Abstract
This paper tests for the presence of nonlinear dependence in the black-market Polish zloty-dollar exchange rate. Using the GARCH-M model, we illustrate use of the Marquardt (Journal of the Society of Industrial and Applied Mathematics, 2, 1963) alternative to the Berndt (Annals of Economical Social Measurement, 4, 1974) iterative nonlinear algorithm for the estimation of such models, and discrimination between estimated models on the basis of the Brock and Potter (Handbook of Statistics, 11, 1993) test for so conditional variance misspecification. We find evidence of a time-varying risk premium such that foreign speculators are compensated for increased exchange rate risk by appreciation which increases the dollar value of zloty holdings, and which is able to account for all of the apparent nonlinearity in the zloty.
| Original language | English |
|---|---|
| Pages (from-to) | 209-214 |
| Number of pages | 6 |
| Journal | Applied Economics Letters |
| Volume | 6 |
| Publication status | Published - Apr 1999 |
Keywords
- EMPIRICAL-EVIDENCE
- HETEROSCEDASTICITY