A new mechanism for product tests à la Consumer Reports, Stiftung Warentest and Co. – Adopting a long-term perspective with endogenous price and quality, and unknown distribution of quality

卖方通常比买方更了解产品质量。为了减少这种信息不对称并帮助购买者，已经建立了诸如消费者报告（美国）或环保测试基金会（德国）之类的产品测试组织。虽然它们提供了可靠的信息，但这些信息可能不是最佳的。由于测试能力有限，必须选择某些产品型号（例如，在所有可用的智能手机型号中选择某些型号）进行测试。例如，Stiftung warrentest主要选择最畅销的产品型号，认为买家最感兴趣的是了解畅销书的质量。然而，在未测试的产品模型中，可能存在主导测试产品模型的产品模型（例如，在相同质量的情况下更便宜）。我们提出了一种新的，能力中立的机制来选择产品模型进行测试。我们的机制允许卖家向产品测试机构申请产品型号，说明价格和质量。然后，产品检测机构根据预先指定的算法从申请人中选择需要检测的产品型号，允许卖家自愿披露其产品质量。在一项初步实验研究中（Vollstädt, Brendel, Ehses-Friedrich, mimeo），我们比较了两种基准治疗：Stiftung warren test当前机制的风格化版本和我们的新机制。我们表明，我们的机制导致在非主导产品模型上具有更多信息的产品测试。因此，消费者剩余和某些卖家的利润都增加了。虽然这些初步结果非常有希望，但它们依赖于几个强有力的假设。本研究计划的目标是通过放松两个主要假设，逐步增加我们的结果的外部有效性。在项目1中，我打算放宽价格和质量是外生的假设。为了实现这一点，我将把初步研究中使用的静态短期博弈扩展到价格和质量是内生的动态长期博弈。在价格和质量内生的情况下，我们的新机制是否会比华伦斯特基金会机制的程式化版本表现得更好？此外，我们的新机制是否会导致更好的产品模型的供应（例如，在相同质量的情况下，更便宜的产品模型）？在项目2中，我打算放宽买方和卖方知道产品质量如何分布的假设。这是所有关于自愿信息披露的相关研究的假设，也是我们初步研究的假设。然而，这一假设实际上既不适用于大多数买家，也不适用于所有卖家。即使产品质量的分布是未知的，我们的新机制是否会比Stiftung waretest机制的风格化版本表现得更好？

Sellers are often better informed about product quality than buyers. To reduce this asymmetric information and to help buyers, product testing organizations like Consumer Reports (US) or Stiftung Warentest (Germany) have been established. Although they provide credible information, this information may not be optimal. Due to limited testing capacities, certain product models (e.g. certain smartphone modells among all available ones) have to be selected for a test. For instance, Stiftung Warentest selects mainly the bestselling product models, arguing that buyers would be interested most in learning about the bestsellers' quality. However, there may be product models among the non-tested ones which dominate the tested product models (e.g. are less expensive given the same quality).We propose a new, capacity-neutral mechanism to select product models for a test. Our mechanism allows sellers to apply to the product testing organization with a product model, stating price and quality. The product testing organization then selects the product models to be tested among the applicants according to a pre-specified algorithm, allowing sellers to voluntarily disclose their product quality. In a preliminary experimental study (Vollstädt, Brendel, Ehses-Friedrich, mimeo), we compared two benchmark treaments: a stylized version of Stiftung Warentest’s current mechanism and our new mechanism. We showed that our mechanism leads to product tests with more information on non-dominated product models. Therefore, consumer surplus as well as profits of certain sellers increase.While these preliminary results are very promising, they rely on several strong assumptions. The goal of the present research proposal is to gradually increase the external validity of our results by relaxing two major assumptions. In project 1, I plan to relax the assumption that price and quality are exogenous. To achieve this, I will extend the static, short-term game used in the preliminary study to a dynamic, long-term game in which price and quality are endogenous. Will our new mechanism, even with endogenous price and quality, perform better than a stylized version of Stiftung Warentest’s mechanism? Moreover, will our new mechanism lead to a supply of better product models (e.g. less expensive product models given the same quality)? In project 2, I plan to relax the assumption that buyers and sellers know how product quality is distributed. This assumption is made in all related studies on voluntary information disclosure and also in our preliminary study. Yet, this assumption does actually neither hold for most buyers, nor for all sellers. Will our new mechanism perform better than a stylized version of Stiftung Warentest’s mechanism even if the distribution of product quality is unknown?