An Evolutionary Algorithm for the Estimation of Threshold Vector Error Correction Models
We develop an evolutionary algorithm to estimate Threshold Vector Error Correction models (TVECM) with more than two cointegrated variables. Since disregarding a threshold in cointegration models renders standard approaches to the estimation of the cointegration vectors inefficient, TVECM necessitate a simultaneous estimation of the cointegration vector(s) and the threshold. As far as two cointegrated variables are considered this is commonly achieved by a grid search. However, grid search quickly becomes computationally unfeasible if more than two variables are cointegrated. Therefore, the likelihood function has to be maximized using heuristic approaches. Depending on the precise problem structure the evolutionary approach developed in the present paper for this purpose saves 90 to 99 per cent of the computation time of a grid search.
|Date of creation:||Feb 2010|
|Date of revision:|
|Contact details of provider:|| Postal: Kleine Märkerstrasse 8, 06108 Halle (Saale)|
Phone: (0345) 7753-60
Fax: (0345) 7753-820
Web page: http://www.iwh-halle.de
More information through EDIRC
When requesting a correction, please mention this item's handle: RePEc:iwh:dispap:1-10. See general information about how to correct material in RePEc.
For technical questions regarding this item, or to correct its authors, title, abstract, bibliographic or download information, contact: (Tobias Henning)
If references are entirely missing, you can add them using this form.