Please use this identifier to cite or link to this item: http://hdl.handle.net/10397/62730
Title: An exact test for the choice of the combination of first differences and percentage changes in linear models
Authors: Ip, WC
Keywords: Constructed variable
Functional form
Box-Cox transformation
Taylor series
Least squares
Issue Date: 2000
Publisher: Pushpa Publishing House
Source: Far East journal of theoretical statistics, 2000, v. 4, no. 1, p. 27-44 How to cite?
Journal: Far East journal of theoretical statistics 
Abstract: Econometric models are often formulated in terms of the first difference or the percentage change, generalized by the Box-Cox difference transformation. The choice of suitable functional forms has relied heavily upon established statistical procedures which adopt primarily the likelihood approach and confine to a single transformation parameter only. We have derived an exact test for the parameter vector of transformation in linear models. By utilizing Taylor series approximations, this reduces to a choice between two regression equations. The test statistic which has an exact F-distribution can be easily calculated from these two regressions by least squares algorithm. Monte Carlo results demonstrate that our proposed test is more capable than the likelihood approach in capturing the correct size yet is as powerful as the latter. It is, therefore, a simple and ready statistical procedure for assessing the suitable choice of the combination of first differences and percentage changes in economic forecast models.
URI: http://hdl.handle.net/10397/62730
ISSN: 0972-0863
Appears in Collections:Journal/Magazine Article

Access
View full-text via PolyU eLinks SFX Query
Show full item record

Page view(s)

15
Last Week
2
Last month
Checked on Jul 9, 2017

Google ScholarTM

Check



Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.