Error correction model: Difference between revisions

Revision as of 06:39, 15 March 2016

Revision as of 16:02, 15 March 2016

Line 23:

While this approach is easy to apply, there are, however numerous problems:

<li>* The univariate unit root tests used in the first stage have low [[statistical power]]</li>

<ul>

<li>* The choice of dependent variable in the first stage influences test results, i.e. we need weak exogeneity for <math>x_t</math>as determined by [[Granger causality]]</li>

<li>The univariate unit root tests used in the first stage have low [[statistical power]]</li>

<li>* One can potentially have a small sample bias</li>

<li>The choice of dependent variable in the first stage influences test results, i.e. we need weak exogeneity for <math>x_t</math>as determined by [[Granger causality]]</li>

<li>* The cointegration test on <math>\alpha </math> does not follow a standard distribution</li>

<li>One can potentially have a small sample bias</li>

<li>* The validity of the long-run parameters in the first regression stage where one obtains the residuals cannot be verified because the distribution of the OLS estimator of the cointegrating vector is highly complicated and non-normal</li>

<li>The cointegration test on <math>\alpha </math> does not follow a standard distribution</li>

<li>* At most one cointegrating relationship can be examined</li>

<li>The validity of the long-run parameters in the first regression stage where one obtains the residuals cannot be verified because the distribution of the OLS estimator of the cointegrating vector is highly complicated and non-normal</li>

<li>At most one cointegrating relationship can be examined</li>

</ul>

===VECM===