Estimating Network and Matching Games of Interfirm Relationships

点击翻译按钮获取中文摘要

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.