Why India's state accounts don't add up, distorting deficits and debt

Misclassification is rife, affecting compliance under FRBM Act, and distorting revenue surplus, fiscal deficit and debt ratios. IMF doesn't like it. Now, CAG is set to clamp down on the bad practic

In 2023-24, Odisha proudly reported capital expenditure of ₹43,273 crore — a number that, on the surface, signalled a state investing aggressively in infrastructure.  

But once the Comptroller and Auditor General of India (CAG) opened the books, the neat picture unravelled. Nearly ₹4,565 crore of what the state had declared as capital spending turned out to be revenue in nature: operation and maintenance costs, routine repairs, and, most significantly, funds transferred to non-government entities for community assets — from wedding halls to temple halls and sports equipment for youth groups. 

After audit correction, Odisha’s capex fell to ₹38,708.75 crore and its capex-to-GSDP (Gross State Domestic Product) ratio dropped from 5.07 per cent to 4.54 per cent. The revenue surplus too had been overstated by the same amount. 

For seasoned observers of state finances, the Odisha example is not an outlier. It is symptomatic of a deeper structural problem: India does not have a coherent, comparable standardised system of state accounting. What should allow clean inter-state comparisons — capex, revenue deficits, development spending — has instead become a patchwork of inconsistent classifications, outdated definitions and accounting choices that vary not by economics but by geography. 

Across 10 states in FY24, auditors identified ₹10,579.19 crore of expenditure that had been misclassified. Some states shifted routine spending into capital accounts, others pushed grants into capex even when the rules explicitly barred it, and several buried operational expenditure under object heads that concealed its true nature. The fiscal picture these states present to markets, rating agencies, policymakers and citizens is not the one that exists on the ground. 

Hidden in plain sight 

The most persistent irregularity concerns Indian Government Accounting Standard-2 (IGAS-2), guidelines designed to remove ambiguity in how governments classify grants. IGAS-2 is unequivocal: grants-in-aid from one level of government to another level or institution must be booked as revenue expenditure in the giver’s accounts, regardless of how the receiving entity uses the funds. Yet Maharashtra, Chhattisgarh, Karnataka, Madhya Pradesh and Rajasthan violated this standard, together misclassifying ₹5,310.97 crore as capital expenditure. 

Maharashtra alone reported ₹3,544.73 crore of grants to institutions as capital, although they qualified unmistakably as revenue. Chhattisgarh followed with ₹1,446.59 crore, most of it State Finance Commission grants to local bodies. Karnataka, Madhya Pradesh and Rajasthan contributed another ₹323.65 crore to this category.  

The fiscal distortion that results is not marginal. 

Maharashtra’s actual revenue deficit was nearly 40 per cent higher than the reported figure; Chhattisgarh’s understatement was 68.8 per cent. The CAG has noted that such misclassification affects compliance under the Fiscal Responsibility and Budget Management (FRBM) Act, distorting revenue surplus, fiscal deficit and debt ratios.  

However, the FRBM Act imposes no direct penalties for violations, and while the CAG highlights non-compliance, it can do little more, creating soft enforcement. Neither does IGAS-2 carry statutory penalties for non-compliance, functioning as an accounting guideline rather than enforceable rule. 

A parallel problem runs through the way states use object heads — the primary units of appropriation in government accounts. With no uniformity across states, identical expenditure items show up under different heads and, in many cases, as entirely different types of spending. For instance, Uttar Pradesh in FY24 misclassified ₹283.39 crore of expenditure across object heads such as vehicle maintenance, computer maintenance, general repairs and outsourcing services, treating all of them as capital instead of revenue.  

Karnataka added ₹374.64 crore of minor works to its capital book, including ₹234.88 crore for afforestation and ₹134.80 crore for protection walls — activities that fall squarely in the revenue category. Bihar recorded ₹22.53 crore of repairs and renovations as capital as well. 

Problems with definitions 

The underlying problem is conceptual. India’s accounting system remains cash-based rather than accrual-based, which means expenditure is recognised when money leaves the treasury, not when the economic activity occurs. A deeper issue lies in how India defines capital expenditure itself. As Anoop Singh — former member of the Fifteenth Finance Commission and author of Managing Public Finances in a New Global Era — points out, India’s capex definition includes transactions that have little to do with building infrastructure. Loans to public sector undertakings to repay old loans are counted as capital expenditure.  Acquisition of financial assets is booked as capital formation. “India defines capital expenditure in an unusual way,” Singh says. “It includes financial asset acquisition — even transfers to PSUs if it’s a loan for the PSU to pay off its own loan. They classify that as capital expenditure.” 

Misclassification and definitional ambiguity feed into a larger governance failure. Without comparable numbers, India cannot produce consolidated general government (centre+states) fiscal accounts. The International Monetary Fund (IMF) noted explicitly in its November 2025 Article IV consultation that India has not published consolidated central and state fiscal accounts since FY16. Nor does the country compile consolidated general government data including local bodies and extra-budgetary funds. The IMF also highlighted another worrying pattern: between FY16 and FY23, both revenue and expenditure were overestimated by about 10-11 per cent in budget estimates, indicating a structural bias towards optimism in state finances. 

Consequences 

This lack of reliable data has consequences far beyond the audit community. Trend analysis becomes unreliable; and comparing capex across states says little when some include repairs and grants while others do not. In addition, credit rating agencies must operate with partial information, never fully clear whether deficits are understated or liabilities hidden off-budget. Policymakers evaluating expenditure efficiency or the developmental impact of fiscal support face the same opacity. India, which has argued in international forums that “data for development” is a global public good, finds its own fiscal data unable to meet the standard. 

Singh argues that these failures cannot be explained away as capacity gaps. “India is a highly developed nation intellectually,” he says. “I don’t think India lacks the knowledge of what is capital. I don’t think it’s because of lack of training or capacity.” The real constraint, he suggests, is definitional inconsistency combined with weak compliance. Audit reports that flag irregularities are supposed to trigger Action Taken Reports in state legislatures. In practice, these ATRs are often mechanical, sometimes only a few lines. The constitutional accountability loop between executive, legislature and auditor is barely functional in several states. 

 

New mandate 

It is against this backdrop that the CAG moved decisively on November 11, invoking constitutional authority under Article 150 to prescribe the form of accounts for both the Union and state governments. The Office Memorandum issued that day lays down a harmonised object-head structure with explicit definitions and uniform treatment of expenditure types across governments. Under the new structure, Object Head 27 will cover minor civil and electrical works — clearly classified as revenue — while Object Head 51 will cover procurement of motor vehicles as capital. Vehicle maintenance, which Uttar Pradesh had treated as capital, will unambiguously fall under Object Head 29, a revenue head. States have been advised to adopt the revised system by 2027-28, allowing a three-year transition period. 

Whether this brings uniformity will depend not on the directive itself but on states’ willingness to apply the rules rigorously. Singh outlines a broader reform agenda that would go well beyond harmonised object heads, including adopting international standards such as the IMF’s Government Finance Statistics Manual (GFSM), publishing multi-year capital project projections and implementing general government reporting that eliminates distortions caused by inter-government transfers.  

Countries like the UK differentiate sharply between financial and non-financial assets in their accounting frameworks. India does not. Most countries present five-year capital project details with financing plans and monitoring mechanisms. India does not. These gaps make it easier for states to hide liabilities or reclassify expenditure. 

The challenge is compounded by India’s weak transparency metrics. The Open Budget Survey 2023 gives India a transparency score of 51 out of 100, and a public participation score of just 6 out of 100 — behind Nepal, Pakistan and Bangladesh. For Singh, this is not merely a compliance issue. 

Transparency, he argues, is essential to efficient public spending, accountability and public trust. 

The CAG’s November directive is a significant attempt to restore coherence. But misclassification has persisted not because rules were unclear but because the incentives to stretch definitions remain strong. FRBM targets look easier to meet when revenue deficits are understated. Community assets funded through transfers can be packaged as capital creation even when the accounting rules say otherwise. 

The next three years will test whether India can finally repair the foundations of its fiscal reporting. If states implement uniform object heads, enforce IGAS-2 and strengthen legislative oversight, India may eventually produce fiscal data that can be compared across governments and trusted by those who rely on it. If not, the country will continue to operate with an accounting landscape where numbers look large, but their meaning remains uncertain — a system where fiscal realism is lost not in policy but in the very form of its accounts.