Table 3 Forecasting improvement significance evaluated by bootstrap-based Diebold–Mariano (1995) statistic.

	h = 1	h = 2	h = 3	h = 4	h = 5
AEX	0.003***	0.028**	0.006***	0.001***	0.030**
ATG	0.035**	0.190	0.090*	0.184	0.154
ATX	0.035**	0.075*	0.051*	0.008***	0.072*
BFX	0.004***	0.002***	0.001***	0.000***	0.009***
FCHI	0.001***	0.021**	0.002***	0.000***	0.009***
FTMIB	0.062*	0.038**	0.014**	0.014**	0.067*
FTSE	0.020**	0.057*	0.003***	0.001***	0.008***
GDAXI	0.044**	0.073*	0.006***	0.008***	0.129
IBEX	0.084*	0.061*	0.040**	0.009***	0.060*
OMXSPI	0.012**	0.052*	0.006***	0.003***	0.035**
OSEAX	0.009***	0.368	0.043**	0.003***	0.178
PSI20	0.099*	0.121	0.022**	0.022**	0.101
RTS	0.033**	0.023**	0.009***	0.006***	0.024**
SSMI	0.053*	0.024**	0.010***	0.021**	0.157
XU100	0.191	0.033**	0.008***	0.002***	0.003***
WIG30	0.066*	0.103	0.005***	0.004***	0.007***
SSEC	0.014**	0.037**	0.067*	0.049**	0.041**
BSESN	0.142	0.049**	0.443	0.066*	0.015**
CSI300	0.005***	0.047**	0.114	0.068*	0.045**
HSI	0.001***	0.144	0.014**	0.002***	0.003***
KLSE	0.051*	0.122	0.120	0.007***	0.007***
KS11	0.136	0.252	0.147	0.006***	0.103
N225	0.041**	0.066*	0.073*	0.088*	0.257
STI	0.117	0.004***	0.010**	0.001***	0.000***
TASI	0.041**	0.023**	0.038**	0.058*	0.139
COLCAP	0.096*	0.001***	0.013**	0.001***	0.004***
DJI	0.000***	0.099*	0.015**	0.001***	0.001***
GSPC	0.000***	0.086*	0.037**	0.007***	0.021**
GSPTSE	0.056*	0.355	0.234	0.002***	0.006***
IXIC	0.003***	0.183	0.172	0.037**	0.095*
MXX	0.044**	0.235	0.139	0.025**	0.108
JTOPI	0.005***	0.071*	0.002***	0.000***	0.005***
AORD	0.017**	0.214	0.041**	0.004***	0.020**
Amount of significant DM statistics	29	22	26	32	24

This table reports the p-value of the statistical significance of the satellite-based container model relative to the prevailing mean model using a bootstrap-based Diebold and Mariano (1995) statistic. We draw inferences by using a stationary bootstrap produce of Politis and Romano (1994) under the null hypothesis that the equity premium is unpredictable. The number of resamples is set as 2000 when bootstrapping, and the block length will be optimally estimated from the data use the selection procedure of Patton et al. (2009). The sample period covers from 2019:01 to 2021:11. The asterisks ***, **, and * indicate statistical significance at the 1%, 5%, and 10% levels, respectively. The last line reports the amount of cases significant at the 10% level.

Search