Bootstrap Tests for Regression Models (Palgrave Texts in by L. Godfrey

By L. Godfrey

An obtainable dialogue interpreting computationally-intensive recommendations and bootstrap tools, offering how you can increase the finite-sample functionality of famous asymptotic assessments for regression versions. This book uses the linear regression version as a framework for introducing simulation-based exams to assist practice econometric analyses.

Show description

Read Online or Download Bootstrap Tests for Regression Models (Palgrave Texts in Econometrics) PDF

Similar econometrics books

Regression Analysis of Count Data (Econometric Society Monographs)

Scholars in either social and common sciences usually search regression how you can clarify the frequency of occasions, comparable to visits to a physician, car injuries, or new patents offered. This publication presents the main entire and up to date account of versions and strategies to interpret such facts. The authors have carried out study within the box for greater than twenty-five years.

Modelling the Riskiness in Country Risk Ratings (Contributions to Economic Analysis)

The significance of state probability is underscored by way of the lifestyles of numerous renowned kingdom possibility ranking businesses. those firms mix information about replacement measures of monetary, monetary and political threat into linked composite threat rankings. because the accuracy of such nation probability measures is open to query, it will be important to examine the enterprise score structures to permit an review of the significance and relevance of enterprise hazard scores.

Applied Macroeconometrics

Until eventually the Nineteen Seventies, there has been a consensus in utilized macroeconometrics, either concerning the theoretical beginning and the empirical specification of macroeconometric modelling, generally called the Cowles fee strategy. this can be now not the case: the Cowles fee strategy broke down within the Nineteen Seventies, changed by means of 3 favourite competing tools of empirical learn: the LSE (London tuition of Economics) technique, the VAR strategy, and the intertemporal optimization/Real company Cycle technique.

Additional info for Bootstrap Tests for Regression Models (Palgrave Texts in Econometrics)

Sample text

Misspecification of the model for autocorrelation and/or heteroskedasticity will, in general, lead to an inconsistent estimator of the covariance matrix of the minimizers of QNLS (β, ψ) and QFGLS (β). 18), because such tests use the estimated covariance matrix to assess the significance of sample outcomes. An investigation of the effects of misspecifying the model for heteroskedasticity is reported in Belsley (2002). It is found that effects can be serious and Belsley concludes that Correction for heteroskedasticity clearly does best when both the proper arguments and the proper form of the skedasticity function are known.

Asymptotic critical values have been provided, using simulation, so that tests can be applied; see Andrews (1993, 2003a). 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.

48) t=j+1 where l denotes the lag truncation value that allows all asymptotically relevant autocorrelations to be taken into account, with l −→ ∞ and l/n −→ 0, as n −→ ∞. 49) when the null hypothesis is true and l is selected in an appropriate way. In the analysis of Newey and West, l is o(n−1/4 ), but it is remarked that, under different assumptions, l being o(n−1/2 ) would be appropriate for asymptotically valid inference; see Newey and West (1987, p. 705). Clearly rules about the asymptotic order of magnitude in n of the truncation value l do not provide real guidance about the choice of this parameter in a practical situation in which a finite sample is employed.

Download PDF sample

Rated 4.37 of 5 – based on 6 votes