Research Article

A Comparative Study of VaR Estimation for Structured Products

Table 7

Summary of accuracy performance of selected time series models through various criterions.

Time series modelBacktesting Multivariate extension of Diebold and Mariano testRMSEMAE
WLHWLHWLHWLH

Panel A: RUD notes

AR( )-ARCH(5)NB BNBNB B B BNBNBNB BNB
AR( )-GARCH(1,1)NB BNB BNBNB BNB BNBNB B
AR( )-GARCH(1,4)NBNBNB BNBNB BNB BNBNB B
AR( )-EGARCH(1,1)NBNBNBNB BNBNB BNB BNBNB

Panel B: SPO notes

AR( )-ARCH(5)NB B BNB BNB B BNBNB B B
AR( )-GARCH(1,1)NBNBNB BNB B B B B BNB B
AR( )-GARCH(2,5) BNB B BNB B B B B BNB B
AR( )-EGARCH(1,1)NBNBNBNBNBNBNBNBNBNBNBNB

Panel C: DJQ notes

AR( )-ARCH(5)NBNBNBNBNB BNBNBNBNBNBNB
AR( )-GARCH(2,1) BNB B B BNB BNB BNB B B
AR( )-GARCH(2,5) BNBNB B BNB BNB BNB B B
AR( )-EGARCH(2,1) BNB BNBNB B B B B BNBNB
AR( )-EGARCH(2,2) BNB BNBNB B B B B BNBNB

Note that W, L, and H stand for the whole period, the low oil price period, and the high oil price period, respectively. NB and B denote “not better” and “better”, respectively. RMSE and MAE demonstrate root mean squared errors and mean absolute errors, respectively.