Collaborative Research: Heterogeneity and Structural Duration Dependence in Macroeconomics

该项目得到了美国国家科学基金会1559459和155922两项奖励的资助。该研究小组试图测试宏观经济学中的久期依赖模型。这项工作直接集中在如何分析来自劳动力市场的持续时间数据，包括失业的持续时间/长度。一些工人经历了长期失业，而另一些同样失业的人很快就找到了新的工作。同样，一些人在同一个雇主工作多年(工作时间较长)，而另一些人在短暂停留后换了一份新工作。PI希望比较导致这些持续时间的不同过程模型。他们希望建立新的、易于处理的模型，并计划估计和测试模型，以查看数据对哪种模型的支持最好。这项研究将帮助我们理解长期失业的性质，并通过为我们预测某些公共政策(如延长失业救济金)对经济的影响提供更好的基础，从而使美国经济受益。该团队计划检验现有简化模型中持续时间相关性的性质，如比例风险模型和具有未观察到的异质性的加速失效时间模型。他们还开发了新的结构模型，包括停止时间模型，以提供对持续时间数据的更经济解释，同时允许任意的未观察到的异质性。他们计划使用现有的关于劳动力市场结果和价格的大型数据集，以估计和测试这些模型。由于持续期模型已被用于研究许多不同的经济现象，包括商业投资、住房成交量、名义价格粘性和耐用消费品购买等决定因素，因此这项研究具有广泛的应用潜力。

This project is funded by NSF awards 1559459 and 155922. The research team seeks to test models of duration dependence in macroeconomics. The work focuses directly on how to analyze duration data from labor markets, including the duration/length of spells of unemployment. Some workers experience long-term unemployment, while others who are also unemployed rapidly find new jobs. Similarly, some people stay with the same employer for many years (a long duration on the job) while others switch to a new job after a short stay. The PIs want compare different models of the processes that lead to these durations. They want to build new and tractable models and also plan to estimate and test the models to see which model is best supported by the data. The research will help us understand the nature of long-term unemployment, and it will benefit the U.S. economy by giving us a better basis for predicting the effects of certain kinds of public policy (such as extending unemployment benefits) on the economy.The team plans to examine the nature of duration dependence in existing reduced-form models, such as the proportional hazard and the accelerated failure times models with unobserved heterogeneity. They also devlop new structural models, including a stopping time model, to offer a more economic interpretation of duration data, while allowing for arbitrary unobserved heterogeneity. They plan to use large existing data sets on labor market outcomes and on prices in order to estimate and test these models. The research has the potential to be useful in a wide variety of applications, since duration models have been used to study many different economic phenomena, including the determinants of business investment, housing turnover, nominal price stickiness, and the purchase of consumer durables as well as others.