Files in this item

FilesDescriptionFormat

application/pdf

application/pdf9990089.pdf (8MB)Restricted to U of Illinois
(no description provided)PDF

Description

Title:Essays in Nonparametric Econometrics
Author(s):Morillo, Daniel Santiago
Doctoral Committee Chair(s):Koenker, Roger W.
Department / Program:Economics
Discipline:Economics
Degree Granting Institution:University of Illinois at Urbana-Champaign
Degree:Ph.D.
Genre:Dissertation
Subject(s):Statistics
Abstract:The work presented here is a collection of three self-contained chapters that explore various aspects of the computation and application of nonparametric methods in Econometrics. All three chapters are linked by the use of Quantile Regression ideas and techniques. The first chapter uses stochastic approximation techniques to develop an algorithm for recursive estimation of a general linear quantile regression model. The asymptotics of the proposed estimator are developed and it is shown that the asymptotic performance of the estimator is equivalent to that of the estimator obtained via standard linear programming methods. A Monte Carlo experiment shows that the proposed estimator is competitive with the standard linear-programming estimator for moderately large sample sizes. The second chapter presents an improvement on existing recursive regression methods for pricing American-type options. The central idea is to reduce bias and variance of existing estimators by combining two option value estimators both of them biased but in opposite direction. A Monte Carlo experiment demonstrates that the proposed method does indeed perform better than existing methods in simple log-normal case. A detailed empirical example uses Quantile regression to capture the dynamics of observed asset price data and to simulate the price paths used to computed option value estimates. Results give evidence that the modeling flexibility provided by Quantile Regression methods can result in pricing performance gains over the standard binomial method. Finally, the last chapter proposes a method for the analysis of income mobility that applies the permanent inequality measures of mobility to simulated data that can be constructed from estimates of the short-term dynamics of income. The method is thus able to track changes in mobility itself while remaining within the general welfare-based framework. Nonparametric quantile regression is used to obtain the transition quantile curves of the income process and then to simulate future income paths. The procedure is applied to income data from the U.S. and Germany. Results show evidence of a significant increase in mobility in the U.S. starting in the early 90's and also provide evidence of a positive relationship between government size and income mobility.
Issue Date:2000
Type:Text
Language:English
Description:144 p.
Thesis (Ph.D.)--University of Illinois at Urbana-Champaign, 2000.
URI:http://hdl.handle.net/2142/85677
Other Identifier(s):(MiAaPQ)AAI9990089
Date Available in IDEALS:2015-09-25
Date Deposited:2000


This item appears in the following Collection(s)

Item Statistics