We adopt the principle of parsimony in heteroscedasticity models on eight insurance stocks. The results showed that the daily returns were stationary but not normally distributed and six out of eight stocks considered for the study showed evidence of ARCH effect. Post estimation and performance evolution metric was evaluated using the RMSE, MAE and MAPE but since the RMSE show a repeated values it is best to adopt the MAE. From the MAE, EGARCH (1, 1) outperformed other volatility model considered in this research. We can now conclude that the parsimony principle should be adopted in the study of volatility model because it captures the important feature of the data.
Download