Research on the Value at Risk of Basis for Stock Index Futures Hedging in China Based on Two-State Markov Process and Semiparametric RS-GARCH Model

Table 10

The Kupiec backtesting of VaR of the basis for stock index futures.

Stock index futures	Long/Short	Model	Confidence level	Failure days	Failure rate	LR statistics
SSE 50		RS-GARCH	95%	22	2.88%	8.480
			97.5%	14	1.83%	1.537
			99%	9	1.18%	0.231
		semi-parametric RS-GARCH	95%	20	2.62%	10.969
			97.5%	13	1.70%	2.246
			99%	5	0.65%	1.050
		RS-GARCH	95%	13	1.70%	23.24
			97.5%	9	1.18%	6.792
			99%	7	0.92%	0.056
		semi-parametric RS-GARCH	95%	12	1.57 %	25.544
			97.5%	6	0.79%	12.533
			99%	3	0.39%	3.6997
CSI 500		RS-GARCH	95%	38	4.97%	0.001
			97.5%	27	3.53%	2.976
			99%	15	1.96%	5.591
		semi-parametric RS-GARCH	95%	36	4.71%	0.136
			97.5%	26	3.40%	2.301
			99%	15	1.96%	5.591
		RS-GARCH	95%	18	2.36%	13.868
			97.5%	12	1.57%	3.112
			99%	7	0.92%	0.055
		semi-parametric RS-GARCH	95%	14	1.83%	21.093
			97.5%	8	1.05%	8.441
			99%	5	0.65%	1.049
CSI 300		RS-GARCH	95%	89	4.51%	1.0466
			97.5%	44	2.23%	0.622
			99%	24	1.22%	0.865
		semi-parametric RS-GARCH	95%	85	4.30%	2.111
			97.5%	42	2.13%	1.1891
			99%	22	1.11%	0.249
		RS-GARCH	95%	77	3.9%	5.434
			97.5%	36	1.82%	4.096
			99%	11	0.56%	4.633
		semi-parametric RS-GARCH	95%	74	3.75%	7.123
			97.5%	35	1.77%	4.770
			99%	11	0.56%	4.633

Note: indicates that the LR statistic falls within the rejection domain.