A Market Failure Framework for Evaluating Public Sector Undertakings

by Arjun Krishnan.

Commentators and investors often judge companies, including state-owned firms, by their profitability. While profit is a useful metric for private firms focused on generating returns for owners, applying the same standard to state-owned enterprises is problematic. Many lament that India’s Public Sector Undertakings (PSUs) incur losses, assuming that losses signal failure and profits signal success. However, this assumption misjudges the real purpose of PSUs. Although some evaluation frameworks expand beyond profit, few explicitly align performance criteria with the specific market failure that the enterprise was established to address. This article argues that assessing PSUs solely on profitability is misguided, and their success should be measured by how well they address the public purpose for which they were created.

This article proposes a two-part framework for evaluating PSUs. The first part poses an ex-ante question about purpose. A PSU is justified only when it aims to correct a market failure that less intrusive instruments cannot correct. The second part poses an ex-post question about performance. Evaluators should then judge a justified PSU on two dimensions: efficiency and effectiveness. Efficiency measures how productively an enterprise converts resources into outputs. Effectiveness captures whether the PSU actually corrects the failure it was created to address. Balance sheets cannot serve as a reliable proxy for either dimension on their own.

Ex-ante: when is a PSU justified?

The justification for any PSU must begin with market failure. When markets function well, they allocate resources efficiently, and the state has no grounds to intervene. Economists identify four situations where markets fail to do so. First, externalities arise when a cost or benefit of an economic activity falls on an unrelated third party. Positive externalities lead to underprovision, and negative externalities to overproduction. Second, public goods are non-excludable and non-rivalrous. Firms cannot easily charge users, and private markets typically underprovide them. Third, information asymmetry occurs when one party to a transaction knows more than the other, distorting decisions and reducing market efficiency. Fourth, market power arises when limited competition allows firms to raise prices or restrict output below socially optimal levels.

A market failure on its own does not justify a PSU. There are three additional tests. First, scale: how many people are affected, and by how much? A localised information asymmetry in a niche market is different from one that excludes millions from credit. Second, persistence: is the failure temporary and self-correcting, or structurally durable? Markets sometimes endogenously mitigate their own failures through competition, reputation, or contracting. Exogenous forces such as technological innovation or institutional adaptation can have the same effect. Non-state mechanisms such as industry associations, cooperatives, or third-party certifiers may emerge to address coordination problems or information asymmetries without government ownership. Even classical public-good cases have been addressed without state ownership. Coase’s (1974) account of English lighthouses is a canonical illustration: what was treated in classical economics as a pure public good requiring state provision was, in fact, supplied for centuries by Trinity House, a private body that collected dues from ships at port. The presence of such adaptive mechanisms weakens the case for a PSU. By contrast, failures that persist despite opportunities for institutional adaptation present a stronger case for public intervention.

Even when a market failure is large-scale and persistent, the state has many ways to respond. The third test, then, is instrument choice: Is ownership the right way to address this failure? The state can regulate, tax, subsidise, or contract with private providers. Ownership is one of the most costly options. Ownership exposes the exchequer to operating losses, creates a vehicle vulnerable to political capture, and softens the budget constraint in ways regulation and subsidy do not. The case for ownership has weakened with experience. Publicly owned natural monopolies in many sectors turned out to deliver less output for a given level of inputs than the textbook treatment had suggested. Regulatory practice has grown more sophisticated, with sector-specific knowledge and administrative law tools that did not exist when many PSUs were created. Public-private partnerships have further narrowed the cases for state ownership, with private firms providing goods and services under contracts that set market structure, pricing, and quality. For every PSU, the central question is why the problem could not be addressed through one of these alternatives.

India operates 291 Central Public Sector Enterprises across sectors as varied as petroleum refining, power transmission, hotel management, and defence manufacturing. The policy debate about this universe has been conducted in terms of profitability. How many are loss-making? What do aggregate losses cost the exchequer? These questions are downstream of a prior one: which market failure, if any, justifies each enterprise. Those that address no market failure have no business existing and should be sold off to buyers or wound down, with assets liquidated. Those that aim to correct a market failure face a harder question: how well do they perform in addressing the failure they are expected to correct?

Ex-post: efficiency and effectiveness

Two questions emerge for judging how well a PSU is performing. The first is efficiency. Efficiency is the ability to derive the greatest possible output from a given quantity of financial, physical, and human resources. The second is effectiveness. Since a PSU is justified only for market failures, we need to evaluate whether it, in fact, corrects the failure it was created to address. A PSU can be efficient at producing what a competitive market would produce anyway, or effective at reaching its target population at three times the cost a regulated private operator would charge. Both outcomes are undesirable. In the first case, the PSU adds little social value. In the second, it imposes unnecessary costs to achieve a legitimate public objective.

Experience with direct public provision over the last half-century has weakened the case for PSUs on efficiency grounds. Publicly owned natural monopolies in many sectors exhibited poor x-efficiency, producing less output for a unit of input than private firms. The Ministry of Road Transport and Highways reports that the revenue-to-cost ratio for the 58 reporting undertakings fell to 63.6% in 2021-22, that state cabinets blocked fare revisions, and that the resulting losses reached Rs 30,192 crore in aggregate. The mobility problem the SRTUs were created to address is real, but the case for state ownership of the operator is much weaker than the case for state involvement in the sector through options like subsidies.

The regulatory and contracting alternatives to ownership have grown more capable over the same decades. The regulatory state now possesses sector-specific tariff and quality regulation, administrative law procedures for rule-making, and incentive-compatible contracting techniques (Laffont and Tirole, 1993). Public-Private Partnership (PPP) models have spread across sectors once considered the natural home of direct provision. Iossa and Martimort (2015) show that bundling construction and operation into a PPP can be efficient when build quality materially lowers operating costs. This structure is common in roads, water systems, and many public utilities. Even classical public goods, including urban streets and water supply, are now routinely constructed and operated through PPP agreements that specify quality and pricing. The Government of India’s disinvestment policy lists market imperfections and public purpose as criteria for retaining a PSU in public hands. The government excludes profitability as a criterion.

For certain types of goods, state ownership may be the preferable option. When contract terms cannot specify aspects such as quality, the case for ownership over contracting strengthens (Hart, Shleifer, and Vishny, 1997). A private operator paid to deliver an output will cut costs along whatever margins the contract leaves unspecified. Where quality is one of those margins, the cost saving comes at the consumer’s expense. A private prison contractor’s contract may specify calorie counts and dietary variety, but regulating food quality or taste may prove impossible. These savings flow to the contractor while the welfare loss falls on inmates. Direct public ownership is preferable in such settings precisely because the public manager’s weaker incentive to cut costs leaves the unspecified quality dimensions intact.

In addition to efficiency, effectiveness needs to be judged. Effectiveness measures whether the PSU is correcting the market failure it was created to address. To illustrate the difference, consider a state-owned bus operator tasked with providing transport connectivity to remote rural areas. An efficient operator minimises the resources needed to run the service. An effective operator ensures that the targeted rural communities are actually connected. A PSU may succeed on one dimension while failing on the other.

The effectiveness criteria take different forms across failure types because the welfare yardstick differs. The market power row needs some additional explanation. A monopolist with declining average costs cannot price at marginal cost without losses. Ramsey-Boiteux pricing sets the loss-minimising alternative: markups above marginal cost should rise as demand elasticity falls, placing the heaviest charges on users whose consumption is least price-sensitive. A markup on inelastic demand reduces output the least and destroys the least surplus per rupee of revenue. A political cross-subsidy structure follows a different logic, allocating markups across user groups by political weight rather than by elasticity. The effectiveness test for a PSU that disciplines market power, therefore, asks whether its markup structure approximates Ramsey-Boiteux rather than political cross-subsidy.

Table 1 sets out the effectiveness criterion for each market failure, with the less intrusive instrument serving as the comparator.

Table 1: Effectiveness criteria for PSUs by type of market failure

Market failure	Market problem	Role of PSU	Effectiveness criterion	Less intrusive instrument
Externalities (positive)	Producers cannot capture the full social benefit, so the private market undersupplies relative to the social optimum.	Produce at a level that accounts for spillovers private producers ignore, or finance investments whose social returns exceed appropriable private returns.	Is the targeted output being delivered, and is the additional supply above the private optimum sufficient to close the externality gap?	Production subsidies, tax credits, intellectual property protection, advance market commitments.
Externalities (negative)	Private producers impose costs on third parties they do not bear, so the market overproduces relative to the social optimum.	Produce at a level that internalises external costs private producers would otherwise externalise.	Has the targeted reduction in harm been achieved, and are emissions per unit of output below the unregulated counterfactual?	Pigouvian tax, tradable permits, command regulation.
Public goods	Private producers cannot exclude users from a non-rivalrous good, so the market undersupplies or fails to supply.	Provide the good where private cost recovery is impossible or inefficient.	Is the service reaching the target population, and is coverage approaching the welfare-maximising level?	Contracting with private providers under a public service obligation.
Information asymmetry (seller knows more)	Private sellers hold information about quality that buyers cannot observe, so low-quality goods crowd out high-quality ones.	Enter the market and disclose costs, quality, and pricing as a benchmark that private sellers would otherwise suppress.	Has the PSU’s presence made quality observable to buyers and sustained transactions that would otherwise have unravelled?	Mandatory disclosure regulation, third-party certification, independent benchmarking authority.
Information asymmetry (buyer knows more)	Buyers hold private information about themselves that sellers cannot verify. Sellers respond by raising prices, rationing, or withdrawing supply.	Offer service to groups private firms avoid because they cannot distinguish high-risk from low-risk customers.	Is the PSU enrolling the high-risk groups private markets exclude, and is the share of high-risk individuals covered higher than under the private counterfactual?	Risk-pooling mandates, mandatory insurance schemes.
Market power	Private firms price above competitive levels or restrict output below the social optimum.	Compete to discipline private pricing. In a natural-monopoly case, price at the welfare-optimal level.	Is the PSU pricing closer to marginal cost than an unregulated monopolist would, and does the markup structure approximate Ramsey-Boiteux rather than political cross-subsidy?	Competition law and antitrust enforcement, sectoral price regulation, separation of monopoly network from competitive services.

A PSU can fail on either dimension or both, and each case calls for a different response. A state-owned bus operator that abandons remote routes for crowded urban corridors is efficiently delivering something the market can deliver. Efficient delivery of a service that the market would have provided is no justification for state ownership. A PSU that effectively corrects a market failure but does so at an unjustified cost may generate more welfare loss through waste than welfare gain from correcting the failure. Reform or contracting out may be the appropriate response.

Soft budget constraints

PSUs operate under what Kornai (1998) called a soft budget constraint. A private firm that runs at a persistent loss is likely to go bankrupt. A PSU does not, because the PSU expects the state to cover the shortfall. The expectation of rescue weakens the discipline that revenues and costs would otherwise impose on managers.

A PSU pursuing a legitimate mandate should not show sustained accounting losses on its own books. Where pricing reflects deliberate policy, including selling output below cost for welfare reasons, the resulting deficit is a transfer from the treasury to the consumer. The right place for that transfer is the government’s expenditure account, recorded as an explicit subsidy and matched by income on the PSU’s accounts. Accounting separation keeps the cost of social policy visible in the budget, where parliamentarians can scrutinise it, rather than in an enterprise’s operating accounts.

The Food Corporation of India illustrates what happens when this discipline breaks down. The Corporation procures grain at minimum support prices set by the Cabinet, stores it, and supplies it to ration shops at prices well below procurement cost. The shortfall is intended to be transferred to FCI as a food subsidy from the union budget. However, for long stretches, the government did not transfer the full subsidy in time, and FCI raised debt, much of it from the National Small Savings Fund, to bridge the gap. This practice was especially prevalent between 2016 and 2021. The losses that appeared on FCI’s books reflected the failure to transfer the subsidy on time. The government used FCI’s balance sheet to delay recognition of expenditure that should have appeared in the budget.

The existence of PSUs can then soften budget constraints in two ways. First, if a PSU knows it can rely on transfers from the treasury to cover any shortfall, its managers have less incentive to contain costs than managers of a private firm would. As a result, the PSU’s budget constraint is softened. Second, PSUs like FCI soften the budget constraint of the government that created them. Since the enterprise absorbs costs that should have appeared in the budget, the state’s social spending is understated. Once this practice exists, the headline profit or loss of a PSU carries less information. A loss-making PSU may be one that delivered the mandate but did not receive the subsidy. A loss-making PSU may also be wasteful. Which case applies cannot be understood based on the profit and loss account. Persistent losses on a PSU’s books are therefore a useful diagnostic. They suggest either operational inefficiency or off-budget accounting through the PSU’s balance sheet.

The framework, so far, treats PSUs as faithfully pursuing their mandates. They often do not. Even an enterprise with a legitimate market failure justification is run by people with their own interests. Governments are themselves made up of self-interested actors, and political objectives can capture an enterprise created to address a market failure. These government failures show up in the performance criteria above: prices far from welfare-optimal levels, investments that do not deliver the promised social benefits, or operations that consume more inputs than the next-best instrument would have required.

Government failure also shapes how the state manages exit from enterprises that have outlived their justification. A framework that identifies when a PSU has no reason to exist is useful only if exit decisions follow honestly from that assessment. Chakrabarty (2023, page 9) finds that around 43% of India’s disinvestment proceeds between 1991 and 2022 involved no actual transfer to private hands. Shares were transferred between public entities, and the proceeds were counted towards the disinvestment target without any change in underlying ownership. Even where the case for a PSU has lapsed, political incentives corrupt the exit process and sustain enterprises that should not exist.

Applying the framework

Consider three applications of the framework. First, the electricity transmission network exhibits natural-monopoly characteristics, with high fixed costs and declining average costs over the relevant output range. State intervention to address market power is justified ex-ante, whether through regulated private ownership or direct public ownership of the grid operator. Power Grid Corporation of India is a PSU that runs the inter-state grid. From the standpoint of allocative efficiency, tariffs should be set close to marginal cost. Because the marginal cost of transmission is below the long-run average cost, such pricing would generate a structural revenue deficit. A deficit of this kind would not necessarily signal operational inefficiency. It can reflect a deliberate tariff policy that expands access and maximises network use, and it would be justified where it represents the least-cost route to the social objective when compared with direct transfers or alternative subsidy mechanisms. India does not price transmission at marginal cost. Tariffs are set under the CERC Tariff Regulations, 2024, which use a cost-plus framework. For new transmission projects, Regulation 30(3) provides a base return on equity of 15%, along with recovery of interest costs, depreciation, interest on working capital, and operating and maintenance expenses. The PSU’s operational record is strong: the transmission system was available 99.85% of the time in FY24 across 1,77,699 circuit kilometres carrying around half of India’s inter-state electricity. Whether the cross-subsidy implicit in cost-plus regulation is the least-cost route to network expansion, or whether a direct transfer would deliver the same access at lower fiscal cost, is a question worth asking. Power Grid no longer builds new lines by default. The regulator auctions each new project to the lowest bidder, and Power Grid competes for these contracts alongside Adani, Sterlite and Tata Power. Private bidders win a growing share. Whether the legacy network would also be cheaper in private hands is a separate question. The framework’s verdict on Power Grid therefore turns on an empirical question: whether the cost-plus regulation of the legacy network delivers cheaper transmission than competitive procurement would.

Second, the India Tourism Development Corporation (ITDC) runs the Ashok Group of Hotels. ITDC was established in 1966 to develop tourist infrastructure, including hotels. The Taj, Oberoi, ITC, Lemon Tree and Marriott chains, among others, operate across the segments and price points ITDC serves. Hotels are not a public good. Private operators can charge customers, competition is adequate, and there are no externalities, information failure, or market power problems that require the state to own a hotel chain. The market failure test fails at the first step, so no ex-post analysis is needed.

A third example concerns the post-independence wave of public investment in heavy industry and infrastructure. The standard defence of state ownership in this period rested on the absence of capital markets: long-term finance was scarce, and only the state could mobilise it. Bhagwati and Desai (1970) contested this claim, arguing that private capital existed and that the licensing regime was producing the shortages it purported to remedy. Whatever the merits of the original argument, India’s capital markets have since deepened. A second defence rests on positive externalities through learning effects and supply-chain spillovers, where social returns exceed what a private investor can appropriate. Underinvestment results from this appropriability gap, even when capital is available. The defence still has limits. Where production subsidies, intellectual property protection, or advance market commitments can close the gap, ownership is the costlier instrument. A surviving PSU founded on these grounds must show that such alternatives remain inadequate.

India’s early integrated steel plants at Rourkela, Bhilai, Durgapur, and Bokaro are concrete cases. The plants were justified on grounds that private firms could not raise the long-term capital required, and that they would generate large downstream spillovers through skilled labour, supplier networks, and engineering capabilities whose full social value private investors could not capture. The appropriability gap was real, and thin capital markets compounded the problem by raising the cost of private investment. Both conditions have since changed. India’s capital markets have deepened, project finance has matured, and Tata Steel and JSW Steel have built modern integrated capacity at scale. The spillovers that public investment was meant to generate now flow through the private steel industry instead. SAIL, the operator of the original plants, produces around 15% of Indian steel and is profitable. Profitability does not save it from the framework’s test. Where private operators produce the same steels at a comparable scale, the market failure has been resolved, and continued public ownership no longer has a justification.

The way forward

Profit is the wrong measure for judging a PSU. It speaks neither to whether the enterprise should exist nor to whether it is doing what it exists to do. For each of India’s 291 central PSUs and more than a thousand at the state level, the question is whether a market failure persists, whether ownership is the cheapest way to address it, and whether the enterprise actually does so. Some profitable PSUs would fail this test. Some loss-making ones would pass.

References

India: Planning for Industrialization: Industrialization and Trade Policies Since 1951 by Bhagwati J and Desai P, 1970, Oxford University Press for OECD Development Centre.

On the Management of Public Monopolies Subject to Budgetary Constraints by Boiteux M, 1971, Journal of Economic Theory, 3(3), 219 to 240.

The Lighthouse in Economics by Coase R H, 1974, Journal of Law and Economics, 17(2), 357 to 376.

The Proper Scope of Government: Theory and an Application to Prisons by Hart O, Shleifer A and Vishny R W, 1997, Quarterly Journal of Economics, 112(4), 1127 to 1161.

The Simple Microeconomics of Public-Private Partnerships by Iossa E and Martimort D, 2015, Journal of Public Economic Theory, 17(1), 4 to 48.

In Service of the Republic: The Art and Science of Economic Policy by Kelkar V and Shah A, 2019, Penguin Allen Lane.

The Place of the Soft Budget Constraint Syndrome in Economic Theory by Kornai J, 1998, Journal of Comparative Economics, 26(1), 11 to 17.

A Theory of Incentives in Procurement and Regulation by Laffont J-J and Tirole J, 1993, MIT Press.

Allocative Efficiency vs. “X-Efficiency” by Leibenstein H, 1966, American Economic Review, 56(3), 392 to 415.

State versus Private Ownership by Shleifer A, 1998, Journal of Economic Perspectives, 12(4), 133 to 150.

Arjun Krishnan is a consultant at the Centre for Civil Society, a Delhi-based think tank. He thanks Sourya Banerjee for the early conversations that inspired this article, Jayana Bedi for her thoughtful feedback during its drafting, and an anonymous referee whose comments considerably sharpened the argument.

Source: https://blog.theleapjournal.org/2026/05/a-market-failure-framework-for.html