Volatility prediction for the energy sector with economic determinants: Evidence from a hybrid model

Yuejing Wang, Wuyi Ye, Ying Jiang, Xiaoquan Liu

Research output: Journal PublicationArticlepeer-review

1 Citation (Scopus)


Given close ties between energies and economic growth and evidence in the literature that fundamental information helps improve the pricing efficiency of energy products, in this study we examine volatility prediction for the U.S. energy sector considering the impact of economic variables. In particular, we develop a hybrid model that combines the GARCH-MIDAS model and LSTM neural network. This particular specification is motivated by the need to simultaneously take a large number of economic predictors into account and allow a flexible volatility component structure with potential nonlinear relation among economic determinants. Based on the sample period from January 1991 to September 2022, our empirical results show that the hybrid model generates statistically more precise volatility forecasts out of sample than a number of alternative models, and this is robust during the energy market turmoil brought by the onset of the COVID-19 pandemic and the Russian–Ukrainian clash. Finally, volatility forecasts from the hybrid model allow mean–variance utility investors to achieve higher economic value.

Original languageEnglish
Article number103094
JournalInternational Review of Financial Analysis
Publication statusPublished - Mar 2024


  • Economic gain
  • Energy market
  • Machine learning technique
  • Subsample analysis

ASJC Scopus subject areas

  • Finance
  • Economics and Econometrics


Dive into the research topics of 'Volatility prediction for the energy sector with economic determinants: Evidence from a hybrid model'. Together they form a unique fingerprint.

Cite this