Table 10

Portfolio construction and testing alpha

ModelAlphaBenchmarkNAdjusted R2
Panel A. Benchmark: GSCI index
SESTM0.008***−0.237***9260.018
(12.132)(−4.192)  
HIV−0.005***0.820***9260.205
(−8.611)(−15.486)  
LM−0.006***−0.882***9260.245
(−10.683)(−17.363)  
VADER−0.005***−0.228***9260.015
(−7.703)(−3.864)  
Panel B. Benchmark: USO ETF
SESTM0.008***−0.157***9260.024
(12.146)(−4.853)  
HIV−0.005***0.475***9260.207
(−8.513)(15.563)  
LM−0.007***−0.492***9260.231
(−10.701)(−16.663)  
VADER−0.005***−0.133***9260.015
(−7.698)(−3.926)  
Panel C. Benchmark: USL ETF
SESTM0.008***−0.192***9260.026
(12.211)(−5.116)  
HIV−0.005***0.513***9260.179
(−8.494)(14.245)  
LM−0.006***−0.572***9260.231
(−10.550)(−16.700)  
VADER−0.005***−0.163***9260.017
(−7.731)(−4.158)  

Note(s): This table presents the results of portfolio performance tests. Panel A reports the results using the GSCI index as the benchmark return, Panel B presents results using the USO ETF as the benchmark, and Panel C provides results using the USL ETF as the benchmark. The significance of alphas is calculated for various sentiment scores calculated from the SESTM model, HIV dictionary, LM dictionary, and VADER algorithm. t-statistics are in parentheses. *, **, and *** denote significance at the 10, 5, and 1% level, respectively

Source(s): Authors’ own work

[ViewLarge]
This Feature Is Available To Subscribers Only

or Create an Account

Close subscription notice
Close access options