Cheap talk with multiple audiences: An experimental analysis

We examine strategic information transmission in a controlled laboratory experiment of a cheap talk game with one sender and multiple receivers. We study the change in equilibrium behavior from the addition of another audience as well as from varying the degree of conflict between the senderʼs and receiversʼ preferences. We find that, as in cheap talk games with just one receiver, information transmission is higher in games with a separating equilibrium, than in games with only a babbling equilibrium. More interestingly, we find clear evidence that the addition of another audience alters the communication between the sender and the receiver in a way consistent with the theoretical predictions. There is evidence of the presence of agents that are systematically truthful as senders and trusting as receivers: deviations from the theoretical predictions, however, tend to disappear with experience, and learning is faster precisely in the games where deviations are more pronounced.     

Perturbed communication games with honest senders and naive receivers

This paper studies communication games in which the sender is possibly honest (tells the truth) and the receiver is possibly naive (follows messages as if truthful). The characterization of message-monotone equilibria in the perturbed games explain several important aspects of strategic communication including sender exaggeration, receiver skepticism and message clustering. Surprisingly, the strategic receiver may respond to more aggressive claims with more moderate actions. In the limit as the probabilities of the non-strategic players approach zero, (i) the limit equilibrium corresponds to a most-informative equilibrium of the limit (Crawford-Sobel) game; (ii) only the top messages are sent.     

Truth,Trust, and Sanctions: On Institutional Selection in Sender–Receiver Games*

We have conducted a laboratory experiment to investigate the impact of institutions and institutional choice on truth‐telling and trust in sender–receiver games. We find that in an institution with sanctioning opportunities, receivers sanction predominantly after having trusted lies. Individuals who sanction are responsible for truth‐telling beyond standard equilibrium predictions, and they are more likely to choose the sanctioning institution. Sanctioning and non‐sanctioning institutions coexist if their choice is endogenous, and the former shows a higher level of truth‐telling but lower material payoffs. Our experimental findings are consistent with logit agent quantal response equilibrium with two distinct groups of individuals: one consisting of subjects who experience non‐monetary lying costs as senders and non‐monetary costs when being lied to as receivers, and the other consisting of payoff maximizers.     

Harmful lie aversion and lie discovery in noisy expert advice games

This study tests whether individuals are reluctant to tell lies, or perhaps only “harmful lies”, in a previously untested environment: an expert sending a message to a decision maker whose interpretation of that message is subject to error, i.e. a noisy sender–receiver game. In the Aligned treatment, the expert can send a “white lie” to the receiver, eliminating the negative effects of noise and improving both parties’ payoffs. In the Conflict treatment, lies are harmful and the inability to commit to truthtelling destroys all meaningful communication in equilibrium unless there is a cost of lying. In the experiment, receivers are overly trusting and experts learn to take advantage of this. As experts gain experience they tell stronger and more frequent lies in both treatments, consistent with models of reinforcement learning. The findings suggest that neither harmful nor universal lie aversion is a factor when communication is noisy, provided individuals have time to discover their personal benefits of lies.     

Why do people tell the truth? Experimental evidence for pure lie aversion

A recent experimental literature shows that truth-telling is not always motivated by pecuniary motives, and several alternative motivations have been proposed. However, their relative importance in any given context is still not totally clear. This paper investigates the relevance of pure lie aversion, that is, a dislike for lies independent of their consequences. We propose a very simple design where other motives considered in the literature predict zero truth-telling, whereas pure lie aversion predicts a non-zero rate. Thus we interpret the finding that more than a third of the subjects tell the truth as evidence for pure lie aversion. Our design also prevents confounds with another motivation (a desire to act as others expect us to act) not frequently considered but consistent with much existing evidence. We also observe that subjects who tell the truth are more likely to believe that others will tell the truth as well.     

Coordination with communication under oath

We focus on the design of an institutional device aimed to foster coordination through communication. We explore whether the social psychology theory of commitment, implemented via a truth-telling oath, can reduce coordination failure. Using a classic coordination game, we ask all players to sign voluntarily a truth-telling oath before playing the game with cheap talk communication. Three results emerge with commitment under oath: (1) coordination increased by nearly 50%; (2) senders’ messages were significantly more truthful and actions more efficient, and (3) receivers’ trust of messages increased.     

Words, Deeds, and Lies: Strategic Behaviour in Games with Multiple Signals

We report the results of an experiment in which subjects play games against changing opponents. In one treatment, "senders" send "receivers" messages indicating intended actions in that round, and receivers observe senders' previous-round actions (when matched with another receiver). In another treatment, the receiver additionally observes the sender's previous-round message to the previous opponent, enabling him to determine whether the sender had lied. We find that allowing multiple signals leads to better outcomes when signals are aligned (all pointing to the same action), but worse outcomes when signals are crossed . Also, senders' signals tend to be truthful, though the degree of truthfulness depends on the game and treatment, and receivers' behaviour combines elements of pay-off maximization and reciprocity.     

Playing the trust game with other people’s money

We experimentally investigate to what extent people trust and honor trust when they are playing with other people’s money (OPM). We adopt the well-known trust game by Berg et al. (in Games Econ. Behav. 10:122–142, 1995), with the difference that the trustor (sender) who sends money to the trustee (receiver) does this on behalf of a third party. We find that senders who make decisions on behalf of others do not behave significantly different from senders in our baseline trust game who manage their own money. But receivers return significantly less money when senders send a third party’s money. As a result, trust is only profitable in the baseline trust game, but not in the OPM treatment. The treatment effect among the receivers is gender specific. Women return significantly less money in OPM than in baseline, while there is no such treatment effect among men. Moreover, women return significantly less than men in the OPM treatment.     

Strategy-proofness in experimental matching markets

We introduce two novel matching mechanisms, Reverse Top Trading Cycles (RTTC) and Reverse Deferred Acceptance (RDA), with the purpose of challenging the idea that the theoretical property of strategy-proofness induces high rates of truth-telling in economic experiments. RTTC and RDA are identical to the celebrated Top Trading Cycles (TTC) and Deferred Acceptance (DA) mechanisms, respectively, in all their theoretical properties except that their dominant-strategy equilibrium is to report one’s preferences in the order opposite to the way they were induced. With the focal truth-telling strategy being out of equilibrium, we are able to perform a clear measurement of how much of the truth-telling reported for strategy-proof mechanisms is compatible with rational behaviour and how much of it is caused by confused decision-makers following a default, focal strategy without understanding the structure of the game. In a school-allocation setting, we find that roughly half of the observed truth-telling under TTC and DA is the result of naïve (non-strategic) behaviour. Only 14–31% of the participants choose actions in RTTC and RDA that are compatible with rational behaviour. Furthermore, by looking at the responses of those seemingly rational participants in control tasks, it becomes clear that most lack a basic understanding of the incentives of the game. We argue that the use of a default option, confusion and other behavioural biases account for the vast majority of truthful play in both TTC and DA in laboratory experiments.

     

MONEY TALKS? AN EXPERIMENTAL INVESTIGATION OF CHEAP TALK AND BURNED MONEY

We experimentally study the strategic transmission of information in a setting where both cheap talk and money can be used. Theoretically, many equilibria exist side by side, in which senders use either costless messages, money, or both. We find that senders prefer to communicate through costless messages. Only when the interest disalignment between sender and receiver increases does cheap talk tend to break down and high sender types start burning money to enhance the credibility of their costless messages. A behavioral model assuming that sellers bear a cost of lying fits the data best.     

Multidimensional cheap talk with sequential messages

This paper studies a two-dimensional cheap talk game with two senders and one receiver. The senders possess the same information and sequentially send messages about that information. In one-dimensional sequential message cheap talk games where the state space is unbounded, the information is fully transmitted under the self-serving belief, as suggested by Krishna and Morgan (2001b). However, this result depends crucially on the structure of the one-dimensional model. It generally does not hold in two-dimensional models. We consider the extended self-serving belief, which implies full information transmission even if the self-serving belief cannot work. Then, we show that the necessary and sufficient condition for the existence of the fully revealing equilibrium is that the senders have opposing-biased preferences.     

Cheap talk with two senders and complementary information

This paper studies a cheap talk model in which two senders having partial and non-overlapping private information simultaneously communicate with an uninformed receiver. The sensitivity of the receiverʼs ideal action to one senderʼs private information depends on the other senderʼs private information. We show that the sendersʼ information transmissions exhibit strategic complementarity: more information transmitted by one sender leads to more information being transmitted by the other sender.     

Informal communication

The typical sender-receiver game studied in the literature assumes that receiver is uninformed. I analyze a model where receiver has private information and sender cares to be perceived as honest. If sender's honesty concerns are strong enough, the model predicts information revelation as a unique equilibrium. This uniqueness result contrasts with the multiplicity of equilibria in games with uninformed receiver. To achieve information revelation as a unique equilibrium, receiver may have to ask for information, which is extraneous for her decision. If sender's honesty concerns are not too strong, asking for extraneous information may create an incentive for lying.     

Who never tells a lie?

I experimentally investigate the hypothesis that many people avoid lying even in a situation where doing so would result in a Pareto improvement. Replicating (Erat and Gneezy, Management Science 58, 723–733, 2012), I find that a significant fraction of subjects tell the truth in a sender-receiver game where both subjects earn a higher payoff when the partner makes an incorrect guess regarding the roll of a die. However, a non-incentivized questionnaire indicates that the vast majority of these subjects expected their partner not to follow their message. I conduct two new experiments explicitly designed to test for a ‘pure’ aversion to lying, and find no evidence for the existence of such a motivation. I discuss the implications of the findings for moral behavior and rule following more generally.     

Optimal truncation in matching markets

Although no stable matching mechanism can induce truth-telling as a dominant strategy for all participants (Roth, 1982), recent studies have presented conditions under which truthful reporting by all agents is close to optimal (Immorlica and Mahdian, 2005, Kojima and Pathak, 2009, Lee, 2011). Our results demonstrate that in large, uniform markets using the Men-Proposing Deferred Acceptance Algorithm, each woman's best response to truthful behavior by all other agents is to truncate her list substantially. In fact, the optimal degree of truncation for such a woman goes to 100% of her list as the market size grows large. In general one-to-one markets we provide comparative statics for optimal truncation strategies: reduction in risk aversion and reduced correlation across preferences each lead agents to truncate more. So while several recent papers focused on the limits of strategic manipulation, our results serve as a reminder that without pre-conditions ensuring truthful reporting, there exists a potential for significant manipulation even in settings where agents have little information.     

Optimal incentive contracts with multiple agents

A setting in which a single principal contracts with two agents who possess perfect private information about their own productivity is considered. With correlated productivities, each agent's private information also provides a signal about the other agent's productivity. In contrast to the setting in which there is only one agent, it is shown that such private information may be of no value to the agents. It is only if the agents are risk-averse that their private information may allow them to command rents. Moreover, when the agents are constrained only to reveal their private information truthfully as a Nash equilibrium, the Pareto optimal incentive scheme may induce the agents to adopt strategies other than truth-telling. This leads to the consideration of truth-telling equilibria that are not Pareto dominated in the subgame played by the agents. Among all such equilibria, the one preferred by the principal restricts one agent to tell the truth as a dominant strategy and the other as a Nash response to truth.     

Morale,performance and disclosure

We study the optimal disclosure policy in a sender–receiver communication game where the receiver's morale, defined as his expected state of the world, affects his performance. The sender observes the state and chooses whether to disclose it to the receiver, who then decides whether to participate in a task. The receiver wins if his performance in the task meets a target. No disclosure is optimal if the receiver wins with average morale in each state. Otherwise, in the threshold disclosure equilibrium that Pareto-dominates full disclosure, the receiver quits as the sender discloses the worst states and wins as the sender withholds the rest. The receiver wins in more states in the Pareto-optimal equilibrium as the sender chooses a non-monotonic disclosure policy. Our theory reveals a trade-off between transparency and efficiency when morale affects performance. It has applications in a broad range of areas including military, family, education and business.     

Bounding preference parameters under different assumptions about beliefs: a partial identification approach

We show how bounds around preferences parameters can be estimated under various levels of assumptions concerning the beliefs of senders in the investment game. We contrast these bounds with point estimates of the preference parameters obtained using non-incentivized subjective belief data. Our point estimates suggest that expected responses and social preferences both play a significant role in determining investment in the game. Moreover, these point estimates fall within our most reasonable bounds. This suggests that credible inferences can be obtained using non-incentivized beliefs.     

Would I lie to you? On social preferences and lying aversion

This paper reinterprets the evidence on lying or deception presented in Gneezy (Am. Econ. Rev. 95(1):384–394, 2005). We show that Gneezy’s data are consistent with the simple hypothesis that people are one of two kinds: either a person will never lie, or a person will lie whenever she prefers the outcome obtained by lying over the outcome obtained by telling the truth. This implies that so long as lying induces a preferred outcome over truth-telling, a person’s decision of whether to lie may be completely insensitive to other changes in the induced outcomes, such as exactly how much she monetarily gains relative to how much she hurts an anonymous partner. We run new but broadly similar experiments to those of Gneezy in order to test this hypothesis. While we also confirm that there is an aversion to lying in our subject population, our data cannot reject the simple hypothesis described above either.

---

Electronic Supplementary Material  The online version of this article () contains supplementary material, which is available to authorized users.      

Strategic information transmission networks

We study a model of multi-player communication. Privately informed decision makers have different preferences about the actions they take, and communicate to influence each others? actions in their favor. We prove that the equilibrium capability of any player to send a truthful message to a set of players depends not only on the preference composition of those players, but also on the number of players truthfully communicating with each one of them. We establish that the equilibrium welfare depends not only on the number of truthful messages sent in equilibrium, but also on how evenly truthful messages are distributed across decision makers.