The global latent factor and international index futures returns predictability

Abstract

This study investigates whether the global latent factor estimated using the three‐pass regression filter (TPRF) contains useful information for the development of a long‐short strategy for international index futures returns. Our results show that the long‐short portfolio strategy based on the TPRF forecast can provide higher excess profits than other competing approaches. We also find that the results remain unchanged within good and bad market conditions, the performance of the long‐short portfolio being better in bad market conditions than in good market conditions. Finally, the results remain unchanged even when the transaction costs are considered in the long‐short portfolio strategy based on the TPRF forecast.

Exploring the impact of oil security attention on oil volatility: A new perspective

“…Therefore, if factors related to proxy variables can be extracted, the accuracy of predictions can undoubtedly be improved. Chang et al (2022) find that the global potential factors constructed based on TPRF contain useful information for formulating long-short strategies for the returns of international index futures. The specifications of TPRF are as follows:…”

“…Therefore, if factors related to proxy variables can be extracted, the accuracy of predictions can undoubtedly be improved. Chang et al (2022) find that the global potential factors constructed based on TPRF contain useful information for formulating long‐short strategies for the returns of international index futures. The specifications of TPRF are as follows: Step 1 : Perform regressions of $${{GSVI}}_{i.t-1}$$ on $${{RV}}_{t}$$ for each keyword $$i$$, where$$\,{{RV}}_{t}$$ denotes the realized volatility. $$${{GSVI}}_{i,t-1}={\alpha }_{i}+{\varphi }_{i}{{RV}}_{t}+{e}_{i,t}.$$$…”

“…Huang et al (2015) confirm the superiority of TPRF. Chang et al (2022) significantly improve the ability to forecast the returns of international index futures by using the global potential factors extracted by TPRF. As a supervised dimensionality reduction technology, the effect of TPRF is significantly better than principal component analysis (PCA).…”