n African Finance Journal - New evidence on the predictability of South African fx volatility in heterogeneous bilateral markets
|Article Title||New evidence on the predictability of South African fx volatility in heterogeneous bilateral markets|
|© Publisher:||AfricaGrowth Institute|
|Journal||African Finance Journal|
|Author||Gordon H. Dash and Nina Kajiji|
|Publication Date||Jan 2003|
|Pages||1 - 15|
|Keyword(s)||C22, C45 and C53|
The purpose of this paper is to model the nonparametric realized volatility of the U.S. based futures contract for dollar exchange with the South African rand (ZAR). We find that the Kajiji-4 Bayesian regularization radial basis function neural network confirms the hypothesis that bilateral mineral alliances contribute to the observed volatility patterns of the ZAR contract. We also confirm the role of conditional volatility, trade-weighted state variables and news effects from the U.S. on the ZAR volatility prediction. Finally, the modelling results provide new evidence to support the heterogeneous trading hypothesis across the bilateral trade dimensions at the daily level.
Article metrics loading...