Oil futures volatility prediction: Bagging or combination?

This paper compares the predictive performance of the bagging method and traditional combination models for forecasting oil futures volatility, using economic policy uncertainty (EPU) indices and macroeconomic variables as predictors. Our empirical findings indicate that the bagging method outperforms the conventional combination models, demonstrating the effectiveness of machine learning combination models. These results are confirmed by different evaluation methods, alternative forecasting methods, and alternative oil futures, and hold up during the COVID-19 pandemic and various business cycles. Furthermore, we show that EPU indices are more useful than macroeconomic variables for forecasting oil volatility during the COVID-19 pandemic. Thus, our analysis provides new insights into combination forecasts.

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

The outbreak of the COVID-19 pandemic led to a slowdown in the world's energy trade and changes in the use of energy resources. Meanwhile, global conditions are complex and can affect fossil energy spot markets, including crude oil, gasoline, heating oil, and natural gas. In this paper, we conduct comparative research to explore the impact of global conditions on fossil energy spot markets during the COVID-19 crisis based on the GARCH-MIDAS framework. We employ a 2010–2022 sample, which we cut off to investigate the differences before and after COVID-19. In-sample estimation shows that all global indicators are significant for forecasting the volatilities of these fossil energy spot prices. Out-sample forecasts reveal that GEPU and GECON outperform GPR and WIP for forecasting these four markets during the pre-COVID-19 period. After the crisis broke out, these global indicators can provide different forecasting information. Hence, this paper can be helpful for decision-makers to formulate and adjust pertinent policies and investments in the case of extreme emergencies in the future. © 2023

Uncertainty indices and stock market volatility predictability during the global pandemic: evidence from G7 countries

This article attempts to examine the predictability of a significant number of uncertainty indices for the G7 stock market volatility based on a Heterogeneous AutoRegressive Realized Volatility (HARRV) model and a combination forecast framework during the global pandemic COVID-19. We include in our analysis the Infectious Disease Equity Market Volatility (IDEMV), the VIX, the Economic Policy Uncertainty (EPU), the Equity Market Volatility (EMV), the Geopolitical risk (GPR) as well as the crude oil futures' realized volatility. Out-of-sample evidence shows that models incorporating all uncertainty indices improve forecasting performance for most stock markets' volatility during a long out-of-sample period and also during the coronavirus period. The results are robust using an alternative volatility model, an alternative realized measure and a recursive window analysis. The predictability of the uncertainty indices is also evaluated through risk management and portfolio loss functions and results suggest that the LASSO combination and a HARRV model including all indices are the most profitable for all stock markets during the global pandemic. © 2023 Informa UK Limited, trading as Taylor & Francis Group.

Volatility forecasting with Hybrid-long short-term memory models: Evidence from the COVID-19 period

Volatility forecasting, a central issue in financial risk modelling and management, has attracted increasing attention after several major financial market crises. In this article, we draw upon the literature on volatility forecasting and hybrid models to construct the Hybrid-long short-term memory (LSTM) models to forecast the intraday realized volatility in three major US stock indexes. We construct the hybrid models by combining one or multiple traditional time series models with the LSTM model, and incorporating either the estimated parameters, or the predicted volatility, or both from the statistical models as additional input values into the LSTM model. We perform the out-of-sample test of our Hybrid-LSTM models in volatility forecasting during the coronavirus disease 2019 (COVID-19) period. Empirical results show that the Hybrid-LSTM models can still significantly improve the volatility forecasting performance of the LSTM model during the COVID-19 period. By analysing how the construction methods may influence the forecasting performance of the Hybrid-LSTM models, we provide some suggestions on their design. Finally, we identify the optimal Hybrid-LSTM model for each stock index and compare its performance with the LSTM model on each day during our sample period. We find that the Hybrid-LSTM models' great capability of capturing market dynamics explains their good performance in forecasting. © 2023 John Wiley & Sons Ltd.

Economic policy uncertainty and environmental governance company volatility: Evidence from China

This study primarily investigates whether China's economic policy uncertainty (EPU) can predict the environmental governance index volatility, which selects companies regarding environmental protection such as sewage treatment, solid waste treatment, air treatment, and energy saving. Empirical results reveal that China's EPU index can predict the environmental governance index volatility. Furthermore, even during periods of fluctuating volatility and the COVID-19 pandemic, China's EPU index can reliably forecast the environmental governance index volatility. This paper tries to provide new evidence regarding the connection between EPU and environmental governance companies' stock volatility. © 2023

Are categorical EPU indices predictable for carbon futures volatility? Evidence from the machine learning method

The purpose of this paper is to explore whether the categorical Economic Policy Uncertainty (EPU) indices are predictable for the volatility of carbon futures, in the mixed data sampling (MIDAS) regression framework. The prediction methods include the MIDAS-RV model, the MIDAS models extended by individual categorical EPU index, combination prediction approaches, the MIDAS models extended by dimensionality reduction techniques as well as the machine learning methods on the basis of MIDAS model and Markov regime switching method. We find firstly that categorical EPU indices are predictable for carbon futures volatility, but the predictive power of individual categorical EPU indices is not robust. Secondly, machine learning methods, especially the machine learning method considering the Markov regime switching structure, help to obtain valid information from multiple categorical EPU indices and produce robust and superior prediction accuracy for carbon futures volatility. The results of the extension analysis also found that machine learning methods, especially the machine learning method considering the Markov regime switching structure help to produce higher investment performance and more accurate long-term carbon futures volatility forecasts. Meanwhile, we also find the advantages of the MIDAS based machine learning methods over the traditional AR based machine learning methods. Finally, the forecasting performance of the machine learning method which considering Markov regime switching structure are superior during both the low and high volatility regimes and even during the COVID-19 pandemic. © 2022 Elsevier Inc.

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

The outbreak of the COVID-19 pandemic led to a slowdown in the world's energy trade and changes in the use of energy resources. Meanwhile, global conditions are complex and can affect fossil energy spot markets, including crude oil, gasoline, heating oil, and natural gas. In this paper, we conduct comparative research to explore the impact of global conditions on fossil energy spot markets during the COVID-19 crisis based on the GARCH-MIDAS framework. We employ a 2010–2022 sample, which we cut off to investigate the differences before and after COVID-19. In-sample estimation shows that all global indicators are significant for forecasting the volatilities of these fossil energy spot prices. Out-sample forecasts reveal that GEPU and GECON outperform GPR and WIP for forecasting these four markets during the pre-COVID-19 period. After the crisis broke out, these global indicators can provide different forecasting information. Hence, this paper can be helpful for decision-makers to formulate and adjust pertinent policies and investments in the case of extreme emergencies in the future.

Forecasting the Asian stock market volatility: Evidence from WTI and INE oil futures

This study investigates whether China's crude oil futures (INE) and West Texas Intermediate (WTI) markets hold valuable information for estimating the realized volatility of seven Asian stock markets. This study has several notable findings. First, China's oil futures can trigger forecast accuracy for three equity indices (Nikkei 225, NSEI, and FT Straits Times), whereas WTI helps forecast the volatility of the two indices (KSE 100 and KOSPI). Second, comparing China's crude oil futures with WTI's crude oil futures, we find that the former could be an effective indicator for all seven Asian stock markets during a high-volatility period, while WTI information is helpful in forecasting the volatility of the KSE 100, NSEI, and FT Strait Times during the low-volatility period. Further, information of both oil futures is ineffective for the Hang Seng and SSEC equity indices. Our results are robust in several robustness checks, including alternative evaluation methods, recursive window approach, and alternative realized measures, even during the COVID-19 pandemic.

Economic policy uncertainty and environmental governance company volatility: Evidence from China

This study primarily investigates whether China's economic policy uncertainty (EPU) can predict the environmental governance index volatility, which selects companies regarding environmental protection such as sewage treatment, solid waste treatment, air treatment, and energy saving. Empirical results reveal that China's EPU index can predict the environmental governance index volatility. Furthermore, even during periods of fluctuating volatility and the COVID-19 pandemic, China's EPU index can reliably forecast the environmental governance index volatility. This paper tries to provide new evidence regarding the connection between EPU and environmental governance companies' stock volatility.

Are categorical EPU indices predictable for carbon futures volatility? Evidence from the machine learning method

The purpose of this paper is to explore whether the categorical Economic Policy Uncertainty (EPU) indices are predictable for the volatility of carbon futures, in the mixed data sampling (MIDAS) regression framework. The prediction methods include the MIDAS-RV model, the MIDAS models extended by individual categorical EPU index, combination prediction approaches, the MIDAS models extended by dimensionality reduction techniques as well as the machine learning methods on the basis of MIDAS model and Markov regime switching method. We find firstly that categorical EPU indices are predictable for carbon futures volatility, but the predictive power of individual categorical EPU indices is not robust. Secondly, machine learning methods, especially the machine learning method considering the Markov regime switching structure, help to obtain valid information from multiple categorical EPU indices and produce robust and superior prediction accuracy for carbon futures volatility. The results of the extension analysis also found that machine learning methods, especially the machine learning method considering the Markov regime switching structure help to produce higher investment performance and more accurate long-term carbon futures volatility forecasts. Meanwhile, we also find the advantages of the MIDAS based machine learning methods over the traditional AR based machine learning methods. Finally, the forecasting performance of the machine learning method which considering Markov regime switching structure are superior during both the low and high volatility regimes and even during the COVID-19 pandemic.

An oil futures volatility forecast perspective on the selection of high-frequency jump tests

This paper examines the forecasting performances of high-frequency jump tests for oil futures volatility from a comprehensive perspective. It contributes to the literature by investigating which jump test is the best for oil futures volatility forecasting under different circumstances and whether the jump component extracted from multiple alternative tests is useful for further improving forecasting performance. Our results show that the jumps of the TOD test (Bollerslev et al., 2013) have satisfactory performance over the medium-term and especially the short-term forecasting horizons. Most importantly, the jump components from the intersection of multiple intraday tests further improve the forecasting performance. A variety of further discussions, including models controlling for stock market effects and considering periods of high (low) volatility and the COVID-19 pandemic period, confirm the conclusions. This paper attempts to shed light on oil futures volatility prediction from the perspective of jump test selection.

Forecasting Crude Oil Future Volatilities with a Threshold Zero-Drift GARCH Model

The recent price crash of the New York Mercantile Exchange (NYMEX) crude oil futures contract, which occurred on 20 April 2020, has caused history-writing movements of relative prices. For instance, the West Texas Intermediate (WTI) experienced a negative price. Explosive heteroskedasticity is also evidenced in associated products, such as the Intercontinental Exchange Brent (BRE) and Shanghai International Energy Exchange (INE) crude oil futures. Those movements indicate potential non-stationarity in the conditional volatility with an asymmetric influence of negative shocks. To incorporate those features, which cannot be accommodated by the existing generalized autoregressive conditional heteroskedasticity (GARCH) models, we propose a threshold zero-drift GARCH (TZD-GARCH) model. Our empirical studies of the daily INE returns from March 2018 to April 2020 demonstrate the usefulness of the TZD-GARCH model in understanding the empirical features and in precisely forecasting the volatility of INE. Robust checks based on BRE and WTI over various periods further lead to highly consistent results. Applications of news impact curves and Value-at-Risk (VaR) analyses indicate the usefulness of the proposed TZD-GARCH model in practice. Implications include more effectively hedging risks of crude oil futures for policymakers and market participants, as well as the potential market inefficiency of INE relative to WTI and BRE.

The role of uncertainty measures in volatility forecasting of the crude oil futures market before and during the COVID-19 pandemic

The purpose of this article is to investigate whether various uncertainty measures provide incremental information for the prediction the volatility of crude oil futures under high-frequency heterogeneous autoregressive (HAR) model specifications. Moreover, by considering the information overlap among various uncertainty measures and fully using of the information in various uncertainty measures, this paper uses two prevailing shrinkage methods, the least absolute shrinkage and selection operator (lasso) and elastic nets, to select uncertainty variables during the entire sampling period, before the COVID-19 pandemic and during the COVID-19 pandemic and then uses the HAR model to predict crude oil volatility. The results show that (i) uncertainty measures can be utilized to predict crude oil volatility under the high-frequency framework in both in-sample and out-of-sample analyses. (ii) Because of the information overlap between various uncertainty measures, adding a large number of uncertain variables to the HAR model may not significantly improve the volatility prediction. (iii) Before and during the COVID-19 pandemic, Chicago Board Options Exchange (CBOE) crude oil volatility (OVX) has the greatest impact on crude oil volatility, infectious disease equity market volatility (EMV) exerts a significant influence on crude oil futures volatility forecasts during the COVID-19 period, and CBOE implied volatility (VIX) and the financial stress index (FSI) have substantial impacts on crude oil futures volatility forecasts before COVID-19.

Forecasting green bond volatility via novel heterogeneous ensemble approaches

Green bonds are powerful tools for fighting against climate change and typically exhibit more volatility than conventional bonds do. However, the volatility forecasting of green bond has received little attention in previous literature. This study proposes two novel heterogeneous ensemble models, which differ from common volatility forecasting in that they are combine advanced tree-based ensemble models and exogenous predictors from other financial and commodity markets to forecast the volatility of green bonds. Validated on multiple green bonds indexes, loss functions, and time horizons, the comparative results show that the incorporation of exogenous predictors can enhance the predictive accuracy of volatility forecasting models, which is also confirmed by the marginal effects illustrated by SHapley Additive exPlanations (SHAP) values. The proposed EX-SEL model significantly outperforms the benchmark models in most cases. The results of the robustness check further indicate that the empirical results are robust to alternative volatility estimators, extreme events such as the COVID-19 pandemic, and alternative selection strategies.

The driving forces of green bond market volatility and the response of the market to the COVID-19 pandemic

The extant literature on green finance is mainly about its contribution to financing the transition to a low-carbon economy and the benefits it has brought to financial market participants’ portfolio diversification. As one of the most profound and important segments of green financial and fixed-income markets, the green bond market is at the centre of recent studies on green finance. This current research analysed the risky side of the green financial market by examining the response of the green bond market to extreme negative shocks, followed by a deep and comprehensive examination of the driving forces of the market’s volatility dynamics. This research contributes to the literature as follows: (1) by illustrating an event study that investigated the response of the green bond market to the shock of the COVID-19 pandemic, (2) by extending the volatility forecasting to the green bond market and (3) by revealing the driving forces of the green financial market’s volatility. The findings showed that (1) huge fluctuations and significant, negative abnormal returns appeared under the shock of the COVID-19 pandemic, (2) a green property or non-pecuniary property of a financial instrument does not help reduce the risk levels of a financial market under an extreme condition, (3) green bond volatility is largely driven by the uncertainty of the traditional fixed-income market, followed by the currency and stock markets and green infrastructure activities, (4) shocks coming from the energy market, including the renewable and non-renewable energy markets, green production and financial uncertainty, cannot spill into the green bond market and (5) markets with higher spillover effects generally produce more accurate volatility forecasts. However, forecasting accuracy decreases when a dynamic correlation becomes volatile.

Forecasting stock market volatility with a large number of predictors: New evidence from the MS-MIDAS-LASSO model.

This paper explores the effectiveness of predictors, including nine economic policy uncertainty indicators, four market sentiment indicators and two financial stress indices, in predicting the realized volatility of the S&P 500 index. We employ the MIDAS-RV framework and construct the MIDAS-LASSO model and its regime switching extension (namely, MS-MIDAS-LASSO). First, among all considered predictors, the economic policy uncertainty indices (especially the equity market volatility index) and the CBOE volatility index are the most noteworthy predictors. Although the CBOE volatility index has the best predictive ability for stock market volatility, its predictive ability has weakened during the COVID-19 epidemic, and the equity market volatility index is best during this period. Second, the MS-MIDAS-LASSO model has the best predictive performance compared to other competing models. The superior forecasting performance of this model is robust, even when distinguishing between high- and low-volatility periods. Finally, the prediction accuracy of the MS-MIDAS-LASSO model even outperforms the traditional LASSO strategy and its regime switching extension. Furthermore, the superior predictive performance of this model has not changed with the outbreak of the COVID-19 epidemic.

Jumps and stock market variance during the COVID-19 pandemic: Evidence from international stock markets.

Based on the work of Buncic and Gisler (2017), this paper investigates whether the roles of jump components will change in forecasting the volatility of international equity markets during the COVID-19 pandemic. Interestingly, in contrast to the conclusions of Buncic and Gisler (2017), we find jump components of the international equity indices are useful to predict the international stock markets' volatility during the COVID-19 pandemic. Our study tries to provide new evidence of jump components in stock markets.

Estimating fluctuating volatility time series returns for a cluster of international stock markets: A case study for Switzerland, Austria, China and Hong Kong

The major aim of this empirical study is to estimate the volatility time series returns for a cluster of international stock markets, such as: Switzerland, Austria, China and Hong Kong. The paper demonstrates statistical modeleling in order to capture volatility clusters and changes in long and short term volatility impact. The econometric approch is based on randomly selected daily closing return collected for the main indices of stock markets in Switzerland, Austria, China and Hong Kong for the sample period January 2003 to September 2021. We used various statistical properties to test normalities based on using GARCH family models for estimating financial market volatility. Moreover, the sampled time interval includes two extreme events such as the global financial crisis (GFC) of 2007–2008 and the recent COVID-19 pandemic. © 2021, Universitatea din Craiova. All rights reserved.

Can dimensional reduction technology make better use of the information of uncertainty indices when predicting volatility of Chinese crude oil futures?

In this paper, we try to forecast the volatility of Chinese crude oil futures (COF) using multiple economic policy uncertainty indicators. MIDAS-RV model is combined with the principal component analysis (PCA), scaled PCA (SPCA) and partial least squares (PLS) techniques in this work, construct the newly MIDAS-RV-PCA, MIDAS-RV-PLS and MIDAS-RV-SPCA models, their prediction performance is compared with the common combination forecasting methods. The in-sample estimation analysis on MIDAS-RV-X models show the that it is necessary to consider multiple economic policy uncertainty indices when predicting the Chinese COF volatility and the in-sample analysis on dimensionality reduction model further demonstrate the rationality of using dimensionality reduction techniques. The out-of-sample evaluation results show that the MIDAS-RV-SPCA is a superior model when forecasting the short-term volatility of Chinese COF using multiple economic policy uncertainty indicators, especially during the periods of high volatility and COVID-19 pandemic. The results also indicates that the DMSPE(0.9) method in the combination forecasting method shows its superior forecasting ability in long-term volatility of Chinese COF, especially during the low volatility and pre-pandemic period.

The role of coronavirus news in the volatility forecasting of crude oil futures markets: Evidence from China.

Based on the high-frequency heterogeneous autoregressive (HAR) model, this paper investigates whether coronavirus news (in China and globally) contains incremental information to predict the volatility of China's crude oil, and studies which types of coronavirus news can better forecast China's crude oil volatility. Considering the information overlap among various coronavirus news items and making full use of the information in various coronavirus news items, this paper uses two prevailing shrinkage methods, lasso and elastic nets, to select coronavirus news items and then uses the HAR model to predict China's crude oil volatility. The results show that (i) coronavirus news can be utilized to significantly predict China's crude oil volatility for both in-sample and out-of-sample analyses; (ii) the Panic Index (PI) and the Country Sentiment Index (CSI) have a greater impact on China's crude oil volatility. Additionally, China's Fake News Index (FNI) have a significant impact on China's crude oil volatility forecast; and (iii) global coronavirus news provides more incremental information than China's coronavirus news for predicting the volatility of China's crude oil market, which indicates that global coronavirus news is also a key factor to consider when predicting the market volatility of China's crude oil.