Subjects
-Tags
-Abstract
This paper compares Johansen's method of maximum likelihood for the reduced rank regression with the method of moments for first differences suggested by Laroque and Salanié. The maximum likelihood estimator is known to be super consistent with a distribution that consists of functionals of the Wiener process while the method of moments estimator is shown to be consistent and asymptotically normal for the common cointegrated model without autocorrelation. A Monte Carlo study for the bivariate case indicates that finite sample properties are consistent with asymptotic results even for samples of less than 100. While the method of moments procedure produces normal distributions so that inference will be straightforward, it has the disadvantage of having larger variability than the maximum likelihood estimators (at least in the case of no autocorrelation considered here).
DOI
10.21608/esju.2001.313816
Keywords
Asymptotic Properties, Single Equation Approach, Monte Carlo
Link
https://esju.journals.ekb.eg/article_313816.html
Detail API
https://esju.journals.ekb.eg/service?article_code=313816
Publication Title
The Egyptian Statistical Journal
Publication Link
https://esju.journals.ekb.eg/
MainTitle
Maximum Likelihood and Method of Moments in Cointegration