From OTT to air & train travel portals: MoSPI eyes new data sources to calculate inflation

Garg told The Indian Express: “For the new CPI series, MoSPI (Ministry of Statistics and Programme Implementation) is expanding its approach by exploring alternative data sources, such as online platforms for airfare, rail fare, OTT platforms and administrative records for price data of petrol, diesel and LPG. Discussions are ongoing with IRCTC, under the Ministry of Railways, and the PPAC under the Ministry of Petroleum and Natural Gas for direct transfer of data for integration in CPI.”

The MoSPI is also exploring the use of scanner data and web scraping with an aim to enhance the accuracy, efficiency, and comprehensiveness of price data collection, Garg said. “The possibility of collecting price data from e-commerce websites is also being considered,” he said.

Among other alternative datasets being tapped by the Ministry, Garg said data of Goods and Services Tax (GST), e-Vahan portal, UPI transactions from National Payments Corporation of India are going to be used for GDP calculation by the National Statistics Office (NSO) in addition to data from Controller General of Accounts, MCA-21, and the Reserve Bank of India. MoSPI has also used the GSTN (GST Network) database for its new service sector survey, Annual Survey of Service Sector Enterprises.

The recent back-to-back Household Consumption Expenditure Surveys for 2022-23 and 2023-24 asked questions on streaming services and air and rail fares. The current CPI basket has a weight of 0.077 per cent for air fare (economy class), 0.185 per cent for railway fare, 0.08 per cent for internet expenses, 0.82 per cent for monthly charges for cable TV connection, and 1.839 per cent for mobile phone charges.

Niti for think-tanks

Meanwhile, speaking on Thursday at the start of a two-day workshop on ‘Using Alternate Data Sources and Frontier Technologies for Policy Making’, Niti Aayog Vice Chairperson Bery questioned if these new areas should be a priority for MoSPI and cautioned that they should “not become an all-consuming task or preoccupation”.

However, he did say that using new data sources and technologies is “very powerful” as it helps give a sense about the direction in which the economy is moving in a fast-moving world for real-time interventions.

“I would also make the point that to some extent the issue of whether all of this should be going on in MoSPI and in the statistical infrastructure or whether it should be going on in our rich network of think tanks is, I think, an appropriate issue. For example, NCAER… has a centre for data analytics,” Bery said.

New Delhi-based think-tank National Council of Applied Economic Research (NCAER) set up a National Data Innovation Centre in 2017 to serve as a “laboratory for experiments in data collection”, among other objectives.

Meanwhile, speaking at the same workshop, Garg said alternate data sources and frontier technology have “reached a stage that we can actually use it for official statistics” and that this augured well for the future of data analysis.

Chief Economic Adviser V Anantha Nageswaran, also speaking at the workshop, pushed for the use of alternative data in official statistics, saying that while surveys, administrative, and national accounts data remain indispensable as inputs while making decisions, the pace, complexity, and granularity required by modern policymaking and the challenges attached to them “have exposed limitations in both the periodicity and dimensionality of such data”.

According to Nageswaran, alternative data such as satellite-based night-time luminosity – used as a proxy for economic activity in areas with delayed or weak statistical reporting – and other satellite data can help policymakers monitor industrial activity, road connectivity, and cropping patterns, among others.

“These insights can inform timely decisions on input provisioning, crop insurance payouts, and regional procurement strategies,” he said, adding that these new types of data capture emergent behaviour, respond faster to shocks, and reflect the “lived experience of economic agents in ways that conventional aggregates sometimes cannot”.

However, the government’s top economist warned that while alternative data can help policymakers move from “retrospective diagnostics to proactive intervention”, they could not replace official statistics and warranted “intelligent integration”.

“Therefore, the mature approach is not to choose between official and alternative data but to design systems where each informs and validates the other, especially in environments where timely action is crucial,” Nageswaran said, adding that enthusiasm must be tempered with prudence as official data still carried a “certain higher sense of authenticity and reliability and accuracy given the years of usage and in-build checks and balances”.

Base revision by MoSPI

In the interview, Garg also spoke about the base revision exercise being undertaken for all major datasets such as CPI-based inflation, Index of Industrial Production (IIP), and Gross Domestic Product (GDP). For GDP, the new series is scheduled to be released on February 27, 2026 with financial year 2022-23 as the new base year, he said.

The revised base year for IIP has also been tentatively identified as 2022 -23, he said, adding that the IIP with the revised base would be released from 2026-27. At present, the base year for GDP and IIP is 2011-12 and for CPI is 2012. For CPI, 2024 has been identified as the revised base year.

“…the item basket and the weightage of the items would be decided based on the NSO’s HCES conducted in 2023-24. The new CPI series is expected to be published from the first quarter of 2026,” Garg said.