Research towards a better understanding of logit type models with fixed effects

许多活动表现出持久性。例如，个人倾向于年复一年地选择相同的保险计划，某一年向某一特定国家出口的公司明年可能更有可能向该国家出口，吸烟者比不吸烟者更有可能在未来吸烟。对于这种持续性，有两种自然的解释：一种是事件本身使事件在未来更有可能发生；另一种是个人或公司的持续性特征导致了活动的持续性。这两种解释可能会产生截然不同的政策含义。本项目将调查何时有可能区分这两种解释，并为此开发工具。研究将主要是理论的，但它也将开发实用的工具来实施将要开发的方法。这项研究项目的结果将有助于制定明智和适当的政策和商业决策，从而改善经济的运作。因此，其结果将提高美国经济的表现。当面板数据集由每个人的少量时间段组成时，该项目将研究具有固定效应的各种离散结果面板数据模型的识别和估计。最近由PI合作撰写的一篇论文开发了一种在简单动态二元结果逻辑模型中搜索力矩条件的方法。本项目将采用相同的一般方法来为许多其他模型构建力矩条件。在可能构造这样的力矩条件的情况下，这些将用于构造模型参数的估计器和测试。这项研究项目的结果将有助于制定明智和适当的政策和商业决策，从而改善经济的运作。因此，其结果将提高美国经济的表现。该奖项反映了美国国家科学基金会的法定使命，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。

Many activities exhibit persistence. For example, individuals tend to choose the same insurance plan year after year, a firm that exports to a particular country one year might be more likely to export to that country the next year, and a smoker is more likely to smoke in the future than a-non-smoker. Two natural explanations for such persistence are that the event itself causes the event to be more likely in the future, or that there are persistent characteristics of the person or firm which lead to persistence in the activity. These two explanations can have very different policy implications. This project will investigate when it is possible to distinguish between the two explanations and to develop tools for doing so. The research will be mainly theoretical, but it will also develop practical tools for implementing the methods that will be developed. The results of this research project will aid in making informed and appropriate policy and business decisions, thus improving the functioning of the economy. The results will therefore increase the performance of the US economy.This project will investigate identification and estimation of a variety of discrete outcome panel data models with fixed effects when the panel data set consists of a small number of time periods for each individual. A recent paper co-authored by the PI developed an approach for searching for moment conditions in simple dynamic, binary-outcome logit models. This project will apply the same general approach to construct moment conditions for a number of other models. In the cases where it is possible to construct such moment conditions, these will be used to construct an estimator and tests of the model parameters. The results of this research project will aid in making informed and appropriate policy and business decisions, thus improving the functioning of the economy. The results will therefore increase the performance of the US economy.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.