Empirical Analysis of Auction Data

拍卖数据实证分析

• 批准号: 9601220
• 负责人: Jean-Francois Richard
• 金额: $ 13.43万
• 依托单位: University of Pittsburgh
• 依托单位国家: 美国
• 项目类别: Continuing Grant
• 财政年份: 1996
• 资助国家: 美国
• 起止时间: 1996-08-15 至 1998-07-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=9601220&HistoricalAwards=false
• 关键词: Empirical; Analysis; Auction; Data

Auctions and procurements are used for a wide range of private and public transactions. They offer conceptually simple mechanisms for price determination in multilateral trading. They are typically under well-defined rules and rich data sets are available (though not necessarily at negligible direct and indirect costs, notwithstanding important issues of confidentiality). Despite an increasing number of innovative empirical studies of auctions, there remains a substantial gap between theory and applications. The reasons for this include: the questionable empirical validity of key assumptions; the fundamental issue of whether the participants in auctions actually are capable of exercising the exceptionally high degree of rationality assumed by theory; the inherent complexity of the mapping between unobservable individual valuations and the related difficulties associated with `structure` estimation of empirical auction models. This research project aims at developing an integrated approach for the empirical analysis of auction data that addresses the gap between theory and applications. The key components of this approach are: The modeling of observed bidding practices in the form of `rules of thumbs` as operational and intuitively appealing alternatives to the complex Nash equilibrium strategies prescribed by theory; The use of Monte Carlo simulation techniques to find those rules of thumb that perform `well` in the context of a specific empirical auction problem; The development of a generic estimation principle capable of accommodating a broad clan of structural models obtained by application of (1) and (2) allowing, in particular, for parameteric, non parametric and semi-parametric stochastic formulations as deemed appropriate in the context of specific applications; and Empirical applications using data sets from the timber industry in the Pacific Northwest region and from the French aerospatial industry.

拍卖和采购广泛用于私人和公共交易。它们为多边贸易中的价格决定提供了概念上简单的机制。它们通常遵循定义良好的规则，并且可以获得丰富的数据集（尽管尽管存在重要的保密问题，但不一定要付出可以忽略不计的直接和间接成本）。尽管关于拍卖的创新实证研究越来越多，但理论与应用之间仍存在很大差距。其原因包括：关键假设的实证有效性存在问题；最根本的问题是，拍卖的参与者是否真的有能力行使理论所假定的高度理性；不可观察的个体估值之间映射的内在复杂性和与经验拍卖模型的“结构”估计相关的困难。本研究项目旨在为拍卖数据的实证分析开发一种综合方法，以解决理论与应用之间的差距。该方法的关键组成部分是：以“经验法则”的形式对观察到的投标实践进行建模，作为理论规定的复杂纳什均衡策略的可操作且直观吸引人的替代方案；使用蒙特卡罗模拟技术来找到在特定的经验拍卖问题背景下表现“良好”的经验法则；开发一种通用估计原理，能够容纳通过应用(1)和(2)获得的广泛的结构模型，特别是在特定应用环境中被认为合适的参数、非参数和半参数随机公式；以及使用来自太平洋西北地区木材工业和法国航空航天工业的数据集的经验应用。