Estimating Network and Matching Games of Interfirm Relationships

企业之间的关系是我们关于工业绩效知识的一个重要而未被充分研究的方面。经济学家利用匹配博弈的理论框架，对风险资本家和创业公司、供应商和零售商、品牌联盟和公司之间的合并进行了实证研究。在企业间关系的形成过程中，企业是竞争对手，与最具吸引力的合作伙伴企业相匹配。换句话说，公司关系在某种程度上是排他性的。以无线电话市场为例，当Sprint与Nextel合并时，Verizon也不太可能与Nextel合并。Sprint和Verizon是Nextel的竞争对手。一个关键的经验观点是，观察者在数据中看到和没有看到的企业关系（匹配）告诉我们很多关于企业对这些关系的目标。了解公司的目标可以让我们了解行业的结构。了解行业结构及其可能发生的变化对于评估拟议的政策是必要的。公司间关系中一个重要的复杂因素，在其他匹配市场（比如婚姻）中不太重要，那就是一家公司的回报往往是竞争对手公司关系的函数。这是因为企业在建立关系后会争夺消费者。举个及时的例子，Cingular与苹果签订了一份多年的独家合同，在美国分销苹果的新款iPhone。在某种程度上，iPhone被证明是受欢迎的，竞争对手的无线运营商的利润将是辛格勒供应商苹果公司的功能。网络游戏使研究人员能够从竞争对手的关系中模拟这些外部性。网络则是企业间关系的集合，结合了不合作企业争夺消费者的紧密程度。本项目提出了新的实证方法来估计企业关系的网络和匹配博弈。一个关键的计算问题是，存在大量可能构成市场均衡的关系组合。本研究引入了计算简单的估计器，并讨论了关于不同类型的均衡结果（匹配、企业之间的货币转移等）的数据如何使经济学家更多地了解企业的目标。更广泛的影响更广泛的影响可以通过网络和匹配估计器的大量现有和潜在应用来最好地评估：风险资本家、合并、频谱拍卖、品牌联盟、住房、农村社区的风险分担团体、知识创造、公立学校教师的劳动力市场、国家强制的公立学区合并和婚姻市场。每个新的经济市场都会以游戏和可用数据的形式带来自己的挑战。该研究使估计网络和匹配博弈成为应用经济学家可用的工具。软件和大量文档可用于统计估计和推断。该研究还指导本科生和研究生的合作者。

Relationships between firms are an important and understudied aspect of our knowledge about industrial performance. Economists have used the theoretical framework of matching games to empirically study venture capitalists and startup companies, suppliers and retailers, brand alliances and mergers between companies. In the formation of interfirm relationships, firms are rivals to match with the most attractive partner firms. In other words, firm relationships are to some degree exclusive. For an example from the wireless phones market, when Sprint merged with Nextel, it made it less likely that Verizon will also be able to merge with Nextel. Sprint and Verizon were rivals to match with Nextel. A key empirical idea is that the firm relationships (matches) that observers see and do not see in the data tell us a lot about the goals firms have for these relationships. Understanding firms' objectives informs us about the structure of the industry. Understanding industry structure and how it might change is necessary to evaluate proposed policies. An important complication in interfirm relationships that is less important in other matching markets (say marriage) is that the payoffs of one firm are often functions of the relationships of rival firms. This is because firms compete for consumers after forming relationships. For a timely example, Cingular has a multi-year exclusive contract to distribute Apple's new iPhone in the United States. To the extent that the iPhone proves popular, the profits of rival wireless carriers will be a function of Cingular's supplier Apple. Network games allow researchers to model these externalities from rivals' relationships. The network is then the set of interfirm relationships combined with how closely noncooperating firms compete for consumers. This project produces new empirical methods to estimate network and matching games of firm relationships. A key computational problem is that there are a large number of combinations of relationships that could be market equilibria. This research introduces computationally simple estimators and discusses how data on different types of equilibrium outcomes (matches, monetary transfers between firms, etc.) allow economists to learn more about the objectives of firms. Broader ImpactThe broader impact can best be assessed by the large number of existing and potential applications of network and matching estimators: venture capitalists, mergers, spectrum auctions, brand alliances, housing, risk sharing groups in rural communities, knowledge creation, labor markets for public school teachers, the state-mandated consolidation of public school districts, and marriage markets. Each new economic market brings up its own challenges in the form of the game and the available data. The research makes estimation network and matching games a usable tool for applied economists. Software and extensive documentation are available for statistical estimation and inference. The research also mentors undergraduate and graduate student collaborators.