Pronab Sen, the first chief statistician of India, in a telephonic conversation with Asit Ranjan Mishra, explains his key concern about the latest gross domestic product (GDP) data and the criticism he doesn’t subscribe to. Edited excerpts:
How do you see the 7.8 per cent GDP growth in the first quarter (April-June/Q1) of 2026-27? Does it pass the smell test?
It’s difficult to say whether it passes the smell test because what we have seen is large downward revisions in the recent past. Remember, the first estimate we get for quarterly GDP is based on very, very weak data. So, one can only go by recent history, and recent history has seen fairly substantial downward revisions. So, I am a little shaky about it because I have questions about the data. My personal hypothesis is that this 10 per cent nominal growth they’ve got is probably right. My question is really about the price deflators, which determine real GDP growth.
The Q1 GDP numbers have also become the most contested in recent times. How do you see the controversy that has erupted surrounding them?
I don’t know what all the criticism is. The only criticism I have read is by (former finance secretary) Subhash Chandra Garg.
So, what do you think about his criticism?
I don’t quite agree. The convention has always been to compare the latest estimate with the latest revised estimate.
Garg is also questioning the sharp downward revision of Q1 of 2025-26 nominal GDP under the new base year?
There can be several reasons for it. When you do the base revision, you make corrections for all the overstatements you made previously. Now, usually, these have not been very large. This time, the revision has been large, 7-8 per cent. And the principal reason for it, I think, is overestimation of GDP in the past. So, they’ve just corrected that.
So, what is the real issue with the new GDP series?
I’m questioning the methodology in the new GDP series. Essentially, the changes they have made are all legitimate, desirable, and known earlier as well. The reason we didn’t do it is that we didn’t have the data.
There are a couple of things, and they all have to do with prices and price indices. What they’ve done this time around is double deflation, which is a good thing to do. But the point is, it is much more data-demanding. I don’t know whether they have that data. When you are doing single deflation, you need data only on the price of the final goods.
Now, you need prices not only for the final goods, but for all inputs, which is a much larger number of goods. I have not seen any evidence that they have that data.
The Ministry of Statistics and Programme Implementation (Mospi) has used the Output Producer Price Index (PPI) in double deflation. But for the input deflator, we do not know whether input PPI has been used because it is at an experimental stage.
I don’t know what they have done. This is my problem because if you are to do the input side, you need input prices. Have they collected them? Have they used them? If so, why have they not been made public? So, my problem is really with the data.
Mospi has said the “sources and methods” will be released within two weeks.
Remember, the new base estimates were released last year. It should have already been released so that we know what was going on.
What are your other concerns?
The second is PPI. We had not been able to calculate the PPI for more than two decades because producers didn’t give us the prices. They said that this is a trade secret. PPI you can only get from the producer. You can’t get it from anybody else, whereas the wholesale price index (WPI) is a market price, which you can go and take quotations. You can’t do that with PPI. The producer has to give it to you.
And quite often, there is no one price because a producer may sell to a large distributor at one price and to a small distributor at a different price. So, he has to give you the average. I don’t know whether they’ve been able to collect this data. And if they have, I don’t know how they’ve done it unless Indian companies have become very cooperative.
Do you think Mospi should have waited for the PPI series to mature before adopting it as a deflator?
The good practice is that when you’re introducing a completely new dataset, you should run the two datasets in parallel. Make the estimates on the basis of the two sets separately and present them to the public. So, you can see what the differences are. But to say, I’m discontinuing this (WPI) and starting this (PPI) abruptly without any further information, what you’re saying, in effect, is, “Trust me”. But trust is only built when you can compare. Now, I can’t compare because I have no overlapping series.
Another contentious issue has been the 9.2 per cent real manufacturing growth, where the deflator has become negative (-1.5 per cent). Is this normal while using double deflation or an aberration?
It can happen with double deflation. With single deflation, it can’t happen.
Mospi has said that there is no reason for the GDP deflator to track the Consumer Price Index (CPI) or WPI because their coverage is different. Do you agree? Questions have been raised about the 2.5 per cent deflator in a quarter when WPI was close to double digits.
It does look suspicious. This is my problem because there are too many things happening. You’re switching from WPI to PPI. You are doing double deflation.
In the UK or Australia, detailed supply-use and national accounts data allow researchers to reconstruct much of the GDP calculation. Why is it not done in India?
Because they don’t release the raw data.
So, are you saying that should be released?
Yes. At a certain level of aggregation, of course. You can’t release the data at the unit level because there are confidentiality concerns.
But why was it not done even during your time at Mospi?
We have never been in that habit. It’s a matter of habit. But at that time, things were straightforward enough. The CPI data was released in the full breakdown. The entire data set is available for WPI.
Tell me a few things that you would like to see that would give you more confidence in the numbers that have been released.
What I would like to see is basically how they’ve calculated GDP. How they’ve got the data for the PPI. If you’ve got it, please release it for a length of time so that we can check its stability.
Secondly, what on earth have you done with double deflation? Where is your input data? You’ve given me no indication of that. To do that properly, you would need a massive Supply and Use Table (SUT). Where is that? They discontinued the SUT, in fact. It was being produced and they stopped it.
Not only should they not have stopped, but they should have actually expanded it. And obviously, we should provide more data so that researchers can reconstruct the index. That is better for transparency purposes.
