Skip to main content

Table 13 Out-of-sample ES forecast evaluation using joint scoring functions

From: Forecasting VaR and ES by using deep quantile regression, GANs-based scenario generation, and heterogeneous market hypothesis

 

FTSE100

N225

SPX500

DAX

SFZN

SFZ0

SFZN

SFZ0

SFZN

SFZ0

SFZN

SFZ0

\(\tau = 0.05\)

        

 GD

3.0831

4.8030

2.7255

2.1201

2.9327

3.1874

4.1507

7.6814

 KDE

3.1668

4.8842

2.7424

2.1253

2.9239

3.1650

4.2828

8.1125

 MCMC

3.1112

4.7895

2.7123

2.1101

2.9078

3.1540

4.1330

7.7470

 LSGAN

2.6262

2.2113

2.9775

2.2288

2.6070

1.7495

3.3790

2.6315

\(\tau = 0.01\)

        

 GD

4.1411

4.0132

2.6054

1.9298

3.8268

3.4683

3.8268

3.4683

 KDE

4.0906

3.9082

2.6791

1.9877

3.6573

3.2680

3.6573

3.2680

 MCMC

4.1067

3.9767

2.6171

1.9374

3.7924

3.4298

3.7924

3.4298

 LSGAN

3.1220

2.3192

2.7616

1.9758

3.2685

2.4135

3.2918

2.9071

  1. Bold text in columns indicates the best method