Repeated Matching Games: An Empirical Framework

经济学家使用匹配模型将关系的形成建模为市场力量的结果。匹配模型解释了婚姻、就业和商业联盟等关系。大多数匹配模型都是在单个时间点应用于关系的。然而，关系随着时间的推移而变化，跟踪这些变化的数据集越来越多。这个项目开发了新的方法，称为重复匹配游戏，用于研究在市场框架内关系如何随着时间的推移而变化。 其中心思想是，在建立伙伴关系时，人们会考虑当前的关系如何影响他们未来的关系。 除了理论之外，该研究项目还开发了新的工具来分析数据以匹配模型。 该研究应用该方法来研究结婚，离婚和再婚，工程师的职业生涯，以了解工作分配和公司之间的流动如何影响晋升到更有声望的角色。 研究人员还应用这个模型来研究早期的风险投资资金，包括加速器的参与，如何提高获得后期资金和最终创业成功的机会。研究结果为匹配决策提供了输入，从而提高了经济绩效。本项目引入了重复匹配博弈的概念，在每个阶段，代理人可以进行新的匹配。前瞻性代理考虑他们今天的比赛将如何影响他们在未来进行富有成效的比赛的能力。重复匹配游戏结合了一个动态的离散选择模型，在一个单独的代理人的水平与每一个时间段，可转移的效用匹配游戏的关系形成。重复匹配博弈具有理想的性质，导致易于处理的均衡计算。该项目扩展了有关这些匹配模型的知识，开发了模型解决方案的计算机算法，并为模型的参数创建了新的估计器。该项目开发了婚姻和再婚，工人转换公司和风险投资家资助初创企业的经验应用程序。 本研究的结果不仅为匹配博弈模型的建立做出了重要贡献，而且为劳动力市场和其他关系政策提供了一些启示。 该奖项反映了NSF的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

Economists model the formation of relationships as the outcome of market forces using matching models. Matching models explain relationships such as marriage, employment, and business alliances. Most matching models are applied to relationships at a single point in time. However, relationships change over time and datasets tracking those changes are increasingly available. This project develops new methods, called repeated matching game, for studying how relationships change over time within a market framework. The central idea is that in forming partnerships, people consider how current relationships affect their future relationships. In addition to theory, the research project also develop new tools to analyze data for matching models. The research applies the method to study marriage, divorce and remarriage, careers of engineers to learn how job assignment and moving among firms affect promotion to more prestigious roles. The researchers also apply this model to study how early rounds of venture capital funding, including accelerator participation, raise the chance of receiving later rounds of funding and eventual startup success. The results provide inputs into matching decisions hence improve economic performance.This project introduces the concept of a repeated matching game, where each period agents can make new matches. The forward-looking agents consider how their matches today will affect their abilities to make productive matches in the future. The repeated matching game combines a dynamic discrete choice model at the level of an individual agent with a per-time period, transferable utility matching game of relationship formation. The repeated matching game has desirable properties that lead to tractable equilibrium computation. The project extends knowledge about these matching models, develops computer algorithms for model solutions, and create new estimators for the model's parameters. The project develops empirical applications to marriage and remarriage, workers switching firms and venture capitalists funding startups. The results of this research makes significant contribution to the matching game models as well provide some insights into labor market and other relationship policies. The results provide inputs into matching decisions hence improve economic performance.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.