Loading field map
Information economics is the branch of microeconomics that studies how the possession, distribution, and quality of information shape economic decisions, market outcomes, and institutional design. Its central departure from classical price theory is the recognition that information is not a free, perfectly shared input but a scarce, costly, and often asymmetrically distributed resource. When some parties know more than others, or when information is simply absent, the standard results of competitive markets—efficient prices, mutually beneficial exchange, optimal allocation—can break down in systematic and predictable ways.
The field's organizing insight is that economic actors rarely share the same information. This condition, called asymmetric information, arises whenever one party to a transaction knows something relevant that the other does not. The party with superior information can exploit that advantage, and the party without it knows this, which changes behavior on both sides before any exchange occurs.
Asymmetric information manifests in two fundamental forms, each with distinct consequences. Hidden characteristics exist before a transaction: a seller knows a used car's defects, a job applicant knows their own ability, an insurance applicant knows their health risks. Hidden actions occur after a transaction begins: an insured person may take more risks, an employee may shirk when unobserved, a borrower may invest in riskier projects than the lender intended. These two forms generate the field's two signature problems: adverse selection (bad types are disproportionately likely to enter a market) and moral hazard (incentives change after an agreement is struck).
The consequences are not merely distributive—they can be efficiency-destroying. In the canonical example, if buyers cannot distinguish good used cars from "lemons," they will only offer a price reflecting average quality. Owners of good cars then withdraw from the market, average quality falls, prices fall further, and the market may collapse entirely. The mere possibility of exchange does not guarantee it will occur, even when both sides would benefit from a fair deal.
Information is also a commodity in its own right, and it behaves unlike ordinary goods. It is non-rivalrous: one person's use does not diminish another's. It is often non-excludable: once disclosed, it is hard to prevent others from using it. These properties make information a public good, which creates a fundamental tension. If information is costly to produce but cheap to copy, producers cannot recover their costs through market prices, so too little information gets produced. But if information is protected by patents, copyrights, or secrecy, it is underused relative to the social optimum. This tension—between incentives for production and efficiency in use—is the central problem of the economics of information as a commodity.
A further failure arises because information is a credence good: its value often cannot be assessed until after purchase, and sometimes not even then. A medical diagnosis, a legal opinion, or a financial forecast is hard to evaluate even after the fact, because the buyer cannot know what a better expert would have said. This makes markets for expertise and advice particularly prone to charlatanism and reputational manipulation.
Much of information economics is organized around the principal–agent model, a general template for any relationship in which one party (the principal) delegates decisions or actions to another (the agent) whose behavior the principal cannot fully observe or verify. The principal wants the agent to act in the principal's interest; the agent has their own interests and private information. The problem is to design a contract, incentive scheme, or governance structure that aligns the agent's incentives with the principal's goals as closely as possible, given that monitoring is costly or impossible.
This framework is remarkably general. It covers the shareholder–manager relationship in corporations, the insurer–policyholder relationship, the employer–employee relationship, the regulator–firm relationship, and the voter–politician relationship. The central result is that when information is asymmetric, the first-best outcome—the one achievable with perfect information—is generally unattainable. The principal must choose among second-best arrangements, trading off the costs of providing incentives against the losses from distorted behavior.
A key distinction within this framework is between hidden information and hidden action, which call for different remedies. Hidden information is addressed through screening (the uninformed party designs a menu of options that induces the informed party to reveal their type through their choices) or signaling (the informed party takes a costly action that credibly communicates their type). Hidden action is addressed through incentive contracts that tie pay to observable outcomes, even when those outcomes are imperfectly correlated with effort.
Screening is the strategy of the uninformed party. The principal offers a set of contracts or choices designed so that different types of agents will select different options, thereby revealing their private information through their own behavior. The classic example is an insurance company offering policies with different deductibles and premiums: low-risk customers will prefer a low premium with a high deductible, while high-risk customers will accept a higher premium for lower deductibles. The menu works because the choices are structured so that each type finds it in their interest to self-select into the option designed for them. The cost is that the menu must be distorted away from what would be offered under full information—the low-risk type must receive a less attractive deal than they would otherwise get, to prevent the high-risk type from pretending to be low-risk.
Signaling is the strategy of the informed party. The informed party takes an observable action that is costly in a way that correlates with their private type, making the action a credible signal. Education is the canonical example: if acquiring education is more costly for low-ability workers than for high-ability ones, then a degree can credibly signal ability even if the education itself adds no productive skill. The signal works only because it is differentially costly—if it were equally costly for all types, it would convey no information. Signaling can be socially wasteful when the signal's only function is to sort people rather than to create value, but it can also be productive when the signal itself (like education) has real benefits.
The relationship between screening and signaling is complementary rather than competitive. They are two sides of the same coin: the uninformed party screens by designing a menu; the informed party signals by choosing an action. In equilibrium, both mechanisms operate simultaneously, and the resulting outcome depends on which party moves first and what commitment power each possesses.
A major theoretical advance was the recognition that many seemingly different institutional arrangements can be analyzed within a unified framework called mechanism design. The revelation principle states that any outcome achievable through any complex game of strategic communication can also be achieved by a direct mechanism in which agents truthfully report their private information and the designer commits to an outcome rule based on those reports. This result is powerful because it simplifies analysis: instead of considering all possible strategic games, the designer can focus on incentive-compatible direct mechanisms—those in which truthful reporting is each agent's best response.
Mechanism design provides a general language for asking what outcomes are feasible when information is private. It yields impossibility theorems as well as constructive results. The most famous is the Myerson–Satterthwaite theorem, which shows that when a buyer and seller each have private information about their own valuation, no mechanism can guarantee efficient trade while also ensuring that both parties voluntarily participate and that the mechanism does not run a deficit. Some inefficiency is unavoidable in any bilateral trade with two-sided private information. This result is not a curiosity; it explains why many potentially beneficial trades do not occur and why real-world trading institutions are imperfect.
Mechanism design also clarifies the limits of decentralization. The Groves–Clarke mechanism shows how to elicit truthful reports of preferences for public goods, but it requires the designer to know the structure of preferences and to be able to make transfers. The Vickrey–Clarke–Groves (VCG) mechanism generalizes this to allocate resources efficiently when agents have private values, but it is vulnerable to collusion and requires the designer to compute optimal allocations—assumptions that fail in many real settings.
A distinct but related tradition studies information not as a fixed asymmetry but as something that agents actively acquire. In search theory, buyers do not know the prices or qualities offered by different sellers and must spend time and effort to discover them. The optimal search strategy involves a reservation price: keep searching until you find a price at or below a threshold, then stop. The existence of search costs explains why identical goods sell at different prices in the same market—a phenomenon that classical price theory cannot accommodate.
This perspective generalizes to the economics of information acquisition more broadly. Agents decide how much information to gather before acting, trading off the cost of information against the expected improvement in decisions. A central result is that information has value only insofar as it can change a decision; information that would not alter the chosen action is worthless regardless of how much it reduces uncertainty. This insight, formalized in the concept of the value of information, provides a rigorous foundation for thinking about when ignorance is rational and when it is not.
Information is not only something agents have or lack; it is something they can manipulate. Information design (also called Bayesian persuasion) studies how an informed party can strategically disclose information to influence the decisions of an uninformed party, even when the informed party cannot lie. The sender chooses a signal structure—a rule that generates messages conditional on the state of the world—and the receiver updates beliefs and acts. The sender's power comes from choosing how much and what kind of information to reveal, not from fabricating facts. A striking result is that the sender can often benefit from committing to a noisy or partial disclosure policy, because the receiver's rational response to that policy can be steered in the sender's favor.
This framework has been applied to advertising, financial disclosure, political campaigning, and regulatory announcements. It differs from signaling in that the sender controls the entire information structure rather than choosing a single costly action, and it differs from screening in that the receiver is not designing a menu but simply responding to disclosed information.
Information economics emerged as a distinct subfield in the mid-twentieth century, though its intellectual roots reach back further. Early discussions of uncertainty and expectations in economics—from Frank Knight's distinction between risk and uncertainty to John Maynard Keynes's treatment of expectations—recognized that economic actors lack perfect knowledge. But these treatments did not systematically analyze how information asymmetries distort market outcomes.
The modern field crystallized in the 1960s and 1970s through a series of foundational contributions. George Akerlof's analysis of the lemons problem showed how adverse selection can destroy markets. Michael Spence's model of job-market signaling demonstrated how costly signals can convey private information. Joseph Stiglitz and Michael Rothschild's work on insurance markets showed how screening menus can partially solve adverse selection problems. These contributions, recognized with the 2001 Nobel Prize in Economics, established the core concepts and the characteristic style of the field: rigorous formal models that derive surprising conclusions from simple assumptions about information asymmetries.
A second wave of development came from mechanism design and contract theory, associated with Leonid Hurwicz, Eric Maskin, Roger Myerson, Jean Tirole, and Oliver Hart, among others. This work generalized the field's insights into a comprehensive theory of institutions and incentives, showing how different contractual and governance arrangements arise as responses to information problems. The 2007 Nobel Prize recognized mechanism design theory, and the 2016 prize recognized contract theory.
A third wave, still ongoing, has extended the field into information design, the economics of data and privacy, and the analysis of digital markets where information is both the input and the output of economic activity. The field has also become more empirical, with researchers testing information-theoretic predictions using field experiments and natural experiments in contexts ranging from microfinance to health insurance to online marketplaces.
Contemporary information economics is characterized by several overlapping research programs rather than a single dominant paradigm. Contract theory continues to develop increasingly sophisticated models of optimal incentive schemes, incorporating dynamics, multiple tasks, and limited commitment. Mechanism design has expanded into algorithmic mechanism design, which studies how to design mechanisms that can be computed efficiently, and into the analysis of large markets such as school choice and organ allocation. Information design has become a major research area in its own right, with applications to financial regulation, platform design, and political economy.
The field has also engaged with behavioral economics, which challenges the assumption that agents process information rationally. Models of inattention and bounded rationality ask what happens when agents cannot or do not process all available information, even when it is free. This work has produced results that differ from classical information economics: for example, inattentive consumers may be exploited by firms that hide fees or use complex pricing, and rational inattention can explain why agents systematically ignore relevant information.
A significant contemporary development is the economics of data and privacy. Data is a peculiar economic good: it is non-rivalrous, it can be copied at zero marginal cost, and its value often depends on who else has access to it. The economics of privacy asks how individuals' control over their own data affects market outcomes, and how data markets should be regulated. This literature draws on information economics but also raises new questions about externalities, property rights, and the social value of information that the classical framework did not address.
The field's methods have also diversified. While formal theory remains central, there is now substantial experimental and empirical work. Laboratory experiments test whether real subjects behave as the models predict; field experiments test interventions designed to address information problems, such as providing information about school quality, health risks, or financial products. This empirical turn has both confirmed and complicated the theoretical results, showing that real-world information problems are often more nuanced than the clean models suggest.
Information economics has been remarkably successful, but it has clear boundaries. Its models typically assume that information problems are the only friction in the market; when combined with other frictions—such as limited enforcement, behavioral biases, or market power—the results can change substantially. The field also struggles with the fact that information is endogenous: the very existence of a market or institution changes what information agents have incentives to acquire and reveal. Models that treat information as fixed may miss important feedback effects.
A deeper limitation is that information economics is better at diagnosing problems than at prescribing solutions. The theory shows that markets fail when information is asymmetric, but it also shows that many proposed remedies—regulation, disclosure requirements, intermediary institutions—create their own information problems. The field has no general theorem stating when intervention improves outcomes, only a collection of models that must be applied case by case.
Finally, the field's treatment of information as a purely instrumental good—valuable only insofar as it improves decisions—leaves out dimensions that matter in practice. Information can have intrinsic value, symbolic value, or value in shaping identity and social relationships. Privacy is not only about protecting information from use; it is also about autonomy and dignity. These considerations lie outside the standard framework, and integrating them remains an open challenge.
Information economics has permanently changed how economists think about markets. The recognition that information is scarce, costly, and unevenly distributed—and that these facts have systematic consequences—is now part of the basic toolkit of economic analysis. The field's models are used not only by economists but also by policymakers designing regulations, by firms designing contracts and platforms, and by anyone who needs to understand why markets sometimes fail and what can be done about it.