Forecasting Power Evaluation: Evidence from Stock Market

High dimensional matrix of financial time-series is typical task for sensible estimation. There are different models for same purpose, such as correlation models to create and evaluate portfolio dynamic conditional correlation (DCC) and dynamic equicorrelation (DECO). The primary purposes of this study are comparing the above models and reducing the volatility of estimation results. Estimation results on the basis of the two models using international crude oil price and same but different three countries’ stock price index data, we investigate the better performance of the time-varying conditional correlation including lower volatility in DECO model than DCC model. There are also, we show that DECO model perform better in the out-of-sample forecasting performance than DCC model.