Public Financial Management System (2009): Strengthening Transparency and Efficiency in Public Expenditure

Policy Update
Divya Natarajan

Background

The Public Financial Management System (PFMS) is a web-based platform developed and operated by the Controller General of Accounts (CGA), Department of Expenditure, Ministry of Finance, to track and manage the flow of government funds in real time. It began in 2009 as the Central Plan Schemes Monitoring System (CPSMS), a Central Sector Scheme under the erstwhile Planning Commission, at a time when the government had no reliable way of knowing how much of the money released for its Plan schemes had actually reached implementing agencies or been spent, since expenditure was typically reported only after the fact, through conventional accounting cycles that could lag actual spending by months. CPSMS was conceived to close that gap, with a narrow original mandate: tracking funds released under Plan schemes and reporting expenditure in real time, rather than months later.                            

The system’s scope expanded considerably after the launch of Direct Benefit Transfer (DBT) on January 1, 2013, which required a reliable digital backbone to move welfare payments, subsidies, and wages directly into beneficiary bank accounts. From April 1, 2015, use of PFMS was made mandatory for the payment, accounting, and reporting of DBT across scholarships, pensions, wage payments, and subsidies. In 2017, the government abolished the distinction between Plan and non-Plan expenditure, further widening PFMS’s role to cover essentially all central government spending routed through the scheme mechanism.

Unlike a scheme with a fixed sunset date, PFMS is infrastructure. Its relevance is not tied to a single welfare programme but to the government’s broader push, since 2013, to make public expenditure traceable at every stage, from the Consolidated Fund of India down to the individual beneficiary’s bank account.

Functioning

PFMS operates as an integrated payment-cum-accounting network, connecting the Reserve Bank of India, scheduled commercial banks, state treasuries, and implementing agencies on a single platform. Its functions extend well beyond DBT, though DBT remains its most visible use case. Three functions in particular define what the platform actually does: controlling how and when public money is disbursed, accounting for incoming revenue, and channelling welfare payments to beneficiaries.

Payment and exchequer control – Article 283(1) of the Constitution requires Parliament to prescribe rules governing how money is paid out of and accounted for in the Consolidated Fund of India, and this provision underpins the Government of India Receipts and Payments Rules. PFMS has effectively codified these rules into a digital workflow, allowing the entire payment and exchequer control process to run online, without the paper-based approvals it replaced.

Accounting of receipts and fiscal reporting –  PFMS is integrated with the systems of the Central Board of Direct Taxes (CBDT) and the Central Board of Indirect Taxes and Customs (CBIC), pulling in near-real-time, transaction-level data on tax receipts. It also operates Bharatkosh (the Non-Tax Receipt Portal, or NTRP), which collects non-tax revenue directly from payees. Because of this integration, the Union Government’s annual accounts are now derived directly from transactions recorded on PFMS through the year, rather than reconstructed after the fact.

Direct Benefit Transfer – As the DBT channel, PFMS enables every central ministry, state government, and implementing agency to transfer funds electronically to individual or institutional beneficiaries. It interfaces with several scheme-specific beneficiary management systems, including PM-KISAN, the National Social Assistance Programme (NSAP), MGNREGASoft, and AwasSoft, which allows scheme-level beneficiary data to sit alongside PFMS’s own payment and accounting records rather than in disconnected silos.

Over time, the government has also layered several structural reforms onto this base to close specific loopholes in how funds were released and monitored.

Table 1: PFMS Reforms Overview

Source: Controller General of Accounts (CGA); Department of Expenditure, Ministry of Finance.

The most consequential of these reforms is the SNA model, introduced through an Office Memorandum in March 2021. Previously, funds for Centrally Sponsored Schemes moved from the Centre to the state treasury and then onward to numerous implementing agencies, often sitting idle in multiple bank accounts along the way, a practice the Finance Ministry had long flagged as inefficient. Under the SNA model, each state must designate a single nodal agency and a single nodal account for every scheme, and funds are released on a Just-in-Time basis, only when an actual expenditure request is made, rather than in advance. According to government data, this has reduced the number of fund-bearing accounts nationally to around 4,800, a sharp contraction from the earlier fragmented structure.

This is, in principle, sound public financial management. It is also, in practice, one of the more contested reforms in recent Centre-state fiscal relations, a tension examined further below.

Underlying the SNA and CNA models are three deeper architectural shifts that the Department of Expenditure has been progressively rolling into PFMS:

• From prescriptive release to Just-in-Time release. Under the older system, funds were released to implementing agencies according to a pre-set schedule, regardless of whether the agency was actually ready to spend them, leaving money idle with banks in the interim. The Just-in-Time model releases funds only when they are about to be used, reducing float and improving cash management for the exchequer.
  
• From credit push to debit pull. Historically, funds were pushed in advance to implementing agencies as a credit, on the assumption that spending would follow. PFMS is moving this toward a debit pull model, where a debit against the central pool is triggered only when a program implementing agency actually issues a payment instruction on the system. This reverses the direction of initiative: instead of the Centre pushing money outward and hoping it is used well, the system now waits for a verified spending request before funds move at all.
  
• From booking releases as expenditure to booking them as transfers. Under conventional government accounting, a fund release to an implementing agency was recorded as final expenditure the moment it left the Centre’s books, even if the money had not yet reached an actual beneficiary or been spent on an actual activity. PFMS is shifting this so that the initial release is booked only as a transfer, with expenditure recognised only once the implementing agency reports actual utilisation. This is a more accurate, and more demanding, standard of fiscal reporting, since it closes the gap between money leaving the Centre’s account and money being genuinely spent.
  

Taken together, these three shifts represent a move away from an accounting system that measured intent to spend, toward one that measures verified, actual spending. This is a meaningful improvement in fiscal reporting integrity, though it also means state and district-level implementing agencies now operate with less flexibility and a smaller cash buffer than before, a trade-off that resurfaces in the federalism debate discussed under Impact and Emerging Issues.

Performance

PFMS’s core deliverable is DBT, and the volume of money and beneficiaries it now processes is substantial, though the year-on-year figures are less linear than official communications sometimes suggest.

Table 2: Performance of PFMS

Source: Press Information Bureau, Ministry of Finance, Year Ender releases (2022, 2024, 2025).

Note: Blank cells reflect gaps in what official Year Ender releases actually reported for each period, not an absence of the underlying data. The FY2021-22 release focused on the disbursement total and did not break out scheme count or transaction volume, while the FY2024-25 release reported scheme count and amount but not transaction count. This inconsistency in what each year’s release chooses to disclose is itself a small but telling illustration of the reporting gaps discussed later in this article.

Figure 1: DBT Disbursement through PFMS

Source: Press Information Bureau, Ministry of Finance, Year Ender releases (2022, 2024, 2025) (compiled by author)

Two things in this data are worth flagging rather than glossing over. First, the number of schemes covered fell from 1,206 in FY 2024-25 to 966 in FY 2025-26, even as the platform is routinely described in official communication as expanding. This could reflect scheme consolidation, discontinuation, or a change in counting methodology, but the public data does not clarify which, and that ambiguity itself is a transparency gap in a system whose entire purpose is transparency. Second, the amount disbursed in FY2024-25 (₹2.23 lakh crore) is well below the ₹6.3 lakh crore disbursed in FY2021-22, despite PFMS covering more schemes and more banks in the interim. This is plausibly linked to policy changes such as the phased withdrawal of certain subsidies (notably LPG subsidies under PAHAL, historically one of the largest DBT line items by value), rather than a decline in the system’s reach. But without a consistent, itemised public breakdown by scheme, that explanation remains an inference rather than a documented fact.

On efficiency gains, the government’s own assessment, based on a 2024-25 quantitative study covering 2009 to 2024, claims cumulative savings of ₹3.48 lakh crore from reduced leakage, a fall in the subsidy share of total government expenditure from 16 percent to 9 percent, and a sixteen-fold rise in beneficiary coverage (from roughly 11 crore to 176 crore, cumulative and non-unique). These are significant claimed gains, but it is worth noting they originate from a single external assessment commissioned and publicised by the government rather than an independent statutory auditor such as the Comptroller and Auditor General, which is a meaningful distinction when evaluating how much weight to place on the ₹3.48 lakh crore figure specifically.

Impact

PFMS’s clearest, least disputable impact is the elimination of duplicate and ghost beneficiaries from welfare rolls. By requiring Aadhaar-seeded, PFMS-validated bank accounts before a payment is released, the system has closed off a long-standing channel for diversion, where funds were nominally disbursed but never reached a genuine recipient. This is a structural fix rather than a one-time clean-up, since every new payment cycle is subject to the same validation.

The second major impact is on the speed and predictability of fund flow to genuine beneficiaries. Large, high-frequency programmes such as PM-KISAN and MGNREGA wage payments now move at a scale and pace that would be difficult to sustain through manual disbursal, with tens of crores of beneficiaries paid in a single transaction cycle. This has tangible welfare consequences, particularly for wage labourers and smallholder farmers for whom payment delays previously meant real financial strain.

The third impact, less discussed but arguably more consequential for how India’s federal fiscal system functions, is centralisation of control over the timing and mechanics of fund release. The SNA and Just-in-Time model gives the Union government far greater real-time visibility into, and effective control over, how and when states actually spend centrally sponsored scheme money. Proponents frame this as basic fiscal discipline: states should not be permitted to let central transfers sit idle in state-level accounts, since idle public money earns no return for the exchequer and delays the very outcomes the scheme was meant to fund. Critics, drawn largely from opposition-ruled states, argue that this reform has gone beyond fund tracking into de facto control over state treasury operations, reducing states from constitutional partners to what one commentary described as data nodes in a Centre-run system. This is not a marginal objection. It has surfaced explicitly in disputes such as Kerala’s, where state officials have linked SNA conditions to broader fights over scheme funding and educational reform, and it echoes older Centre-state friction over Centrally Sponsored Schemes and Article 282 more generally.

Both readings can be true at once. PFMS has almost certainly reduced leakage and improved traceability. It has also almost certainly shifted real fiscal leverage toward the Centre in ways that were not, and could not have been, resolved by a purely technical accounting reform, because the underlying dispute is constitutional and political, not just administrative.

Emerging Issues

Theatreisation-style challenges in PFMS should be understood as a set of interconnected gaps rather than isolated technical glitches, since each one bears on whether the system’s stated efficiency gains can actually be trusted and sustained.

First, data transparency gaps persist in the government’s own reporting. The drop in schemes covered between FY2024-25 (1,206) and FY2025-26 (966) is not accompanied by any public explanation, despite PFMS being marketed as a transparency instrument. It is unclear from available data whether this reflects scheme consolidation under a smaller number of umbrella schemes, discontinuation of specific programmes, or a change in how “schemes covered” is counted from one year’s report to the next. The same problem applies to disbursement figures: PFMS and the Department of Expenditure publish aggregate annual totals, but not a consistent, itemised, scheme-wise breakdown that would let an outside analyst independently verify whether a year-on-year change reflects policy shifts, seasonal effects, or something else. A system built explicitly to make government spending traceable should, at minimum, make its own reporting traceable in the same way.

Second, the headline efficiency claims rest on a single, government-commissioned savings estimate. The widely cited ₹3.48 lakh crore savings figure, along with the claim that the subsidy share of expenditure fell from 16 percent to 9 percent, comes from one external quantitative assessment commissioned and then publicised by the government. It has not, to date, been separately validated by the Comptroller and Auditor General or another independent statutory body with access to underlying scheme-level data. This does not mean the estimate is wrong. It does mean that a figure repeated in official communication as a settled achievement is, more precisely, a single study’s finding, and that distinction matters for anyone relying on this number for downstream analysis or advocacy.

Third, Centre-state friction over fiscal autonomy remains the most consequential issue facing PFMS today, and the one least likely to be resolved through further technical refinement of the platform. The SNA model and the shift to Just-in-Time, debit-pull fund release have been challenged by several opposition-ruled state governments as an erosion of their control over state treasury functions under Article 266 of the Constitution. Kerala’s dispute over PM SHRI conditionalities, where SNA-linked fund controls became entangled with a broader fight over education policy, is one visible example, but the underlying grievance, that states have gone from holding a working float of central funds to receiving money only against verified, pre-approved expenditure, predates that specific case and recurs across states of different political affiliations. Framing this purely as a technology or compliance issue understates what is, at its core, a dispute about the constitutional division of fiscal authority between the Centre and the states.

Fourth, technical integration across states remains uneven. PFMS’s national architecture assumes a reasonably functional Integrated Financial Management Information System (IFMIS) at the state level to interface with it. In practice, state IFMIS maturity varies considerably, and states with weaker digital treasury infrastructure struggle to meet PFMS’s reporting and reconciliation requirements at the same pace as more digitally advanced states. This creates a two-speed system in practice, even though PFMS is formally uniform in design, meaning the platform’s stated efficiency gains are unevenly realised across the country.
Finally, operational friction persists at the last mile. Even where the architecture works as intended, beneficiaries continue to report problems, including bank account and Aadhaar mismatches, delayed or rejected Fund Transfer Orders, and periods of technical downtime, through PFMS’s own grievance redressal (CRM) system. These are not new or unusual issues for a system of this scale, but their persistence indicates that the platform’s architectural sophistication has outpaced the consistency of its execution at the point where it actually meets an individual beneficiary.

In the end, PFMS’s credibility will depend less on the elegance of its fund-flow architecture and more on whether its reporting is genuinely as transparent as its stated purpose, and whether its efficiency gains hold up under scrutiny from bodies with no institutional stake in the outcome.

Way Forward

Strengthening PFMS in its next phase should be treated as an ongoing exercise in transparency and institutional trust-building rather than a one-time technical upgrade. The goal should be ensuring that the system’s efficiency gains are independently verifiable and that its fund-flow reforms are matched by equal clarity in reporting and equal consultation with the states that operate within it.

First, India needs consistent, itemised, scheme-wise disbursement data published on a regular basis, rather than aggregate figures released around anniversaries. The Department of Expenditure already has this data internally; extending the existing SNA Dashboard model to cover overall DBT performance, not just fund-flow status, would allow year-on-year changes, including shifts in scheme counts and disbursement value, to be understood in context rather than inferred by external analysts working from incomplete summaries.

Second, the government’s efficiency claims should be subjected to independent audit before being treated as settled fact. The widely cited ₹3.48 lakh crore leakage-reduction figure currently rests on a single commissioned quantitative assessment. This is not a challenge to the underlying claim so much as a recognition that self-reported savings, however plausible, carry more weight once verified by a body such as the CAG, which has no institutional stake in the outcome.

Third, Centre-state friction over the SNA and Just-in-Time reforms needs a dedicated resolution channel. General Centrally Sponsored Scheme negotiations are typically scheme-specific and not well suited to addressing disputes over the fund-release architecture itself. A standing consultative forum between the Department of Expenditure and state finance departments, focused specifically on SNA implementation, exemption criteria, and dispute resolution, would give states a defined process to raise concerns before they escalate into high-profile political disputes, as seen in states such as Kerala, without requiring any dilution of the underlying efficiency objective.

Fourth, technical capacity gaps between states need to be closed rather than left to individual state budgets and administrative capacity. PFMS’s national effectiveness is only as strong as its slowest-integrating state, and a targeted capacity-building programme, potentially routed through Finance Commission grants for governance reform, could help states with weaker IFMIS infrastructure reach parity with more digitally advanced ones.

Finally, last-mile execution should be tracked as a measurable performance standard rather than absorbed informally at the state or district level. Converting the CRM grievance mechanism into a management tool, with public, aggregated reporting of resolution timelines by state, scheme, and issue type, would make it possible to see whether PFMS is actually functioning well for the people it is meant to serve, not just whether its architecture is sound on paper.

In the end, how well PFMS succeeds as a public financial management reform will depend less on the sophistication of its fund-flow architecture and more on whether its efficiency claims are independently verifiable and its Centre-state fiscal relationships remain workable over time.

References 

Controller General of Accounts, Ministry of Finance, Government of India. Public Finance Management System (PFMS). https://cga.nic.in/Page/Public-Finance-Management-System-PFMS.aspx 

Department of Expenditure, Ministry of Finance, Government of India. PFMS. https://doe.gov.in/pfms 

Press Information Bureau, Ministry of Finance. (2026, January). Ministry of Finance Year Ender 2025: Department of Expenditure. https://www.pib.gov.in/PressReleasePage.aspx?PRID=2212546 

Press Information Bureau, Ministry of Finance. (2024, December). Ministry of Finance Year Ender 2024: Department of Expenditure. https://www.pib.gov.in/PressReleasePage.aspx?PRID=2088438 

Press Information Bureau, Ministry of Finance. (2025). India’s DBT: Boosting Welfare Efficiency. https://www.pib.gov.in/PressReleasePage.aspx?PRID=2123192

 Press Information Bureau, Ministry of Finance. (2022, August). India transferred on an average >9mn DBT payments per day (in FY 2021-22). https://www.pib.gov.in/PressReleasePage.aspx?PRID=1856116 

National Institute of Public Finance and Policy (NIPFP). (2025). Public Financial Management Reforms in India, Working Paper No. 438. https://www.nipfp.org.in/media/documents/WP_438_2025.pdf 

The Week. (2025, November 1). NEP’s backdoor entry? Cracks emerge in Kerala’s ruling LDF coalition over PM SHRI adoption. https://www.theweek.in/theweek/statescan/2025/11/01/neps-backdoor-entry-cracks-emerge-in-kerala-ruling-ldf-coalition-over-pm-shri-adoption.html

 Drishti IAS. CSS and Fiscal Federalism. https://www.drishtiias.com/daily-updates/daily-news-analysis/css-and-fiscal-federalism

About the Contributor

Divya Natarajan is a Research and Editorial Intern at IMPRI and a recent Economics graduate from Stella Maris College, Chennai. She has experience in public policy research, editorial writing, and policy analysis through internships with think tanks and research organisations. Her interests include public policy, governance, and development economics.

Acknowledgements

I would like to extend my gratitude to IMPRI for this opportunity. I also extend my sincere thanks to Sandra Menon and Prisha Sachdeva for their constructive reviews and editorial support. 

Disclaimer

All views expressed in the article belong solely to the author and not necessarily to the organisation.

Read more at IMPRI:

Viksit Ude Desh ka Aam Naagrik (UDAN) (2026): A Policy Initiative for Regional Air Connectivity

Andhra Pradesh RTGS: Reinventing Governance Through Data