Original scientific paper
ECONOMETRIC APPROACH TO DIFFERENCE EQUATIONS MODELING OF EXCHANGE RATES CHANGES
Josip Arnerić
orcid.org/0000-0002-2901-2609
; Faculty of Economics Split, University of Split, Split, Croatia
Lana Kordić
orcid.org/0000-0001-5398-4152
; Faculty of Economics Split, University of Split, Split, Croatia
Abstract
Time series models that are commonly used in econometric modeling are autoregressive stochastic linear models (AR) and models of moving averages (MA). Mentioned models by their structure are actually stochastic difference equations. Therefore, the objective of this paper is to estimate difference equations containing stochastic (random) component. Estimated models of time series will be used to forecast observed data in the future. Namely, solutions of difference equations are closely related to conditions of stationary time series models. Based on the fact that volatility is time varying in high frequency data and that periods of high volatility tend to cluster, the most successful and popular models in modeling time varying volatility are GARCH type models and their variants. However, GARCH models will not be analyzed because the purpose of this research is to predict the value of the exchange rate in the levels within conditional mean equation and
to determine whether the observed variable has a stable or explosive time path. Based on the estimated difference equation it will be examined whether Croatia is implementing a stable policy of exchange rates.
Keywords
difference equations; time series; econometric estimation; exchange rate
Hrčak ID:
94973
URI
Publication date:
22.12.2010.
Visits: 4.019 *