Two players with common interests exchange information to make a decision. Their communication is scrutinized by an observer with different interests who understands the meaning of all messages and may object to the decision. We show how the players can implement their optimal decision rule using a back and forth conversation. Such a subversive conversation reveals enough information for the players to determine their best decision, but not enough information for the observer to determine whether the decision was against his interest. Our results provide a theory of conversations based on deniability in the face of possible public outrage.

The social psychology literatures on esteem, impression management, achievement motivation, and self-handicapping have long shown that a fear of embarrassment affects behavior involving risk. We reformulate these early insights as a Bayesian model of embarrassment aversion -- or risk aversion with respect to estimated skill. Embarrassment aversion can rationalize the assumptions of classic models, and in particular implies choices consistent with the prospect theory anomalies of loss aversion, probability weighting, and framing. Loss aversion arises because losing any gamble, even a friendly bet with little or no money at stake, reflects poorly on the decision maker's skill. Probability weighting emerges because winning a gamble with a low probability of success is a strong signal of skill, while losing a gamble with a high probability of success is a strong signal of incompetence. Framing matters by altering equilibrium beliefs, such as daring someone to take a risk rather than admit to a lack of skill. Since the predictions of embarrassment aversion depend on the social and information context, the theories make diverging predictions in specific situations.

If a biased sender can distort some of the news, is it more persuasive to make relatively good news look even better, or to make relatively bad news look less bad? We show that when the news is mostly good, shoring up relatively bad news is most persuasive since it makes the good news appear more consistent and hence more credible. But when the news is mostly bad, exaggerating relatively good news is most effective since it makes the bad news appear less consistent and hence less damaging. We test for such selective news distortion by examining the consistency of reported segment earnings across different units in firms. As predicted by the model, managers appear to manipulate segment earnings to boost underperforming segments when firm earnings are above expectations and to boost overperforming segments when firm earnings are below expectations. More generally, we show how Bayesian updating leads managers and other biased senders to have “mean-variance news preferences” that differ from traditional mean-variance preferences in that more variance sometimes helps and a higher mean sometimes hurts.

High prices can credibly signal product quality if a low quality seller would rather sell a large quantity at a low price than pretend to be high quality by setting a high price. But the literature shows that stopping such mimickry requires high quality sellers to set a distortionary high price that can greatly reduce demand. We reexamine this problem of pure price signaling with the only difference that we allow the seller to sell multiple products. We show that relative prices can credibly signal relative qualities with little or none of the distortion required for a single-product firm. Such comparative price signaling solves the information problem that just ranking the different products' qualities, e.g., "good, better, best", is not in general credible by itself. We apply our results to versioning and crimping environments where the multi-product firm uses differential prices to screen buyers based on their preferences for quality, and buyers rationally use prices to infer which version is the higher quality. More generally the results show how comparative signaling on multiple dimensions can reduce signaling distortions.

Is it always wise to disclose good news? We find that the worst sender with good news has the most incentive to disclose it, so reporting good news can paradoxically make the sender look bad. If the good news is attainable by sufficiently mediocre types, or if the sender is already expected to be of a relatively high type, withholding good news is an equilibrium. Since the sender has a legitimate fear of looking too anxious to reveal good news, having a third party disclose the news, or mandating that the sender disclose the news, can help the sender. The predictions are tested by examining when economics faculty at different institutions use titles such as "Dr" and "Professor" in voicemail greetings and course syllabi.

When can you trust an expert to provide honest advice? We develop and experimentally test a recommendation game where an expert helps a decision maker choose among two actions that each benefit the expert and an outside option that does not. For instance, a salesperson recommends one of two products to a customer who may purchase nothing. Behavior is largely consistent with predictions from the cheap talk literature. For sufficient symmetry, recommendations are persuasive in that they benefit the expert by lowering the chance that the decision maker takes the outside option. If the expert is known to be biased toward either action, such as a salesperson receiving a higher commission on one product, the decision maker partially discounts a recommendation for it and is more likely to take the outside option. If the bias is uncertain, then biased experts lie even more, while unbiased experts follow a political correctness strategy of pushing the opposite action so as to be more persuasive. Even if the expert is known to be unbiased, when the decision maker already favors an action the expert panders toward it, and the decision maker partially discount the recommendation. The comparative static predictions hold with any degree of lying aversion up to pure cheap talk, and most subjects exhibit some limited lying aversion. The results highlight that transparency of expert incentives can improve communication, but need not ensure unbiased advice.

Certifiers of quality often report only coarse grades to the public despite having measured quality more finely, e.g., "Pass" or "Certified" instead of "73 out of 100". Why? We show that coarse grades result in more information being provided to the public because the coarseness encourages those of middling quality to apply for certification. Dropping exact grading in favor of the best coarse grading scheme always reduces public uncertainty because the extra participation outweighs the coarser reporting. In some circumstances, the coarsest meaningful grading scheme, pass-fail grading, is the most informative.

Sellers often make claims about product strengths without providing evidence. Even though such claims are mere puffery, we show that they can be credible because talking up any one strength comes at the implicit tradeoff of not talking up another potential strength. Puffery pulls in buyers who value product attributes that are talked up or emphasized, while pushing away other buyers who infer that the attributes they value are not the product's strengths. When the initial probability of a sale is low there are more potential buyers to pull in than to push away so puffery is persuasive. This persuasiveness requires some buyer privacy about their preferences so that the seller does not completely pander to them. More generally the results show how cheap talk by an expert to a decision maker can be credible and persuasive in standard discrete choice models used throughout marketing and economics.

When can you cheat some people without damaging your reputation among others? In a trust game between a firm and a series of individuals from two groups of different sizes, the firm has more incentive to cheat minority individuals because trade with the minority is less frequent and the long-term benefits of a reputation for fairness toward the minority are correspondingly smaller. If the majority is sufficiently large it gains nothing from a solidarity strategy of punishing opportunism against the minority, so the firm can continue doing business with the majority even if it cheats the minority. When some firms have a preference-based bias against the minority, the interaction with reputation effects gives all firms a stronger incentive to cheat the minority, and discrimination is the unique equilibrium for firms of intermediate patience.

When consumers are unsure of the exact standard that a quality certificate or label represents, they must infer the difficulty of the standard in part from observing which firms adopt the label. Key results from the certification and disclosure literatures are thereby altered. First, consumers are more suspicious of a label if a firm with a bad reputation adopts it, so certification can be less appealing to bad firms than average firms or good firms. Second, as the number of available labels increases, the informativeness of certification decreases rather than increases. Third, adoption of a label by one firm need not increase pressure on other firms to also adopt it. Instead, a label can be either "legitimitized" or "spoiled" for use by other products depending on whether a product with a favorable or unfavorable reputation is certified first. These problems are mitigated if certification is mandatory or if some standards are "focal", even if standards remain uncertain. The model is applied to eco-labels and is also revelant for nutrition labels, academic journals, club memberships, diplomas, and related environments.

We consider the credibility, persuasiveness, and informativeness of multidimensional cheap talk by an expert to a decision maker. We find that an expert with state-independent preferences can always make credible comparative statements that trade off the expert's incentive to exaggerate on each dimension. Such communication benefits the expert -- cheap talk is "persuasive" -- if her preferences are quasiconvex. Communication benefits a decision maker by allowing for a more informed decision, but strategic interactions between multiple decision makers can reverse this gain. We apply these results to topics including media bias, advertising, product recommendations, voting, and auction disclosure.

More intensive copyright enforcement reduces piracy, raises prices, and lowers consumer surplus. We show that these results do not hold regarding the extent rather than intensity of enforcement. When enforcement is targeted at high-value buyers such as corporate and government users, the copyright holder charges super-monopoly prices, thereby encouraging low-value buyers to switch to inferior pirated copies. Extending enforcement down the demand curve broadens the copyright holder’s captive market, reduces prices toward the monopoly level, and increases sales of legitimate copies. Since more extensive enforcement increases both profits and consumer surplus over some range, the standard tradeoff between the incentive to generate intellectual property and the welfare cost of monopoly power is avoided. Technologies such as digital rights management which allow for more extensive copyright enforcement can sometimes benefit consumers even apart from their effects on the incentive to generate intellectual property.

When are comparative statements credible? We show that simple complementarity conditions ensure that an expert with private information about multiple issues can credibly rank the issues for a decision maker. By restricting the expert’s ability to exaggerate, multidimensional cheap talk of this form permits communication when it would not be credible in a single dimension. The communication gains can be substantial with even a couple of dimensions, and the complete ranking is asymptotically equivalent to full revelation as the number of issues becomes large. Nevertheless, partial rankings are sometimes more credible and/or more profitable for the expert than the complete ranking. Comparative cheap talk is robust to asymmetries that are not too large. Consequently, for sufficiently many independent issues, there are always some issues sufficiently symmetric to permit comparative cheap talk. (Formerly titled "Ordinal Cheap Talk".)

Should a seller with private information sell the best or worst goods first? Considering the sequential auction of two stochastically equivalent goods, we find that the seller has an incentive to impress buyers by selling the better good first because the seller's sequencing strategy endogenously generates correlation in the values of the goods across periods. When this impression effect is strong enough, selling the better good first is the unique pure-strategy equilibrium. By credibly revealing to all buyers the seller's ranking of the goods, an equilibrium strategy of sequencing the goods reduces buyer information rents and increases expected revenues in accordance with the linkage principle.

In equilibrium play of a two-round tournament we find that underdogs exert more effort in the opening round while favorites save more effort for the final. Ability differences between players are therefore compressed in the opening round so upsets are more likely, and amplified in the final so blowouts are more likely. Measures that reduce the need to strategically allocate effort across games make for a more exciting final but a less exciting opening round. Consistent with the model, introduction of a one-day rest period between regional semi-final and final matches in the NCAA men's basketball tournament was found to increase the favorite's victory margin in the semi-finals by about five points. Non-sports applications of the model include the allocation of resources across primaries and general elections by candidates and the allocation of resources across a career ladder by managers.

When workers can bargain over wages and employment, standard insider-outsider models predict underemployment if the number of insiders is small and overemployment if the number of insiders is large. For instance, a union will restrict employment growth in an expanding firm and oppose layoffs in a contracting firm. This paper shows that employee stock ownership can solve both problems and that the necessary ownership share is often relatively small. The results are compared with related results in the literatures on profit-sharing and labor-managed firms.

Bargaining over two issues as a bundle permits credible cheap talk about their relative importance even when interests are directly opposed on each issue. The resulting communication gains can exceed the gains from bundling previously identified in the monopoly pricing literature.

In signaling environments ranging from consumption to education, high quality senders often shun the standard signals that should separate them from lower quality senders. We find that allowing for additional, noisy information on sender quality [i.e., two-sided private information] permits equilibria where medium types signal to separate themselves from low types, but high types then choose to not signal or "countersignal". High types not only save costs by relying on the additional information to stochastically separate them from low types, but countersignaling itself is a signal of confidence which separates high types from medium types. Experimental results confirm that subjects can learn to countersignal. [Unedited version with counterculture, advertising, donation behavior, fashion cycles, and other fun but insufficiently convincing examples.]

Concern for individual consumption relative to per capita consumption can induce a fear of falling behind, which increases precautionary savings. This fear increases as societal income growth increases, allowing for a positive effect of growth on savings.

Sellers benefit on average from revealing information about their goods to buyers, but the incentive to exaggerate undermines the credibility of seller statements. When multiple goods are being auctioned, we show that revealing a complete or partial ordering of the different goods by value can be credible. Ordinal cheap talk of this form is not susceptible to exaggeration because it simultaneously reveals favorable information about some goods and unfavorable information about other goods. Any informative ordering increases revenues in accordance with the linkage principle, and the complete ordering is asymptotically revenue-equivalent to full revelation as the number of goods becomes large. These results provide a new explanation in addition to bundling, complementarities, and versioning for how a seller benefits from the sale of multiple goods.

People are sometimes risk-averse in gains but risk-loving in losses. Such behavior and other anomalies underlying prospect theory arise from a model of local status maximization in which consumers compare their wealth with other consumers of similar wealth. This social explanation shares key features with the psychological explanation offered by Kahneman and Tversky.

Equity ties between businesses change the division of the firms' joint profits, thereby affecting incentives for relation-specific investments and other strategic actions. Depending on which side owns the equity and how readily the equity can be resold, we find that the changed incentives can resolve all four types of holdup-related problems: underinvestment, overinvestment, undercooperation, and sabotage. Equity stakes indirectly affect bargaining over the joint profits by making the bargaining positions of the firms dependent on each other. For instance, if one firm is made unprofitable by a breakdown in negotiations, the other firm loses as well. Since bargaining positions are linked, each firm benefits less from attempts to grab a larger share of the joint profits by strengthening its relative bargaining position, and benefits more from actions that increase joint profits. While both firms can gain from increased efficiency due to the equity stake, firms in many cases should only acquire an equity stake if they can bargain for a discounted price.

Since the early 1980s China has witnessed a rapid increase in its national savings rate to one of the highest rates in the world. Unlike the socialist period when consumption was repressed in order to redirect resources to investment, most of these savings are voluntary. This paper addresses why China's savings rates are so high and the implications for China's economy.

(Chinese Characters: A Genealogy and Dictionary)

If you take Xu Shen's etymological dictionary from 2000 years ago, put all the connections into a computer, and generate trees showing the connections, what do you find? That every part of every character can be traced back to less than 200 root characters (wen). This is not the bushou system (literally "section-heading" but often mistranslated as "radical") which only connects one part of each character with 214 characters. Rather, this new zipu system shows how every part of every character is itself a character. As a method for organizing a dictionary, it generalizes the bushou system by allowing any character to be found if the reader knows any component of the character or knows any character which shares the same component. Students can quickly locate characters while also better remembering the relations between characters.