New insights into price drivers of crude oil futures markets:Evidence from quantile ARDL approach

Hao-Lin Shaoa, Ying-Hui Shaob,∗, Yan-Hong Yangc

aDepartment of Statistics, Graduate School of Arts and Sciences, Columbia University, New York, NY 10027, USAbSchool of Statistics and Information, Shanghai University of International Business and Economics, Shanghai

201620, ChinacFaculty of Education, East China Normal University, Shanghai 200062, China

Abstract

This paper investigates the cointegration between possible determinants of crude oil futures pricesduring the COVID-19 pandemic period. We perform comparative analysis of WTI and newly-launched Shanghai crude oil futures (SC) via the Autoregressive Distributed Lag (ARDL) modeland Quantile Autoregressive Distributed Lag (QARDL) model. The empirical results confirm thateconomic policy uncertainty, stock markets, interest rates and coronavirus panic are importantdrivers of WTI futures prices. Our findings also suggest that the US and China’s stock marketsplay vital roles in movements of SC futures prices. Meanwhile, CSI300 stock index has a signifi-cant positive short-run impact on SC futures prices while S&P500 prices possess a positive nexuswith SC futures prices both in long-run and short-run. Overall, these empirical evidences providepractical implications for investors and policymakers.

Keywords: Crude oil futures; Stock markets; COVID-19 pandemic; Economic policyuncertainty; Quantile ARDLJEL classification:C5, Q4, Q43

1. Introduction

Crude oil futures is the most traded commodity in the world. It has been increasing crucial tofinancial markets and global economy. While there are multiple types of crude oil, the primaryinternational benchmark crude is West Texas Intermediate (WTI) which is traded continuously onthe New York Mercantile Exchange (NYMEX) from 6:00 pm to 5:00 pm EST. It is classified aslight sweet crude oil and traded in US dollars.

Understanding the determinants of the crude oil future market is of great interest. Previousresearch mainly examines the contributions of supply and demand fundamentals (Pan et al., 2017;Meng and Liu, 2019), macroeconomic factors (Aloui et al., 2016; Meng and Liu, 2019), financialfactors (Kyrtsou et al., 2016; Lu et al., 2020) and political factors (Chen et al., 2016; Qin et al.,

∗Corresponding author.Email address: yinghuishao@126.com (Ying-Hui Shao )

Preprint submitted to Elsevier October 7, 2021

arX

iv:2

110.

0269

3v1

[ec

on.E

M]

6 O

ct 2

021

2020b) to crude oil markets. Among extensive studies on the price drivers of crude oil market,much attention has focused on developed markets, not emerging market.

While crude oil prices used to be quoted in US dollar, China has been attempting to launcha yuan-denominated oil futures and internationalize the RMB. As the largest importer and majorconsumer of crude oil, China has a strong desire to establish an Asian crude oil benchmark andbecome an oil price setter. On March 26, 2018, China launched its crude oil futures on the Shang-hai International Energy Exchange (INE), which is traded in Chinese yuan. Shanghai crude oilfutures contracts (SC) is China’s first commodity derivative open to international investors. Dif-ferent from WTI crude, SC futures is medium sour crude oil traded on weekdays in day sessions(9:00 am–11:30 am and 1:30 pm–3:00 pm BJT) and night sessions (9:00 pm-2:30 am BJT).

With the potential for arbitrage trade with other financial markets, SC has a notable increase ininternational involvement. It has been the third most traded crude oil futures contract in the world.And the Chinese crude oil futures contract helps to reflect dynamic changes in crude oil supplyand demand of Asia-Pacific region. Hence, we conduct a comparative empirical analysis of WTIand SC to complement studies on crude oil price drivers. We reconsider the influence of economicpolicy uncertainty (EPU), stock market, and interest rate on crude oil futures.

Besides the above factors, crude oil futures price remains extremely fragile to extreme events.Since December 2019, the COVID-19 pandemic swept across the world and triggered investorpanic. The shock caused by the coronavirus created a turmoil in crude oil futures market. Theunprecedented and dramatic movements in crude oil market have got ample attention (Mensi et al.,2020b; Atri et al., 2021). Nevertheless, a few studies have addressed the long-run and short-run asymmetric impacts of COVID-19 pandemic on crude oil futures markets. To investigatethe temporary and long-lasting influences of shocks and other factors on crude oil markets, weemploy the Autoregressive Distributed Lag (ARDL) model (Pesaran et al., 1999) and the QuantileAutoregressive Distributed Lag (QARDL) model (Cho et al., 2015) during COVID-19 pandemicperiod. The methods can simultaneously measure the potential asymmetric effects of variables oncrude oil prices.

This paper makes the following extensions to the current literature. We examine how WTIand the newly shipped SC market reacts to factors in the long and short term. First, We findthat in the short-run, stock markets of China and US are strongly and positively correlated withcrude oil markets. However, in the long-run, the fluctuations in domestic stock market have limitedinfluence on crude oil prices. Correspondingly investors and regulators should watch stock marketsperformance closely. Second, we find that except stock market, other price drivers of WTI futureshave no effect on SC futures. This indicates the determinants of Chinese crude oil futures marketremain to be determined.

This paper is organized as follows. Section 2 reviews the relevant literature on determinationsof crude oil price. Section 3 reports source of data and describes the methodology. Section 4presents the empirical results of ARDL model and QARDL model. Section 5 reports the conclu-sion.

2

2. Literature Review

An array of studies have examined the time-varying effects of supply and demand fundamen-tals. Some studies point that the influence of supply shocks decrease over time. Supply shocks arenot the primary factors that drive the movements of crude oil prices (Baumeister and Peersman,2013). Oil demand plays a much more important role than the oil supply shocks (Kilian, 2009).After 2008 financial crisis, the roles of non-fundamentals have attracted considerable attention.The financial crisis changed market participants’ expectations and mechanisms of price forma-tion. Non commercial traders begin to play a vital role in crude oil markets (Joo et al., 2020).Therefore, our study do not concentrate on the direct effects of supply and demand shocks.

Constructed by Baker et al. (2016), economic policy uncertainty (EPU) index through affectingparticipants’ beliefs and macroeconomic activities to influence crude oil price movements (Alouiet al., 2016). USEPU plays a crucial role in determining other countries’ EPU and long-run WTIspot prices (Yang, 2019). Compared with other countries’ EPU indexes, Global EPU and USEPUare more informative in forecasting volatility of WTI prices (Wei et al., 2017). Existing studieshave not reached a consistent conclusion in the impact of EPU on crude oil markets. Some studiesfind USEPU is negative correlated with WTI returns (Zhang and Yan, 2020). The positive rela-tionship between USEPU and WTI prices has also been found (Lei et al., 2019). In addition, therelationship between EPU index and crude oil prices might vary over time. Lei et al. (2019) revealthat before 2008 financial crisis, USEPU has a negative impact on WTI returns. After the financialcrisis, higher USEPU is accompanied with higher WTI returns.

With continuous development of international financial markets, more and more financial in-stitutions have entered commodity markets. Financial factors play indispensable roles in crude oilrelated topics. Plenty of studies have found strong evidence of dependence between stock mar-kets and crude oil markets. Standard & Poor’s 500 (S&P500) index has been widely employedin crude oil forecasting models (Lu et al., 2020). In addition, a persistent lead-lag relationshipbetween S&P500 and WTI futures has been found (Kyrtsou et al., 2016). In recent years, Chinesegovernment attempts to internationalized mainland China’s financial market and attract more for-eign capitals. China’s stock market also has been observed a close relationship with WTI crudeoil prices (Guo et al., 2021). Using China Securities Index 300 (CSI300) and S&P500 as a rep-resentative financial index for China and US respectively, Zhu et al. (2021) find significant riskspillovers from China and US stock markets to crude oil markets.

Interest rates are often used as a proxy of macroeconomic activities in crude oil studies. Com-pared with short-term interest rates, crude oil prices are more sensitive to variations in long-terminterest rates. Furthermore, the influence of long-term interest rates is growing with time (Aroraand Tanner, 2013). There are some possible explanations of the linkage between crude oil marketsand interest rates. One of explanation is based on the theory of storage. Low interest rates wouldrequire more inventories and further change supply chain of products. Thus, interest rates arerelated with commodity prices. An alternative explanation is that interest rates can affect the de-mand of crude oil (Prest, 2018). More specifically, interest rates reflect country’s monetary policyand are highly correlated with inflation and economic prospects. Low interest rates would changeparticipants’ expectations and motivate companies to increase their spending (Mensi et al., 2020a).Another explanation is related to financialization of commodity markets. Fluctuations in interest

3

rates make market participants change their investment portfolios and choose to make investmentsin crude oil (Arora and Tanner, 2013).

COVID-19 pandemic has caused severe disruptions in global financial and commodity mar-kets. A plenty of studies have examined the effects of COVID-19 on crude oil markets. Duringpandemic period, crude oil markets become more inefficient (Mensi et al., 2020b). Pandemic-related news and reports spread people’s panic and increase investors’ anxiety. It has been provedthat the number of deaths and coronavirus panic have significant and negative impacts on WTIprices (Atri et al., 2021). The influence of COVID-19 pandemic on the volatility of WTI futureprices has exceeded 2008 financial crisis (Zhang and Hamori, 2021). There are relatively few stud-ies focus on the impact of COVID-19 pandemic on SC futures market. By applying high-frequencyheterogeneous autoregressive model, Niu et al. (2021) find coronavirus news can predict volatilityof SC futures prices.

Many studies have examined the derminants of WTI crude oil futures prices. However, thepotential nonlinear and asymmetric relationships between explanatory variables and WTI crudeoil prices are often ignored. In addition, SC futures is a newcomer of international crude oilmarkets. There are few studies concentrate on it. In this study, we apply the linear ARDL modeland QARDL model to investigate determinants of crude oil prices during COVID-19 pandemicperiod. The QARDL model has been widely employed in energy economics (Benkraiem et al.,2018; Apergis et al., 2021; Guo et al., 2021). This model can simultaneously address long-run andshort-run relationships between crude oil futures and explanatory variables across quantiles. Theresults of QARDL model are more robust than traditional linear models.

3. Data and Methodology

3.1. DataOur sample data comprises of daily observations of WTI futures prices, SC futures prices, US

economic policy uncertainty (USEPU) index, China economic policy uncertainty (CEPU) index,S&P500 prices, CSI300 prices, US 10-year treasury bond rate (Tbill), 1 week-Shanghai InterbankOffered Rate (SHIBOR) and Panic index. We use the closing prices of each futures and stockindex. For Panic index data series, each country’s first non-zero date is different. Therefore, thepanel data start from the earliest day that all observations are available. For US panel, the sampleperiod ranges from January 16th, 2020, to May 26th, 2021. For China panel, the sample periodranges from January 2nd, 2020, to May 26th, 2021. As the trading time of different markets varies,we eliminate the missing data and non-match data of each panel. The US economic policy uncer-tainty series come from the EPU website (http://www.policyuncertainty.com). Baker et al. (2016)’sCEPU index is measured at monthly frequency. Thus, we choose to use a new released CEPU in-dex which is constructed by Huang and Luk (2020). It is a daily index based on China mainlandnewspaper and is sourced from https://economicpolicyuncertaintyinchina.weebly.com. The US10-year Treasury Yield rate is gatherd from US department of the treasury website (https://www.treasury.gov). Panic index is obtained from Ravenpack (https://coronavirus.ravenpack.com). Thisindex measures the level of news that mentions hysteria or panic with coronavirus. Data of othervariables are collected from Wind database. In addition, we transfered all data series into naturallogarithmic form.

4

3.2. MethodologyTo examine the asymmetric and nonlinear effects of economic policy uncertainty, stock mar-

kets, interest rates and coronavirus panic on crude oil futures markets, we employ the ARDLmodel and QARDL model. The QARDL model is developed by Cho et al. (2015), it extends theARDL-ECM model into quantile regression context.

The traditional linear ARDL model is written as follows:

Oilt = α +

p∑i=1

φiOilt−i +

q1∑i=0

ωiEPUt−i +

q2∑i=0

λiS &P500t−i +

q3∑i=0

θiCS I300t−i +

q4∑i=0

ψiInterestt−i

+

q5∑i=0

δiPanict−i + εt

(1)where Oilt refers to WTI futures prices and SC futures prices, EPUt−i refers to USEPU and CEPU,and Interestt−i denotes Tbill and SHIBOR, respectively. εt is the error term which is defined asOilt − E[Oilt/Ft−1)]. Ft−1 is the smallest σ-field formed by {EPUt, S &P500t, CS I300t, Interestt,Panict, Oilt−1, EPUt−1, S &P500t−1, CS I300t−1, Interestt−1, Panict−1,...}. p, q1, q2, q3, q4, q5 arethe orders of lag. The ARDL model has some advantages over other normal linear model. Itallows mixed orders of integration (I0 and I1) and the lag lengths of variables can be different.In addition, with appropriate lags, the ARDL model can correct the problem of endogeneity andserial correlation (Pesaran et al., 1999). However, the linear ARDL model can only examine theaverage effects among variables. Therefore, following Cho et al. (2015), the linear ARDL modelis extended into quantile regression context as following equation:

QOilt = α(γ) +

p∑i=1

φi(γ)Oilt−i +

q1∑i=0

ωi(γ)EPUt−i +

q2∑i=0

λi(γ)S &P500t−i +

q3∑i=0

θi(γ)CS I300t−i

+

q4∑i=0

ψi(γ)Interestt−i +

q5∑i=0

δi(γ)Panict−i + εt(γ)

(2)

Eq. (2) is the basic form of QARDL(p, q) model, where γ is the quantile index relates to {0.1,0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8 and 0.9}. εt(γ) = Oilt − QOilt(γ/Ft−1) where QOilt(γ/Ft−1) refers toγ-th conditional quantile of QOilt (Kim and White, 2003). To avoid the potential serial correlationproblem, we rewrite Eq. (2) into the following form:

Q∆Oilt = α + ρOilt−1 + φEPU EPUt−1 + φS &P500S &P500t−1 + φCS I300CS I300t−1 + φInterestInterestt−1

+ φPanicPanict−1 +

p−1∑i=1

φi∆Oilt−i +

q1−1∑i=0

ωi∆EPUt−i +

q2−1∑i=0

λi∆S &P500t−i +

q3−1∑i=0

θi∆CS I300t−i

+

q4−1∑i=0

ψi∆Interestt−i +

q5−1∑i=0

δi∆Panict−i + νγ

(3)where ∆ refers to the first difference operator. In Eq. (3), ∆EPUt, ∆S &P500t, ∆CS I300t, ∆Interestt

5

and ∆Panict might have a contemporaneously correlation with νγ. To overcome this issue, we em-ploy the following projection of νγ on ∆EPUt, ∆S &P500t, ∆CS I300t, ∆Interestt and ∆Panict:

νγ = γEPU∆EPUt + γS &P500∆S &P500t + γCS I300∆CS I300t + γInterest∆Interestt + γPanic∆Panict + εt

Now, εt is uncorrelated with explanatory variables. Then, we apply this projection into Eq. (3) andobtain the following QARDL-ECM model:

Q∆Oilt = αγ + ργ(Oilt−1 + βEPU(γ)EPUt−1 + βS &P500(γ)S &P500t−1 + βCS I300(γ)CS I300t−1

+ βInterest(γ)Interestt−1 + βPanic(γ)Panict−1) +

p−1∑i=1

φi∆Oilt−i +

q1−1∑i=0

ωi∆EPUt−i

+

q2−1∑i=0

λi∆S &P500t−i +

q3−1∑i=0

θi∆CS I300t−i +

q4−1∑i=0

ψi∆Interestt−i +

q5−1∑i=0

δi∆Panict−i + εt(γ)

(4)The long-run parameters are defined as: βS &P500 = −

φS &P500ρ

, βCS I300 = −φCS I300

ρ, βEPU = −

φEPUρ

,βInterest = −

φInterestρ

, βPanic = −φPanicρ

. By applying the delta method, the cumulative short-run pa-rameters are defined as: φ∗ =

∑p−1i=1 φi, ω∗ =

∑q1−1i=0 ωi, λ∗ =

∑q2−1i=0 λi, θ∗ =

∑q3−1i=0 θi, ψ∗ =

∑q4−1i=0 ψi,

δ∗ =∑q5−1

i=0 δi. The ECM parameter is presented as ρ∗. The estimated value of ρ∗ should be negativeand significant.

Table 1: Descriptive statistics of daily logarithmic level series.

Minimum Maximum Mean Std. Dev Skewness Kurtosis Jarque-BeraPanel A: USWTI 2.2039 4.1954 3.7556 0.3352 -1.2664 2.2342 152.0261***USEPU 4.0337 6.6040 5.3833 0.5543 -0.1385 -0.6899 6.9764**S&P500 7.7131 8.3506 8.1339 0.1354 -0.5350 -0.0800 15.1719***CSI300 8.1691 8.6669 8.4358 0.1245 -0.3534 -1.0902 21.8000***Tbill -0.6539 0.6098 -0.0514 0.3668 0.3851 -1.3020 29.6462***Panic -4.6052 2.4230 1.0287 0.6199 -4.0816 32.9603 15292.9583***

Panel B: ChinaSC 5.4935 6.2275 5.8183 0.1681 0.3345 -0.9193 17.0280***CEPU 3.4809 6.1040 4.8158 0.4797 -0.1100 -0.0576 0.6741S&P500 7.7131 8.3506 8.1346 0.1332 -0.5498 0.0380 16.3667***CSI300 8.1691 8.6669 8.4347 0.1233 -0.3209 -1.0909 21.1308***SHIBOR 0.3954 1.1613 0.7549 0.1122 -0.4614 1.8765 60.1344***Panic -4.6052 3.1041 1.5333 0.9571 -3.0674 13.2414 2889.3626***

Notes: ***, ** and * indicate significance at 1%, 5% and 10% levels, respectively.

6

4. Result

Table 1 illustrates the descriptive statistics of daily logarithmic level series. The mean of WTIfutures prices is greater than mean of SC futures prices. The standard deviation of WTI futures is0.3352 while that statistics of SC futures is 0.1681. These results indicate that volatility for WTIfutures market is higher than SC market. Tbill and SC prices are left skewed while other dataseries are right skewed. The kurtosis statistics of two coronavirus panic series are greater than 3.The results of Jarque-Bera test depict that all of the variables except CEPU deviated from normaldistribution. These statistics results qualify QARDL model is an appropriate method for furtheranalysis.

The ARDL and QARDL model are preferable when variables are integrated of either I(0) orI(1). Therefore, we employ Augmented Dickey-Fuller (ADF) test and Phillips Perron (PP) test toconfirm none of variables is integrated of I(2). The results of unit root test are presented in Table 2.The findings show that all variables are stationary at the first difference. Hence, we can proceedour analysis.

Table 2: Results of unit root test.

ADF(level) ADF(∆) PP(level) PP(∆)Panel A: USWTI -2.9957 -5.5761*** -13.1414 -329.9850***USEPU -3.2244* -8.7544*** -64.4737*** -401.5527***S&P500 -3.2079* -5.9811*** -16.4971 -454.1951***CSI300 -2.1788 -7.2543*** -12.7325 -305.6564***Tbill -3.0488 -8.1289*** -8.5860 -274.7621***Panic -2.8220 -9.4032*** -71.4335*** -412.4296***

Panel B: ChinaSC -2.8159 -6.8910*** -7.7497 -299.8715***CEPU -6.4799*** -11.1268*** -262.3458*** -335.1428***S&P500 2.7251 -6.0310*** -13.3521 -464.9241***CSI300 -2.5621 -7.3213*** -12.1749 -311.1095***SHIBOR -3.1304* -8.6182*** -36.1086*** -264.3322***Panic -7.3652*** -8.6817*** -49.9165*** -393.1931***

Notes: ADF is the Augmented Dickey-Fuller test of unit root, PP is PhillipsPerron test of unit root. The null hypothesis for the ADF and PP is that seriesis non-stationary. ***, ** and * indicate significance at 1%, 5% and 10% levels,respectively.

The results of linear ARDL and QARDL model estimation are presented in Table 3 and Table 4.We report the error correction parameters (ρ∗), the long-run cointegrating coefficients (β) andshort-run coefficients (φi, ωi, λi, θi, ψi, δi), along with their standard errors in brackets. Fig. 1 andFig. 2 show the dynamic trends of estimated parameters with a 95% confidence interval acrossvarious quantiles. The figures indicate that most of estimated parameters perform differently ateach quanitles.

7

Figure 1: 95% confidence intervals of the QARDL model parameters (US). Horizontal axis indicates the quantilelevels (0.1, 0.2, ..., 0.8, 0.9). Vertical axis indicates coefficient estimates of parameters.

8

Figure 2: 95% confidence intervals of the QARDL model parameters (China). Horizontal axis indicates the quantilelevels (0.1, 0.2, ..., 0.8, 0.9). Vertical axis indicates coefficient estimates of parameters.

9

The ECM parameter measures how strongly crude oil futures prices react to the deviationsfrom the long-run equilibrium relationships. The findings of linear ARDL model show that theECM coefficient is -0.1082 for US panel and -0.0330 for China panel which means the adjustmentspeed is 10.82% and 3.3%, respectively. Therefore, the adjustment speed of SC futures is slower.For US panel, S&P500 and Tbill are positively correlated with WTI futures prices in long-run.Furthermore, it can be observed that βPanic and δ0 are significant. These results indicate an inversecorrelation between WTI futures prices and Panic index both in long run and short run. In terms ofChina panel, the short-run parameters of S&P500 and CSI300 are significant. More specifically,a 1% increase in S&P500 would cause SC futures prices increasing by 23.83%. For every 1%increase in CSI300, SC futures prices would increase by 46.73%. These results indicate thatdomestic stock market has a stronger effect on SC futures prices than US stock market. However,the linear ARDL model cannot show the asymmetric relationships between variables. Therefore,we employ the quantile ARDL model.

The results of QARDL estimation explicate that ρ∗ is significant and negative in five out of ninequantiles for WTI futures and four out of nine quantiles for SC futures. The empirical findings alsoindicate that φ4 is positive at 20% quantile for US panel. It becomes negative at higher quanitles. Inother words, the past changes in WTI futures prices exhibit a heterogeneous and opposite influenceon current price levels at extreme quantiles. For China panel, the previous variations in crude oilprices have no effect on current price levels.

Turning to the results of EPU parameters, we observe that βUSEPU is positive and significant atthe highest quantile. It indicates an upward trending long-run relationship between extreme WTIfutures price and USEPU. The short-run parameter ω0 is negative at 10% quantile and positive at90% quantile. Our results are partly consistent with Qin et al. (2020a). They demonstrate thatthe effects of USEPU on WTI prices are both negative and positive. Sudden events or policieswould change people’s sentiments and drive them to change their investment strategies. For Chinapanel, both βCEPU and ω0 are statistically insignificant across all quantiles. It reveals that duringour sample period, the changes in CEPU have no impact on SC futures prices. Our results arenot aligned with prior research by Yi et al. (2021). They use Baker et al. (2016)’s monthly CEPUindex and report that CEPU has a significantly negative impact on the long-term volatility of SCfutures.

In term of stock markets, for US panel, the empirical findings show that the short-run parameterλ0 is significant at all quantiles. However, in the long-run, S&P 500 has no impact on WTI futuresprices. Furthermore, S&P500 plays an important role in SC markets both in short-run and long-run. λ0 is positive and significant across seven quantiles. βSP500 is significant at quantiles 10% to30% and 60%. These findings highlight the strong correlation between US stock market and crudeoil markets. The linkage between S&P500 and crude oil markets has been widely observed andconfirmed in the previous studies (Kyrtsou et al., 2016; Zhang and Wang, 2019). In the case ofCSI300, for US panel, βCSI300 is positive and significant at 70% quantile and 90% quantile. θ0 issignificant at low and medium quantiles and indicates a long-run equilibrium between CSI300 andWTI futures prices. CSI300 is also positively correlated with SC prices in short-run. However, thelong-run parameter of CSI300 is insignificant. In summary, our results confirm volatility spilloversfrom stock markets to crude oil markets. In the short-run, crude oil prices are more sensitive to thechanges in domestic stock market. However, in the long-run, domestic stock market have limited

10

influence. The price movements of WTI futures are driven by CSI300. Similarly, SC futures pricesare driven by variations in S&P500.

Our results also indicate that current changes in Tbill are positively correlated with WTI futuresprices while past changes in Tbill have a negative impact on it. The co-movements betweeninterest rates and crude oil prices are sensitive to abnormal global and macro-economic events(Mensi et al., 2020a). The nexus between US interest rates and WTI prices have been reported inprevious literature (Arora and Tanner, 2013). Interest rates might change levels of inventory andmarket participants’ expectations (Wang and Chueh, 2013; Mensi et al., 2020a). Consequently,interest rates would affect the supply and demand fundamentals and drive oil prices. For Chinapanel, SHIBOR is not correlated with SC prices in any case. Due to the outbreak of COVID-19pandemic, global oil prices dropped sharply on 2020. At that time, Chinese companies importeda large volume of crude oil. Therefore, variations in interest rates have no substantial effect oninventory levels of China.

Furthermore, the results exhibit that coronavirus panic has a short-term negative impact onWTI futures prices. Our findings are consistent with results reported by Atri et al. (2021). COVID-19 pandemic has a profound influence on supply and demand of crude oil. In 2020, global aggre-gate demand for oil dropped significantly due to travel restrictions. The crude oil producers of USreduced drilling activity in response to rapid decline in oil demand. Therefore, there is a negativelinkage between Panic index and WTI futures prices. Interestingly, βPanic and δ0 are insignificantin China panel. It means that COVID-19 pandemic has limited influence on SC futures prices.These results do not support Niu et al. (2021) who find coronavirus panic has significant influ-ence on volatility of SC prices. Our results are supported by the fact that China’s total oil importsincreased by 7.3% in 2020. As the largest oil-importing country, the sharp decline in global oildemand has a relatively small impact on China. Therefore, Panic index cannot drive SC futuresprices. In summary, our results demonstrate that oil-importing country and oil-exporting mightrespond differently to COVID-19 pandemic shocks.

11

Tabl

e3:

Lin

earA

RD

Lan

dQ

AR

DL

estim

atio

nre

sults

forU

S.

Lin

earA

RD

Lα

ρ∗

βU

SEP

Uβ

S&

P50

0β

CS

I300

βTb

illβ

Pani

cφ

1φ

2φ

3φ

4ω

0λ

0λ

1θ 0

ψ0

ψ1

δ 0-1

.031

6***

-0.1

082*

**0.

0029

0.05

150.

1249

0.00

98-0

.046

7***

-0.0

669

-0.0

843

-0.1

915*

**0.

1553

***

-0.0

062

0.97

38**

*-0

.293

90.

2012

0.12

95**

-0.0

017

-0.0

315*

*(0

.396

9)(0

.026

3)(0

.015

1)(0

.094

9)(0

.081

4)(0

.017

6)(0

.015

8)(0

.058

3)(0

.053

2)(0

.055

0)(0

.054

9)(0

.012

7)(0

.224

0)(0

.234

8)(0

.264

8)(0

.065

8)(0

.065

3)(0

.015

6)

QA

RD

Lα

ρ∗

βU

SEP

Uβ

S&

P50

0β

CS

I300

βTb

illβ

Pani

cφ

1φ

2φ

3φ

4ω

0λ

0λ

1θ 0

ψ0

ψ1

δ 00.

1-0

.989

60.

0304

0.10

38-5

.370

71.

7290

2.16

461.

1672

-0.1

173

0.02

31-0

.172

50.

2186

-0.0

219*

0.99

13**

0.29

510.

3407

0.08

470.

0209

-0.0

232

(0.9

017)

(0.0

867)

(0.5

450)

(19.

5559

)(5

.469

7)(5

.716

5)(3

.624

9)(0

.258

2)(0

.173

3)(0

.135

9)(0

.138

3)(0

.011

5)(0

.419

0)(0

.332

6)(0

.284

0)(0

.098

6)(0

.086

9)(0

.017

8)0.

2-1

.219

7**

-0.0

114

0.21

9614

.519

4-0

.874

1-4

.176

6-3

.658

70.

0504

-0.0

878

-0.1

506

0.25

63**

-0.0

095

0.95

05**

*0.

0534

0.37

35*

0.06

27-0

.010

6-0

.026

4**

(0.5

684)

(0.0

523)

(1.2

591)

(61.

6178

)(7

.036

1)(2

0.05

38)

(16.

1686

)(0

.184

7)(0

.114

3)(0

.098

1)(0

.117

3)(0

.007

9)(0

.310

2)(0

.278

0)(0

.198

2)(0

.088

7)(0

.072

8)(0

.010

6)0.

3-0

.870

6**

-0.0

314

0.16

873.

3048

0.50

60-0

.535

7-0

.966

30.

0070

-0.0

719

-0.0

825

0.12

140.

0011

0.93

40**

*0.

0753

0.23

650.

0916

-0.0

070

-0.0

173*

(0.4

280)

(0.0

439

(0.3

186)

(4.1

220)

(1.7

316)

(1.1

910)

(1.2

182)

(0.1

514)

(0.0

922)

(0.0

862)

(0.1

193)

(0.0

065)

(0.2

904)

(0.2

387)

(0.1

634)

(0.0

849)

(0.0

727)

(0.0

094)

0.4

-0.9

479*

*-0

.067

20.

0646

1.55

160.

6253

-0.0

787

-0.4

167*

-0.0

457

-0.0

331

-0.0

821

0.08

22-0

.000

80.

6741

***

-0.1

560

0.29

99*

0.15

34**

-0.0

046

-0.0

204*

*(0

.383

0)(0

.041

2)(0

.116

1)(1

.068

5)(0

.738

3)(0

.274

0)(0

.225

2)(0

.128

8)(0

.082

0)(0

.087

1)(0

.112

6)(0

.006

0)(0

.243

3)(0

.221

8)(0

.160

4)(0

.077

3)(0

.065

9)(0

.009

4)0.

5-0

.860

2**

-0.1

027*

*0.

0440

1.09

330.

3887

0.14

46-0

.218

8**

-0.1

235

-0.0

805

-0.0

233

0.03

960.

0012

0.64

20**

*-0

.122

30.

2491

*0.

1565

**-0

.008

0-0

.016

6*(0

.415

1)(0

.045

8)(0

.078

6)(0

.674

1)(0

.441

1)(0

.140

2)(0

.101

1)(0

.136

9)(0

.080

7)(0

.094

4)(0

.114

1)(0

.005

6)(0

.208

4)(0

.225

7)(0

.145

5)(0

.074

9)(0

.068

9)(0

.009

8)0.

6-0

.795

7*-0

.114

2**

0.09

520.

7590

0.51

390.

291*

*-0

.180

8**

-0.1

909

-0.0

899

-0.0

525

0.04

370.

0055

0.77

60**

*0.

0325

0.16

320.

1111

0.00

73-0

.016

3(0

.418

3)(0

.049

1)(0

.073

0)(0

.592

7)(0

.362

7)(0

.119

0)(0

.086

9)(0

.139

3)(0

.080

1)(0

.103

2)(0

.119

2)(0

.005

7)(0

.181

7)(0

.236

1)(0

.135

4)(0

.073

0)(0

.064

4)(0

.009

9)0.

7-0

.839

5*-0

.171

6***

0.06

000.

4940

0.54

03**

0.33

19**

*-0

.104

1-0

.109

2-0

.081

0-0

.140

8-0

.074

10.

0051

0.68

01**

*0.

0931

0.07

260.

1580

**-0

.022

9-0

.012

6(0

.439

3)(0

.049

7)(0

.049

7)(0

.376

2)(0

.233

5)(0

.093

2)(0

.064

5)(0

.134

2)(0

.089

9)(0

.106

4)(0

.131

1)(0

.006

6)(0

.201

8)(0

.234

1)(0

.126

9)(0

.071

0)(0

.056

7)(0

.010

6)0.

8-0

.664

0-0

.212

1***

0.03

570.

5127

0.32

340.

3733

***

-0.0

228

-0.0

206

0.03

83-0

.119

5-0

.230

7*0.

0028

0.47

29**

-0.0

418

0.04

650.

1500

**-0

.099

8*0.

0010

(0.4

303)

(0.0

445)

(0.0

437)

(0.3

208)

(0.2

005)

(0.0

854)

(0.0

521)

(0.1

196)

(0.0

844)

(0.0

950)

(0.1

226)

(0.0

077)

(0.2

131)

(0.2

500)

(0.1

337)

(0.0

703)

(0.0

567)

(0.0

108)

0.9

-0.7

632*

-0.2

255*

**0.

0849

*0.

4057

0.43

92**

0.39

18**

*-0

.102

8**

-0.1

038

0.02

07-0

.093

1-0

.293

1***

0.01

34*

0.51

07**

0.04

800.

0415

0.11

55-0

.066

9-0

.017

3(0

.453

2)(0

.039

8)(0

.043

4)(0

.331

8)(0

.214

0)(0

.084

6)(0

.049

5)(0

.102

4)(0

.081

8)(0

.090

0)(0

.097

1)(0

.007

9)(0

.217

2)(0

.247

1)(0

.164

7)(0

.072

1)(0

.058

1)(0

.011

3)

Not

es:T

hest

anda

rder

rors

are

inpa

rent

hese

s.**

* ,**an

d*

indi

cate

sign

ifica

nce

at1%

,5%

and

10%

leve

ls,r

espe

ctiv

ely.

12

Tabl

e4:

Lin

earA

RD

Lan

dQ

AR

DL

estim

atio

nre

sults

forC

hina

.

Lin

earA

RD

Lα

ρ∗

βC

EP

Uβ

S&

P50

0β

CS

I300

βSH

IBO

Rβ

Pani

cφ

1φ

2ω

0λ

0λ

1θ 0

θ 1ψ

0δ 0

-0.1

025

-0.0

330*

**0.

0010

0.03

370.

0028

-0.0

082

-0.0

020

0.08

430.

0334

-0.0

023

0.23

83**

*0.

0442

0.46

73**

*0.

0226

-0.0

283

-0.0

009

(0.1

090)

(0.0

123)

(0.0

037)

(0.0

287)

(0.0

269)

(0.0

145)

(0.0

019)

(0.0

573)

(0.0

535)

(0.0

028)

(0.0

768)

(0.0

786)

(0.0

986)

(0.0

971)

(0.0

248)

(0.0

025)

QA

RD

Lα

ρ∗

βC

EP

Uβ

S&

P50

0β

CS

I300

βSH

IBO

Rβ

Pani

cφ

1φ

2ω

0λ

0λ

1θ 0

θ 1ψ

0δ 0

0.1

-0.1

093

-0.0

411

-0.0

590

2.22

2*-1

.192

50.

2322

-0.0

755

0.19

260.

0598

-0.0

076

0.27

060.

0805

0.32

92*

0.25

32-0

.035

00.

0006

(0.2

497)

(0.0

273)

(0.1

494)

(1.1

532)

(-1.

2320

)(0

.733

5)(0

.103

4)(0

.123

2)(0

.104

8)(0

.006

0)(0

.179

8)(0

.190

1)(0

.194

6)(0

.196

6)(0

.062

2)(0

.004

9)0.

2-0

.083

5-0

.042

3**

0.11

071.

5280

**-0

.720

90.

7007

-0.0

205

0.09

93-0

.025

1-0

.000

70.

2836

**-0

.026

70.

3972

***

0.12

900.

0185

0.00

06(0

.177

4)(0

.019

9)(0

.148

9)(0

.749

1)(0

.864

1)(0

.540

9)(0

.082

7)(0

.094

9)(0

.066

9)(0

.004

0)(0

.122

8)(0

.131

4)(0

.149

4)(0

.133

3)(0

.038

2)(0

.003

5)0.

3-0

.008

2-0

.042

2**

0.12

251.

3163

*-0

.699

80.

6060

-0.0

748

0.04

200.

0474

0.00

020.

2882

**-0

.004

90.

3212

**0.

1497

0.02

43-0

.002

5(0

.181

5)(0

.016

8)(0

.157

6)(0

.717

0)(0

.737

7)(0

.455

2)(0

.067

5)(0

.088

9)(0

.072

6)(0

.003

8)(0

.117

4)(0

.117

3)(0

.152

2)(0

.135

5)(0

.031

6)(0

.003

1)0.

4-0

.078

9-0

.031

2**

0.03

471.

2800

-0.2

925

0.25

81-0

.080

9-0

.023

5-0

.011

0-0

.000

50.

4080

***

0.13

180.

4449

***

0.07

11-0

.004

8-0

.002

5(0

.162

4)(0

.013

9)(0

.180

8)(0

.943

6)(0

.993

2)(0

.514

9)(9

0.07

60)

(0.0

795)

(0.0

674)

(0.0

038)

(0.1

075)

(0.1

047)

(0.1

565)

(0.1

253)

(0.0

276)

(0.0

030)

0.5

-0.2

095

-0.0

202

0.15

901.

2000

0.68

50-0

.042

0-0

.086

3-0

.048

1-0

.040

40.

0010

0.33

50**

*0.

1301

0.53

31**

*0.

0438

-0.0

217

-0.0

036

(0.1

271)

(0.0

141)

(0.2

793)

(1.3

088)

(1.6

873)

(0.7

563)

(0.1

107)

(0.0

728)

(0.0

645)

(0.0

033)

(0.0

991)

(0.0

944)

(0.1

562)

(0.1

003)

(0.0

245)

(0.0

027)

0.6

-0.2

087*

-0.0

273*

0.04

522.

1777

**-0

.490

2-0

.207

7-0

.032

4-0

.056

4-0

.028

0-0

.000

30.

3625

***

0.12

460.

4881

***

0.07

28-0

.027

7-0

.003

1(0

.124

4)(0

.015

6)(0

.157

2)(1

.013

6)(0

.853

5)(0

.543

3)(0

.068

7)(0

.069

0)(0

.062

9)(0

.003

1)(0

.084

2)(0

.095

4)(0

.123

6)(0

.107

9)(0

.025

2)(0

.002

5)0.

7-0

.182

3-0

.019

00.

1570

2.67

22-0

.736

3-0

.537

3-0

.051

1-0

.063

50.

0018

-0.0

019

0.38

07**

*0.

1209

0.44

38**

*0.

0725

-0.0

202

-0.0

033

(0.1

311)

(0.0

186)

(0.2

539)

(1.6

998)

(1.2

355)

(1.2

106)

(0.1

046)

(0.0

757)

(0.0

783)

(0.0

032)

(0.0

989)

(0.1

102)

(0.1

178)

(0.1

273)

(0.0

265)

(0.0

025)

0.8

-0.1

067

-0.0

217

0.11

031.

5666

-0.1

232

-1.1

455

-0.0

310

-0.0

062

-0.0

718

-0.0

017

0.28

08*

0.11

910.

4101

**-0

.051

6-0

.034

7-0

.002

7(0

.179

2)(0

.025

2)(0

.246

6)(1

.527

1)(2

.047

6)(2

.066

1)(0

.099

0)(0

.105

1)(0

.094

6)(0

.004

4)(0

.149

2)(0

.147

9)(0

.168

2)(0

.195

9)(0

.032

2)(0

.003

2)0.

9-0

.199

9-0

.019

90.

2227

-0.5

307

2.71

74-3

.058

6-0

.112

00.

0879

-0.0

342

-0.0

010

0.14

950.

1028

0.29

90-0

.155

1-0

.060

7-0

.003

2(0

.244

2)(0

.029

1)(0

.417

7)(4

.181

0)(6

.651

7)(5

.220

9)(0

.184

3)(0

.128

4)(0

.108

8)(0

.005

6)(0

.192

6)(0

.236

2)(0

.231

9)(0

.208

3)(0

.037

3)(0

.004

7)

Not

es:T

hest

anda

rder

rors

are

inpa

rent

hese

s.**

* ,**an

d*

indi

cate

sign

ifica

nce

at1%

,5%

and

10%

leve

ls,r

espe

ctiv

ely.

13

5. Conclusion

This study employs the ARDL model and QARDL model to explore the impact of economicpolicy uncertainty, stock markets, interest rates and coronavirus panic on crude oil markets duringCOVID-19 pandemic period. This paper finds some interesting results of these price drivers.First, compared with SC futures market, WTI futures market is more sensitive to the fluctuationsin economic policy uncertainty. USEPU has a significant impact on extreme WTI futures priceswhile CEPU has no influence on SC futures prices. Second, our results support the presence ofspillover effects between stock markets and crude oil markets. More specifically, in the short-run,crude oil market is positively and strongly influenced by domestic stock market. In the long-run,WTI futures prices are driven by variations in CSI300. Similarly, the movements of S&P500plays a crucial role in determining changes in SC futures prices. Third, the results indicate thatWTI futures prices are driven by US interest rates both in long-run and short-run. However,SHIBOR does not have impact on SC futures prices. Fourth, we observe a significant negativelinkage between coronavirus panic and WTI futures prices. However, panic is insignificant forChina panel over all quantiles.

In conclusion, the determinants of WTI futures and SC futures prices are quiet different. Thereexist several possible explanations for the results. First, the US is an oil-exporting country whileChina is an oil-importing country. Therefore, the direct (such as oil production) and indirect (suchas interest rates) effects of supply and demand shocks are different in US and China. Second,WTI futures and SC futures have diverse properties, including different trading regulations andinvestment environment. As a new-developed market, SC has limited pricing power and is lessefficient in market operation mechanism (Zhang et al., 2021). Third, our explanatory variablesare chosen based on determinants of mature crude oil markets. These factors might have limitedeffects on the newly launched market during our sample period. For these reasons, some factorsplay prominent roles in movements of WTI futures prices but have no influence on SC market.Further study need to be done to examine other factors’ impact on SC futures, especially theeffects of supply and demand fundamentals.

Our study provides practical and significant implications for investors and policymakers. WTIfutures market is driven by economic policy uncertainty, stock markets, interest rates and coro-navirus panic. Therefore, portfolio managers and individual investors should pay attention tofluctuations in stock markets to manage risk efficiently. When adjusting monetary policies, regu-lators should concern financial stability and transmission mechanism. According to World HealthOrganization, impacts of COVID-19 pandemic would be long-lasting. To maintain the stabilityof crude oil futures market, US government should restore investors’ sentiment and oversee risksfrom financial markets.

Although Chinese stock market has a strong positive effect on Shanghai crude oil futures in theshort-run, the market participants also need to concern risks and volatility of US stock market. SCfutures still has a long way to become an Asian benchmark. China government should improve theinvestment environment and trading mechanism of SC futures market. Further research is requiredto better understand the price mechanism of SC futures as well as its relationship with internationaloil benchmarks.

14

References

Aloui, R., Gupta, R., Miller, S.M., 2016. Uncertainty and crude oil returns. Energy Economics 55, 92–100. doi:10.1016/j.eneco.2016.01.012.

Apergis, N., Hayat, T., Saeed, T., 2021. Us partisan conflict uncertainty and oil prices. Energy Policy 150, 112–118.doi:10.1016/j.enpol.2020.112118.

Arora, V., Tanner, M., 2013. Do oil prices respond to real interest rates? Energy Economics 36, 546–555. doi:10.1016/j.eneco.2012.11.001.

Atri, H., Kouki, S., imen Gallali, M., 2021. The impact of covid-19 news, panic and media coverage on the oil and goldprices: An ardl approach. Resources Policy 72, 102061. doi:https://doi.org/10.1016/j.resourpol.2021.102061.

Baker, S.R., Bloom, N., Davis, S.J., 2016. Measuring Economic Policy Uncertainty. The Quarterly Journal ofEconomics 131, 1593–1636. doi:10.1093/qje/qjw024.

Baumeister, C., Peersman, G., 2013. Time-varying effects of oil supply shocks on the us economy. AmericanEconomic Journal: Macroeconomics 5, 1–28. doi:10.1257/mac.5.4.1.

Benkraiem, R., Lahiani, A., Miloudi, A., Shahbaz, M., 2018. New insights into the us stock market reactions to energyprice shocks. Journal of International Financial Markets, Institutions and Money 56, 169–187. doi:10.1016/j.intfin.2018.02.004.

Chen, H., Liao, H., Tang, B.J., Wei, Y.M., 2016. Impacts of opec’s political risk on the international crude oil prices:An empirical analysis based on the svar models. Energy Economics 57, 42–49. doi:10.1016/j.eneco.2016.04.018.

Cho, J.S., hwan Kim, T., Shin, Y., 2015. Quantile cointegration in the autoregressive distributed-lag modeling frame-work. Journal of Econometrics 188, 281–300. doi:10.1016/j.jeconom.2015.05.003.

Guo, Y., Li, J., Li, Y., You, W., 2021. The roles of political risk and crude oil in stock market based on quantilecointegration approach: A comparative study in china and us. Energy Economics 97, 105198. doi:10.1016/j.eneco.2021.105198.

Huang, Y., Luk, P., 2020. Measuring economic policy uncertainty in china. China Economic Review 59, 101367.doi:10.1016/j.chieco.2019.101367.

Joo, K., Suh, J.H., Lee, D., Ahn, K., 2020. Impact of the global financial crisis on the crude oil market. EnergyStrategy Reviews 30, 100516. doi:10.1016/j.esr.2020.100516.

Kilian, L., 2009. Not all oil price shocks are alike: Disentangling demand and supply shocks in the crude oil market.American Economic Review 99, 1053–1069. doi:10.1257/aer.99.3.1053.

Kim, T.H., White, H., 2003. Estimation, inference, and specification testing for possibly misspecified quantile regres-sion. Emerald Group Publishing Limited. doi:10.1016/S0731-9053(03)17005-3.

Kyrtsou, C., Mikropoulou, C., Papana, A., 2016. Does the s&p500 index lead the crude oil dynamics? a complexity-based approach. Energy Economics 56, 239–246. doi:10.1016/j.eneco.2016.02.001.

Lei, L., Shang, Y., Chen, Y., Wei, Y., 2019. Does the financial crisis change the economic risk perception of crude oiltraders? a midas quantile regression approach. Finance Research Letters 30, 341–351. doi:10.1016/j.frl.2018.10.016.

Lu, Q., Li, Y., Chai, J., Wang, S., 2020. Crude oil price analysis and forecasting: A perspective of “new triangle”.Energy Economics 87, 104721. doi:10.1016/j.eneco.2020.104721.

Meng, F., Liu, L., 2019. Analyzing the economic sources of oil price volatility: An out-of-sample perspective. Energy177, 476–486. doi:10.1016/j.energy.2019.04.161.

Mensi, W., Rehman, M.U., Al-Yahyaee, K.H., 2020a. Time-frequency co-movements between oil prices and interestrates: Evidence from a wavelet-based approach. The North American Journal of Economics and Finance 51,100836. doi:10.1016/j.najef.2018.08.019.

Mensi, W., Sensoy, A., Vo, X.V., Kang, S.H., 2020b. Impact of covid-19 outbreak on asymmetric multifractality ofgold and oil prices. Resources Policy 69, 101829. doi:10.1016/j.resourpol.2020.101829.

Niu, Z., Liu, Y., Gao, W., Zhang, H., 2021. The role of coronavirus news in the volatility forecasting of crude oilfutures markets: Evidence from china. Resources Policy 73, 102173. doi:10.1016/j.resourpol.2021.102173.

15

Pan, Z., Wang, Y., Wu, C., Yin, L., 2017. Oil price volatility and macroeconomic fundamentals: A regime switchinggarch-midas model. Journal of Empirical Finance 43, 130–142. doi:10.1016/j.jempfin.2017.06.005.

Pesaran, M.H., Shin, Y., Smith, R.P., 1999. Pooled mean group estimation of dynamic heterogeneous panels. Journalof the American statistical Association 94, 621–634. doi:10.2307/2670182.

Prest, B.C., 2018. Explanations for the 2014 oil price decline: Supply or demand? Energy Economics 74, 63–75.doi:10.1016/j.eneco.2018.05.029.

Qin, M., Su, C.W., Hao, L.N., Tao, R., 2020a. The stability of u.s. economic policy: Does it really matter for oilprice? Energy 198, 117315. doi:10.1016/j.energy.2020.117315.

Qin, Y., Hong, K., Chen, J., Zhang, Z., 2020b. Asymmetric effects of geopolitical risks on energy returns and volatilityunder different market conditions. Energy Economics 90, 104851. doi:https://doi.org/10.1016/j.eneco.2020.104851.

Wang, Y.S., Chueh, Y.L., 2013. Dynamic transmission effects between the interest rate, the us dollar, and gold andcrude oil prices. Economic Modelling 30, 792–798. doi:10.1016/j.econmod.2012.09.052.

Wei, Y., Liu, J., Lai, X., Hu, Y., 2017. Which determinant is the most informative in forecasting crude oil marketvolatility: Fundamental, speculation, or uncertainty? Energy Economics 68, 141–150. doi:10.1016/j.eneco.2017.09.016.

Yang, L., 2019. Connectedness of economic policy uncertainty and oil price shocks in a time domain perspective.Energy Economics 80, 219–233. doi:10.1016/j.eneco.2019.01.006.

Yi, A., Yang, M., Li, Y., 2021. Macroeconomic uncertainty and crude oil futures volatility–evidence from china crudeoil futures market. Frontiers in Environmental Science 9, 21. doi:10.3389/fenvs.2021.636903.

Zhang, Q., Di, P., Farnoosh, A., 2021. Study on the impacts of shanghai crude oil futures on global oil market and oilindustry based on vecm and dag models. Energy 223, 120050. doi:10.1016/j.energy.2021.120050.

Zhang, W., Hamori, S., 2021. Crude oil market and stock markets during the covid-19 pandemic: Evidence from theus, japan, and germany. International Review of Financial Analysis 74, 101702. doi:10.1016/j.irfa.2021.101702.

Zhang, Y.J., Wang, J.L., 2019. Do high-frequency stock market data help forecast crude oil prices? evidence from themidas models. Energy Economics 78, 192–201. doi:10.1016/j.eneco.2018.11.015.

Zhang, Y.J., Yan, X.X., 2020. The impact of us economic policy uncertainty on wti crude oil returns in different timeand frequency domains. International Review of Economics & Finance 69, 750–768. doi:10.1016/j.iref.2020.04.001.

Zhu, P., Tang, Y., Wei, Y., Lu, T., 2021. Multidimensional risk spillovers among crude oil, the us and chinese stockmarkets: Evidence during the covid-19 epidemic. Energy 231, 120949. doi:10.1016/j.energy.2021.120949.

16