Download Bootstrap Tests for Regression Models by L. Godfrey PDF

By L. Godfrey

ISBN-10: 0230202314

ISBN-13: 9780230202313

ISBN-10: 0230233732

ISBN-13: 9780230233737

An available dialogue studying computationally-intensive concepts and bootstrap tools, offering how one can enhance the finite-sample functionality of famous asymptotic exams for regression types. This e-book makes use of the linear regression version as a framework for introducing simulation-based exams to aid practice econometric analyses.

Show description

Read Online or Download Bootstrap Tests for Regression Models PDF

Best econometrics books

Random Regret-based Discrete Choice Modeling: A Tutorial

This instructional provides a hands-on advent to a brand new discrete selection modeling procedure in keeping with the behavioral proposal of regret-minimization. This so-called Random remorse Minimization-approach (RRM) varieties a counterpart of the Random software Maximization-approach (RUM) to discrete selection modeling, which has for many years ruled the sector of selection modeling and adjoining fields resembling transportation, advertising and environmental economics.

An Introduction to Order Statistics

This ebook offers the idea of order records in a fashion, such that novices can get simply conversant in the very foundation of the speculation with no need to paintings via seriously concerned concepts. even as more matured readers can money their point of figuring out and varnish their wisdom with definite information.

Structural Change in Macroeconomic Models: Theory and Estimation

This e-book grew out of a 'Doctorat D'Etat' thesis offered on the collage of Dijon-Institut Mathematique Economiques (lME). It goals to teach that amount rationing thought presents the technique of enhancing macroeconometric modelling within the learn of struc tural adjustments. The empirical effects offered within the final bankruptcy (concerning Portuguese economic climate) and within the final Appendix (con cerning the French economy), even supposing initial, urged that the trouble is worthwhile and will be endured.

Statistics and Data Analysis for Financial Engineering: with R examples

The recent variation of this influential textbook, geared in the direction of graduate or complex undergraduate scholars, teaches the facts valuable for monetary engineering. In doing so, it illustrates ideas utilizing monetary markets and financial facts, R Labs with real-data workouts, and graphical and analytic equipment for modeling and diagnosing modeling mistakes.

Additional resources for Bootstrap Tests for Regression Models

Example text

It is hoped that the results obtained from simulation experiments are of value in practical situations, but this property can never be guaranteed. It is very common to refer to such experiments as Monte Carlo experiments; see Davidson and MacKinnon (1993, ch. 21) for an excellent discussion. However, given that “Monte Carlo” will often be used below to refer to a particular type of simulationbased test, this terminology will not be adopted in order to avoid any confusion. There are three sets of simulation experiments, each reported in its own subsection.

The choice can be based simply on convenience, given the set of error distributions to be used. The distributions that are used to obtain the terms of (u1 , . . , u27 ) are: N(0, 1) as a benchmark; t(5) as an example of a symmetric distribution that is far from being Normal; and the heavily-skewed χ 2 (2) distribution. Drawings from these distributions are adjusted, when necessary, so that, without loss of generality, they come from a population with zero mean and variance equal to one. ) In order to derive fairly precise estimates of rejection probabilities, 25,000 sets of artificial data with n = 27 are generated for each error distribution, that is, the number of replications is R = 25,000.

However, there is evidence from simulation experiments which indicates that these asymptotic critical values may not be accurate approximations to actual finite sample values; see, for example, Diebold and Chen (1996). The second example of a non-standard procedure is also based upon a test described in Chow (1960). As well as proposing a test of the claim that all regression coefficients are constant, Chow explains how to carry out a test of the hypothesis that prediction errors have a zero mean.

Download PDF sample

Rated 4.01 of 5 – based on 25 votes