How macro-variables drive crude oil volatility? Perspective from the STL-based iterated combination method

Challenges for volatility forecasts of US fossil energy spot markets during the COVID-19 crisis

“…We chose the GARCH-MIDAS framework as our research method for two reasons. First, the main goal of this paper is to study whether exogenous variables of monthly sampling frequency can provide effective information for predicting daily volatility And a large number of works have confirmed that the GARCH-MIDAS framework has the ability to reasonably construct a model of mixed-frequency data when matching such low-frequency variables (such as weekly, monthly, and quarterly) with high-frequency daily fluctuation indicators of asset prices ( Wang, Wu, et al, 2022 ; Zhang, Wang, et al, 2022 ). If the high-frequency information of daily prices is transformed into monthly comprehensive information, then a large amount of effective information will be lost, which will also bias the parameter estimation and model prediction ( Engle et al, 2013 ; Engle & Rangel, 2008 ).…”

“… defines realized volatility. Motivated by Zhang, Wang, et al (2022) , we can replace RV with by the following equation: …”

“…Motivated by Xu et al (2021) , we employ another widely used test for robustness checks. The MCS test has received much attention in the existing research ( Liang et al, 2020 ; Wang, Wu, et al, 2022 ; Zhang, Wang, et al, 2022 ). According to the MCS testing results in Table 7 , GECON can also provide the most useful information for predicting these energy markets before the COVID-19 outbreak.…”