When Markets Fail — The Moments a Price Cannot Tell the Whole Story

What Is Missing on a Fried Chicken Receipt

Suppose you order a whole fried chicken at ten o'clock at night. The price is 20,000 won, and once both you and the shop are satisfied, the transaction is over. The price speaks only of the business between these two people.

Surrounding this scene, there are four things that do not appear in the price.

The noise of a delivery motorcycle passing through an alleyway is heard by the person living upstairs. The upstairs neighbor neither received nor paid a price in this transaction.

The streetlamp in that alley helps the delivery worker find the way, even though no one paid for it.

If there is only one chicken shop in the neighborhood, there is nowhere to compare whether 20,000 won is a fair price.

The brake condition of the used bicycle bought last weekend was known only to the seller.

This article asks about these four things. It follows what happens when something is left out of the price, and looks through the literature at what it takes to fix it. This explains general concepts only and does not recommend any actions.

Prices Only Speak for the Value of Those Who Traded

Market failure is the name given to cases where prices do not capture all values, so the outcome of a trade falls short of what society could achieve. Few papers reviewing development experience point out that price tags do not always carry enough information for economic decisions (DOI: 10.1257/jep.4.3.25).

Let us recall the points made in previous pieces one by one. Opportunity cost is the value of the best alternative given up, and it does not appear on a receipt, as covered in "Scarcity and Opportunity Cost: What Do You Give Up When You Choose?". The idea that people change their behavior based on prices and rewards is in "Incentives: How Do People Respond to Prices and Rewards?". The equilibrium price is the price where the quantity buyers want to buy equals the quantity sellers want to sell, which "Supply and Demand: Who Sets the Price?" covered.

The place where these three stories connect is right here. People who trade decide their actions by looking at the values loaded into prices. Values that are not loaded do not enter into those actions.

The scope pointed to by this name is narrow. Trades where prices capture values well are the trades covered in previous pieces. Market failure only points to places where there are values that are not captured.

Values that are not captured appear in four shapes: costs paid by people outside the trade, benefits for which no one receives payment, the power to raise prices, and information known to only one side.

The Price Paid by People Outside a Transaction: Externalities

An externality is a cost or benefit that spills over to a third party outside the transaction, and is not included in the price. When harm is done, it is called a negative externality; when a benefit is conferred, it is called a positive externality.

The noise of a chicken delivery motorcycle is a negative externality. The upstairs neighbor loses sleep, and that loss is not in the price of the chicken. The value of the sleep that the upstairs neighbor gave up is also the opportunity cost we saw earlier. Only, the person paying that cost is outside the transaction.

The social cost is the price paid by the people in the transaction plus the price paid by people outside the transaction. When there is a negative externality, the cost felt by the people in the transaction is smaller than the social cost. For the side that does not pay that difference, the reason to reduce noise is weak. A paper dealing with environmental damage notes that this external cost is shouldered by society, not by the producer. It argues that without measures to pull that cost inside, profit-seeking producers will not reduce it on their own (DOI: 10.47968/gapgyan.31007).

The paths of correction mentioned in the literature are several, and many paths list trade-offs separately.

A Pigovian tax is a method of putting the social cost into the side's costs by levying a tax on the side causing the externality, unit by unit, in an amount equal to the damage (DOI: 10.2139/ssrn.6341260). It is like making the shop pay as much as the noise from each motorcycle ride inflicts in damage. This idea came from Pigou's 1920 book, and the price put on carbon is cited as an example (DOI: 10.1080/10370196.2020.1827759).

The trade-off of a Pigovian tax arises from matching the amount. There is a mathematical model showing that if other taxes like wage taxes or commodity taxes already exist, the right size changes. It also changes depending on whether the damage is constant or changing, and whether household preferences are the same or different (DOI: 10.31357/vjm.v8ii.5603).

When pairs that can negotiate and pairs that cannot are mixed, and the regulating side cannot separate the two, the optimal tax rate becomes lower than the Pigovian benchmark. There is a theoretical model that shows this (DOI: 10.2139/ssrn.6992218 · a working paper stage model).

There is also a model where, when the regulating side does not know well the entanglement between individual demand and externalities, controlling the trading volume based on the worst-case scenario is optimal. It is an upper limit for negative externalities and a lower limit for positive externalities. If the direction of the entanglement is known, a uniform tax or subsidy can be optimal (arXiv:2603.15832 · pre-peer-reviewed manuscript, model).

Coasean bargaining is a method where the side causing the harm and the side receiving it bargain directly to sort it out. It is like the upstairs neighbor and the shop reaching an agreement on the timing and noise of night deliveries. Conditions are often attached here: rights must be clear and bargaining costs must be low. How well those conditions hold up in reality is being debated, but a survey paper evaluates it as a practical solution for problems that are difficult to solve by other methods even if the scope is narrow (arXiv:2004.04247). Meanwhile, there is a chapter arguing that because civil liability for damages already acts as regulation, the scope of government regulation justified by the market failure argument becomes much narrower (DOI: 10.1093/9780197784631.003.0003 · author's argument).

There are also tools that lean on rights. An emission trading system is a system that sets a total allowable emission cap and lets people buy and sell those rights. An introductory paper outlines the shift toward such rights-based tools, placing common and collective ownership, government taxes and regulations, and more sophisticated private property rights side by side. This introductory paper is written from the author's perspective that both resource and environmental problems stem from incompletely defined and enforced property rights (DOI: 10.1111/j.1467-8489.2007.00425.x).

Using Things Without Paying: Public Goods

A public good is something that is difficult to stop people who haven't paid from using, and where one person's use does not reduce the share left for others. Streetlights in an alleyway are like this. It is hard to sort out only the people who paid among those using the light.

Free-riding is getting the benefit without paying. If you are asked to split the cost of the streetlight, you think, "The light will turn on even if I don't pay." If you get the same benefit without paying, there is no reward coming back to the side that pays. The discussion of Incentives from the previous part applies right here. There is a model showing that when contributions to a shared benefit are left to voluntariness, they tend to fall short (arXiv:1910.11337 · preprint, model).

Public goods connect to positive externalities in that the benefit spreads to anyone outside the transaction, and there is no way to charge a price for it.

What does not get captured in prices is a benefit. Because there is no way to charge for the benefit of a streetlight, sellers have a hard time stepping up. For a good whose effects spread to many, when one person takes action, they pay all the cost and get back only a part of the share. Therefore, there is a theory that things held together are under-supplied and things suffered together are under-deterred (DOI: 10.2139/ssrn.7485638 · working paper stage theory).

In the laboratory, there are experiments that shrink this situation down. Each person puts money into a joint account or keeps it, and the money in the joint account is multiplied and divided equally among everyone. One experimental paper notes that this structure creates the incentive for free-riding, which is the background for contributions decreasing over time. In the same paper, conversation and continuous history showed positive effects on contributions (DOI: 10.3390/s19122786).

The ways to fix this are also several.

One way is for the government to provide it, funded by taxes. A paper discussing service supply writes within that discussion that pure public goods must be supplied by the government even if the government has weaknesses, and views most serious policies as lying between pure public goods and pure private goods (DOI: 10.35536/lje.2013.v18.isp.a1 · author's assertion).

One theoretical paper concludes that voluntary methods are prone to suffering the exact failure they are trying to fix, so coercion is necessary. Both coercion and voting are needed, but when matched with market principles, several contradictions arise, and the paper calculates those costs (DOI: 10.2139/ssrn.7485638 · working paper stage theory).

There is a way to punish people who do not contribute. Punishment (disadvantages given by participants to other participants) increased contributions. However, the sum of the participants' shares increased or decreased depending on the conditions (arXiv:2508.17151 · large-scale experiment, preprint). The figures will be looked at later.

There is a way for communities to make rules on their own. Common-pool resources—where one person's use reduces the share left for others, like a fishery or pasture—deviate from the second condition of the definition above. Still, because they share the point that it is difficult to block people who haven't paid, the same literature covers them together. In case studies of forests, irrigation, pastures, and fisheries, communities managed resources together through self-governing rules. The authors conclude that for new shared forms in urban and digital environments, that framework must be revised, but some findings remain valid (DOI: 10.1515/humaff-2019-0007). There is also a model showing that when there is no coercion, a structure where multiple voluntary consultative bodies emerge maintains cooperation better than a single comprehensive system (arXiv:1910.11337 · model).

The Power to Raise Prices: Monopoly

A monopoly is a state where a single seller completely occupies the market, so that even if they raise prices, customers cannot easily walk away. This is like having only one fried chicken shop in the neighborhood. Even if prices go up, customers have nowhere else to go.

It is said that competing markets produce lower prices, better quality, diverse choices, and more innovation. It is also said that sellers may try to weaken that competition (DOI: 10.1093/actrade/9780198860303.001.0001 · the primer).

What fails to be captured in the price under a monopoly is a lost transaction. Suppose the shop raises its price so that some customers stop buying. If the cost of the chicken was lower than the price those customers were willing to pay, the gain from that transaction goes to neither the shop nor the customers. Deadweight loss is the share of potential gains from trade that ends up going to no one and simply vanishes. It occurs because resources are allocated inefficiently, representing a loss that neither the consumer nor the supplier captures (DOI: 10.1002/9781118785317.weom080247).

Let us look at the paths for correction.

There is the path of protecting competition. Collusion (an agreement among competitors to fix prices), anti-competitive agreements, monopolies, and mergers are cited as challenges that block competition. Enforcement involves separate policy considerations. The primer notes that the relationship between free markets and government intervention is delicate. The abstract of this primer does not separately list the costs of individual measures (DOI: 10.1093/actrade/9780198860303.001.0001).

There is the path of regulating prices or profits. A model estimating regulation in a public utility estimates underinvestment and calculates the value of regulators keeping their promises. Depending on the political environment, it also calculates the trade-off where reducing one type of loss causes another type of loss to grow—arising because the regulated party stops striving to cut costs (DOI: 10.1086/695474 · model estimation).

There are also experiments showing that trading rules can temper a monopolist's power. In a classroom experiment where participants experience various market structures in sequence, an auction-based trading system prevented monopolists from exercising their power to raise prices. The authors conclude that reference prices formed in earlier markets weakened the power of the monopolist in later markets (DOI: 10.1111/jpet.12504 · single study).

When One Side Knows More: Asymmetric Information

Asymmetric information is a state where only one of the two parties in a trade knows the true condition of the item or person. If only the seller knows whether the brakes on a used bicycle work, that is this state.

Adverse selection is when only the lower-quality side remains in a transaction because of a trait the other party does not know. Buyers cannot tell a good bicycle from a bad one, so they try to offer only an average price. Owners of good bicycles will not sell at that price and leave. A market where good items drop out and only low-quality items (a 'lemon') remain is called a lemon market (DOI: 10.1002/9781118785317.weom080245). This name comes from the title of a classic paper dealing with quality uncertainty and market operation (DOI: 10.1093/oso/9780199253906.003.0002).

What the price fails to capture is the quality of each individual item. The price only speaks for the average. In insurance, a similar problem arises when a policyholder knows their own risk better than the insurer does. It is established that the coverage received by high-risk and low-risk individuals becomes misaligned, making resource allocation inefficient under any contracting scheme (DOI: 10.1002/9780470015902.a0026687).

There is also literature showing the limits of this theory. There are conditions under which low-quality items enter but do not drive good items out of the market, and instead are sold alongside them (DOI: 10.1057/s41599-020-00658-w). There are models where sellers can offer multiple contracts and buyers have other choices. In these models, even a small mixture of very bad types can trigger a large-scale cascade of exits. As information about types becomes more precise, the market becomes more vulnerable to such cascades (DOI: 10.17016/feds.2022.013 · working paper model).

Let us look at the paths to correction.

Signaling is when the side that knows the quality lets the other party know in a credible way (DOI: 10.23943/princeton/9780691202754.003.0017). This is like the owner of a bicycle showing a recent inspection record. Means cited in the literature, alongside signaling, include information disclosure, third-party certification, risk adjustment, and compensation design (DOI: 10.70267/icfmb.202619553560).

The limits are also noted. One paper summarizes that signaling and methods by which the uninformed side digs for information reduce the impact, but neither completely solves the outcome, and each method has flaws (DOI: 10.54254/2754-1169/2025.21130).

A deductible is the share of a loss that the policyholder pays first. In insurance, one paper concludes that mandatory enrollment, subsidies, deductibles, monitoring, and regulatory reform are essential for market stability, fairness, and sustainability (DOI: 10.54536/ajase.v5i1.6099). On the other side, there is a book arguing that adverse selection in insurance markets is weaker than commonly said, and that a certain degree of adverse selection is actually necessary to expand the scope of loss compensation (DOI: 10.1017/9781316178843 · author's claim).

It Costs Money to Fix Things Too: Government Failure

Government failure is when an intervention meant to fix a market failure misses its mark or creates new costs. If we want to tax the noise from upstairs, someone has to measure the damage. Measuring costs money, and a mismeasurement makes the tax too high or too low. This is the exact difficulty of matching the amount that we saw in the Pigovian tax section just before.

Two papers on government failure both ask what we are comparing against what. One paper argues for a comparative institutional approach that weighs real market outcomes against real government intervention outcomes. It notes that collective choice brings problems of information and incentives, and that an intervention can create a worse problem than before (DOI: 10.1177/092137409500700313 · author's argument).

Another paper evaluates policies in light of the constraints and tools that governments actually have. It sees the capacity to enforce and monitor policies as a constraint, and argues that we should prioritize policies that endure under varied conditions over those sensitive to ideal enforcement. It also notes that all policies create interest groups that bind later decisions (DOI: 10.35536/lje.2013.v18.isp.a1 · author's argument, service delivery discussion).

The concept of market failure has long served as a rationale that governments should step in whenever markets fail to allocate efficiently. Some papers question whether that rationale is sufficient. Drawing on a case of national science and technology support, one paper argues that the concept provides neither an adequate basis nor a proper guide, and has instead been used to justify politically predetermined decisions (DOI: 10.1080/08109028508628976 · author's argument).

Taxes cost money too. When we levy a sales tax on goods without externalities, it drives a wedge between the price buyers are willing to pay and the price sellers are willing to accept. Some transactions that would have happened simply disappear (DOI: 10.3386/w6852). A tax meant to fix an externality differs in that the transactions it aims to curb are harmful ones. Even so, if the tax amount misses the actual damage or piles on top of other taxes, the exact same kind of deadweight loss can occur. One point-in-time data-based estimate suggests that taxing a newly spreading technology slows its entry into the market, and when you add that loss, the efficiency cost becomes much larger than standard models suggest (DOI: 10.2202/1538-0645.1505).

The flip side of the coin appears in this same literature. The service delivery paper we saw earlier notes that neither the harm that private market failures inflict on society nor the difficulty of intervening is frequently measured or discussed. A 1973 paper estimating monopoly losses speculates that even larger allocation losses may stem from government activities such as import quotas, water diversion projects, and common-property conservation standards (DOI: 10.2307/1056875 · author's speculation). Which side bears the heavier cost cannot be settled by these papers alone.

Checking Against Reality

Here we collect a few papers that check the stories of the previous sections with numbers. The targets and conditions measured differ for each number.

First, there are policies that put a price on carbon. One study analyzed the earliest large-scale emissions trading system at the level of regulated manufacturing plants. It estimates that this system reduced emissions from those manufacturing plants by an average of 14 to 16% and that no contraction in economic activity was detected (the time period is not in the abstract). The authors found no evidence that emissions moved to unregulated sides and concluded that it led to a global reduction in emissions. They explain this with a model where, when a price is attached, companies spend fixed costs on equipment to save energy (DOI: 10.1093/restud/rdae055).

The conclusion of a review collecting ex-post evaluations is different. The oldest emissions trading system mentioned in the review is the same system that the previous study looked at. The 37 evaluation papers collected in the review are mostly concentrated on one continent. Most studies see the total reduction as limited and generally at 0 to 2% per year. The review summarizes that deviations by sector are large, and in terms of emission reduction, a carbon tax (a tax levied per unit of emission) generally reported better results than an emissions trading system. Studies looking at the same system reported 0 to 1.5% per year, and the review concludes that the impact of carbon prices is limited (DOI: 10.1088/1748-9326/abdae9). This article judges that it is difficult to directly match the two numbers because the previous study reads as talking about regulated manufacturing plants, while the review talks about the policy as a whole. The two papers did not explain that gap.

Second, there are public account experiments. In an experiment collecting 147,618 decisions made by 7,100 people across 360 conditions, punishment consistently increased contributions. When punishment was present, the sum of the participants' shares spread from cases where it increased by up to 43% compared to when there was no punishment to cases where it decreased by up to 44%. Whether or not communication was possible best predicted the result (arXiv:2508.17151 · preprint).

Another sample is university economics lectures (different groups of students each semester). 67 to 76% contributed to the public account in the semester before the pandemic, 71% right after the first lockdown, and 43% seven months later. Whether this drop was statistically distinct depended on how biases in the experimental design were handled (arXiv:2011.09189 · preprint, single sample). Even in experiments with incentives to free ride, a considerable number of people chose to contribute, and it is confirmed that this proportion moved depending on the situation.

Third, there is the deadweight loss of monopoly. Estimates vary widely. An old pioneering estimate was at the level of 0.1% of gross national product, while another indicator produced about 10% of gross corporate product. The denominators also differ. Two things are cited as the main causes of the difference. One is how to handle the degree to which prices deviate from the cost of making one more unit at the economy-wide level. The other is how to handle the effect that the power to raise prices has on corporate costs. Some point out that because analytical frameworks differ by study, it is difficult to separate them by cause (DOI: 10.2307/2098390).

A study estimated using data from a single country in the 1950s and 1960s expected before the research that this loss would be quite large, but obtained a much lower value. The abstract does not disclose what standard the size is based on. The author concluded that there is almost no basis for the common belief that there are large losses. However, they added that measurement could be the problem (DOI: 10.2307/1056875). A 1972 paper points out the loss (managerial slack) where a monopoly company spends more than necessary on costs because there is no pressure from competition. It argues that if this is included, the loss can be much larger than typical calculations (DOI: 10.1007/978-1-349-15486-9_2 · author's argument).

Fourth, there is adverse selection in insurance. A review that sorted out whether the prediction that people who buy more insurance are riskier is correct summarizes that the correlation differs depending on the insurance market and the subscriber group. Because the same correlation can be explained by behavioral changes where people become less careful after signing up, a separate method is needed to separate the two (DOI: 10.1111/j.1539-6975.2009.01337.x). In one working paper looking at auto insurance, high coverage and more accidents appeared together only among new subscribers with three or more years of driving experience. The author reads this as a result consistent with the side where asymmetric information exists (DOI: 10.2139/ssrn.289966 · working paper).

Where Else to Look

There were four places where prices do not capture the full value. Most of the correction paths for each place have their own separate costs or limits written down, and the literature shows together the places where the correction worked and where it did not. It does not cover every branch, and each branch has its own separate, lengthy literature.

The situation where people inside the company know more than those outside leads to a misalignment between those who manage and those who own, as discussed in "What You Actually Own When You Own a Share — And Where You Stand in Line". "Financial Statements — What Each of the Three Ledgers Is Asking" covers one branch of the information that a company puts out to the outside world, which is the ledger.

Source Papers

• "Market Failure and Government Failure" (1990) — DOI: 10.1257/jep.4.3.25 — A review of 40 years of development experience: market prices do not always carry enough information for economic decisions, and institutions matter.
• "Market Failure and Government Support for Science and Technology: Economic Theory versus Political Practice" (1985) — DOI: 10.1080/08109028508628976 — The concept of market failure has often been used as grounds for government intervention, but the author argues, using a country's case, that this concept provides neither sufficient grounds nor proper guidance, and has been used to justify politically predetermined decisions.
• "Working of Pollution-Tax and Tradable-Permits to Control Environmental Pollution" (2020) — DOI: 10.47968/gapgyan.31007 — Pollution is an external cost borne by society rather than the producer, and without measures to internalize it, producers will not reduce it.
• "Pigouvian Taxation" (2026) — DOI: 10.2139/ssrn.6341260 — Definition: A tax that charges an amount equal to the marginal damage for each unit of pollution, incorporating public costs into private costs.
• "One Hundred Years Ago. The Book That Inspired the Carbon Price: Pigou's The Economics of Welfare" (2020) — DOI: 10.1080/10370196.2020.1827759 — Pigou's 1920 book introduced externalities and the Pigouvian tax to correct them, with carbon pricing being one example.
• "A Comparative Mathematical Study of the Relationship Between Marginal Social Cost and Pigouvian Tax in the Presence of Commodity and Wage Taxes" (2020) — DOI: 10.31357/vjm.v8ii.5603 — Mathematical model: When other taxes exist, the magnitude of the Pigouvian tax increases depending on whether damage is treated as a variable and whether preferences are heterogeneous.
• "Unifying Pigouvian Taxation and Coasean Bargaining in Externality Regulation" (2026) — DOI: 10.2139/ssrn.6992218 — Theoretical model of a working paper: If regulators cannot distinguish between pairs that can and cannot bargain, the optimal tax rate is lower than the Pigouvian benchmark.
• "Prices vs. Quantities: Robust Regulation" (2026) — arXiv:2603.15832 — Model in a preprint: Under a worst-case criterion, the optimal policy is quantity control (an upper limit for negative externalities), and knowing the sign of the correlation can make a uniform tax-subsidy optimal.
• "Applications of the Coase Theorem" (2020) — arXiv:2004.04247 — Survey: The practical applicability of the Coase theorem is debated; its scope is limited, but it is pragmatic for problems that are difficult to solve by other means.
• "Market Failure, Regulation, and Invisible Gorillas" (2024) — DOI: 10.1093/9780197784631.003.0003 — A book chapter: The author's argument that incorporating the regulatory effects of civil liability makes the scope of government regulation justified by market failure arguments much smaller.
• "The tragedy of the commons: property rights and markets as solutions to resource and environmental problems" (2008) — DOI: 10.1111/j.1467-8489.2007.00425.x — Feature introduction paper: Compares common and collective property rights, government taxes and regulations, and private property rights, explaining the shift toward rights-based instruments like emission permits.
• "Coalition-structured governance improves cooperation to provide public goods" (2019) — arXiv:1910.11337 — Model in a preprint: Shared gains tend to fall short when left to voluntary contributions, and multiple voluntary coalitions sustain cooperation better than a single comprehensive regime.
• "The Design Space of Market Provision for Public Goods" (2026) — DOI: 10.2139/ssrn.7485638 — Theory in a working paper: Goods with diffusing effects lead actors bearing the full cost while receiving only a fraction back, resulting in under-provision and under-deterrence.
• "Predicting Group Contribution Behaviour in a Public Goods Game from Face-to-Face Communication" (2019) — DOI: 10.3390/s19122786 — The structure of public goods games creates free-riding incentives that reduce contributions, while face-to-face communication reduces free riding.
• "Balancing Market and Government Failure in Service Delivery" (2013) — DOI: 10.35536/lje.2013.v18.isp.a1 — The author's argument that pure public goods must be supplied by the government, but most policies lie in between; advocating for policies that endure, and pointing out that the losses of market failure and the difficulties of implementation are rarely measured and discussed.
• "Integrative Experiments Identify How Punishment Impacts Welfare in Public Goods Games" (2025) — arXiv:2508.17151 — Preprint: Across 7,100 people and 360 conditions, punishment increased contributions, but performance ranged from a maximum 43% improvement to a maximum 44% deterioration, with the presence of communication being the best predictor.
• "Contemporary commons: Sharing and managing common-pool resources in the 21st century" (2019) — DOI: 10.1515/humaff-2019-0007 — Examples of community self-governance rules (forests, irrigation, pastures, fisheries) and the conclusion that frameworks must be adapted for new urban and digital sharing forms, though some remain valid.
• "Competition and Antitrust Law: A Very Short Introduction" (2021) — DOI: 10.1093/actrade/9780198860303.001.0001 — Primer: Competition generates low prices, quality, choice, and innovation, with collusion, monopoly, and mergers being challenges, and the relationship between the free market and government intervention being delicate.
• "Deadweight Loss" (2015) — DOI: 10.1002/9781118785317.weom080247 — Encyclopedia entry: The net loss of total surplus that is captured by neither consumers nor suppliers.
• "Dynamic Natural Monopoly Regulation: Time Inconsistency, Moral Hazard, and Political Environments" (2017) — DOI: 10.1086/695474 — Dynamic model estimation: Estimates underinvestment, calculates the value of regulator commitments, and computes the trade-offs where reducing one loss increases another depending on the political environment.
• "Controlling monopoly power in a double-auction market experiment" (2021) — DOI: 10.1111/jpet.12504 — Classroom experiment: Auction trading institutions prevented monopolists from exercising power, and benchmark prices in upstream markets weakened downstream monopoly power.
• "Lemons Market" (2015) — DOI: 10.1002/9781118785317.weom080245 — Encyclopedia entry: When sellers know more about quality than buyers, buyers lower their price expectations, causing good items to leave the market and leaving only low-quality items behind.
• "The Market for 'Lemons': Quality Uncertainty and the Market Mechanism" (2005) — DOI: 10.1093/oso/9780199253906.003.0002 — Republished classic paper: Addresses the interaction between quality differences and uncertainty.
• "Adverse Selection and Insurance" (2017) — DOI: 10.1002/9780470015902.a0026687 — Encyclopedia entry: Adverse selection occurs when insurance companies do not know risks while individuals do, leading to inefficient resource allocation.
• "On the market for 'Lemons': quality provision in markets with asymmetric information" (2020) — DOI: 10.1057/s41599-020-00658-w — Empirically meaningful conditions under which low-quality goods do not drive out high-quality goods but are sold together.
• "Moldy Lemons and Market Shutdowns" (2022) — DOI: 10.17016/feds.2022.013 — Discussion paper model: More accurate information makes markets more vulnerable to cascades of exits.
• "Adverse selection and market signaling" (2020) — DOI: 10.23943/princeton/9780691202754.003.0017 — Textbook chapter: The solution to adverse selection is signaling, where the informed party takes action to reveal information.
• "A Literature Review on Theories and Empirical Evidence Under Asymmetric Information" (2026) — DOI: 10.70267/icfmb.202619553560 — Review: Identifies signals, information disclosure, third-party certification, risk adjustment, and compensation design as means to mitigate asymmetric information.
• "Asymmetric Information: Insights Across Economic and Market Contexts" (2025) — DOI: 10.54254/2754-1169/2025.21130 — Signaling and screening reduce effects, but neither fully resolves the outcome.
• "Adverse Selection and Moral Hazards: Information Asymmetry as a Source of Market Failures in Insurance and Finance" (2026) — DOI: 10.54536/ajase.v5i1.6099 — Authors' conclusion: Mandatory participation, subsidies, deductibles, monitoring, and regulatory reforms are essential for market stability.
• "Loss Coverage" (2017) — DOI: 10.1017/9781316178843 — Book: The author's argument that adverse selection in insurance markets is weaker than commonly stated and is necessary to some extent to expand the scope of loss coverage.
• "Market Failure and Government Failure" (1995) — DOI: 10.1177/092137409500700313 — Advocates for institutional comparison weighing the actual outcomes of real markets versus real government interventions, suggesting that interventions can create worse problems than before.
• "New Estimates of the Welfare Loss to Monopoly, United States: 1956-1969" (1973) — DOI: 10.2307/1056875 — Monopoly loss estimates were much lower than expected prior to the study (the benchmark is not in the abstract), with the author adding the possibility of measurement issues and speculating that larger allocative losses may exist in government activity.
• "Three Sides of Harberger Triangles" (1998) — DOI: 10.3386/w6852 — Explains the loss (deadweight loss) where sales taxes wedge between willingness to buy and willingness to sell, causing some transactions to disappear.
• "The Value of Broadband and the Deadweight Loss of Taxing New Technology" (2006) — DOI: 10.2202/1538-0645.1505 — Estimation: Incorporating fixed costs of entry makes the efficiency cost of taxes on new technology more than twice that of standard models.
• "Does Pricing Carbon Mitigate Climate Change? Firm-Level Evidence from the European Union Emissions Trading System" (2024) — DOI: 10.1093/restud/rdae055 — Emissions from regulated manufacturers fell by an average of 14 to 16%, and no contraction in economic activity was detected.
• "Does carbon pricing reduce emissions? A review of ex-post analyses" (2021) — DOI: 10.1088/1748-9326/abdae9 — 37 ex-post evaluations: The majority report that aggregate reductions are limited to 0 to 2% per year, with large variations across sectors and carbon taxes generally performing better.
• "Cooperation in the Age of COVID-19: Evidence from Public Goods Games" (2020) — arXiv:2011.09189 — Preprint: Contributions to a shared account fell from 67 to 76% down to 43% seven months later, and statistical significance varied depending on bias treatment.
• "Welfare Losses in Oligopoly and Monopolistic Competition" (1985) — DOI: 10.2307/2098390 — Monopoly loss estimates range from around 0.1% of gross national product to about 10% of corporate gross output, with differences arising from the analytical framework and the treatment of cost impacts.
• "Allocative Efficiency, X-Efficiency and the Measurement of Welfare Losses" (1972) — DOI: 10.1007/978-1-349-15486-9_2 — The author's argument that incorporating the reduction in cost efficiency caused by monopolies can make total losses much larger than standard calculations.
• "Testing for Adverse Selection in Insurance Markets" (2009) — DOI: 10.1111/j.1539-6975.2009.01337.x — Review: Coverage-risk correlations vary across insurance markets and pools and can also be explained by moral hazard, necessitating methods to separate the two.
• "Asymmetric Information and Learning: Evidence from the Automobile Insurance Market" (2003) — DOI: 10.2139/ssrn.289966 — Working paper: High coverage and more accidents appeared together only among new policyholders with three or fewer years of driving experience.