For over a decade, Direct Benefit Transfer (DBT) has been sold as India’s great administrative upgrade: cut out the middleman, move money straight into the citizen’s account, and watch leakages fall. The promise is elegant, even moral. If the State can place the right amount into the right person’s bank account at the right time, with minimal friction, welfare becomes both humane and fiscally responsible.
But a hard truth is now staring us in the face: adopting technology and trimming intermediaries does not automatically deliver that promise. In fact, when governance is weak, digitisation can merely make the pipeline faster—without making it cleaner.
The Odisha “Smoking Gun”
The immediate spark for the present debate is the CAG’s audit of DBT implementation for Odisha’s Post Matric Scholarship (PMS) and Medhabruti schemes. The findings read like a manual on how to defeat a digital system: money routed through non-existent institutions, payments to students in ineligible institutes, and double payments to the same students across multiple courses. Even more damaging was the discovery that most beneficiary bank accounts were not Aadhaar-seeded—turning the “JAM trinity” into an article of faith rather than an enforceable control.
This is not a story about one state’s moral failure. It is a story about the national temptation to treat DBT as a software installation rather than a public-finance system that must be designed, stress-tested, monitored, and independently audited.
Cutting Intermediaries Without Fixing the Process
DBT was meant to reduce human discretion, shorten delays, and create a clean audit trail. Yet the Odisha audit shows that the old intermediaries did not disappear; they simply shifted roles. Educational institutions, district welfare offices, and departmental back offices still sit between eligibility and payment. Where workflows remain manual, approvals remain discretionary. Where datasets are poor, verification becomes cosmetic. Where accountability is unclear, responsibility evaporates.
In such a setting, “direct” becomes a misnomer. The transfer may be electronic, but the system remains vulnerable at precisely the stages that decide who qualifies, who is verified, and who is paid.
The Real Leak: Verification, Not Transfer
India has become impressively good at moving money. The hard part is determining—at scale, repeatedly, and with integrity—who should receive it.
Scholarships and pensions are especially prone to failure because eligibility is multi-dimensional and time-sensitive: income thresholds, category certificates, attendance requirements, course continuity, institutional accreditation, and sometimes disability status or other conditions. If these inputs are weak, the output will be wrong—no matter how “direct” the final click looks on a dashboard.
This is why DBT’s weakest points are not always at the bank’s end. They are at the State’s end, where verification must be ruthless, continuous, and automated wherever possible.
Aadhaar Seeding: Compulsory, But Only After Certification
It is tempting to treat Aadhaar seeding as a magic bullet: seed every account, de-duplicate beneficiaries, and the leakages will vanish. The Odisha evidence shows why this instinct exists—non-seeded ecosystems are porous. Yet a more dangerous error is to assume that seeding, by itself, guarantees identity.
Aadhaar seeding should indeed be compulsory for DBT schemes—but only after Aadhaar has been properly verified and certified to belong to the beneficiary. Otherwise, we risk replacing one kind of fraud with another: wrong Aadhaar linked to the right bank account, or right Aadhaar linked to the wrong person through sloppy enrolment, coercion, or misrepresentation. A system that mandates seeding without first guaranteeing the authenticity of the identity link merely hardens error and makes it harder to unwind. The State’s duty is not merely to “seed”; it is to certify that the seed is genuine.
In practical terms, this means designing a clear, time-bound “identity assurance” step before a beneficiary enters the payment pipeline: verified Aadhaar ownership, verified bank-account ownership, and verified scheme eligibility—each with recorded checks, exception-handling, and an appeal mechanism for genuine beneficiaries who face mismatches or biometric failures.
The Myth of the Perfect Digital Trail
A common defence of DBT is that digitisation creates an audit trail. In theory, yes. In practice, many DBT systems generate data without generating accountability.
Consider what happens when bank accounts are inactive, incorrectly linked, or not properly mapped; when payment status does not flow back to the scheme portal; when failed transactions are treated as expenditure; when “serving accounts” exist on paper but cannot be reconciled in reality. At that point, the digital trail becomes a fog—lots of records, little truth.
Even worse, digitisation can compress timelines in a way that benefits fraudsters. A ghost institute can register students rapidly. Multiple applications can be filed at machine speed. A scheme that pays on bulk approvals can be gamed long before a human auditor arrives—unless the architecture itself is designed to trigger alerts, stop anomalous patterns, and force resolution.
The “Manual DBT” Paradox
If anyone thinks these are abstract risks, they should look at how “DBT” is sometimes implemented on the ground. In some audited cases of social security pension disbursement, beneficiary lists were generated as spreadsheets and sent to banks through email or even via pen drives—an arrangement that defeats the very logic of directness and leaves weak audit trails. When a welfare system runs on spreadsheets, the citizen’s dignity is hostage to the weakest clerk, the loosest process, and the least accountable handover.
This is not an argument against technology. It is an argument against pretending that technology is governance.
The Silo Problem: One Government, Many Databases
The most dangerous DBT failures are no longer petty. They are systemic. They arise because the State does not behave like a single entity. Ministries, departments, divisions, and even joint secretaries often operate as independent islands, each guarding its beneficiary list, each defining eligibility in its own way, and each building portals that do not talk to one another.
When databases do not converge, de-duplication becomes impossible across schemes. The same person can draw overlapping benefits. The same ineligible household can pass through multiple filters because the filters are not connected. The same fraud pattern can repeat, undetected, in scheme after scheme. Technology cannot cure institutional fragmentation; it only makes fragmentation more expensive.
Why Audits Must Become a Design Feature
Traditionally, audits were post-mortems: they told you what went wrong after the money was gone. Today, with data analytics and faster scrutiny, audits can function like an immune system—detecting anomalies early, identifying clusters and patterns, and forcing corrective action while damage is still containable.
But that requires a philosophical shift. Audit trails must be built into scheme software by design, not bolted on as an afterthought. Independent experts—auditors, data scientists, domain specialists, and process engineers—must be part of the architecture, not merely the clean-up crew.
Most importantly, the people who actually run the schemes must be at the design table. Too many systems are built by technical teams who know platforms but not ground realities: how institutions verify, where certificates are forged, what incentives distort behaviour, why approvals get delayed, and how citizens experience “hassle” as humiliation.
What Reform Should Actually Look Like
India does not need to abandon DBT. It needs to stop romanticising it.
First, verification must be treated as the heart of DBT, not a peripheral checklist. Aadhaar seeding should be mandatory—but preceded by verified and certified identity linkage to the beneficiary, along with verified account ownership and a recorded eligibility check.
Second, real-time feedback loops must be non-negotiable. If a payment fails, the system must know quickly, the beneficiary must be informed, and resolution must be tracked—not buried.
Third, institutional onboarding must be made as rigorous as beneficiary onboarding. If an institute is not real, not accredited, or not eligible, it should never receive credentials that allow it to inject data into the pipeline.
Fourth, database convergence must become a governance priority. A unified, deduplicated beneficiary registry—properly governed, privacy-respecting, and accessible across departments—would do more to stop leakage than a thousand dashboards.
Finally, silo-breaking must be institutionalised. Coordination cannot depend on goodwill. If divisions within the same ministry do not share a common database and common controls, the State is practically inviting duplication, fraud, and political misuse.
The Citizen’s Test—and the Treasury’s Risk
A Warning to Heed, Not a Report to Dismiss
CAG’s report—more accurately, CAG’s warning—should be read in the right earnest: not as a nuisance document to be managed, but as a trigger for nationwide systemic reform. It would be easy, and politically convenient, to dismiss it as yet another “auditor’s alarm”—the way some once tried to brush aside the then CAG Vinod Rai’s audit findings by branding them a smokescreen of astronomical losses in the 2G spectrum valuation and coal block allocation cases. That would be a grave mistake, because the Odisha findings are not about a dramatic headline number; they are about the quiet, repeatable mechanics of leakage in welfare delivery.
The ultimate test of DBT is not how much money was “transferred”. It is whether the right beneficiary received the right amount at the right time, with minimum friction and maximum dignity.
Odisha’s scholarship audit is a warning flare, not an isolated scandal. It shows how digitisation can offer a sleek façade while the plumbing underneath leaks—and how, sans safeguards, DBT can slide into a “devious bleeding of the treasury”, with public funds siphoned off through ghost institutions, duplicate claims, and unverified identity linkages.
If DBT is to remain a credible reform rather than a political slogan, the State must treat welfare delivery as critical financial infrastructure: designed jointly by domain administrators and technologists, built with audit trails as a default feature, governed through shared databases rather than departmental silos, and subjected to continuous scrutiny by independent experts. Only then will “direct” become an assurance in practice, not merely an aspiration in rhetoric.