This paper studies how an information designer can influence experimentation with a new product of unknown quality. The designer commits to a history-contingent recommendation policy. A sequence of consumers arrives, not knowing their positions in the sequence or past outcomes, observing only the designer's recommendations. The designer learns from a noisy signal of quality generated from each purchase. When the designer maximizes expected consumer surplus, the first-best level of experimentation is achievable under general conditions. With alternative objectives—such as internalizing externalities on non-consumers or maximizing profit—the first best may fail, and at least one consumer incentive constraint must bind. We characterize the optimal second-best mechanisms under perfect good news and perfect bad news learning processes. With perfect good news, the designer can increase or decrease experimentation relative to the consumer-welfare benchmark until the relevant incentive constraint binds, and optimal policy may involve mixing before stopping. With perfect bad news, a designer with weaker experimentation incentives cannot do better than delaying experimentation, while a designer with stronger experimentation incentives cannot extend experimentation but may increase adoption by occasionally recommending a product already revealed to be of low quality.
This paper studies dynamic competition in quality and price when consumers are heterogeneous. Firms face exogenous horizontal differentiation from consumer preferences and endogenous vertical differentiation from evolving product quality. The firms investment decisions are non-monotonic in the investment cost, and the two dimensions of differentiations exhibit dynamic substitution / complementary relations depending on the investment cost.
This paper considers optimal information structures in restricting the information available to the Sender in a cheap talk game. Monotone partitional information structures might not obtain optimality, and some signals that the Sender receives are required to be partially garbled together in order to relax the Sender's incentive constraints. We provide a partial characterization of the optimal structure and show that it belongs to a class of structures that are more general than monotone partitional, but that only exhibit certain local non-monotonicity of the information structure. Our analysis builds on the graphical characterization of feasible information structures in Getzkow and Kamenica (2016) and shows that this methodology can be used in complex information design problems with large state and action spaces. As an extension we use the same approach to solve problems with capacity constraints on communication.
In judicial practice, blockchain evidence is frequently regarded as self-authenticating by virtue of its decentralized storage architecture and cryptographic safeguards. Nevertheless, these very characteristics may also lead judges to adopt a superficial standard of review, thereby both relaxing the admissibility threshold for blockchain evidence and attenuating its probative assessment within the broader fact-finding process. Based on empirical analysis of 2,741 judgments, we find that there are two distinct paths behind the seemingly high admissibility rate of blockchain evidence: a logic-driven path of substantive scrutiny and a cue-driven path of superficial scrutiny. Deeper analysis reveals that “notarization or forensic examination” and “explicit objections” both significantly increase the likelihood of substantive scrutiny, indicating that judges tend to construct chains of fact through multiple, interrelated pieces of information. While such external verification can indeed promote substantive scrutiny, the primary responsibility must rest with judges rather than litigants. Accordingly, future reforms should aim to reduce over-reliance on peripheral cues and, by strengthening courtroom communication, enhancing the effectiveness of cross-examination, and refining evidentiary rules, foster a shift toward substantive scrutiny of blockchain evidence.
It is really easy to cheap talk with social media, and extreme events happen more often. In this paper, we consider how much information can be transmitted by cheap talk when truth is no longer constrained to a finite scope. More technically, we study the classic Crawford and Sobel (1982) with an unbounded state space. We found that the information transmission is still effective if the prior is a thin-tail distribution, but information transmission is limited if the prior has a heavy tail. This suggests cheap talk can be less useful in a world with more extreme events.
While the truth is always complicated, people may not always be able to paint the full picture. We consider a sender-receiver game in which the truth state is high-dimensional, while Sender can only communicate low-dimensional information.