India loses 4-7% of welfare budget to poor data quality: Report

India loses an estimated four to seven per cent of its annual welfare spending to poor-quality government data, according to a NITI Aayog assessment cited in a new report that warns fragmented data systems will undermine the country’s artificial intelligence ambitions unless addressed first.

The report by the Tony Blair Institute for Global Change, released on Saturday, argues that while India has built digital infrastructure at a scale few countries can match, the data powering government systems remains siloed across departments and states. This gap between digital scale and usable data creates what the report calls a delivery failure, not merely an administrative inconvenience.

Past clean-up exercises illustrate the scale of the problem. Removing 17.1 million ineligible names from the PM-KISAN farmer support scheme saved an estimated ₹90 billion. Eliminating 35 million bogus LPG connections saved ₹210 billion over two years. Dropping 16 million fake ration cards is saving roughly ₹100 billion annually.

The report notes that Indian states generate enormous volumes of administrative data daily, including beneficiary lists, health records, land registers, school enrolments and tax filings. However, these data are collected to satisfy one department’s reporting requirement, not to answer the cross-cutting questions that actually drive policy, the report states.

Data quality roadmap

The paper sets out seven recommendations for states, built around what it calls a state data operating model. These cover leadership and mandate, data ownership and stewardship, common standards, shared data capabilities, independent quality assurance, legal and procurement safeguards, and a State Data Balance Sheet to help leaders manage data assets and risks the way they manage finances.

Vivek Agarwal, country director, Tony Blair Institute for Global Change and co-author of the paper, said the report asks a basic question first: is the data good enough to trust an AI system, or a human official, to act on it?

“Right now, in most states, the honest answer is not yet. That’s not a reason to slow down on AI, it’s a reason to be deliberate about what comes first,” Agarwal said. “States that get this foundation right won’t just be AI-ready, they’ll make better decisions, serve citizens faster, and catch problems before they become crises.”

The report complements India’s ongoing push for data harmonisation spearheaded by the Ministry of Statistics and Programme Implementation. It analyses key government initiatives including NMDS 2.0, SQAF, the AI-Readiness Framework, the QPR Portal and the Model Data Sharing Framework.

State-level programmes

The study draws on state-level programmes including Karnataka’s Kutumba social registry, Odisha’s Social Protection Delivery Platform and Rajasthan’s Pehchan Portal, alongside initiatives in Andhra Pradesh, Uttar Pradesh, Tamil Nadu and Telangana.

India’s digital infrastructure has scaled further and faster than almost any country’s. Aadhaar, the national biometric identity system, has enrolled over 1.4 billion people and supported more than 24.5 billion e-KYC transactions, a process where identity is verified electronically using Aadhaar data.

The Unified Payments Interface now processes over 20 billion transactions a month, an estimated 49 per cent share of all real-time payments globally. DigiLocker has issued and verified nearly 10 billion documents for over 685 million citizens.

Formal banking inclusion rose from roughly 25 per cent in 2008 to over 80 per cent by 2023, partly supported by Aadhaar-linked verification.

But the report argues that none of this is the same as making that data usable across departments and levels of government.

“Without fixing this first, AI deployed on top of fragmented data will simply amplify existing blind spots rather than solve them,” the report states.

Ott Velsberg, former government chief data officer, Estonia and co-author of the paper, said India’s first digital achievement was scale that few countries can match.

“The next test is different. It is about whether data can be trusted, connected and used to deliver better outcomes. As the report argues, India has not yet built this capability with the same consistency as its digital infrastructure,” Velsberg said.

The report concludes that states which address data quality will not simply end up with better dashboards. They will be better placed to catch problems early, direct resources where they are genuinely needed and build public trust in the numbers the government publishes.

As India moves toward wider AI adoption in governance, that foundation will determine whether technology strengthens delivery or simply scales up the same blind spots faster, stressed the report.

Save article