Search

Experimental studies of social dilemmas have shown that while the existence of a sanctioning institution improves cooperation within groups, it also has a detrimental impact on group earnings in the short-run. Could the introduction of pre-play threats to punish have enough of a beneficial impact on cooperation, while not incurring the cost associated with actual punishment, so that they increase overall welfare ? We report an experiment in which players can issue non-binding threats to punish others based on their contribution levels to a public good. After observing others' actual contributions, they choose their actual punishment level. We find that threats increase the level of contributions significantly. Efficiency is improved, but only in the long run. However, the possibility of sanctioning differences between threatened and actual punishment leads to lower threats, cooperation and welfare, restoring them to levels equal to or below the levels attained in the absence of threats.

I study how an informed manager can benefit from cheap-talk communication by strategically forming diverse groups of uninformed, heterogenous workers. Optimal grouping, or partition, of workers trades off the benefit from gaining credibility through having greater diversity in each group against the cost from the inability to tailor communication among workers in the same group. Public and private communication are restrictions on the partitions the manager can adopt, and the manager prefers semi-public communication whenever there is a possibility to benefit from differently diverse groups of workers. I show that it is optimal for the manager to separate workers who need to be persuaded from those that do not, and I derive further properties of optimal partitions when workers are single minded and when workers have preferences over a spectrum of project types.

In majoritarian legislatures, minority parties cannot directly influence policymaking. However, they may serve as whistleblowers against the majority parties' harmful policymaking. This study aims to explore whether a minority party's monitoring mitigates a political agency problem in a majoritarian legislature. For this purpose, we construct a two-period election model in which voters face information asymmetries regarding each party's type and the state of the world. After the majority party's policymaking, the minority party sends a cheap-talk message about whether to oppose the policymaking. We show that the monitoring by the minority party is not useful, even if we allow the possibility that the party is the truth-telling type. As long as the initial popularity of the minority party is not too high, its opposition is informative, but does not change the electorate's voting behaviors. Consequently, it does not improve electoral accountability at all. This irrelevance of monitoring is resolved when the minority party can directly influence policymaking, for example, through obstruction.

This paper presents a simple application and experimental test of a theoretical model of discriminatory social preferences. First, it establishes conditions under which a specification of “defection aversion” can support an equilibrium in which costless messages engender cooperative strategies in a one-shot Prisoners’ Dilemma (PD) game. The main result is a signaling equilibrium with full separation of signals, and cooperation whenever the social distance is less than a certain threshold level. The second part of the paper is an experimental application and test of a discrete version of the theory. Using the minimal group paradigm, subjects are put into a set of interlocking groups. 116 subjects played 4 games each (within-subjects) under different informational (treatment) contexts. Results are broadly consistent with the theory: people cooperate more when faced with group members. They are also less “honest” with their messages when messages are sequential than when they are simultaneous, which suggests a strategic motivation to the messages. A somewhat unexpected empirical result shows that the threshold social distance supporting cooperation is lower when the group information is based on simultaneous cheap talk than when it is exogenously revealed. Noting that the empirical rates of honest messaging are also rather far from the 100% predicted by the model, a theoretical extension shows that smaller social neighborhoods are in fact a straightforward prediction when the truth of messages is uncertain.

Scientific models structure our perception of reality. This paper studies how we choose among them under expert advice. Scientific models are formalised as probability distributions over possible scenarios. An expert is assumed to know the most likely model and seeks to communicate it to a decision maker, but cannot prove it. As a result, communication about models is a cheap talk game. The decision maker is in a situation of model-uncertainty and is ambiguity sensitive. I show that information transmission depends on the strategic misalignment of players and, unlike similar models in the literature, a form of consensus among scientific models. When science is divided, there is an asymmetry in information transmission when the receiver has maxmin expected utility preferences. No information can be conveyed about models above a certain threshold. All equilibria of the game are outcome equivalent to a partitional equilibrium and the most informative one is interim Pareto dominant.

Recent politics has been characterized by politicians' harsh anti-immigration appeals and backlash against immigrants. I present a novel explanation for this backlash that hinges on politicians' ability to make such appeals credible. The starting point is a cheap talk model in which a politician (sender) is aligned with one of two opposed groups (receivers) and seeks to communicate her preferences to win support. Importantly, an increase in the weaker group's capacity may enable credible communication by the opposed type of politician, ironically making the weaker group worse-off. Illustrating the model, I discuss how Donald Trump credibly communicated alignment with anti-immigration groups in 2016 through harsh messaging against immigrants, whose power was increasing. More broadly, the model and case show how the behavior of strategic actors can underpin realignments, with shifts in relative group power proving crucial in enabling politicians to assemble novel political coalitions.

In a model with cheap talk, employers can send messages about their willingness to pay for higher ability workers, which job-seekers can use to direct their search and tailor their wage bid. Introducing such messages leads—under certain conditions—to an informative separating equilibrium which affects the number of applications, types of applications, and wage bids across firms. This model is used to interpret an experiment conducted in a large online labor market: employers were given the opportunity to state their relative willingness to pay for more experienced workers, and workers can easily condition their search on this information. Preferences were collected for all employers, but only treated employers had their signal revealed to job-seekers. In response to revelation of the cheap talk signal, job-seekers targeted their applications to employers of the right “type” and they tailored their wage bids, affecting who was matched to whom and at what wage. The treatment increased measures of match quality through better sorting, illustrating the power of cheap talk to improve market outcomes. Institutional subscribers to the NBER working paper series, and residents of developing countries may download this paper without additional charge at www.nber.org.

This paper addresses two related questions: How can we capture different degrees of sharing a language with pre-specified meanings? And, how do we expect such a language to be used?

Disclosure of climate-related financial risks greatly helps investors assess companies' preparedness for climate change. Voluntary disclosures such as those based on the recommendations of the Task Force for Climate-related Financial Disclosures (TCFD) are being hailed as an effective measure for better climate risk management. We ask whether this expectation is justified. We do so with the help of a deep neural language model, which we christen ClimateBert. We train ClimateBert on thousands of sentences related to climate-risk disclosures aligned with the TCFD recommendations. In analyzing the disclosures of TCFD-supporting firms, ClimateBert comes to the sobering conclusion that the firms' TCFD support is mostly cheap talk and that firms cherry-pick to report primarily non-material climate risk information. From our analysis, we conclude that the only way out of this dilemma is to turn voluntary reporting into regulatory disclosures.

The current study investigates the usage of virtuous language in the management discussion and analysis (MD&A) section of SEC filings (10-K Form) and the prognostic power of such language for takeover performance. The sample under study, obtained from Bloomberg, is comprised of a large number of M&A deals from the United States that took place between January 2000 and October 2016. The empirical results, based on textual analysis, reveal that trust is negatively associated with long-term takeover performance, suggesting that managerial virtuous talk is, in practicality, an indication of lower post-acquisition gains for the acquirers in the long run. Furthermore, takeover returns are found to reflect textual information on trust with a delay, owing to general inattention and inability of investors to process soft cues inherent in textual content and to managers purposefully lulling investors to keep them from paying attention and identifying managerial misconduct. Quite interestingly, the significance of virtuous talk becomes more evident in the post-crisis period due to relatively higher uncertainty linked with evaluating such kind of deals on the basis of hard information alone. Finally, an inflated virtuous talk when coupled with pessimistic tone, the ability of managerial ‘good talk’ to create a trustworthy image and to distract investors reduces and the predictive power of managerial trust talk increases even more. Overall, it is concluded that managerial virtuous talk should not be regarded as a ‘cheap talk’. It is, in fact, very pertinent for predicting future takeover returns in the long run.

We analyze a model of political competition, in which the uninformed median voter chooses whether to follow or ignore the advice of the elite that forms endogenously to aggregate information. In equilibrium, information transmission is possible only if voters trust the elite's endorsement of potentially biased candidates. When inequality is high, the elite's informational advantage is minimized by the voters' distrust. When inequality reaches a certain threshold, the trust, and thus the information transmission, breaks down completely. Finally, the elite size and thus the extent of information aggregation depends on the amount of trust they can maintain.

We present a version of Baliga and Sjostrom’s (2012a) conflict games with two asymmetric players. The players contemplate whether to take an active engagement action to compel the leader of a neighboring state (an extremist) to give up his risky weapons. We show that a player with greater damage from the extremist is more likely to choose an active engagement action than a player with lesser damage. Furthermore, we examine cheap-talk communication equilibria with the extremist. The likelihood of both players choosing the active engagement action decreases by a hawkish extremist who can send a provocative message, if both players are coordination types. If both players are opportunistic types, a dovish extremist can send an appeasement message that causes one player to be more active while another to be more inactive. Lastly, we show that there does not exist any other communication equilibrium for either kind of extremist, for any other combination of player types.

Governments often make public announcements that call into question oligopolistic behavior. Yet little is known about how firms respond to them. We study gasoline retailers' price responses to antitrust announcements shaming them for price gouging during the COVID-19 pandemic. Price effects are identified through a high-frequency event-study that leverages unique real-time station level price data and well-identified, discrete antitrust announcements. We find evidence of announcement effects that depend on firms' pre-announcement margins and hence exposure to being publicly shamed. Our results provide evidence against a "cheap talk" interpretation of government antitrust announcements.

We show that experience good sellers facing myopic buyers can solve the inherent moral hazard problem by communicating their observation of quality before trade, provided that communication is part of their public track record. Such cheap-talk communication, if trusted, allows market prices to reflect the actual value created, thus providing an immediate reward for the seller's effort which complements the conventional, reputational incentives. Pre-trade communication achieves maximal efficiency when truthful and the full efficiency as the noise in the seller's observation vanishes. We fully characterize the conditions for communication to improve efficiency and the extent to which it does so.

Cheap-talk communication between parties with conflicting interests is common in many business and economic settings. Two distinct behavioral economics theories, the trust-embedded model and the level-k model, have emerged to explain how cheap talk works between human decision makers. The trust-embedded model considers that decision makers are motivated by nonpecuniary motives to be trusting and trustworthy. In contrast, the level-k model considers that decision makers are purely self-interested but limited in their ability to think strategically. Although both theories have been successful in explaining cheap-talk behaviors in separate contexts, they point to contrasting drivers for human behaviors. In this paper, we provide the first direct comparison of both theories within the same context. We show that, in a cheap-talk setting that well represents many practical situations, the two models make characteristically distinct and empirically distinguishable predictions. We leverage past experiment data from this setting to determine what aspects of cheap-talk behavior each model captures well and which model (or combination of models) has better explanatory power and predictive performance. We find that the trust-embedded model emerges as the dominant explanation. Our results, thus, highlight the importance of investing in systems and processes to foster trusting and trustworthy relationships in order to facilitate more effective cheap-talk interactions. This paper was accepted by Charles Corbett, operations management.

In procurement auctions, suppliers' senior managers in charge of bidding often rely on advisors, such as project managers and technical specialists, to evaluate project costs. These advisors may be biased for inflating the costs due to agency or career concerns. We study the procurement agency's choice between sealed-bid static auctions and a dynamic Dutch auction in this setting. If the number of competing suppliers is small, static auctions result in lower expected costs than the Dutch auction. This ranking is reversed if procurement is sufficiently competitive. The Dutch auction is always more efficient regardless of the number of suppliers.

We investigate a novel continuous-time mechanism in a public-goods game. Within a fixed period, a clock ensures that contributions increase simultaneously for every player, and players can choose when to stop, while their actions are observed by others in real time. We show both theoretically and experimentally that this mechanism is very effective in improving the contribution. Three critical factors may play a role: cheap talk, incremental commitment, and the clock. We further decompose these factors and find that cheap talk alone is not effective, while incremental commitment and the clock each account for 60% and 40% of the total effect, respectively.

The debate over how firm stakeholder engagement is tied to preserving shareholder wealth has received growing attention in recent years, especially in the wake of the COVID-19 crisis. Against this backdrop, we examine the relation between corporate social responsibility (CSR) and stock market returns during the COVID-19 pandemic-induced market crash and the post-crash recovery. Using a sample of 1,750 U.S. firms and two major sources of CSR ratings, we find no evidence that CSR affected stock returns during the crash period. This result is robust to various sensitivity tests. In additional cross-sectional analysis, we find some supporting evidence, albeit weak, that the relation between CSR and stock returns during the pandemic-related crisis is more positive when CSR is congruent with a firm’s institutional environment. We also find that Business Roundtable companies, which committed to protecting stakeholder interests prior to the pandemic, do not outperform during the pandemic crisis. We conclude that pre-crisis CSR is not effective at shielding shareholder wealth from the adverse effects of a crisis, suggesting a potential disconnect between firms’ CSR orientation (ratings) and actual actions. Our evidence suggests that investors can distinguish between genuine CSR and firms engaging in cheap talk.

We experimentally investigated whether feedback received by participants improves the effort level in a minimum effort game involving continuous-time cheap talk. Continuous-time cheap talk allows participants to freely change and constantly monitor other group members’ messages. We used two treatments: Full Feedback, in which participants are informed of the distribution of effort levels in the group, and Limited Feedback, in which participants are informed of only the minimum effort level in the group. Our main findings are as follows. The average and minimum effort level in the Full Feedback treatment are not significantly different from those in the Limited Feedback treatment. In addition, participants who choose the lowest effort level are not likely to increase the effort level in the next round, despite receiving feedback.

A central benefit of manufacturers’ sales reps is their ability to provide sellers with information about prospective buyers. Yet sales reps often selectively under-report this information, which is viewed as a major problem by sellers and academics alike. We feel this is only part of the story. Beyond their economic roles of passing along information, sales reps also manage social tensions between sellers and buyers that can threaten exchange. We show selective under-reporting can be beneficial to facilitate exchanges that would be impossible with fully truthful reports. We develop a game-theoretic model in a cheap-talk setting that adds the sales reps’ social role, and show when and which piece(s) of information sales reps will under-report to induce buyers and sellers to agree to a deal. We show this makes all three parties better off, generating a Pareto-improving “Selective Under-reporting” equilibrium. Our work suggests that half-truths aren’t always half-bad and can serve as an essential social lubricant for exchange. This broadens our understanding of the complexity of sales reps’ work and sales managers’ toolkit for managing under-reporting, and suggests new understandings of third parties’ roles in a world increasingly focused on information and big data.

In procurement auctions, bidders are usually better informed about technical, financial, or legal aspects of the goods and services procured. Therefore, the buyer may include a dialogue in the procurement procedure which enables the suppliers to reveal information that will help the buyer to better specify the terms of the contract. This paper addresses the question of the value added of letting the sourcing process consist of both an auction and a negotiation stage, theoretically and in a laboratory experiment. Our theoretical results suggest that in a setting where the buyer and the suppliers have aligned interests regarding the terms of the contract, allowing the winning supplier to communicate with the buyer after the auction is beneficial to the buyer compared to no communication and ex-ante communication. In a setting where the buyer and the winning supplier have misaligned interests regarding the terms, the buyer benefits from ex-ante communication relative to no communication and ex-post communication. Our experimental data provide strong evidence for the predictions in the aligned-interest setting. In the misaligned-interest setting, we do not observe significant differences between the three mechanisms. Our experimental findings offer several managerial implications for the appropriate design of sourcing processes.

Venture capital (VC) investors have been criticized for lax monitoring and for be- ing too “founder-friendly.” We identify an overlooked benefit of such behavior: entrepreneurs lie less to friendly VCs. The entrepreneur is privately informed about project success, enjoys private benefits of control, and recommends a project to the VC. The VC chooses the project and can intervene in the interim. The equilibrium features a “monitoring trap”: possible intervention leads the entrepreneur to lie, which prompts further intervention. However, both are better off if the VC commits to intervene less. We characterize implications for information acquisition, control rights, outside directors, and stage financing.

To truly create new value, you need a theory. A theory helps you see what others can’t see. Contrarian or unique beliefs provide the underlying raw material of a firm’s theory of value. But beliefs are just rhetoric and cheap talk, unless you identify the specific obstacles and problems that stand in the way of realizing your contrarian beliefs. This puts problem-formulation and problem-solving at the very heart of your theory of value. Identifying the right problems allows you to spot valuable solutions, to find asset bargains and to make the right investments. Importantly, a theory allows you to be more scientific about value creation and to perform the right type of experiments. A theory allows startups and businesses to learn the right lessons from their experimentation and, if needed, to pivot accordingly. Overall, in the context of strategy, you can’t rely on generic approaches or best practices. You need to develop your own theory. It will allow you to see the world differently, to out-think competitors and to create value in unique ways.

Decision makers are often accessible via a variety of communication channels (e.g. direct vs. intermediated, formal vs. informal, in-person vs. electronic, etc.), and the channel selected by an informed party can signal information about the sender's preferences. This paper uses a model of strategic communication to study how and why partisan and nonpartisan communication channels arise in organizations. Polarized organizations use only partisan channels and information transmission occurs primarily via cheap talk (channel selection). On the other hand, discussion equilibria include partisan channels for moderates, while those with important information instead use nonpartisan or discussion channels. Welfare analysis provides applications to various communication practices observed within organizations, particularly the near-ubiquitous use of both informal (e.g. gossip and rumors), and formal (e.g. business meetings or seminars) channels of communication in organizations and communities.

Is tough-on-China campaign rhetoric cheap talk or a signal of policy attention? Analyzing China-related campaign advertisements during the 2010 midterm elections and subsequent cosponsorship of China-related bills, we find that campaign ads are a noisy predictor of legislative attention. Challengers who attacked on China were more likely to cosponsor China-related legislation, while incumbents who were attacked for being soft on China took tougher positions on China after reelection. By demonstrating the correspondence between anti-China campaign appeals and subsequent legislative attention, our findings add to a growing body of evidence linking campaign rhetoric to members’ legislative agendas. This research note provides the first evidence demonstrating the connection between campaign appeals and legislative attention on a foreign policy issue like China.

This paper analyzes the welfare impact of a cap on commissions paid by product providers to intermediaries who advise consumers. In contrast to the extant literature, with a downward sloping demand capped commissions have a direct impact on product providers' margins and consumers' prices. I show that a general ban is not welfare optimal as higher commissions do not necessarily lead to higher consumer prices. Starting from a general ban, allowing (marginally) higher commissions leads to lower prices as positive commissions make intermediaries wary to recommend more expensive products to consumers, thus making demand more elastic with respect to price.

An uninformed sender chooses a publicly observable experiment and sends a message to a receiver after privately learning the experimental outcome. To design the optimal experiment, the sender faces a tension between acquiring more information and alleviating the conflict of interest. In the benchmark model, the optimal experiment generates a conclusive signal (conclusive good news) about the state in which the two parties’ interests coincide. When the choice of experiment is not publicly observable and the sender cannot commit to it, an informative equilibrium exists if and only if there exists an equilibrium where the sender chooses to become perfectly informed.

This paper studies when introducing verifiable communication choices between agents in a cheap-talk benchmark setting, with social tie, is beneficial to welfare. In our model agents have two ways to communicate their private information: either through a costly verifiable information (hard) link or through a low-cost cheap talk (soft) link. We identify that the appearance of hard links in the pure cheap talk setting has two opposing effects on welfare: (i) a positive effect stems from the information improvement and (ii) a negative effect arises from crowding out soft communication with costly verifiable communication. Surprisingly, the final welfare outcome of the two opposing forces depends on the cost structure. If only one party bears the cost of a hard link, then the positive (informational) effect always dominates the negative (crowding out) effect, and thus introducing hard links is beneficial to welfare. In contrast, if the cost of a hard link is shared by both parties, then allowing for verifiable communication can be detrimental to welfare. We also derive several testable implications about introducing hard links in corporate governance, and demonstrate the robustness of our findings in face of heterogenous costs, general signal structures, as well as the case where cost is endogenizied via negotiation about how to split the costs.

We consider the problem of a decision-maker who seeks for advice from reputation-concerned experts. The experts have herding incentives when their prior belief about the state of the world is sufficiently strong. We address the following question: Should experts be allowed to exchange their information before providing advice ("collective expertise") or not ("independent expertise")? Allowing for such information exchange modifies the herding incentives in a non-trivial way. The effect is beneficial for the quality of advice when there is low prior uncertainty about the state and detrimental in the opposite case. We also show that independent expertise is more likely to be optimal when the decision-maker has a valuable enough "safe" option with a state-independent payoff. Finally, collective expertise is more likely to be optimal as the number of experts grows.

I study a platform-mediated two-sided matching with incomplete information. Platform users voluntarily disclose their taste information to the platform, and the platform aggregates and redistributes the collected information back to each user before they form matches. I characterize conditions under which the platform can (or cannot) achieve an optimal matching outcome by the exchange of information. It is shown that a simple rule of recommendation can induce full disclosure of user information and an optimal matching under a certain condition. I also study an optimal customized pricing schedule and show that full extraction of user surplus is possible. Lastly, as an application, I study a two-way communication protocol by cheap-talk and demonstrate that communication strictly improves efficiency in any circumstances.

Better-informed individuals are often unable to ignore their private information when forecasting others’ beliefs. We study how this bias, known as “the curse of knowledge,” affects communication and investment within a firm. A principal utilizes an informed manager’s recommendations when investing. The curse of knowledge leads the manager to over-estimate his ability to convey information, which hampers communication and decreases firm value. However, this same misperception increases the manager’s information acquisition and can increase value when endogenous information is indispensable (e.g., R&D). Finally, we characterize settings where the principal delegates the investment decision only if the manager is cursed.

An expert advises a decision maker over time. With both the quality of advice and the extent to which it is followed remaining private, the players have limited information with which to discipline each other. Even so, communication in and of itself facilitates cooperation, the relationship evolving based on the expert’s advice. We show a formal equivalence between our setting and one of cheap talk with capped money burning. This enables an exact characterization (at fixed discounting) of the expert’s attainable payoffs. While an ongoing relationship often helps, our characterization implies that relational incentives alone can never restore commitment power.

Does public cheap talk by a biased expert benefit voters? The answer depends on the nature of democratic institutions and the extent of communication possibilities. Expert endorsements induce office-seeking parties to serve the expert's interests, hurting voters. Expert advocacy makes policies respond to information, helping voters. Together, policy advocacy and partisan endorsements are often better than either alone. Their interaction creates a delegation benefit of indirect democracy. Voters may prefer this institution to one where policymaking is geared to serving the public interest. Direct expert capture of one party is another form of delegation and the best institution for voters.

I present a cheap talk model of information leaks in takeovers. Takeover targets strategically leak information about the value of synergies to attract an additional bidder. Targets with high synergies leak exaggerated information, cause a runup, and are sold in an auction. Targets with low synergies do not leak and are sold to a single acquirer. Traditional takeover defenses backfire, by increasing the likelihood that the target is sold to a single acquirer, while toeholds increase the acquirer's revenue by discouraging leaks. More stringent enforcement reduces the likelihood of leaks, reduces runups, and increases efficiency in the takeover market.

When agents with private information compete for resources from a principal and are biased towards their own favored projects (e.g., a CEO decides which division manager’s project to fund) an agency problem arises. However, possible future interaction can mitigate this problem even without reputational concerns, since an agent who induces acceptance of a low-valued project today consumes resources that crowd out even better opportunities that may arrive in the future. Using the theoretical model from Schmidbauer (Games & Economic Behavior, 2017), we devise experiments to address this organizational environment. Specifically, we study the incentives of competing agents to strategically communicate about their own favored project’s value (high or low) to an uninformed decision maker when new projects arrive over time. After observing all advice from agents, the decision maker decides which project to adopt, if any. If no project is adopted subjects enter the next period with new independently drawn projects and continue indefinitely until one project is adopted. We hypothesize that truth telling is easier to support the larger is the benefit from a high-quality project or the less likely it is to occur, but harder to support as agent competition grows. Our findings are broadly consistent with these hypotheses.

CEO activism refers to corporate leaders taking a public stand on issues such as race relations, gender equality or climate change not directly related to their business. In this paper, we investigate under what conditions CEO activism can create firm value. We develop a model where consumers care about the type of firm they buy from. Taking a stand raises two issues. First, not all consumers share the same viewpoint, so while some may be more eager to buy from a firm with an activist CEO, others may be put off. Second, consumers may discount CEO communications if they perceive them to be profit motivated. We show that credibility requires CEO communications to be public and sufficiently controversial. CEO activism is more likely to create firm value when competition is strong, consumers care a lot about "symbolic" value and polarization is high. CEO activism is associated with niche product market strategies and high prices, while "strategic ambiguity" (not taking a stand) is associated with mass market strategies and low prices. The model sheds light on the costs and benefits of intrinsically motivated CEOs and the limits of corporate governance.

The paper studies incentives of low-quality sellers to disclose negative information about their product. We develop a model where one’s quality can be communicated via cheap-talk messages only. This setting limits ability of high-quality sellers to separate as any communication strategy they pursue can be costlessly imitated by low-quality sellers. Two factors that can incentivize low-quality sellers to communicate their quality are buyers’ risk-attitude and competition. Quality disclosure reduces buyers’ risk thereby increasing their willingness to pay. It also introduces product differentiation softening the competition. We show that equilibria where low-quality sellers separate exist under monopoly and duopoly. Even though low-quality sellers can costlessly imitate high-quality sellers, equilibria where high-quality sellers separate can also exist but under duopoly only.

Political elites increasingly express interest in evidence-based policymaking, but transparent research collaborations necessary to generate relevant evidence pose political risks, including the discovery of sub-par performance and misconduct. If aversion to collaboration is non-random, collaborations may produce evidence that fails to generalize. We assess selection into research collaborations in the critical policy arena of policing by sending sincere requests to discuss research partnerships to roughly 3,000 law enforcement agencies in 48 states. A host of agency and jurisdiction attributes fail to predict affirmative responses to generic requests, alleviating concerns over generalizability. However, across two experiments, mentions of agency performance in our correspondence depressed affirmative responses --- even among top-performing agencies --- by roughly eight percentage points. Many agencies that indicate interest in transparent, evidence-based policymaking are likely engaging in cheap talk, and recoil once performance evaluations are made salient. These dynamics can inhibit valuable policy experimentation in many communities.

Should a peer-to-peer platform set prices for the products on the platform, or should it let sellers set their own prices while providing price recommendations? Centralized prices allow a platform to use demand information it observes, while price recommendations allow for competition in which sellers set prices based on their private information. On sharing economy platforms, for example, we observe a myriad of such pricing regimes. We investigate the implications of each pricing regime for the profits of platforms, buyers and sellers. When a platform recommends prices, it effectively plays the role of a sender in a multi-receiver cheap-talk game. Platforms are not always better off by centralizing pricing. When the variance of aggregate demand is large, price recommendations can be sustained in equilibrium and are often more profitable for the platform. Otherwise, a price recommendation is not credible. High (low) quality sellers have a stronger (weaker) preference for centralized pricing than the platform. Buyers, in contrast, receive lower surplus when the platform provides price recommendations, and prefer centralized pricing or competition without price recommendations. The results provide tools for platform designers and policy makers to assess the impact of different pricing regimes in markets with platforms. Although price recommendations might seem to encourage lower prices among sellers through increased competition, this is not always the case.

I study incentives for information transmission and collective decision-making between multiple players in an alliance. Within an alliance, i) players have private information; ii) actions exhibit substitutability; iii) action sets are constrained; and iv) preferences are heterogeneous. The main result derives conditions for full information aggregation under any public communication protocol. Full information revelation ensues as long as players' biases are sufficiently cohesive with respect to constraints on the action set. I derive precise conditions on the minimal action set under full information revelation. Differences in the size of action sets across players is exacerbated by two variables: i) individual biases over outcomes, and ii) degree of interdependence between the actions of players. The results are discussed in the context of collective decision-making in international organizations and climate change alliances.

We investigate models of cheap talk, information disclosure, and Bayesian persuasion, in a unified experimental framework. Our umbrella design permits the analysis of models that share the same structure regarding preferences and information, but differ in two dimensions: the rules governing communication, which determine whether information is verifiable; and the sender’s commitment power, which determines the extent to which she can commit to her communication strategy. Commitment is predicted to have contrasting effects on information transmission, depending on whether information is verifiable. Our design exploits these variations to explicitly test for the role of rules and commitment in communication. Our experiments provide general support for the strategic rationale behind the role of commitment and, more specifically, for the Bayesian persuasion model of Kamenica and Gentzkow (2011). At the same time, we document significant quantitative deviations. Most notably, we find that rules matter in ways that are entirely unpredicted by the theory, suggesting a novel policy role for information verifiability.

In many areas where strategic communication models are applied (e.g., lobbying by interest groups), the relevant uncertainty about the sender is not over the direction of their bias, but their priority over different issues. Within a model of multidimensional cheap talk, if the salience of different issues to the sender is unknown by the receiver, the expert may use the dimensions that are of less importance (to them) in order to achieve further manipulation of the policy-maker. The sender will do this unless he is sufficiently unbiased such that he has little conflict of interest with the receiver. This is inefficient from the policy-maker’s perspective, as she will be taking a loss (relative to babbling) from the dimensions on which she is receiving “false” information. Moreover, even with two senders, one is not able to recreate the reporting mechanism of Battaglini (2002), and instead potentially ends up with less information transmission than with a single sender. I propose a novel linear tax on lobbying along different dimensions in order to align the incentives of the receiver and sender when it comes to information acquisition and cross-dimensional manipulation.

We consider the trade-off between imperfect control and communication in organizations. A principal anticipates receiving private information and hires an agent to take an action for her. She has the ability to contractually tie the agent's action to the state, but this control is incomplete. States not covered by a contract induce a communication game. Close alignment of interests favors communicating and, thus, ceding authority to the agent, and vice versa. Contracting increases the number of actions that can be induced through communication. Optimal contracts that do not cover all states both substitute for and facilitate communication.

We present a new framework for analyzing defamation liability that serves both to clarify and complicate understandings of the law’s consequences for speakers, victims, and the marketplace of ideas. In addition to the familiar deterrence and chilling effects, we show how defamation liability can generate a “warming effect,” making statements more credible and potentially raising both the quality and quantity of speech. We also explain how a more plaintiff-friendly liability regime may exacerbate harms to defamation victims. We end by considering the possibility of “self-tailored” defamation law, with victims or speakers selecting the defamation liability regime that applies to them.

Entrepreneurs organize into a variety of collective arrangements: partnerships, cooperatives, franchises, joint ventures, and so on. We study the efficiency and viability of these collective organizations. This is done by analyzing communication and the allocation of decision making among members with opportunity costs and knowledgeability differences. Junior members are defined as having a lower value of their outside opportunities than Senior members. Consistent with the previous literature, we find that members' heterogeneity implies costs by making communication coarse and breeding conflict in decision making. However, heterogeneity has also a bright side since heterogeneous partners benefit from learning from each other. We demonstrate that the costs of heterogeneity are allocated by the governance structure, while the benefits of learning are determined by the market. Governance structure and market are, therefore, intertwined. Viable collective entrepreneurship requires that the probability of learning of the non-deciding member is sufficiently low in order to have the benefits of learning outweigh the costs of heterogeneity. One of the implications is that Junior partnerships do not emerge in equilibrium when the probability of learning is lower for Junior than Senior members. More generally, the viability of collective entrepreneurship is problematic in many settings. The identity of the unique, efficient decision maker is characterized by her preferences being congruent with the market and a sufficiently high probability of knowledgeability. Finally, repeated interactions improve viability, which is due to the equilibrium messages being less coarse and the richer strategies in a repeated game.

We consider a large decentralized freelance platform where buyers with private information about their quality preferences are matched with freelancers that differ in quality. When posting their job requests, buyers can report their quality preferences via cheap talk, which influences freelancers' application and pricing strategies. By exaggerating one's quality preference, a buyer attracts not only more applications from freelancers, but also those with higher quality, at the cost of a higher expected price. We find that it is always an equilibrium for the buyers to report their quality preferences truthfully when they cannot renegotiate with freelancers on their asking prices after getting matched. On the other hand, when post-match renegotiation is allowed and buyers have relatively high bargaining power, low-type buyers may strategically exaggerate their quality preferences, and subsequently after getting matched, costly signal their true type and bargain for lower prices. From a platform design perspective, our analysis implies that the option of renegotiation, designed to facilitate post-match information transmission, may backfire by giving rise to buyers' pre-match opportunistic behaviors of information distortion.

Experiments typically rely on small payments to incentivize participants. This works if participants view these payments as fungible with their own money, but if participants view the payments as a windfall, they may behave differently in experiments than in real life. We modify standard risky choice protocols by making participants earn their money at risk by completing manual tasks such as peeling potatoes. This leads to less risk-taking and to choices more consistent with those online survey respondents anticipate making with their own money. When realistic levels of risk aversion are important, experiments should require participants to earn their stakes.

Cheap-talk lies are at risk of being detected. We investigate the implications of this fact in a communication game in which players have no common interests and messages are cheap, but deceit is detectable with positive probability. In particular, when the sender makes a claim that is not true, with positive probability the receiver learns that the claim is a lie. We show that partially informative equilibria exist because lie detection generates an endogenous cost from lying consisting of being confused with lower types in equilibrium. In any informative equilibrium, the moderate and the highest types tell the truth, while the lowest types lie claiming to be the highest types. We show that more information may be revealed if the sender is given an opportunity to prepare a lie in advance and thereby decrease its detectability. We analyze an extension in which the sender may make multiple attempts at convincing the receiver, and show that if lie detectability is high, the receiver may benefit from committing to listening to the sender only once. And finally, we analyze a two-sender version of the model, and show that senders will exaggerate their claims only if the state disadvantages them sufficiently.

We study group competition with a single public good prize, perfectly discriminating contest success function, and the weakest-link effort technology, in which the marginal cost of effort for each player is private information. We focus on pure strategy Bayes-Nash equilibria and show that teammates always employ symmetric strategies. Various degrees of coordination are possible, ranging from all cost types coordinating on a single effort level to every cost type choosing a distinct effort level. Such coordination may not enhance welfare. If groups are symmetric except for group size, players in the smaller group bid more aggressively than those in the larger group, but when the asymmetries are along multiple dimensions, no clear-cut conclusions can be made with respect to the effects of group sizes and valuations. As an additional avenue for cooperation, we investigate incentives to share private information via cheap talk among teammates, who then coordinate on the effort level most preferred by the player with the largest cost. A single group sharing information does better. But when players within-group cooperate in this fashion, all within-group gains are lost to increased competition between groups.

How should a manager optimally choose transfers to incentivize multiple agents both to collect and to share costly information? To answer this question we study a simple model with a principal and two agents. The agents can obtain costly signals and communicate with each another via non-verifiable messages (cheap talk). A principal offers a contract which is separable in the performances of the agents. We characterize the optimal transfers and show a surprising result that for sufficiently correlated information and not too high costs of information acquisition an agent's optimal transfer should depend mainly on the performance of the other agent.