Expert Explains | Better informal sector data behind lower nominal GDP estimates, says Statistics Ministry

What is the main reason for the reduction in nominal GDP in previous years, not just April-June 2025?

The primary reason for the downward revision in nominal GDP across all overlapping years (2022-23 to 2024-25 and subsequent quarters) is the transition from proxy-based estimates in the old series to direct, empirical annual surveys for the informal sector in the new series.

In the previous series, because regular surveys for the unorganised sector did not exist, MoSPI calculated Gross Value Added (GVA) for unorganised sector by multiplying estimated workforce counts by the Value Added per Worker derived from decadal surveys. This base year GVA was projected forward for over a decade using proxies like formal corporate growth rates, inter-survey growth rates, or historical tax collections as no regular data was available.

With the availability of annual survey evidence, it has become possible to better capture the distinct growth patterns of the informal sector. This has provided a stronger empirical basis for updating the estimates and improving their alignment with observed economic activity.

Therefore, the single largest driver of the revision in GDP numbers is the improved measurement of India’s informal services sector. By replacing corporate proxies with actual empirical data from the newly institutionalised Annual Survey of Unincorporated Sector Enterprises (ASUSE) and Periodic Labour Force Survey (PLFS), MoSPI transitioned to direct annual measurements of unorganised productivity and workforce size.

One criticism that has come from former Chief Economic Advisor Arvind Subramanian is that the GDP data does not capture the informal economy properly and we are extrapolating the performance of listed companies. Are we not using ASUSE data for quarterly GDP data?

We are using it. We now have ASUSE on a quarterly basis and PLFS on a monthly basis. So, we are using them; we are not using proxies.

Even in the earlier series, we weren’t using only proxies for the informal sector; that was used only between the quinquennial surveys as you had to use something until the next five years. And that was from the previous base years, so that overhang was continuing.

So, the previous 2011-12 series was overestimating GDP?

No, that’s the point I am trying to make: it’s not an overestimation. GDP is an estimation and you make estimates based on the best data available. When the 2011-12 and 2004-05 series were estimated, certain indicators were used and later changed, leading to better estimates. Now you have even better estimates. Saying it was an overestimation gives the impression there was a systematic bias. That’s alluding a bit much.

A question raised by Pronab Sen in an interview is the use of producer price data. Is the appropriate data available?

We had shifted the method of calculation to producers in the 2011-12 series of the WPI itself. That’s the old series.

The only differences between WPI [Wholesale Price Index] and PPI [Producer Price Index] are that WPI does not have exports and imports and includes taxes and trade margins to some extent. Apart from that, our WPI was akin to a PPI. But we have made those corrections now.

So, the data for the past 10 years has been collected from industry and that change was done way back.

The PPI being used for double deflation was released only in June. Has it stabilised?

It is important to note that the stability of the output PPI was assessed before the launch of the official PPI series as the Commerce Ministry had been compiling an experimental output PPI during that period, which provided an opportunity to assess its behaviour over time. In addition, MoSPI examined the feasibility of using the output PPI in national accounts prior to its official launch.

The prices tracked in WPI are already ‘ex-factory’ prices for manufactured goods (excluding transport charges and GST), which is conceptually identical to the ‘Basic Price’ required for the PPI framework. Empirical validation of the DPIIT (Department for Promotion of Industry and Internal Trade) datasets showed that the month-on-month and annual movements of the new PPI and the legacy WPI at the highly detailed two-digit National Industrial Classification level are highly consistent and stable. As a result, both WPI and output PPI depict similar kinds of movement. So, formally, when output PPI with base 2022-23 is launched, based on our evaluation and consultation with experts, it has been adopted for deflation in the new series as it is conceptually more aligned to the principle of national accounts as noted in the UN System of National Accounts manual.

If the input PPI is currently at a trial stage, what was used to adjust input prices using the double deflation process?

MoSPI did not use the experimental, trial-stage input PPI for compiling the national accounts. Instead, double deflation is executed entirely using the newly released 2022-23 series of Output PPI.

While labeled as an ‘Output PPI’, the comprehensive item basket released by the DPIIT is heavily enhanced and includes both final consumption goods and intermediate goods that act as inputs across industries. The index tracks ex-factory, ex-mine, and ex-mandi basic prices of commodities such as wheat, milk, bauxite, coking coal, electricity, refined palm oil, cotton yarn, cement and semi-finished steel products. Therefore, robust price indices for input items are available within the Output PPI dataset. The input structure is taken from the latest Annual Survey of Industry and such input items are mapped to the concerned items of output PPI.

The statistical framework followed in the new series aligns with the framework followed in the advanced countries such as the US, European Union, Japan, etc.

People want the Sources and Methods document to understand the changes and new data.

That’s a point I want to make: the Sources and Methods will only compile the methodology which we have already put out in the public domain. There will be no (new) data in Sources and Methods: PPI, IIP [Index of Industrial Production] data, everything is there.

Data was released for the 2022-23 GDP series in February. So, I am surprised at these comments that all the data is not there.

The Sources and Methods will only compile the three reports we had brought out on the new methodology. This publication will deliver the precise technical comparison and mathematical bridges required for researchers to evaluate and reconcile the differences between the two series.

Another criticism aimed at the GDP data is that this level of growth is not being ‘felt’ on the ground.

That, I think we need to recognise, is a larger question. I think the GDP is a very important indicator, no doubt. But there are a lot of other factors, uncertainties, and the global situation which determine how you ‘feel’. It’s like when someone mentions, ‘We feel inflation (rate) should be higher’. So, individually, how you see the prices might be different from when they are aggregated across the country and across different areas. I really wouldn’t want to comment on that issue.

There have been calls to give researchers access to databases such as MCA-21, MGT-7/7A, and GSTN in appropriate anonymised format for them to make their own assessment. Will this be made available?

They give it at an aggregate level. But at what granular level they can give it, that’s a separate issue that we will have to discuss with those ministries. Obviously, it can’t be very granular.

Would you say that this new GDP series is by far the most superior so far?

Absolutely, without doubt. Every series, at that time, is the best series. There is no doubt (about that) because we are using the best available data.

Has there been any criticism of the new GDP series that you think is fair?

Obviously, I think this debate is unnecessary. Yeah, that’s a good question. I really can’t think (of any) because we have made the best use of data, digital data, and surveys. I don’t think anyone has pointed out something obvious that we must do in the next (GDP series).

What does this whole GDP debate show? Have people not understood how GDP is calculated or how complex it is to estimate?

Maybe my sense is that no one expected a 7.8% number. When your a priori assumption does not meet the data, and the data is showing that there is no reason to doubt the numbers, perhaps we should not dismiss the GDP data simply because it does not conform to prior expectations.

Forget the GDP data for a moment and look at the real economy data. And that’s partly because government policy interventions have helped support real economic activity and thereby strengthened economic resilience.