What is information asymmetry and market design?

The short answer

Akerlof’s 1970 paper used the used-car market as illustration. Sellers know whether their car is a “lemon” (low quality) or a “peach” (high quality); buyers cannot reliably distinguish before purchase. Buyers offer prices reflecting average quality; high-quality sellers withdraw from this average price; remaining quality drops; buyers offer lower prices; the market spirals toward dysfunction or non-existence. The model demonstrated formally what informal observation already suggested.

The intuition is that markets require some baseline of trust or signal-credibility to clear. When neither exists, the rational response is to either not transact or to transact only in narrow circumstances where direct observation substitutes for missing information.

The complication is that Akerlof’s model predicted some markets would disappear that empirically continue to exist. The market for organ transplants, where matching errors are catastrophic, exists through institutional design (Roth’s matching algorithms). Used cars still sell despite the lemon problem because mechanic inspections, warranties, and reputation systems provide partial information. Asymmetric information sometimes destroys markets and sometimes provokes institutional innovation — which response occurs depends on the design space available.

→ New to information economics? Financial education hub

What the data shows

The empirical record on information asymmetry spans both sides of the theoretical question.

The figures (academic literature, market microstructure data, 1970-2024):

• Akerlof’s lemons paper has been cited over 40,000 times across economics, law and management literature, making it among the most influential single papers in modern economics
• Used-car price discounts for similar models from private sellers vs dealers (where dealer warranties partially close the asymmetry) document approximately 10-15% lemon discounts in standard markets
• Roth’s matching algorithms have been deployed for school choice in New York City (since 2003), Boston, and other systems, with documented improvements in matching efficiency and stability
• Bid-ask spreads in equity markets — a direct empirical proxy for information asymmetry — vary by orders of magnitude across stocks, with the most opaque securities showing spreads 10-100x those of the most transparent

The asset-management variant of asymmetry is also well-documented. Investors face fund managers whose true skill is hard to distinguish from luck over short samples; managers face investors whose true horizons are hard to distinguish from stated horizons. The two-sided asymmetry shapes industry structure (rebalancing penalties, lock-ups, liquidity provisions).

The exception worth flagging: in some markets institutional design has resolved asymmetry so completely that the asymmetry is no longer apparent. Spectrum auctions (Vickrey-Clarke-Groves mechanisms, FCC implementations since 1994) extract information from bidders through auction design rather than relying on disclosure. The market design literature shows that asymmetry is sometimes a problem to be solved by institutional architecture rather than by improving information per se.

→ Dataset: VIX volatility index dataset

Why it happens — the macro mechanism

Three channels explain the persistence of information asymmetry effects.

Channel 1 — Adverse selection. When low-quality participants disproportionately accept the going price, the average quality of the pool drops. Insurance markets are a textbook example: those most likely to file claims have strongest incentives to buy coverage, raising premium-funded payouts. The standard responses are screening (medical exams, deductibles, pricing tiers) that partially restore separation between high- and low-risk participants.

Channel 2 — The institutional-design solution. This is the angle most overlooked. Market design treats asymmetry as a constraint to engineer around rather than as a fact to lament. Roth’s organ-exchange algorithms allow patients to receive kidneys despite incompatibility with their would-be donor by chaining swaps; the design extracts cooperation from informed parties without requiring full information disclosure. School-choice mechanisms (deferred-acceptance algorithms) similarly allocate scarce educational resources without requiring families to truthfully reveal preferences they could strategically misrepresent under simpler systems.

The corollary is that whether asymmetric information destroys a market depends on whether the institutional design space is available and whether designers have access to it. Markets that historically failed due to asymmetry (organ exchange in the 1990s) now function under modern designs.

Channel 3 — Reputation and signalling. Spence’s signalling model (1973) showed that informed parties can convey information through costly signals. Education credentials, brand investment, and warranty length are partial substitutes for direct quality observation. The signal works only when its cost is correlated with the underlying quality being signalled — pure cheap-talk does not solve the asymmetry problem.

Synthesis by regime: in pre-Akerlof economic theory (pre-1970), markets were assumed to clear under standard conditions; post-Akerlof (1970-2000), the focus was on documenting market failures from asymmetry and proposing regulatory disclosures as a partial fix; post-Roth (2000s onward), market design has provided a third response — institutional engineering that extracts cooperation from informed parties without requiring full disclosure. Three regimes, three approaches to the same underlying problem.

Information asymmetry destroys some markets and provokes the invention of others — Akerlof showed the problem; Roth showed that institutional design is sometimes the solution.

→ Framework: Financial education framework

What it means for different economic actors

Retail investors face asymmetry in nearly every direction — versus market makers (who see order flow), versus insiders (who know corporate developments before disclosure), versus high-frequency traders (who see microsecond information advantages). Regulatory frameworks (insider-trading rules, public disclosure requirements, best-execution obligations) partially close these gaps.

Institutional investors face their own asymmetry stack. They have access to research and execution unavailable to retail, but face information advantages held by issuers, by larger institutions, and by intermediaries with greater flow visibility. The arms race for information has produced a documented increase in research spending across the industry without obvious returns at the aggregate level — a Red Queen effect where running faster maintains relative position.

Issuers bear costs of asymmetry through higher capital costs when they cannot credibly signal quality. Costly disclosure, third-party audit, credit ratings and analyst coverage are partial signalling mechanisms — costly in themselves, justified by the cost reduction in capital they make possible.

A common error is to assume that more disclosure always reduces asymmetry. Disclosure that exceeds users’ processing capacity can entrench rather than reduce asymmetry, since sophisticated users extract more from voluminous filings than less-sophisticated ones can. The shape of disclosure matters as much as the volume.

Practical observation

Go deeper

Questions associées

• What is the principal-agent problem in finance?
• Why is financial literacy education ineffective?
• What are the mechanical differences between ETFs and mutual funds?
• Does smart beta really outperform?

Frequently asked questions

Why doesn’t more disclosure simply solve information asymmetry?

Disclosure has well-documented limits. First, disclosed information must be processed by recipients with finite cognitive capacity; complex disclosures can advantage sophisticated readers over simpler ones. Second, disclosure is itself shaped by disclosers — what is disclosed and how it is presented carries a strategic dimension. Third, some information is genuinely private and not directly observable (intentions, future actions). The market-design literature treats these limits seriously, designing mechanisms that work even when full disclosure is infeasible.

What did Spence’s signaling theory add to Akerlof’s framework?

Akerlof showed that asymmetric information can destroy markets through adverse selection. Spence (1973) showed that informed parties can voluntarily convey information through costly actions whose cost varies systematically with quality. Education was his original example: high-ability students find education less costly (in time and effort) than low-ability students, so completing education credibly signals ability. The key requirement is that the signal’s cost be correlated with the underlying quality — costless signals provide no information. Together, Akerlof and Spence frame how markets fail and how they can be rebuilt around credible signals.

Has algorithmic trading reduced or increased financial market asymmetry?

Both, depending on the metric. Algorithmic trading has reduced bid-ask spreads measurably for liquid securities, suggesting reduced asymmetry between market makers and natural traders. It has also increased the speed advantage of co-located high-frequency systems over slower participants, suggesting that asymmetry has migrated from a quality dimension to a latency dimension. The empirical literature finds net effects that vary by market, time horizon, and event type — a pattern consistent with asymmetry being shifted rather than eliminated by technological change.

Last updated — 30 July 2026