really need to know how they have gone about it because that’s not entirely clear to me. By and large, what one would have expected, is that since you are doing until 2004-05 and 2011-12, the period for which you are backcasting, you would hold 2004-05 and 2011-12 constant and you would simply bring a portion, the discrepancies between the old series and new series, over that period. If that had happened, you shouldn’t have seen reduction in the growth rate of this magnitude.