EXPONENTIAL SMOOTHING TECHNIQUE IN CORRELATION STRUCTURE FORECASTING OF VISEGRAD COUNTRY INDICES

The paper provides evidence on correlation structure forecasting techniques. We use index model parameters to forecast significant parts of securities returns volatility – systematic risk and specific risk. Except of this static perspective we have suggested, improved and dynamised these techniques with exponentially weighted moving average or EWMA variance and covariance forecast, which enables us to model time-varying beta coefficient as a significant part of systematic risk That enables us to construct correlation structure of national index returns we have selected for the explanation of our methodology. As we discuss in final part of the paper some adjustment in EWMA Beta and other variables have to be done through application of Blume or Vasicek techniques. So we can forecast an appropriate correlation structure of returns which enable us to be better aware about potential exposures in market returns.