Is India's Q2 GDP Surge Believable?

The Gross Domestic Product or GDP growth rate for quarter two of 2025-2026 has come at a whopping 8.2% – which is a six-quarter high – much faster than experts expected. The Reserve Bank of India had also expected a growth rate of 7%. This is surprising because the expected GST reduction impacted production and consumption of various items in August-September. The demand boost came after September 22 – which left just a week before the close of Q2. Reports have come in of many investment projects being withdrawn or curtailed and of net FDI becoming negative. These are not the signs of a robust economy.

IMF pronouncement

The same day, a shocker came from the International Monetary Fund. In its 2025 Article IV Consultation Report on India, it has given a 'C' for assessment of quality of data used for national accounts. The rating is mentioned in Annex VII on Data Issues (page 66). What does a 'C' imply?

“The data provided (by India) to the Fund have some shortcomings that somewhat hamper surveillance”.

In plain terms, it means that Indian official data is not up to the mark for the IMF team to come to a correct assessment of India’s GDP. This assessment is also valid for the current estimate of GDP for Q2 of 2025-26. The shortcomings pointed by the IMF team are:

1. Use of an outdated base year (2011/12), 
2. Use of wholesale price indices as data sources for deflators due to the lack of producer prices indices, 
3. Excessive use of single deflation, which may introduce cyclical biases, 
4. At times sizeable discrepancies between production and expenditure approaches that may indicate the need to enhance the coverage of the expenditure approach data and the informal sector, 
5. Lack of seasonally adjusted data and room for improvement of other statistical techniques used in the quarterly national accounts compilation, and
6. Lack of consolidated data on states and local bodies after 2019.

These points have been raised by several analysts since demonetisation in November 2016, followed by faulty Goods and Services Tax rollout in 2017, the non-banking financial company (NBFC) crisis in 2018 and the pandemic in 2020.

Base-year issue

While the IMF has raised the issue of outdated base year, the problem is deeper. The series for GDP with base year 2011-12 announced in 2015 faced several criticisms: 

1. In 2015 when it was announced, the back series for comparison was not given.
2. It was reworked by a government committee but rejected since it showed higher growth for the United Progressive Alliance period than the National Democratic Alliance period.
3. The NITI Aayog reworked the GDP series, but it is not the appropriate agency for announcing such a series. It showed a higher growth for NDA than the UPA period and was accepted.
4. A former Chief Economic Advisor noted that the GDP was overstated by 2.5% or more.
5. During demonetisation, 3 lakh of the 18 lakh companies in the Ministry of Corporate Affairs data base were removed for being shell companies. As this author pointed out then, this should have impacted GDP calculation. Shell companies are used for under and over invoicing so that value added is not shown originating in the regular companies but in the shell companies. So, removal of the shell companies should have impacted GDP.
6. Further, a survey of the remaining companies for a different purpose (services sector index) found 35% of them were missing from their address. These were possibly fake companies giving fake data. This put a question mark on GDP data.
7. Finally, the black economy is not accounted for in GDP calculations. This has multiple implications.

Other data issues

The base year issue is also flagged by the IMF report for the Consumer Price Index (CPI). The basket used in its calculation is dated, based on the 2011-12 data. Since then the consumption pattern has substantially changed in the country due to changes in the distribution of income and availability of newer goods and services in the economy. These changes are not accounted for when the 2011-12 base data is used.

Importantly, a Census has not been conducted since 2011. It should have been done in 2021 but was delayed due to the pandemic. It could have been done in 2022 or 2023 as has happened in other countries. It is now planned for 2026. In the meantime, demographic changes have occurred since 2011 but they cannot be captured till the Census is conducted. So, the surveys that use the Census data to draw the sample are dated and not representative. This impacts the reliability of the surveys and their results.

Further, government has been systematically rejecting adverse data. Not only was the GDP series changed as indicated above, but the consumer survey of 2017-18 was not released and the unemployment data showing it to be at a 45-year high in 2019 was withheld. Government’s rejection of adverse official data suggests that it is only concerned with painting a rosy picture. Similarly, poverty reduction claim is based on faulty multi-dimensional deprivation data. It compares data for 2015-16 with that of 2019-21. 2020 was a pandemic year and deprivation dramatically increased during that year so how can it be that deprivation (and poverty) decreased as a whole?

Clearly, not only is the Indian GDP data – as now flagged by the IMF – flawed, the reliability of other crucial Indian data is also suspect.

Discrepancies 

The IMF report points to the sizeable ‘discrepancies’ between the two methods of estimating GDP – the production and the expenditure approaches. The former captures the incomes earned and the latter, the incomes spent in the economy. By definition, these two should be equal since they are the two sides of the same coin. But, because these two estimates are arrived at by different methods, they do not match and the difference between the two is called ‘discrepancy’.

As pointed out in Reference iii, both estimates have errors because of lack of independent data for the unorganised sector. This impacts the two estimates differently so the discrepancy changes from year to year. Prior to demonetisation, they used to be small, less than 1%. Now, no more so. They are swinging wildly from positive to negative and back. This causes confusion about which estimate to depend on and how much error even that may have.

Further, the official document giving the methodology of quarterly estimation states:

1. “The production approach used for compiling the QGVA estimates is broadly based on the benchmark-indicator method.” 
2. “In general terms, quarterly estimates of Gross Value Added (GVA) are extrapolations of annual series of GVA.”

These points clarify that for quarterly estimates of GDP, for the production approach, which the officials say is the more reliable method, most current data are not available. That is why proxies are used. But when the growing organised sector is used as proxy for the declining unorganised sector, there is over-estimation of GDP. Effectively, the higher the growth rate of the organised sector, the higher the mis-estimation of the unorganised sector.

Further, when extrapolations of annual series of GVA are used in the estimation for a year in which the economy experiences a shock, it will lead to overestimation. The actual decline of the economy would not be captured. No wonder in the year of demonetisation, 2016-17, evidence suggested that trade and production came to a halt and recovered gradually. It is likely that the economy had a negative growth for the year as a whole while official data still showed a high growth of 8%.

Further, due to the decline in the unorganised sector, the expenditure side estimate also has errors. The unorganised sector largely produces consumption goods and services, so its over-estimation implies an over estimation of consumption in the economy. Further, most of the investment takes place in the organised sector (it is capital intensive), which is underestimated due to overestimation of the unorganised sector, so, investment would get underestimated. These two errors would not cancel out due to their source being different. And, these errors are different from the errors in the production approach. Hence the discrepancy. 

Latest Q2 2025-26 data

Given the various infirmities pointed to above, how reliable can the surge in GDP shown by the latest data Q2 2025-26 data be?

The infirmities pointed out in the Q1 data apply here as well. Thirteen (17 for Q1) out of the 22 high frequency data used to project from the previous year to the current one are lower than in 2024-25. In brief, if the methodology used in the estimation is accepted, the rate of growth in Q2 2025-26 cannot have accelerated compared to 2024-25.

Conclusion

Some analysts argue that if data is not available, as is the case with the unorganised sector, then assumptions have to be made to estimate the data. While this is correct, it does imply that the estimates are conditionally correct. Shocks like demonetisation invalidate earlier assumptions and they require change. Since 2016-17, this change has not been brought about, so official data is out of line with the reality. Estimates are not only incorrect but non-comparable with data for earlier years.

Finally, the higher the growth rate of the organised sector, the higher the error in estimation of GDP because of the mis-estimation of the unorganised sector. Reliability of data is not just a fad but an essential requirement for sound policy making. That is what has forced the IMF to acknowledge the lack of reliability of Indian GDP data.

Arun Kumar was professor of economics at JNU and author of Indian Economy’s Greatest Crisis: Impact of the Coronavirus and the Road Ahead, 2020. 

The Wire is now on WhatsApp. Follow our channel for sharp analysis and opinions on the latest developments.