Page 250 - Contributed Paper Session (CPS) - Volume 1

CPS1297 Chin W.C. et al.
outcomes suggested that the average forecasts gathered all the advantages
of each individual forecast and provided the best forecasts by averaging them.
In other words, it is worthy to implement the average forecasts in order to
obtain a more accurate forecast result.

The value-at-risk determination
For the application in finance, we computed the market risk for the
Mexican IPC using the value-at-risk approach. Three student-t models namely
the HAR(RV)-TGARCH, HAR(BV)-TGARCH and TGARCH are used for the
purpose of comparisons. In this specific evaluation, we only calculated the
one-day-ahead forecast and the student-t distributed return is obtained by
the AR-TGARCH model. According to the dynamic forecast evaluations in
Table 1, we can find that, in most of the times, combination forecasts present
better forecast outcomes compared to the individual models. Therefore, the
forecasts come from these methods may be more reliable for investors. The
overall market risks results are presented in Table 2.

Table 2: Value-at-Risk Determination based on actual BV (Dynamic forecast)
HAR(RV) - HAR(BV) - TGARCH Simple mean Simple Least-squares
TGARCH TGARCH median
(1) 0.00003490 0.00002320 0.00001250 0.00002035 0.00001789 0.00002324
2 ̂
VaR
5% -0.00956950 -0.00778527 -0.00569012 -0.00728579 -0.00682484 -0.00779196
quantile
5% VaR -95969.4991 -7785.2711 -5690.1161 -7285.7887 -6824.8422 -7791.9619
VaR
1% -0.01477851 -0.01203231 -0.00880755 -0.01126353 -0.01055407 -0.012014261
quantile
1% VaR -14778.5077 -12032.3112 -8807.5509 -11263.5322 -10554.0659 -12042.6092
Note: Value-at-risk calculates with $1 million of capital

4. Conclusion
This study uses a modified heterogeneous autoregressive model with
various high frequency realized volatility to re-examine the heterogeneous
market hypothesis for the Mexican stock market. The empirical discoveries
show that the jump-robust volatility outperformed the standard realized
volatility and ARCH-type volatility in the forecast evaluations. For better
forecast outcomes, the combination forecasts from three models are used and
the forecasts are utilized in determining value-at-risk. In conclusion, the study
enhances the literature on market information efficiency analysis especially in
the empirical case study of high frequency heterogeneous market hypothesis.
The empirical results offer an alternative way to forecast and determining
market risk particularly in the analysis of investment portfolio strategy and risk
management.

239 | I S I W S C 2 0 1 9