专家服务市场的欺诈、信任与效率--基于社会偏好和空谈博弈的视角

Fraud,Trust and Efficiency in the Expert Service Market:From the Perspectives of Social Preferences and Cheap Talk Games

本文从空谈博弈和专家的社会偏好角度来分析价格外生给定的专家服务市场的低效率现象。区别于已有文献,本文认为如果市场中的部分专家存在诚信行为,他们将与逐利专家形成高信号(如患者病情较重)区间混同、低信号(如患者病情较轻)区间分离的均衡,同时导致欺骗、服务失当以及信任缺失等一系列现象。逐利专家会以概率分布的方式模仿诚实专家的行为,向消费者随机发送均值高于其真实需求的信号。同时,低于某个阈值的信号会获得消费者完全的信任,而消费者对高于此阈值的信号的信任度会随着信号的上升而下降,所以逐利专家的信号策略本质上是在获取信任与绝对收益之间的权衡。在均衡中低需求消费者获得过多的服务(如小病大治),而高需求消费者获得的服务不足(如大病小治)。最后,本文发现引入无搜索成本的竞争可以解决以上市场效率低下的问题。

Summary ： In an expert service market, consumers know little about the level or quality of the services they need, so they rely on the advice of experts. We model a situation in which experts can freely send diagnostic signals to consumers, informing them of the level of their demands, and recommend verifiable treatments accordingly. The credence goods literature typically studies the effect of price menus as a signal of market equilibrium. Nevertheless, when prices are exogenously determined, the cheap talk signal from expert diagnosis becomes the only signal available, and few studies pay attention to this scenario, We analyze equilibrium in the expert service market through a cheap talk game where expert diagnosis is the only available signal and two types of experts coexist： honest and selfish experts. We assume honest experts always send truthful signals and provide the most desirable treatment of consumers. In a monopolistic expert service market, the cheap talk signals are partially informative. The perfect Bayesian equilibrium in this market has three important characteristics. （ 1 ） The equilibrium is pooled at the top and separated at the bottom. Selfish experts do not send signals lower than a certain threshold value of demand. Their signals are continuously distributed over the upper-end of the space above the threshold. （2） When the signal is lower than the threshold, consumers fully trust it. However, when the signal is above the threshold, consumers＇ beliefs about their demands are not affected by cheap talk signals. （3） If the signaling strategy is a monotone （the expected value of the random signal rises as its true state increases） , then selfish experts will randomly exaggerate the true state of the demand and are more likely to send higher signals. Consumers, however, trust less as the signal goes higher. As a result, those who have low demands are over- treated, and those who have high demands are under-treated. Pooling at the top and separation at the bottom are caused by selfish experts trying to imitate the signals sent by honest experts, but only when the profits are sufficiently high. As a signal goes higher, the profit increases once a consumer accepts the treatment; however, the probability of a consumer believing the signal is real also falls. The tradeoff between profit and trust restrains the selfish experts＇ signaling strategy to a probability distribution that gives an equal expected profit to each signal within its support. In addition, by imposing a condition that selfish experts send signals from the same distribution point regardless of the true state, the selfish experts＇ signaling strategy has a tractable solution. Comparative statistics suggest that an increase in the proportion of honest experts hurts low-demand consumers more when they encounter selfish experts. Furthermore, price regulation such as lowering the service price ceiling results in selfish experts sending increasingly higher signals, which in turn lowers the consumers＇ trust level. Policymakers should be wary of these adverse effects when they consider measures such as cultivating professional conduct and price regulations. We further discuss the effects of introducing a competitive market structure and lowering consumer search costs. Obtaining a second opinion should provide consumers with more information about their true type, and we find that the fraudulent behavior of experts diminishes when consumers can search for other advice more easily.