Nonparametric Estimation of Econometric Models

计量经济模型的非参数估计

• 批准号: 9110039
• 负责人: Whitney Newey
• 金额: $ 17.32万
• 依托单位: Massachusetts Institute of Technology
• 依托单位国家: 美国
• 项目类别: Continuing Grant
• 财政年份: 1991
• 资助国家: 美国
• 起止时间: 1991-08-01 至 1995-01-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=9110039&HistoricalAwards=false
• 关键词: Nonparametric; Estimation; Econometric; Models

Semiparametric estimators are used intensively in applied work as a way of avoiding detailed assumptions on model specifications and as a way of avoiding difficult computational problems. The contribution of this project comes from developing a general asymptotic theory for semiparametric estimators, investigating methods for nonparametric estimation of structural equation models, and using nonparametric estimators of consumer surplus for an empirical analysis of gasoline demand. Knowledge of the asymptotic variance of an estimator is important for large sample inference, efficiency, and as a guide to the specification of regularity conditions. This research develops a general formula with a relatively simple set of regularity conditions for the asymptotic variance of semiparametric estimators. This is an important contribution because at present asymptotic variance calculations for semiparametric estimators are model specific and usually involve a great deal of tedious detail. This project will make semiparametric estimators much easier to use. Nonparametric estimators have gained wide attention in the past few years in econometrics because they impose even fewer restrictions than semiparametric estimators. But nonparametric models are difficult to interpret and their usefulness in applied research has been demonstrated in a limited number of cases. This project applies nonparametric regression models to estimation of demand curves of the types most often used in applied research. Estimators of exact consumer surplus and deadweight loss will be derived from the demand curve estimators. The research includes an application to gasoline demand. This represents an important contribution because estimators of consumer surplus and deadweight loss are the most widely used welfare measures in applied areas of economics such as public finance. The results from the empirical research will provide new insights into the magnitude of welfare loss from carbon taxes that are currently being considered as one of the policy tools for reducing emissions of greenhouse gases.

半参数估计器在应用工作中被广泛使用，作为一种避免对模型规格进行详细假设的方法，也是避免困难计算问题的一种方法。该项目的贡献来自于发展了半参数估计的一般渐近理论，研究了结构方程模型的非参数估计方法，并将消费者剩余的非参数估计用于汽油需求的实证分析。关于估计量的渐近方差的知识对于大样本推理、效率以及作为规范正则性条件的指导是重要的。本文给出了半参数估计渐近方差的一般公式和一组相对简单的正则性条件。这是一个重要的贡献，因为目前半参数估计的渐近方差计算是特定于模型的，通常涉及大量繁琐的细节。这个项目将使半参数估计器更容易使用。在过去的几年里，非参数估计在计量经济学中得到了广泛的关注，因为它们施加的限制甚至比半参数估计更少。但非参数模型很难解释，它们在应用研究中的有效性已在有限的案例中得到证明。该项目应用非参数回归模型来估计应用研究中最常用的类型的需求曲线。准确的消费者剩余和无谓损失的估计值将从需求曲线估计器中得到。这项研究包括对汽油需求的应用。这是一项重要贡献，因为消费者剩余和无谓损失的估计值是公共财政等经济学应用领域中使用最广泛的福利衡量标准。实证研究的结果将对碳税造成的福利损失的规模提供新的见解，目前碳税被视为减少温室气体排放的政策工具之一。