Information asymmetry arises when one side of a transaction knows materially more than the other: the seller of a used car knows its true condition, and the founder of a venture knows its real prospects, while the buyer or backer sees only what is shown. When the informed party cannot credibly communicate its private information, the predicted result is adverse selection — uninformed parties rationally discount every offer toward the average, high-quality sellers exit because the discounted price no longer compensates them, and the market fills with low-quality goods, or lemons.

Three broad remedies have been studied empirically and theoretically: voluntary disclosure, in which sellers publish their private information; reputation mechanisms, which record past behavior and expose it to future trading partners; and information design, in which a designer deliberately chooses what information participants observe. Evidence from online marketplaces, laboratory experiments, and contest theory shows that each can mitigate the problem, though none eliminates it entirely.

Voluntary disclosure and the lemons problem

Used cars are the canonical lemons market: since Akerlof's 1970 analysis, information asymmetry between seller and buyer has been expected to cause adverse selection, which made the rapid growth of online used-car auctions — where buyers cannot inspect the vehicle at all — something of a puzzle. Studying transaction data from eBay Motors, Lewis (2011) resolves part of that puzzle through voluntary disclosure: sellers post their private information directly on the auction page in the form of detailed descriptions and photographs. Disclosure works because the auction page defines a precise implicit contract — deliver the car as shown for the closing price — which protects buyers from adverse selection. Empirically, these online disclosures are important determinants of sale prices, and disclosure costs matter: they affect both how much sellers disclose and the prices realized, so lowering the cost of credible disclosure improves market outcomes.

A natural objection is that only high-quality sellers should want to disclose. Shapiro and Huh (2021) examine the opposite case: when do low-quality sellers voluntarily reveal negative information about their own products, in a setting where the only communication channel is cheap talk — no repeated purchases, no reputation, no warranties, and no inference of quality from prices? Cheap talk limits high-quality sellers' ability to separate, since any message they send can be costlessly imitated. Two forces nevertheless push low-quality sellers toward honesty: buyers' risk and loss aversion, because disclosure removes quality uncertainty and raises willingness to pay; and competition, because disclosure differentiates the product from rivals and softens price competition. Sellers thus trade off the quality advantage of pooling with better products against the information advantage of revealing their true type. Equilibria in which low-quality sellers honestly separate exist under both monopoly and duopoly, and the paper's motivating examples — voluntary flaw disclosure on eBay and on Craigslist, which has no reputation mechanism at all — suggest such honesty occurs in practice. This inverts the usual focus of the disclosure literature, in which high-quality sellers reveal and low-quality sellers conceal.

Reputation mechanisms

Where disclosure addresses hidden information, reputation systems address hidden action — the moral hazard of a trading partner who could take payment and underdeliver. Bolton, Katok, and Ockenfels (2004) tested electronic feedback mechanisms of the kind used on eBay-like platforms in a laboratory experiment comparing three market institutions: a market with a feedback mechanism, a strangers market with no feedback, and a partners market in which the same pairs interact repeatedly. The feedback mechanism induced a substantial improvement in transaction efficiency — more trust and more trustworthy behavior — relative to the no-feedback strangers market. Reputation markets nonetheless fell short of partners markets. The reason is that reputational information has a public-goods character: the benefits of trust and trustworthiness accrue to the whole community rather than being captured by the individual who builds the reputation, so incentives are weaker than under bilateral repeated interaction. Electronic reputation mechanisms therefore only partially substitute for long-run relationships, a finding from which the authors draw design implications for improving online feedback systems.

The information-design perspective

A complementary literature treats information itself as a design variable: rather than asking how participants can credibly reveal what they know, it asks what a market or contest designer should let participants observe. Makris and Renou (2023) extend the core tool of this literature — Bayes correlated equilibrium, due to Bergemann and Morris (2016) — from static settings to multi-stage games, in which at each stage players receive private signals about past and current states, past actions, and past signals before choosing actions. They fully characterize the distributions over actions, states, and signals attainable in any (sequential) communication equilibrium of any expansion of the game — that is, under any additional information players might be given. This revelation-principle-style result bounds what any designer-chosen information structure can achieve in a dynamic strategic setting, without enumerating candidate structures one by one.

Design choices also shape how informative equilibrium behavior is about hidden types. In selection contests — grades, promotions, retention, job assignments — observed performance is a strategically manipulable signal of ability, because contestants can choose how risky their performance distribution is. Fang and Noe (2022) show that when many contestants compete for few slots, strategic contestants adopt high-risk gambling strategies, weakening the correlation between performance and true ability: competition can undermine meritocracy. Paradoxically, uncompetitive contest designs — lax standards (“lowering the bar”) and restricted candidate pools (“limiting the field”) — can make selection more meritocratic by damping risk-taking incentives, and policies that permit promoting candidates unlikely to be the very best strengthen the link between selection outcomes and genuine merit.

Relevance to Caper

A launchpad is a lemons market: founders know more about a project's quality and intentions than prospective backers, and adverse selection predicts that backers will discount every raise toward the expected average unless credible information intervenes. The remedies surveyed above map directly onto Caper's structure. Transparent on-chain treasuries lower the cost of credible disclosure by making a caper's finances observable rather than merely claimed — the condition Lewis identifies for disclosure to improve market outcomes. Founders raising funds can disclose voluntarily, and the experimental evidence on feedback systems suggests that the public record of a team's proposals and their execution can act as a reputation mechanism — substantially improving trust among strangers, while remaining an imperfect substitute for direct long-run relationships.

References

1. Gregory Lewis (2011). Asymmetric Information, Adverse Selection and Online Disclosure: The Case of eBay Motors. American Economic Review, 101(4), 1535–1546.
2. Dmitry Shapiro and David Seung Huh (2021). Incentives of Low-Quality Sellers to Disclose Negative Information. Journal of Economics & Management Strategy, 30(1), 81–99.
3. Gary E. Bolton, Elena Katok, and Axel Ockenfels (2004). How Effective Are Electronic Reputation Mechanisms? An Experimental Investigation. Management Science, 50(11), 1587–1602.
4. Miltiadis Makris and Ludovic Renou (2023). Information design in multi-stage games. Theoretical Economics, 18(4), 1475–1509.
5. Dawei Fang and Thomas Noe (2022). Less Competition, More Meritocracy? Journal of Labor Economics, 40(3), 669–701.