The rebasing of GDP to 2022-23 received predictable attention when the Ministry of Statistics and Programme Implementation released the new national accounts series on February 27, 2026. Yet, the headline revision obscured a much larger story.
For years, official statistics relied heavily on periodic surveys, benchmark studies and extrapolations. That approach served India well when data were scarce, but it struggled to keep pace with a rapidly formalising and digitising economy. MoSPI's reforms seek to replace that model with one built on continuous administrative data drawn from GST filings, e-Vahan registrations, MCA-21 company filings and the Public Financial Management System.
The changes span four areas: GDP estimation, informal sector measurement, enterprise statistics and public data infrastructure. Individually, each addresses a longstanding weakness. Together, they represent the most significant transformation of India's statistical system in two decades.
GDP Methodology
The revised GDP series introduces several methodological improvements.
For the unincorporated sector, the Effective Labour Input approach has been replaced by the Labour Input method. From 2022-23 onwards, household sector output is estimated directly using the Annual Survey of Unincorporated Sector Enterprises (ASUSE) and the Periodic Labour Force Survey (PLFS), reducing reliance on fixed coefficients derived from older surveys.
Corporate sector measurement has also become more granular. Instead of assigning diversified companies to a single dominant industry, MoSPI now uses activity-wise revenue shares from MGT-7 and MGT-7A filings to allocate Gross Value Added across multiple sectors.
Double deflation has replaced single deflation in agriculture and manufacturing, while the Proportional Denton method has replaced pro rata benchmarking for Quarterly National Accounts. GST collections now serve as a high frequency indicator of manufacturing and services activity. A back series linking the revised methodology with earlier years is expected by December 2026.
These are not merely technical refinements. Single deflation can distort value added when input and output prices move differently, particularly in agriculture and manufacturing. The revised methodology therefore improves both the accuracy and the policy relevance of GDP estimates, especially in an economy where the informal sector still accounted for roughly 45% of GDP in 2022-23.
Informal Economy
Improving GDP estimates is only one part of the exercise. MoSPI is also attempting something India has never achieved before: a comprehensive framework for measuring the informal economy.
In January 2025, Secretary Saurabh Garg convened an inter ministerial consultation to develop a new measurement framework using administrative datasets including GST records and digital payment systems.
The challenge is not statistical technique but economic reality. India's informal economy spans manufacturing, retail, transport, gig work, construction and seasonal employment. A single annual survey cannot adequately capture such diversity, while individual administrative datasets reveal only part of the picture.
MoSPI's solution is to combine multiple sources. ASUSE provides annual enterprise and employment data, PLFS supplies monthly labour market signals, GST records validate business activity and digital payments offer transaction level insights. The framework also proposes incorporating specialised databases such as PM SVANidhi for street vendors, Pehchan Cards for artisans, sectoral boards and even satellite imagery for agricultural activity.
ASUSE 2025 identified 79.2 million unincorporated non agricultural establishments employing 128.1 million workers. Their GVA reached about ₹19.92 trillion, growing 10.87% over the previous year.
For the first time, policymakers will have the ability to monitor large parts of the informal economy continuously rather than infer trends from intermittent surveys.
Business Register
Another longstanding weakness has been India's incomplete picture of its business population.
The last Economic Census was conducted in 2013. Since then, official surveys have relied on sampling frames that require periodic rebuilding, making it difficult to track how businesses are created, expand or exit.
MoSPI's proposed National Statistical Business Register aims to close that gap. Announced by Secretary Garg in June 2026, the register will integrate data from EPFO, the Ministry of Corporate Affairs, GST and the MSME database into a continuously updated database of establishments.
More importantly, it will become the common sampling frame for future Economic Census exercises, ASI and ASUSE surveys, with updates planned every quarter or six months.
A reliable business register would significantly strengthen industrial policymaking. It would finally allow governments to assess whether schemes such as Production Linked Incentives encourage new entrants or merely strengthen existing firms. Today, many such assessments rely primarily on administrative claims rather than an independent statistical baseline.
Data Harmonisation
Better statistics also require better coordination across governments.
Recognising this, MoSPI convened the National Deliberative Summit on Harmonising Administrative Data for Governance in Bhubaneswar in April 2026, bringing together more than 300 participants from 31 states and Union Territories.
The summit produced a four phase roadmap. By December 2026, states are expected to prepare comprehensive dataset inventories, adopt the National Metadata Structure 2.0, appoint Data Custodians or Chief Data Officers and develop machine readable data catalogues.
Subsequent phases envisage dynamic data catalogues, automated quality assessment, API based sharing and, eventually, interoperable AI ready administrative systems linked through 17 common identifiers.
Importantly, the summit concluded that technology is not India's biggest obstacle. Inconsistent definitions, weak metadata and uneven data quality pose far greater challenges than the absence of digital infrastructure.
The emphasis on common standards before interoperability reflects a recognition that poor quality data cannot be fixed simply by sharing it more quickly.
Public Access
Public access to official statistics has been restructured alongside production.
The e-Sankhyiki portal now hosts 21 statistical products containing over 136 million records. In February 2026, MoSPI introduced a beta Model Context Protocol server that enables AI platforms such as ChatGPT and Claude to access seven official datasets directly, including PLFS, CPI, ASI, IIP and National Account Statistics. A semantic search feature allows users to query the portal using natural language.
Alongside this, MoSPI is developing a Common Data Platform to consolidate official datasets currently scattered across departments into a single AI enabled interface. The GoIStats mobile application, launched June 2025, recorded 14,948 downloads at an average rating of 4.8. The Microdata Portal, developed in collaboration with the World Bank, recorded 8.8 million visits since January 2025. Survey publication timelines have shortened: monthly results release within 15 to 30 days, quarterly results within 45 to 60 days, and annual results within 90 to 120 days of survey completion.
These initiatives are intended to make official statistics more accessible, more timely and easier to use for researchers, businesses and policymakers alike.
Taken together, these reforms could materially improve economic policymaking.
For fiscal and monetary authorities, administrative data should narrow the gap between preliminary and revised GDP estimates, reducing uncertainty around growth and inflation assessments.
For industrial policy, the National Statistical Business Register could provide the first continuous picture of enterprise creation, survival and sectoral shifts, enabling more credible evaluation of PLI schemes, MSME support and formalisation initiatives.
For labour and social policy, integrating PLFS, ASUSE, GST and digital payment data offers the possibility of tracking informal sector employment and output far more frequently than has previously been possible.
Better statistics do not automatically produce better policy. But they substantially improve the quality of policy choices.
Remaining Challenges
The reforms are ambitious, but they do not eliminate every concern.
Administrative datasets are created for compliance rather than statistical analysis. Changes in GST filing behaviour, enforcement practices or portal design can introduce breaks in time series that are difficult to distinguish from genuine economic shifts. As administrative data assumes a larger role in estimation, methodological oversight will become increasingly important.
Institutional safeguards also need strengthening. Although the National Statistical Commission continues to exist, its role in overseeing these reforms has not been clearly articulated. With India expected to adopt the System of National Accounts 2025 framework in the next base year revision around 2029-30, independent quality review will become even more important.
The greatest near term challenge, however, lies with implementation. States must establish dataset inventories, adopt common standards and appoint Data Custodians within demanding timelines. MoSPI itself acknowledges that many departments still lack a complete understanding of the data they already possess. Financial support under the revamped Support for Statistical Strengthening scheme should help, but administrative capacity will ultimately determine how quickly these reforms translate into better statistics.
GDP rebasing may have dominated the headlines, but it is only one element of a much broader transformation. MoSPI is replacing a statistical system built around periodic surveys with one increasingly anchored in continuous administrative data, interoperable databases and digital public infrastructure.
The gap between what the reform architecture promises and what state-level capacity delivers within the stated timelines determines whether this transition produces better evidence for policymaking, or simply more data, faster, inside the same institutional constraints.
.png)
