essay

Morality and Market Failures: Asymmetry of Information

Xavier Landes and Pierre-Yves Néron

Introduction

The central idea defended in this article is that significant parts of state reg- ulation could be explained and justified by invoking the concept of information asymmetry (IA hereafter), that is, “situations in which not all parties to a po- tential exchange are equally well informed,”1 and that such a concept should be the target of more sophisticated normative and conceptual scrutiny. The reason for more scrutiny is that IA offers a moral vocabulary as well as a concept to think about appropriate, that is, morally justified, state intervention. Moreover, we argue that a normative account of IAs also needs to include moral concerns (equality, distributive and relational concerns, vulnerability, and so on) that are not directly related to efficiency (which is the primary focus of the use of IAs in economics: to identify and spell out the loss of efficiency, in the insurance indus- try for instance). To put it crudely, we argue that ethics of market regulations is (partly) a normative theory of market failures and, as such, it ought to offer an analysis of the wrongness of IAs, along with addressing other market failures such as externalities or adverse selection.

Of course, the idea that market failures, more than justifying state interven- tion, are the basis for a more complex regulatory architecture, is not new. It is the cornerstone of welfare economics.2 But we argue that there is a need for an explicitly normative account that could offer support for a fruitful ethics of regu- lation. As Rutger Claassen observes, few authors have attempted to bring market failures theorists together with moral and political philosophers.3 Following the development of the market failures approach to business ethics, most notably by Joseph Heath or Wayne Norman, we will attempt to offer a philosophical and normative account of the concept of IA.4

The idea embodied in the market failure approach to regulation is to consider that states are often more efficient than market and private actors for delivering various goods and services. This advantage is partly due to states’ superior orga- nizational and financial capacities, its power (police, law, tribunals, and so forth), its capacity to bind future generations, or to respond to collective action problems by enforcing collective solutions (e.g., by forcing contribution to public goods through taxation).5

An example is social insurance. If industrialized states, in particular in Europe, provide social insurance (i.e., mostly health insurance, unemployment benefits, public pensions) it is not only for redistributive purposes or out of

© 2018 Wiley Periodicals, Inc. DOI: 10.1111/josp.12260 JOURNAL of SOCIAL PHILOSOPHY, Vol. 49 No. 4, Winter 2018, 564–588.

bs_bs_banner

Morality and Market Failures: Asymmetry of Information 565

egalitarian motives (e.g., using unemployment benefits or public pensions for re- ducing socioeconomic inequalities), but also for efficiency reasons.6 In many in- stances, states have more capacities than private organizations for collecting and analyzing all sorts of data (e.g., morbidity, lifestyles, eating habits, exposure to pollutants) of a given population.7 Therefore, they are more able to calculate prob- abilities for various risks. Furthermore, states have access to populations large enough for producing reliable statistics (in virtue of the Law of Large Numbers8). In addition, states have broader resources for facing running costs, exceptional costs (e.g., those induced by a sudden economic downturn or a major environ- mental catastrophe), or deep structural changes (e.g., aging population). States are also better equipped for handling moral hazard: they can impose stringent regulations on policyholders (e.g., increased care requirements). In his lecture “Politics as a Vocation,” Max Weber describes how states have a monopoly on legitimate violence, which implies that they can legitimately force their citizens to adopt specific behaviors (e.g., decrease smoking). They also have the power to closely monitor behaviors through governmental agencies. No other agent can put into action such a combination of power and authority.

In other words, the state is better equipped than markets and private actors for handling market failures. This material superiority does not mean that the state is necessarily more efficient in all situations, it simply means that the state has some initial advantages by comparison with markets and private actors. This also implies that political theorists and normative thinkers ought to pay attention and that investigating such initial advantages should be a topic. Of course, exter- nalities have already started receiving some attention from moral and political theorists.9 Moreover, externalities are in the background of John Stuart Mill’s harm principle.10 Negative externalities (i.e., costs imposed on agents who are not part of the exchanges that generate these costs) are harms to others, and, as such, are subject to regulation from a Millian perspective.11

The discussion of the role of public institutions in handling negative exter- nalities is hardly new. However, market failures are not limited to externalities. They also include monopoly power, missing or incomplete markets, and so on. Among these failures, the question of information is central, as highlighted by research in economics. As Joseph Stiglitz argued, some of the most important developments in economics in the last forty years are related to information.12

And IA is one of the major sources of market failures. IAs generate failures when different parties in a transaction do not possess the same level of informa- tion relevant for the transaction and when such difference results in having one side unduly advantaged in the transaction. (Our task is precisely to clarify in which sense advantages can be unwarranted and, therefore, morally problematic.)

Economists use the terms information asymmetry or asymmetric information to describe situations in which buyers and sellers are not equally well informed about the characteristics of products or services. In these situations, sellers are typically much better informed than buyers, but sometimes the reverse is true.13 As a result, sellers have the opportunity to take advantage of buyers. IAs are potential

566 Xavier Landes and Pierre-Yves Néron

sources of market failures because they impair the law of supply and demand; they lead some buyers to overpay for some goods and services, diminishing the re- sources available for other expenses. As such, buyers do not maximize their utility.

IAs are pervasive in doctor–patient relations (what is referred in the eco- nomic literature as “physician agency”14) because doctors usually know more about diagnosis, diseases, morbidity, treatments, and so forth than their patients. (Nevertheless, this does not mean that there is something morally wrong. More on this point below.) Another example is the market for used cars. The seller usu- ally has important information about the car (e.g., past incidents, usage) that he could conceal from potential buyers for gaining a negotiation leverage (i.e., if he wants to obtain a better price).15

Two important caveats before pursuing. First, there are many instances of informational issues that do not involve asymmetries. An example is when de- fective or missing information impairs the optimization of the utility function by individuals. Bob wants to buy a used car; he bought one that was more expensive than another sold in the neighborhood. He did not know, nor did the seller. In this case, it is not an issue of IA because there is no asymmetry, and, in any case, no side in the transaction was unduly advantaged.

Another useful example is defective treatment of information due to cogni- tive biases and heuristics. Psychology and behavioral economics convincingly demonstrate how individual decisions are deeply affected by biases such as over- estimation and loss aversion.16 Nonetheless, perceptive and cognitive biases in themselves (i.e., as long as no agent takes advantage of them during a transaction) are not IAs because the whole population is assumed to be equally exposed to them, everything else being equal. Therefore, there is no unwarranted advantage.

Two points are worth making about the widespread nature of cognitive bi- ases and heuristics. First, it does not mean that in a given interaction, one agent cannot have access to more information as a result of having addressed her own biases and heuristics. However, such IA is not inherent to biases and heuristics. It emerges because some agents have become aware of these shortcomings and decided to “fix” them. Second, it does not mean that some agents cannot take advantage of widespread biases and heuristics (it is actually one of the pillars of marketing and consumer behavior), in which cases there are IAs. However, in themselves biases and heuristics are not IAs, though they might give rise to IAs.

There are also cases that could be labeled as IAs, but do not raise any prima facie moral problem. For instance, it is not morally objectionable per se that your family physician has more medical knowledge than you do. Furthermore, it is arguably beneficial to your health as long as your physician respects his profes- sional deontology. Of course, your physician can always abuse his position, which is the very reason why guardrails are necessary such as a code of deontology (derived from the Hippocratic Oath) and a self-regulatory body in the form of a medical college. The point is that it is not the IA itself between your physician and you that is morally problematic, but specific instances in which specific phy- sicians use IAs for abusing specific patients.

Morality and Market Failures: Asymmetry of Information 567

Therefore, our focus in this piece is on morally relevant IAs, which does not include situations where IAs do not offer any advantage17 to a party to the trans- action. For instance, Eva perfectly knows all about the Volvo cars. Peter is selling a used Renault. Prima facie, Eva’s knowledge does not give her any advantage be- cause Peter is the only one who possesses the relevant knowledge about this used car. To be morally relevant, asymmetries need to be taking place in the relevant set of information for the considered transaction.

Second, for the sake of this article, we assume that efficiency concerns are cen- tral justifications for markets, and therefore we put aside freedom-based justifica- tions. Since our main concern is IAs, we will therefore work under this assumption and will not try to show why efficiency might trump freedom. Thus, the purpose of our article is threefold. First, it is to offer criteria for identifying morally relevant IAs and moral issues conveyed by IA. Second, it is to determine in which sense these moral issues justify public regulation, and under which form. We propose then a typology of the institutional tools for dealing with IAs, and for each of them we highlight the reasons why public institutions may arguably be better equipped than private actors for addressing morally relevant IAs. Finally, our article is a contribu- tion to debates about public regulation. We propose research avenues for elaborating a robust normative theory of public institutions based on market failures.

1. Moral Issues

The goal of our normative inquiry is to flesh out IA’s ethical dimensions, that is, the moral aspects of the relations between agents who have unequal access to information. Also, because IA is often used for justifying public regulation or arguing that an institutional arrangement is defective, one needs to clarify what, if anything, is wrong with IA and why it calls for correction. In other words, it is essential to lay down the conditions of morally relevant IAs.

The normative issues raised by IA could be sorted into two categories, which partially overlap. The first category is efficiency. The second is equality. Morally relevant IAs (not IAs in general) are informational imbalances that generate ef- ficiency losses (by comparison with at least one feasible alternative) that illegit- imately benefits one agent while burdening another (cross-subsidization). Stated differently, morally relevant IAs are efficiency issues (in the sense that they pre- vent Pareto improvements18), which embody distributive dimensions (in the sense that agents extract illegitimate surplus from exchanges and interactions19 at the expense of other agents).20 Although the efficiency and distributive dimensions are intertwined, they remain conceptually distinct.

1.1. Human Cooperation and Efficiency

First of all, it is worth beginning with a discussion of the idea that mor- ally relevant IAs generate a problem of efficiency. IA may undermine various cooperative mechanisms by perverting their operation, modifying the terms at

568 Xavier Landes and Pierre-Yves Néron

the expense of some parties, or lowering the collective benefits being produced. When it is so, IA alters cooperative efficiency and reduces the social product. Among the mechanisms negatively affected by IA, two deserve special attention: market transactions and risk pooling. These mechanisms are central because they cover a vast range of human activities in industrialized economies.

1.1.1. Market transactions

IAs block Pareto improvements because one of the parties in a transaction extracts an undeserved surplus, which distorts the price system and, therefore, the structure of incentives (i.e., suboptimal activities appear more profitable due to IAs, thus resources get channeled and wasted in suboptimal activities). Of course IAs create Pareto suboptimal situations because some agents become worse off than they would have been with information symmetry. This definitely embodies a distributive issue21: some cooperative gains are redistributed from the not well informed to the better informed without solid justification outside the initial, uneven, distribution of information. If we leave aside the fact that the extracted surplus is undeserved (and therefore the moral legitimacy of the redistribution op- erated by IAs), this redistribution is economically inefficient because it rewards a market imperfection, namely concealing crucial information.

1.1.2. Risk Pooling

Another issue is the alteration of the conditions under which individuals pool their risks when they join an insurance scheme. When individuals decide to pool their risks, they do so on the basis of how they evaluate their expected losses (the product of each possible event probability by each possible event outcome).22 In other words, there is the need for an accurate actuarial calculation of the differ- ent risks subject to pooling, their probability and the potential losses they may incur. IAs jeopardize this calculation because some individuals conceal crucial information about their risk profile. In other words, they conceal some informa- tion about their risk profile that allow them to get better terms for their insurance (e.g., lower premiums or higher coverage) by comparison to a situation where all information about their risk profile would be available to other policyholders.

Adverse selection and moral hazard are common market failures for insur- ance schemes. Adverse selection happens when individuals underestimate their risk exposure at the time they insure themselves for obtaining lower premiums than those they would have paid if other policyholders would have access to a more accurate knowledge of their risk profile (i.e., their expected losses).

Adverse selection deeply affects insurance mechanisms: risks are under- estimated and, thus, premiums do not cover for overall risks. Imagine that ten merchants operate one vessel each for trading overseas. They initially face the same risk (.1) of losing their vessel and cargo. Each cargo is 1,000 euros worth. Merchants have two options. They can face risk alone, which is comparable to a

Morality and Market Failures: Asymmetry of Information 569

lottery where the outcome is either 1,000 (.9) or 0 (.1). Alternatively, they can pool their resources and pay a premium of 100 euros for compensating the unlucky merchant.

In the first case (individual lottery), the outcome is either 1,000 or 0, while in the second case (insurance) it is 900. In both cases, the expected outcome is 900; the only difference is the certainty of the result. Whereas the lottery produces probabilistic outcomes, insurance leaves no uncertainty. In technical language, the standard deviation (from the expected outcome) has been reduced. In the case of the merchants, the standard deviation becomes zero, i.e., the mean outcome is the only possible outcome.

Now imagine that five merchants buy a defective vessel that has twice more probabilities (.2) to be wrecked than a normal vessel. Additionally, imagine that the five merchants know that their vessel is defective, but decide to conceal the information to save on their insurance premiums. Premiums are now insufficient to cover the expected losses that have been shifted from 1,000 to 1,500 euros per turn (it could be that two ships, instead of one, are lost every other turn). As it stands the insurance fails to fulfill its function since the initially agreed compen- sation (900) does not cover any more for the expected losses.

After several turns, merchants would most likely realize that the insurance is defective in the sense that it fails to cover all losses. The risk of a sunk vessel is not .1, but .15 for the whole insured pool. If merchants are not able to precisely evaluate the defective vessels, the expected losses will be spread evenly on every- one. Therefore, each will pay a premium of 150 euros for being covered for 850 euros of loss.

The problem becomes a redistributive one since the 150 euros premiums does not accurately reflect the risk profile of every member of the insured pool. Those who operate normal vessels overpay their insurance, since they pay more than their expected losses (100 euros), while those who operate defective vessels underpay their insurance, contributing less than their expected losses (200 euros). The result is a net transfer of (expected) wealth from the former to the latter.

It should be noted that this only applies to cases where the alteration of prob- abilities has been partly identified (collectively but not individually). If risks are not correctly assessed, IA may lead to the insolvency of the insurance pool due to a lack of resources (i.e., collected premiums) for making up for the losses.

The second information issue—moral hazard—characterizes situations where individuals, once being insured, adopt riskier behaviors in comparison with their before-insurance behaviors.23 The problem is that moral hazard under- mines original risk calculation because of the increase of probabilities of adverse events and/or attached losses posterior to insurance enrollment. In addition to offsetting the economic sustainability of insurance, moral hazard is a form of hid- den cross-subsidization: some policyholders paying for others. And usually, the issue is perceived as carrying a moral dimension (the unfairness of the cross-sub- sidization) since prudent policyholders are forced to pay for part of the losses of reckless policyholders.24

570 Xavier Landes and Pierre-Yves Néron

Morally relevant IAs take place when policyholders are not aware of the proper risks (i.e., probabilities and losses associated with adverse events) except for the bold and when the bold uses his knowledge for taking advantage of the situation, without a collective benefit, which runs against cooperation and ef- ficiency. Other policyholders believe that risks are pooled under fair terms al- though these terms are actually altered outside their knowledge (and without their consent). Furthermore, cooperation is less efficient because, as long as the altered behaviors are not addressed, the bold is likely to transfer part of her expected losses to other policyholders. In addition to being potentially unfair, this imbal- ance distorts the structure of incentives by handicapping less risky activities (be- cause of higher premiums than expected losses) and promoting riskier activities (because of lower premiums than expected losses).

IA raises two cooperative issues. First of all, it raises an issue of (actuarial) fairness since some material responsibility is shifted from agents who benefit from the IA to other agents (while material gains flow the opposite way). Second, it raises an issue of efficiency because it disrupts insurance mechanisms. IA pres- ents a deeper challenge to social cooperation: because they shift part of their expected losses, agents who benefit from IA have no incentive to reduce their risk exposure (or they have a lower incentive than if they will face the full—ex- pected—costs implied by their risk profile). As a result, the community produces more risks than in the absence of IA. This is a case of moral hazard.

In sum, IAs permeate various situations. Our article deals with two cooper- ative settings that are located at the core of market economies: transactions, that is, when two or more individuals exchange goods and services, and risk pooling, that is, when two or more agents gather their resources for facing risks. In these situations, morally relevant IAs threaten the terms of social cooperation by low- ering the social product or undermining risk calculations (and therefore the via- bility of insurance). But, as seen in this section, morally relevant IAs have another characteristic: they illegitimately redistribute resources among agents, which is the focus of our next subsection.

1.2. Distributive Issues

Beyond threatening cooperation and efficiency, IA also raises distributive issues, that is, issues related to the fairness in the repartition of the gains and bur- dens of cooperation and market transactions, as we already noticed with adverse selection for insurance. Since IA characterizes situations where there is an imbal- ance between two or more agents concerning access to crucial information, the equality dimension seems obvious. It is actually possible to use the vocabulary of equality to formulate the moral issues carried by the redistribution of gains and burdens of cooperation and market transactions. Now, there is the need to specify which equality is at stake.

According to a preliminary interpretation, the problem is one of equality of opportunities. The fact that two (or more) agents have different levels of knowledge

Morality and Market Failures: Asymmetry of Information 571

gives an advantage that can be cast in terms of opportunities. Therefore, IA re- allocates opportunities among agents and such reallocation constitutes a moral issue.

Nevertheless, it could be that such reallocation is morally justifiable. It is not because two agents have different information sets (one being more complete than the other) that it is necessarily morally wrong. For instance, Eva has per- haps spent more time investigating market conditions, prices, and so on, while Peter has spent his time playing tennis. As a result, Eva knows more about the product, and proper price, than Peter but it can be hardly considered as an issue of equality, especially of opportunities. Both Eva and Peter had originally the same opportunities, but their different levels of efforts lead to different levels of knowledge that appear to produce subsequent benefits. If there are inequalities of opportunities, they derive from choices for which individuals could be held responsible. (Formulated as such, the issue overlaps with many of the discussions surrounding luck egalitarianism, most notably discussions involving gardeners and tennis players.25)

Other unproblematic cases are when differences of knowledge result from education, division of labor, and specialization, differences that constitute the basis of the capitalist system as described by, among others, Adam Smith.26 In complex society individuals interact on the basis of different degrees of more or less specialized knowledge without that being a moral issue per se (which does not prevent moral issues in particular situations). For instance, the fact that your physician and you do not share the same level of medical information is the result of different educative and life choices. Your physician decided to study medicine, you decided to study something else or nothing. It is also the result of living in complex societies where individuals are put in highly specialized func- tions that require very specialized knowledge. In such cases, the social division of labor does not necessarily convey an objectionable occupational inequality, a problematic distribution of power and prerogatives, but simply a useful occu- pational differentiation, which is partly an epistemic differentiation, based on agents having highly specific and different sets of information, knowledge, and expertise.27

To be sure, in the case of your physician and you, there is an asymmetry that may lead to morally problematic situations (medical ethics is full of such cases). There is a potential for morally relevant IAs. For instance, your physician may take advantage of your poor medical knowledge to enroll you in a test for hazardous drugs or could prescribe expensive or unnecessary medicine sold by a pharmaceutical company in which she has a commercial interest or stocks. The situation is exactly the same with your car mechanic or the person fixing your bi- cycle. They both have a knowledge that you probably do not have. Therefore, they can use this asymmetry for their own benefit, for example by overcharging you. This is what our distinction between IAs and morally relevant IAs try to capture.

A strict distinction needs to be made for identifying more precisely instances of morally bad IAs. On the one hand, there is the IA and, on the other hand,

572 Xavier Landes and Pierre-Yves Néron

there is the advantage that is extracted from the IA. In the case of dishonest or deceptive physicians and mechanics, the fundamental problem is not the IA, but the objectionable advantage that is extracted from the IA. In other words, the challenge is not to put an end to all instances of IA, but to be sure that agents in the advantageous situation in regard to the opportunities open to them do not push their advantage beyond a fair and decent point, that is, transform a morally relevant IA into a morally bad one.

“Fine,” you might think, “but what is ‘a fair and decent point’?” The answer is not straightforward and, to a large extent, it is contingent. Despite this context dependence, few landmarks can be established. First, the IA should not be the result of manipulation or deception, which means that the acquisition of enhanced opportunities should not be tainted by immoral or illegal means. IA should re- sult from the normal process of (self or institutional) education or specialization. Insider trading is an example of an unfair use of IA, leading to a moral issue (personal gains at the expense of other market agents).

Second, the advantage stemming from enhanced opportunities should not be pushed to the point where it is detrimental to the party that has less knowledge. Pushing an advantage beyond a decent and fair point happens when IA generates an interaction with a net loss that would not occur if the other party would have adequately known the outcomes. Individuals should not suffer from IA, which is the case when a physician enrolls a patient for a hazardous experiment while concealing the conditions and implications of such experiment.

Third, the advantage cannot be pushed up to the point where it is used to rip off all gains from interaction or cooperation. IA cannot lead to a capture of all social benefits of cooperation. For instance, an agent cannot exhaust all surplus of risk pooling mechanisms (i.e., the money collected for paying out the losses due to adverse events) for financing her risky activities (once included the social benefits generated by these activities).

The threshold of when an advantage is pushed too far ahead is probably even lower than “rip off all gains from interaction or cooperation.” There are plenty of situations where one agent does not “rip off all gains from interaction or coop- eration,” but nonetheless rip an amount large enough for creating a moral issue. Without expanding this point further, a general theory of cooperation is obviously needed for determining and justifying the legitimate share that various agents can extract from cooperative mechanisms, especially in regards to the distribution of information. But this exceeds the ambition of this article.

In sum, IA (which generates enhanced opportunities) in itself (except if re- sulting from manipulation or deception) is not the problem. In addition, it should be noted that IA often appears in situations where the object of the market trans- action is the knowledge itself, that is, that agents are remunerated for the surplus of information they hold. For instance, if you pay your physician or mechanics, it is partly for getting a diagnosis, that is, an appropriate use of information on a given problem.

Morality and Market Failures: Asymmetry of Information 573

To conclude, if IA should represent a problem of equality of opportunities, it is certainly not in itself, but in relation to the use some agents make of their initial advantage.

Another domain of equality at stake with IA is power. The very fact that Eva knows facts that are ignored by Peter could represent a problem of equality of power. But that being so, Eva should be able to change the nature of the asym- metry, that is, transform information into power. Apart from considering that all information is power, additional qualifications with regard to information are necessary.

The fact that your physician or mechanic has an extra knowledge unknown to you does not inherently create a moral issue. As indicated above, it becomes an issue when the agent on the upper side of the asymmetry extracts undue sur- plus or an extra advantage. At that very moment, it becomes useful to frame the advantage in terms of power.28 This is not to say that all asymmetries of power are morally reprehensible; only some are. The goal of ethical reasoning consists in setting the conditions and criteria for distinguishing morally problematic from unproblematic power asymmetries (as we have been doing so far with the very concept of IA).

Here, various theoretical frameworks can be used for delineating the fron- tiers of acceptable asymmetries of power and spotting unacceptable ones. Neo- republicanism is particularly fruitful in the sense that objectionable power is at the very heart of, for instance, Philip Pettit’s formulation: power is objectionable when it could lead to arbitrary interference.29 To be objectionable, power does not need to be exercised. It only takes an agent being able to arbitrarily interfere in the life of others (i.e., to dominate others). In other words, asymmetry of power is morally problematic because it grounds domination, independently of such dom- ination being exercised (or not).

Therefore, if we reframe our issue using a neo-republican vocabulary, ob- jectionable IA takes place when it creates a situation of domination, that is, when an agent has the capacity to use the additional information to which she has access for arbitrarily interfering in the activities of other agents.30 What is objectionable according to this line of thought is the transformation of IA into un- checked capacity to arbitrarily interfere in someone else’s life. Therefore, neo-re- publicanism offers a concept (domination) to clearly locate the origin of the moral issue with specific instances of IA.

An example is when your physician has some knowledge about your health he hides from you that could be later used to her advantage. Or when your me- chanic does not completely fix your car because it may be more expensive (and profitable for him) to fix later. In both cases, your physician and your mechanic do not take immediate advantage of the IA, but the additional (concealed) knowl- edge gives them the possibility to arbitrarily interfere in your future (which they will).

Briefly stated, it seems that IAs could induce asymmetries of power (and therefore become morally relevant) when no ex ante check exists for preserving

574 Xavier Landes and Pierre-Yves Néron

against potential transformation of knowledge into power or when no ex post redress mechanism (that could reverse some detrimental outcomes of accrued power) is available. For instance, where physicians are not subject to a deontolog- ical code enforced by a professional order, where the law does not protect patients against unconsented and unbeneficial experiments, where doctors are not legally liable for their actions, IAs raise an issue of domination in the neo-republican sense.

Finally, IA could be a symptom of more fundamental underlying inequalities and vulnerabilities. People may suffer from the consequences of IAs because they are worse off on other dimensions. Often IAs are more the symptoms of, or predates on, underlying inequalities or injustices than they are inequalities or injustices in themselves. This is blatant in the case of credit markets, in which the most epistemically vulnerable agents are often the most vulnerable ones from an economic and social point of view.31

This is not banal. These dynamics were clearly at play during the 2007–2008 financial crisis. The subprime crisis partly emerged because some lenders tar- geted borrowers with low income and assets (ironically described as “NINJAs,” “No Income, No Job and no Assets”) for predatory mortgage loans. Lenders were searching to maximize the commission they received on signed loans with little consideration of the ability of borrowers to repay their loans. Moreover, they often concealed information from borrowers or obfuscated the details and timeline of mortgage repayments, making it more difficult for the latter to make appropriate decisions. In addition, borrowers were usually lured into contracting a mortgage loan by a very low initial repayment rate, which quickly climbed. To make things worse, NINJA borrowers belonged to vulnerable socioeconomic groups with low resources, education, or knowledge of lending practices. In short, some predatory lenders took advantage of both IAs and borrowers’ poor economic and social conditions.

Furthermore, IA may build on or reinforce elitism, nepotism, and other dem- ocratic defects. Some individuals may acquire highly relevant information for specific market transactions because they belong to privileged social categories, professional networks, and so forth. Qua members of these groups, they may have the opportunity of extracting undue advantages (what economists characterize as “rents”) from market transactions with agents who do not belong to these privi- leged groups.32 A typical case of this is insider trading.

Here, IA does not (only) express distributive problems that could be fixed by redressing the outcomes of specific interactions or cooperative mechanisms, but also structural ones, that is, looming issues where some individuals are con- stantly at a disadvantage because of some socioeconomic characteristics. The fact that individuals are constantly disadvantaged in their access to information because they belong to certain groups highlights a deeper problem: the existence of structural inequalities within the society and the tendency of such inequalities to “contaminate” many dimensions of social life such as relations to agents with a specific knowledge (e.g., bankers, mortgage brokers, physicians, and mechanics).

Morality and Market Failures: Asymmetry of Information 575

Again, IA in itself is not the most pressing issue here. It is only the symptom of some deeper, more worrying, problems.

2. IA and Public Regulation

Once we have shown that there are instances of IAs which are morally prob- lematic and we have identified different layers at which IAs are problematic, two questions are still begging:

1. Why could the state legitimately tackle IA? One thing is to demonstrate that IAs could be morally problematic; another quite different question is to identify the state as the appropriate institution to deal with them. Answering this question is to frame the principle of public regulation.

2. How is it legitimate to do so? In addition, it is important, once the principle of public regulation has been generally justified, to delve into the possible forms of public regulation.

2.1. The Principle of Public Regulation

First, it should be noted that to identify morally relevant IAs is different from justifying public regulation as the proper response (if there should be any “proper response”). A different option could be to let unregulated markets take care of IA. Such an option would benefit from justifications for free markets that emphasize the efficiency of unconstrained markets for handling and coordinat- ing knowledge. Other options include dealing with these issues through the court system or implementing self-regulatory regimes (like labels). Our point is that public regulation requires more extensive justification than it is often assumed.

For example, Friedrich Hayek considers that the complexity and abundance of local knowledge is the main argument for unregulated markets.33 Because knowledge is complex and spread over a multitude of agents, and because markets are precisely about knowledge creation and transmission (through prices), it is preferable for public institutions to stay clear of market interactions and to impose as little constraint as possible on individual free interactions. In this scheme, any public intervention for altering prices or quantities translates into disequilibria and, ultimately, welfare losses (on the top of unjustified restrictions to freedom of exchange).

A second point is that public regulation does not mean that public institutions should replace private initiatives in any instance where IAs (could) appear, for example by nationalizing health care or turning mechanics into civil servants. It simply means that rules, more stringent than those free markets could possibly set, should be established. It also means that some external control, that is, control that is external to markets and private actors, should be exercised. Finally, it also implies that external coercion should be available.

576 Xavier Landes and Pierre-Yves Néron

So why appeal to public regulation? We have seen that most of the problems are not due to IA per se, but to the capacity of agents to exploit IAs by extracting additional benefits, acquiring illegitimate power that creates domination (i.e., the capacity of an agent to arbitrarily interfere in the life of another agent), and so forth. Left to themselves, it is very likely that free markets and agents will leave such a capacity untouched. In relation to this shortcoming, public institutions have several advantages over private institutions such as markets.

First, public institutions operate from outside markets, meaning that they can alter the conditions under which agents enter markets or the outcomes of market transactions.34 This last point does not mean that public institutions should not be limited in their capacity to interfere in individual transactions.

Second, public institutions, in virtue of being public, that is, state-admin- istered, have extensive powers for imposing disclosure or demanding comple- mentary information in specific contexts (e.g., informed consent for medical interactions). They can use legal provisions to alter the rules of the game, for example by forcing the disclosure of important information (basis of consumer law). Through specific agencies, they can monitor market agents and transactions to be sure that they respect some standards with regard to the diffusion of strate- gic information or conflicts of interest. In addition, they can punish agents who unduly benefit from IA.

Third, liberal democratic public institutions, in virtue of being public, that is, founded on popular will and subject to the rule of law, are less prone to arbi- trariness. (There is still the risk of the tyranny of the majority, but overall the risk of arbitrariness is less important in political regimes founded on popular will than in alternatives such as oligarchy, monarchy, despotism, totalitarianism, and so on.) Of course, this condition assumes that private interests have not captured public institutions,35 that the legislative power is issued from the popular will, that the executive power is accountable in some meaningful sense, and that state activities are checked against the background of fundamental rights and a legal apparatus.

Except for radical libertarians, the principle that public institutions may be (and are often as a matter of fact) more efficient for addressing IAs is uncontro- versial.36 Thus, we do not discuss further whether public institutions could be more efficient than private actors or free markets, but assume this general pos- sibility for the rest of our article. Instead, we discuss the conditions under which public regulation is preferable to self-regulation, that is, regulation by market ac- tors themselves within free, minimally constrained markets.

An obvious answer is to pay attention to contingent comparative advantages, that is, to situations where public institutions are more efficient for taming mor- ally relevant IAs than private counterparts. However, as said, this answer is very simple, ad hoc (limiting itself to a case-by-case analysis by comparison with identifying deep reasons and broad mechanisms for state superior efficiency) and, furthermore, it does not acknowledge that the comparative advantage of public institutions may not be self-evident.

Morality and Market Failures: Asymmetry of Information 577

Nevertheless, if we put aside the question of comparative efficiency, there are also principled reasons for favoring public regulation under certain condi- tions. As we mentioned, IAs raise problems of equality between well-informed agents and epistemically vulnerable ones.37 Under these circumstances, there might be good reasons for regulating such interactions. Moreover, the princi- pled reason one may mobilize to justify state intervention is somewhat a logical consequence of one’s views of the general responsibilities of the state. For in- stance, if one considers that it is a state’s responsibility to guarantee equality of opportunities or power and that morally relevant IAs are shown to undermine equality of opportunities or power, then a prima facie case for state regulation has been made. However, it also means, as we argue in this article, that a nor- mative account of IAs (and market failures in general) cannot go without rein- troducing broader normative categories to sharpen the principled case for public regulation.

It is important to highlight this. We argue in favor of something slightly dif- ferent from Heath’s normative account of market failures, although we are sym- pathetic to his approach. Heath suggests that a clear account of market failures helps us to articulate what he calls the “implicit morality” of the market, namely a set of requirements that are embedded in the key conditions (like perfect compe- tition and information) that a market economy needs to satisfy in order to produce efficiency.38

But he somehow stops there. According to his view, a clear account of market failures is enough to unveil the implicit morality of the market and the justifica- tions for its regulation. We suggest that it goes both ways. If the economic theory of market failures sheds light on the implicit normative requirements of the good functioning of markets, political philosophy also provides a moral vocabulary to think more clearly about both what kind of problem markets failures like IAs represent and the institutional implications of attempting to tackle them.

However, that also means that the arguments advanced as well as our discus- sion below assume that our readers assign some duties to public institutions with regard to efficiency and equality (under various forms). In this article, we have no ambition to convince readers who would deny such duties.

2.2. The Forms of Public Regulation

Once the principle of public regulation is justified, the next question is how to regulate IA. Again, it is worth insisting that the search for justifying public regulation does not imply that the state should substitute itself for private actors. Often it means that public institutions should step in and impose rules on interac- tions and transactions. Public institutions should determine the rules of the game and control the respect of the rules by various market agents.

We list below the main forms that could be used by public institutions in order to address morally relevant IAs. For each form of regulation, we indicate in which sense they address the morally problematic aspect of IAs.

578 Xavier Landes and Pierre-Yves Néron

2.2.1. Compulsory disclosure

Although private regimes of labeling might represent an appealing alterna- tive, only public institutions can (de jure and de facto) force market agents to disclose all or the main bits of information they use when transacting. Such re- quirement is already imposed by consumer law in many countries with regard to the composition of products (e.g., if they contain allergens such as nuts, ge- netically modified organisms, colorants, conservators, chemicals, and so on) and, often, their origin (e.g., for the traceability of meat in Europe).

Compulsory disclosure aims at redressing the original informational imbal- ance and setting on a relatively equal footing the agents engaged in a transaction or interaction. In other words, it aims at fostering equal opportunities by reducing IAs.

Compulsory disclosure is a privileged tool for public institutions because cor- porations or individuals freely interacting on markets have many, usually self-re- garding, reasons for not disclosing all relevant information about their products, services, personal attributes, or risk profile. They might not be willing to reveal they have been using cheaper, but more hazardous, components in their products or that they have severe preexisting medical conditions. In addition, state agen- cies are often the sole agents to be able, by law, to force and monitor information disclosure. Finally, it is worth emphasizing that compulsory disclosure has the virtue of being usually minimally invasive, since it does not require a change in behavior.39

Addressing morally relevant IAs through compulsory disclosure, however, has a serious shortcoming: for the information to matter, it is important that in- dividuals have adequate cognitive abilities and sufficient time for processing the information and to decide on this basis. (Hence the importance of the cooling-off period in consumer law: humans are sometimes slow at making up their minds, even after the sale.) If they do not have adequate abilities or enough time, the detrimental aspect or consequences of IAs, in terms of opportunities and power, will remain.40

2.2.2. Complementary/alternative information

Public institutions may provide alternative channels of information available to all at a minor cost (because of the scale effect for the information collected and distributed by the state and its broader resources for undertaking this task). That could help counterbalance some asymmetries. The difference with the previous instrument is that it is not the producers or direct partners that are forced to reveal information. Information is in that case produced by either the government itself or third parties.

In the former case, government provides information through specific agencies (e.g., customer information), public announcements, or public media (e.g., radio or television broadcasting). In the latter case, it is usually the role of

Morality and Market Failures: Asymmetry of Information 579

independent research institutions (universities, think tanks), monitoring bodies, media, nonprofit organizations, and so on. Still, governments can play an indirect role here by funding these institutions.

Again, complementary/alternative information aims at redressing the infor- mational imbalance, that is, at promoting more equal opportunities. An aspect is also to provide more variety in the sources of information available, for example to consumers, which is a good thing.41 Then, individuals have access to more comprehensive information, in addition to a plurality of views and opinions (some of it supported by the state) that could help them to better integrate this volume of information.

The comparative advantage of public institutions is that, through taxation, they can levy great resources for funding various information channels and ini- tiatives. They have a higher power of investigation and, also, can reach a broader public (potentially, all inhabitants of the territory where state authority prevails). They can also impose that given activities and practices are monitored by inde- pendent bodies.

The use of complementary/alternative information may encounter difficul- ties. The main one is the independence of these initiatives. Of course, in a market economy structured by powerful economic interests it may prove challenging for independent bodies to do their job without being influenced. There is also the risk that public institutions themselves color the information they distribute, influence the work of these independent initiatives and then the quality of information they diffuse. Nonetheless, such risk does not constitute an argument against state pro- vision or support of information. It is an argument in favor of as objective as pos- sible public provision of information and minimum (and controlled) interference in the work of independent bodies and initiatives.

2.2.3. Educating individuals

IA is sometimes harmful because individuals lack the ability to search for and process the right kind of information (i.e., accurate and relevant for their situ- ation). They may lack adequate cognitive abilities, as convincingly shown by re- search in psychology over recent decades (e.g., the work of Daniel Ariely, Daniel Kahneman, and Amos Tversky), which is the shortcoming of the two previous instruments. Through general or more specialized education, public institutions then can help citizens to acquire the necessary skills to become more efficient at compensating initial IA.

Unlike the two previous instruments, education does not enhance the quality of information. It enhances the capacity of individuals to process information. In that respect, it may be argued that education plays a deeper role in strengthening equality of opportunities and more. Because it equips individuals with cognitive tools for finding, understanding, and acting on adequate information, education is important for lifting barriers in the transformation of opportunities into outcomes. Public institutions are particularly fit for the task, especially if one considers the

580 Xavier Landes and Pierre-Yves Néron

success of public education in industrialized nations. It is not to say that cognitive abilities cannot be developed in private settings (e.g., private schools) or voluntary initiatives (e.g., self-education). The point here is to underline the importance of the state for citizens’ education, through publicly funded schools and universities and for supporting educational campaigns and instruments (e.g., TV programs, public debates) that enhance individuals’ cognitive and critical thinking skills.

The advantage of public institutions can be understood by, again, a scale effect (i.e., the broader, the more capable, the less costly, the more efficient, and so on). States have great resources for educating citizens, subsidizing high quality media, and so on. At first blush, states are also less likely to give way to particular interests (e.g., corporate interests) when they educate people. Effective state neu- trality toward the diverse conceptions of the good life (even if public policies can never be totally value-free42) as well as the promotion of science are advantages when it comes to building initiatives for improving citizens’ cognitive skills.

Nonetheless, public intervention has not always been supportive of the Enlightenment enterprise of enhancing cognitive and critical abilities. History is full of examples of states using education for subjugating citizens, promot- ing blind obedience, enforcing conformism, and silencing critical thinking. Nonetheless, it should be acknowledged that modern democratic states are less exposed to these shortcomings that other political regimes thanks to political pluralism, institutional checks and balances, fundamental rights, public debates, free press, and other ingredients of what could be labeled as “democratic culture” (or, more prosaically, the development of democratic institutions). In any case, education highlights the importance of democratic institutions.

2.2.4. Constraining and monitoring transactions and interactions

Public institutions are essential for controlling market transactions and out- side market interactions (e.g., between patients and doctors within the public healthcare sector). They can monitor the terms of market transactions and other interactions to spot cases where one party exploits the other party through IA. Such monitoring can be performed ex ante by constraining what individuals can do during transactions or interactions43 or ex post by reviewing the outcomes of past transactions and interactions.

Monitoring transactions and interactions helps to prevent and redress some of the negative effects in terms of unjustified advantages or imbalance of power stemming from morally relevant IAs. At the difference of the first three instru- ments, which have an internal dynamic, monitoring is external: instead of trying to influence individual behavior from within by improving informational or cog- nitive skills, it externally constrains and checks transactions as well as interac- tions. Monitoring is about making sure that agents refrain from exploiting IAs.

The advantage of public institutions is, again, an affair of greater—financial and legal—resources (the scale effect mentioned above). The state defines the legal rules for everyone operating on its territory. Private actors have less reach.

Morality and Market Failures: Asymmetry of Information 581

They cannot enact rules for all citizens, they have fewer resources for monitoring, and less (democratic) legitimacy for undertaking these tasks.

Public regulation encounters challenges when it comes to altering the context of choice and monitor particular interactions and transactions.44 There is the risk of ill-intentioned, undemocratic state manipulation. Because the architecture of choice is manipulative by definition, there is a risk of sliding from justifiable to unjustifiable intervention. The context of choice can be altered for either reducing or increasing morally relevant IAs (e.g., by blocking any alternative source of information).

Another challenge is the potentially detrimental effect of state intervention on individuals’ morality and psychology. By excessively interfering in transactions and interactions, states can nurture psychological mechanisms that undermine in- dividual moral traits (at least so the argument goes). For instance, individuals may increasingly rely on public institutions and become less autonomous, they might eschew their responsibility by blaming public institutions for adverse events in their life, and so on. The corroding effect of state intervention is omnipresent in the literature on paternalism (e.g., John Stuart Mill and the claim that paternalism threatens individuality45). However, pointing at the risks carried by public inter- vention does not invalidate the principle of public regulation or the comparative advantages of public institutions. But it calls for close democratic scrutiny and clear rules delineating public action. In other words, it calls for a proper ethics of public regulation.

2.2.5. Controlling asymmetries

Public institutions may also engage with broader asymmetries, for example asymmetries of power. For instance, they can devise mechanisms for limiting ac- cumulation of power. By doing so, they lower vulnerabilities. The difference with previous instruments is that asymmetries are indirectly controlled by engaging deeper imbalances that may not be about information per se.

There is a strong assumption at the core of this fifth instrument, namely the idea that morally relevant IAs are to a significant extent the result of background imbalances that could be efficiently addressed by public institutions. Such a view can find support in political theories such as neo-republicanism or Marxism, and it raises questions about the nature of background imbalances, the kind of influ- ence they exert on IAs, and so on.

Without claiming that morally relevant IAs are solely symptoms of deeper asymmetries, it is plausible to argue that morally relevant IAs are perhaps gen- erated, or at least accentuated, by deeper asymmetries. As a matter of fact, it is difficult to deny that socioeconomic inequalities drastically impact individu- als’ access to information (through cultural capital), but also their ability to pro- cess information and use it to gain and retain various advantages (e.g., Pierre Bourdieu’s habitus or how individuals can transform cultural capital in deeply ingrained habits, skills, and dispositions).46 If we accept this idea that morally

582 Xavier Landes and Pierre-Yves Néron

relevant IAs are partly the consequence of profound socioeconomic structures and dynamics and if we agree that IAs present a real moral problem, it becomes legitimate to consider the means for lessening or correcting these background imbalances.

Here again, the advantages of public institutions are obvious. Thanks to the size of their material, legal and human resources, states can adequately address background imbalances. If the problem of IAs is due to (increasing) economic inequalities, a solution could be to adopt a more aggressive taxation policy on income and wealth. In other words, the state can be more efficient than private initiatives for addressing background imbalances. Even if these “background” in- terventions cannot be properly labeled as direct forms of market regulation, there are good reasons to think that an ethics of market regulation should be equipped to make sense of them.

Of course, the shortcomings are the same as for any social engineering. There is a serious risk of states tampering with social structures in a way that does more harm than good. Extreme cases involve the Communist experience of the twentieth century when the state carried on an ambitious project of redesign- ing a “new man” through completely revamped social structures leading, in most places, to undermine civil society (e.g., citizens’ capacity of mobilization) and jeopardize social cooperation (e.g., trust). However, addressing background in- equalities does not imply the complete recomposition of the structure of society. And again, the success of taming asymmetries depends on effective democratic control and strict rules imposed on public regulation. It is another plea for a solid ethics of public regulation.

2.2.6. Direct provision

Despite all these tools, some markets might be so crippled by IAs that it is difficult or impossible to regulate them efficiently while leaving the supply side in private hands. Therefore, public institutions may legitimately offer goods and services for bypassing IA. The main difference with previous tools is that direct provision is properly about public institutions replacing private actors.

Public health insurance is an obvious example. Despite serious informational problems, namely adverse selection and moral hazard, the state offers health in- surance directly to individuals. The justification is public institutions’ higher effi- ciency at imposing rules on agents (such as automatic disclosure of their medical conditions), nudge them by promoting less risky behaviors (such as less fat diet or more exercise), evaluate more accurately their expected losses (due to the size of the insured population and the collection of nationwide statistics), and spread losses.

Direct provision is probably the most obvious remedy against market failures: when markets fail (i.e., when they fail to produce the outcomes they were expected to, which could manifest through distorted prices or rents, ex- cess of demand or supply), the solution appears to replace competition by public

Morality and Market Failures: Asymmetry of Information 583

monopolies. To be clear, public monopolies remedy morally relevant IAs in two different ways. First, direct provision of goods and services addresses the issues in a straightforward way because it allows governments to ensure that all con- sumers are adequately informed. Second, public monopolies have this effect of removing the need to make choices by substituting fixed options for free choice. In such a case, there is no need to worry about the information that is supposed to inform choice.

Public monopoly, however, does not offer in itself any guarantee against the exploitation of IAs. Moreover, it does not protect against abuses of state power. It is not because public institutions provide goods and services that consumers will be fully informed about the nature of the products, their composition, their effects, and so on. It is an independent issue. Also, public provision opens, again, the possibility that the state will use such provision to subject people, as shown by history’s multiple examples. Direct provision underscores the necessity of an ethics of public regulation.

3. Toward a Normative Theory of Public Institutions Based on Market Failures

Whereas the first section of our article reviewed the main ethical dimensions of morally relevant IAs and the second section tried to unfold the main venues for public regulation, this third section delineates few general principles that are useful for designing an ethical theory of public regulation. Such a theory seems to us intimately intertwined with a theory of public institutions based on market failures.

In brief, the question is twofold: what are the advantages of adopting a mar- ket failure approach? Which ethical principles of public regulation are underlined by an approach based on market failures?

To answer this question, it is useful to try to identify the different normative commitments that emerge from an analysis of morally relevant IAs. First of all, it suggests that a general condemnation of the exploitation of morally relevant IAs is a key feature of well-functioning markets and good public policies. To put it otherwise, the nonexploitation of morally relevant IAs is a crucial part of what Heath calls the implicit morality of the market.47

From this general principle, two more specific moral commitments seem to emerge from a closer attention to morally relevant IAs.

(a) The need to reduce morally relevant IAs. In some situations, the presence of IA calls for a more equitable distribution of information between economic agents. This dimension is mostly an institutional duty (i.e., one that applies to and is performed by institutions). (b) The need to refrain from exploiting IAs. In some situations, the presence of IAs calls for economic agents to refrain from exploiting them, which can be achieved by devising control mechanisms that assure that agents who possess

584 Xavier Landes and Pierre-Yves Néron

extra knowledge does not take too much advantage of it. This second dimension is more a general duty applying across the board to institutions and individual actors.

Most regulatory measures, like the ones identified in the second section, aim to achieve these two tasks by either reducing IAs or making sure agents will re- frain from exploiting them. However, this distinction is important because it also suggests guidelines, not only for an ethics of market regulation, but also for an ethics of market actors. From such a point of view, where regulation is incomplete or imperfect, firms ought to either reduce IAs (by voluntarily disclosing infor- mation, for example) or refrain from exploiting it (by avoiding certain advertising strategies, for instance).

Hence, one of the implications of this approach is, as Norman puts it, to cre- ate more symmetry between the language, principles, and tools we use to justify market regulation and the language, principles, and tools we can use to justify ethical constraints and beyond-compliance obligations in the conduct of business firms.48

This is crucial because it is, according to Heath, a clear advantage of the market failures approach.49 It allows us to identify problematic or straightfor- wardly immoral behaviors of economic actors without relying on a theory of “general morality,” like Kantian or Aristotelian ones. Indeed, instead of applying somewhat clumsily the general principles of those theories to market interac- tions, it starts from the underlying conditions for achieving market efficiency and the normative requirements that are embedded in those conditions. It then proceeds by identifying practices that contradict those requirements and under- mine those conditions, such as the exploitation of IAs. From this point of view, what is wrongful about business practices such as deceptive advertising, and what precisely grounds their regulation, is not that it represents a violation of a Kantian duty or a departure from the virtue of honesty, but that it represents an exploita- tion of a market failure, namely IAs.

Whereas we share Heath’s insight and ambition, we nonetheless hope to have shown in this article that such an approach cannot go without reintroduc- ing broader normative categories such as social inequality, vulnerability, power, arbitrary interference, and so on.50 Even if one takes Heath’s guidance from a normative reading of market failures, that is, one that gives us a sense of the nor- mative dimensions of economics in general and markets in particular, it seems nonetheless to be both useful and inevitable to reintroduce broader concerns, like egalitarian ones.51 IAs, like other market failures, are not purely isolated. They emerge alongside other issues and can actually be useful in order to assess the various qualities of specific interactions. And a sophisticated ethics of market regulation should account for this.

That being said, it might be useful here to say a few words about how the market failures approach differs from other ways of thinking normatively about markets and their limits. Take, for instance, the “corruption” theory of the limits

Morality and Market Failures: Asymmetry of Information 585

of markets, developed by Michael Sandel and others.52 According to this view, markets reach their limits when they degrade the nature of specific goods, like honors, health, and so on. And it is why markets should be, if not forbidden, care- fully monitored and restricted, or at least “put in their place.”

Our analysis probably has a lot in common with this approach. However, it is useful to remark that our account of IAs focuses less on the nature of spe- cific informational goods, like knowledge, expertise, cultural production, and so on, than on the objectionable nature of some economic interactions or relations. Hence, even though, at first sight, our analysis seems to rely on analysis of a spe- cific good, namely information, it ultimately gives a lot of weight to interactional concerns. Normatively, the analysis of morally relevant IAs is more about objec- tionable interactions or relations than it is about specific goods.

In sum, the asymmetries at the core of IAs express imbalances between in- dividuals and, often, categories of individuals within a given society (e.g., so- cioeconomic status, membership in privileged groups, and access to relevant information). In many cases, morally relevant IAs are not only informational asymmetries, they are relational asymmetries. They overlap and nurture differ- entials of status among citizens. In other words, their unchecked existence threat- ens democratic equality in the sense given to the concept by an author such as Elizabeth Anderson.53 Future research might attempt to highlight the relations between forms of market failures and the egalitarian concern for equal relations. After all, both share the common goal of improving the quality of our social and economic relations.54

Notes

1Robert H. Frank and Ben S. Bernanke, Principles of Microeconomics (New York: McGraw-Hill Ir- win, 2009), 341; Joseph Heath, “Three Normative Models of the Welfare State,” Public Reason 3, no. 2 (2011): 13–43.

2William J. Baumol, Welfare Economics and the Theory of the State (London: London School of Economics and Political Science; New York: Longmans, Green, 1952).

3Rutger Claassen, “Externalities as a Basis for Regulation: A Philosophical View,” Journal of Institu- tional Economics 12, no. 3 (2016): 541–63.

4See Joseph Heath, Morality, Competition and the Firm: A Market Failures Approach to Business Ethics (Oxford: Oxford University Press, 2014).

5David Moss, When All Else Fails: Government as the Ultimate Risk Manager (Cambridge: Harvard University Press, 2002); Heath, “Three Normative Models of the Welfare State.”

6François Ewald, L’État Providence (Paris: Grasset, 1986); Joseph Heath, “The Benefits of Coopera- tion,” Philosophy and Public Affairs 34, no. 4 (2006): 313–51.

7An example is the system of personal number and health card in place in Scandinavian countries that give public institutions access to large amounts of personal data about health.

8The Law of Large Numbers is the “mathematical premise stating that the greater the number of exposures, (1) the more accurate the prediction; (2) the less the deviation of the actual losses from the expected losses (…); and (3) the greater the credibility of the prediction (…). This law forms the basis for the statistical expectation of loss upon which premium rates for insur- ance are calculated. Out of a large group of policyholders the insurance company can fairly accurately predict not by name but by number the number of policyholders who will suffer the

586 Xavier Landes and Pierre-Yves Néron

loss.” Harvey W. Rubin, Dictionary of Insurance Terms (Hauppauge, NY: Barrons Educational Series, 2008), 272–73.

9Claassen, “Externalities as a Basis for Regulation.” 10John Stuart Mill, On Liberty (1859). 11Frank applies Mill’s harm principle for justifying the regulation of positional externalities; e.g.,

Robert H. Frank, “Should Public Policy Respond to Positional Externalities?,” Journal of Public Economics 92, no. 8–9 (2008): 1777–86.

12Joseph Stiglitz, “Information and the Change in Paradigm in Economics,” The American Economic Review 92 (2002): 460–501.

13Frank and Bernanke, Principles of Microeconomics, 333. 14Thomas G. McGuire, “Physician Agency,” in Handbook of Health Economics, Volume 1A, ed.

A. J. Culyer and J. P. Newhouse (Amsterdam: North-Holland, 2000), 461–536. 15George A. Akerlof, “The Market for ‘Lemons’: Quality Uncertainty and the Market Mechanism,”

Quarterly Journal of Economics 84, no. 3 (1970): 488–500. 16George Ainslie, Breakdown of Will (Cambridge: Cambridge University Press, 2001); George A.

Akerlof and Robert J. Shiller, Animal Spirits (Princeton, NJ: Princeton University Press, 2009); Daniel Ariely, Predictably Irrational (New York: HarperCollins, 2008); Daniel Kahneman, Thinking, Fast and Slow (New York: Farrar, Straus and Giroux, 2011); Robert J. Shiller, Irratio- nal Exuberance (Princeton, NJ: Princeton University Press, 2005).

17An open question is the nature of the advantage that is morally relevant for this discussion. The classic interpretation is of actual advantage: morally relevant IAs are situations where advan- tages are actually extracted. A concurrent interpretation is of potential advantage: morally relevant IAs are situations where advantages could be potentially extracted.

18A situation is Pareto superior (or represents a Pareto improvement) to a default situation when there is a possibility to improve the situation of at least one agent without worsening the situa- tion of any other agent.

19Unless stated otherwise, we use exchanges for qualifying market exchanges while interactions define more general inter-individual relations that do not imply the trade of goods or services in exchange of money.

20If an IA is efficient, meaning that it leads to a Pareto superior situation by comparison to a situ- ation characterized by no IA, then it would still be hypothetically possible to compensate the disadvantaged part (Kaldor-Hicks criterion). If an IA is inefficient in the Pareto sense, it means that it generates a situation where at least one agent is disadvantaged (in equal proportion to the advantage obtained by the better-informed agent) by comparison with a situation without IA. Thus, both efficiency and redistribution contribute turning IAs into morally relevant IAs.

21Both parties may benefit from such a transaction. Thus, it is not needed that the transaction is detrimental to one of the parties. Most of the time IAs do not produce a net loss, but a relative loss by comparison with a situation without IA and, more importantly, such a loss is without legitimate reasons. We will return to this point below.

22Xavier Landes, “Insurance,” in Encyclopedia of Corporate Social Responsibility, ed. Samuel O. Idowu, Nicholas Capaldi, Liangrong Zu, and Ananda Das Gupta (New York: Springer, 2013), 1433–40.

23Xavier Landes, “Moral Hazard,” in Encyclopedia of Corporate Social Responsibility, ed. Samuel O. Idowu, Nicholas Capaldi, Liangrong Zu, and Ananda Das Gupta (New York: Springer, 2013), 1715–22.

24For a complete discussion of the moral dimensions of moral hazard, the reader can refer to Benjamin Hale, “What’s so Moral About Moral Hazard?,” Public Affairs Quarterly 23, no.1 (2009): 1–26.

25Will Kymlicka, Contemporary Political Philosophy: An Introduction (Oxford: Oxford University Press, 2002), 72–74.

26Adam Smith, An Enquiry into the Wealth of Nations (1776): Book I. chapters 1–3. 27For these notions, see Samuel Arnold, “The Difference Principle at Work,” Journal of Political

Philosophy 20, no.1 (2012): 94–118.

Morality and Market Failures: Asymmetry of Information 587

28For an analysis of power and informational issues in principal-agent theory, see Terry M. Moe, “Power and Political Institutions,” Perspectives on Politics 3, no. 2, (2005): 215–33.

29Philip Pettit, Republicanism (Oxford: Oxford University Press, 1997). 30An important distinction here is between the potential leverage that any IA gives to any agent who

holds extra knowledge and the cases described here in which specific IAs come with asymmet- ric power. The difference lies in the presence or the absence of institutional checks limiting the potential transformation of knowledge in actual power such as a medical college for physicians or any professional self-regulatory body.

31Concerning the relations between financial crises and economic inequalities, see of R.G. Rajan, Fault Lines (Princeton, NJ: Princeton University Press, 2011); Michael Kumhof and Romain Rancière, Inequality, Leverage and Crises, IMF Working Paper (2011) 10/268; 3–37.

32This kind of rent seeking is basically Stiglitz’s explanation for the rise of inequalities in United States. Joseph E. Stiglitz, The Price of Inequality (New York: W.W. Norton, 2012).

33Friedrich A. Hayek, “The Use of Knowledge in Society,” American Economic Review 35, no. 4 (1945): 519–30.

34A stronger formulation could be made: public institutions actually create markets. Without prop- erty rights guaranteed by public force and law enforced by courts, no durable and functioning market is possible. Daron Acemoğlu and James A. Robinson, Why Nations Fail (New York: Crown Business, 2012).

35The capture is one of the major concerns of the literature on winner-take-all politics. Larry M. Bar- tels, Unequal Democracy (Princeton, NJ: Princeton University Press, 2008); Jacob S. Hacker and Paul Pierson, Winner-Take-All Politics (New York: Simon & Schuster, 2010).

36Most economics textbooks include discussions of this idea and treat it as being fairly uncontrover- sial; e.g., Frank and Bernanke, Principles of Microeconomics.

37For an interesting account of epistemic injustices, see Miranda Fricker, Epistemic Injustice: Power and the Ethics of Knowing (Oxford: Oxford University Press, 2007).

38See Joseph Heath, Morality, Competition and the Firm: A Market Failures Approach to Business Ethics (Oxford: Oxford University Press, 2014), chap.7. For an attempt to develop further the idea of an implicit morality of the market, see Pierre-Yves Néron, “Rethinking the Ethics of Corporate Political Activities: Political Equality, Corporate Citizenship and Market Failures,” Journal of Business Ethics 136, no. 4 (2016): 715–28.

39We would like to thank one anonymous reviewer for urging us to make this point. 40Oren Bar-Gill, Seduction by Contracts (Oxford: Oxford University Press, 2012). 41Ibid. 42The emphasis on the illusion of a state fully neutral on moral and cultural matters is arguably one

of the most important contributions of the debates on multiculturalism that have taken place since the 1990s.

43Frank adopts a similar position when he argues in favor of an incremental tax on consumption. His rationale is the following: since individuals are trapped in unhealthy patterns of over-con- sumption, i.e., in a collective action problem, it is necessary to alter the context of choice by rendering consumption steeply more expensive. Robert H. Frank, Luxury Fever (New York: The Free Press, 1999). For a critical discussion, refer to Xavier Landes, “Why Taxing Consumption: Justifications, Objections and Social Cooperation,” in Philosophical Explorations of Justice and Taxation: National and Global Issues, ed. Helmut P. Gaisbauer, Gottfried Schweiger, and Clemens Sedmak (New York: Springer, 2015), 101–17.

44For a general discussion of the political challenges of architecture of choice, one may refer to the, by now, broad literature on nudge and architecture of choice; e.g., Bryan Jones, Politics and the Architecture of Choice (Chicago: University of Chicago Press, 2001); Cass R. Sunstein, Why Nudge? (New Haven, CT: Yale University Press, 2014); Richard H. Thaler and Cass R. Sunstein, Nudge (New Haven, CT: Yale University Press, 2008).

45Peter de Marneffe, “Avoiding Paternalism,” Philosophy and Public Affairs 34, no. 1 (2006): 68–94. 46See Pierre Bourdieu, Distinction: A Social Critique of the Judgement of Taste (Cambridge: Har-

vard University Press, 1987).

588 Xavier Landes and Pierre-Yves Néron

47Heath, Morality, Competition and The Firm, chap. 7. 48Wayne Norman, “Business Ethics as Self-Regulation: Why Principles that Ground Regulations

Should Be Used to Ground Beyond-Compliance Norms as Well,” Journal of Business Ethics 101 (2012): 43–57.

49See Heath, Morality, Competition and The Firm, chap. 3–4. 50For an attempt to do something similar, see Abraham Singer, “Justice Failure: Efficiency and

Equality in Business Ethics,” Journal of Business Ethics (2016): https://doi.org/10.1007/ s10551-016-3086-x

51Singer, “Justice Failure.” 52Michael Sandel, What Money Can’t Buy: The Moral Limits of Markets (New York: Farrar, Straus

and Giroux, 2012). For a similar but different account, see Debra Satz’s excellent book, Why Some Things Should Not Be for Sale: The Moral Limits of Markets (Oxford: Oxford University Press, 2010).

53Elizabeth Anderson, “What Is the Point of Equality?,” Ethics 109, no. 2 (1999): 287–337. 54See Pierre-Yves Néron, “Rethinking the Very Idea of Egalitarian Markets and Corporations:

Why Relationships Might Matter More than Distribution,” Business Ethics Quarterly 25, no. 1 (2015): 93–124.