Wang, K.-H., & Su, C.-W. (2021).

Asymmetric Link Between COVID-19 and Fossil Energy

Prices. Asian Economics Letters.

Peer-reviewed research

Asymmetric Link Between COVID-19 and Fossil Energy Prices

a
Kai-Hua Wang 1, Chi-Wei Su 1
1 School of Economics, Qingdao University, China
Keywords: bootstrap fourier granger causality test, energy prices, covid-19
10.46557/001c.18742

Asian Economics Letters

This paper investigates the asymmetric relationship between COVID-19 and fossil energy
prices through bootstrap Fourier Granger Causality test in quantiles. The empirical
results indicate that COVID-19 influences oil and natural gas prices in high volatile
quantiles. Meanwhile, no significant causal link is discovered in the coal market.
Therefore, we provide some important policy implications in terms of results.

I. Introduction had a significant negative effect on the performance of en-

ergy companies.
This paper examines the asymmetric causal relationship The second group of studies analyzes how the COVID-19
from COVID-19 to fossil energy prices. The emergence of pandemic is related to the energy market. Ertuğrul et al.
COVID-19 has challenged the global economic and financial (2020), for instance, find that volatility dynamics of the
systems (Iyke, 2020b; Phan & Narayan, 2020; Sharma & Turkish diesel market starts rising with the announcement
Sha, 2020). On March 11, the World Health Organization of the first COVID-19 case and peaks at the end of May 2020.
(WHO) announced that the COVID-19 pandemic had be- Salisu & Adediran (2020) find that the new uncertainty in-
come a “global pandemic” and cumulative confirmed cases dex includes infectious diseases and is a good predictor of
had reached 42 million in more than 200 nations in October energy market volatility in both in-sample and out-of-sam-
2020. This pandemic brought more serious repercussions ple tests. Devpura & Narayan (2020) show that COVID-19
1
than SARS in 2003, MERS in 2012 and Ebola in 2014. The cases and deaths led to an increase in daily oil price volatil-
crude oil market suffered due to high uncertainties because ity by between 8% and 22%. Gil-Alana & Monge (2020)
of the COVID-19 pandemic and political instabilities during demonstrate that oil price is mean reverting. Narayan
this period (Ali et al., 2020; Devpura & Narayan, 2020; (2020) shows that there is a threshold effect of COVID-19
Narayan, 2020). infections after which COVID-19 has had a bigger effect on
COVID-19 is also a major cause of the historic demand oil prices.
shock for the global gas market, which is expected to fall The third group of studies tests the link between energy
4% year-on-year in the first half of 2020 and represents the and financial markets. Prabheesh et al. (2020), for instance,
largest annual drop on record (IEA (International Energy find evidence of a positive co-movement between oil price
Agency), 2020). Meanwhile, natural gas spot prices plum- returns and stock price returns during the COVID-19 period.
meted to their lowest levels in at least a decade across all Huang & Zheng (2020) show that there is a structural
major gas-consuming regions. As an important coal ex- change in the relationship between investor sentiment and
porter, Indonesian coal exports dropped below average in crude oil futures price due to COVID-19. Liu et al. (2020)
February and March 2020 and is expected to decline further show that the COVID-19 pandemic cannot exert a negative
in April, which reached the lowest level since October 2010. effect but has a statistically significantly and positive effect
The main message is that the pandemic has significantly on crude oil returns and stock returns.
impacted the energy market. We add to this energy-COVID-19 literature by focusing
Previous studies have carried out considerable work on on oil, natural gas, and coal prices. We apply a Bootstrap
the effects of the COVID-19 pandemic on the energy in- Fourier Granger Causality test in Quantiles (BFGC-Q), and
dustry. One group of studies focuses on COVID-19 and en- find that COVID-19 cases influence oil and natural gas
ergy firms. Iyke (2020a), for instance, finds that the US oil prices in high quantiles. In other words, only an increase
and gas firms have heterogenous reactions to the COVID-19 in COVID-19 cases would influence fossil energy market.
pandemic. Polemis & Soursou (2020) indicate that the pan- When compared to previous studies, our contributions can
demic influenced returns of the majority of Greek energy be summarized as follows. First, the existing studies mainly
listed firms. Fu & Shen (2020) point out that COVID-19 has focus on the link between COVID-19 and oil market (Jeffer-

a Corresponding author: School of Economics, Qingdao University, 78, Kedazhi Rd., Qingdao, Shandong, China. E-Mail: cw-
[email protected].

1 SARS and MERS are abbreviations for the Severe Acute Respiratory Syndrome and the Middle East Respiratory Syndrome, respectively.
 Asymmetric Link Between COVID-19 and Fossil Energy Prices

son, 2020; Mensi et al., 2020; Salisu et al., 2020). However,

natural gas and coal also play important roles in the global
energy consumption and demand structure. Therefore, we
take natural gas and coal into consideration in our study to
where is the vector of estimated coefficients of
understand the impacts on their pricing due to the pandem-
ic. Second, previous studies just show that COVID-19 influ- quantile, is the consistence estimator of variance co-
ences energy prices (Liu et al., 2020; Prabheesh et al., 2020). variance matrix of the .
However, they ignore the COVID-19 heterogeneity (in terms
of cases and deaths) effect on energy price. The empirical
III. Data and empirical results
results demonstrate that COVID-19 influences oil and gas
We follow Liu et al. (2020) and adopt the COVID-19 data
prices in quantiles [0.6, 0.8] and [0.8, 1].
from the official website of WHO. The sample spans the
Following this introduction, the rest of the paper is
period January 13, 2020 to November 30, 2020, consisting
structured as follows: Section II introduces the research
of 230 daily observations. The starting date corresponds
method; Section III shows data and empirical results; and
to the date that the Chinese authorities officially reported
Section IV provides conclusions and policy implications.
its COVID-19 cases to the WHO. We further employ the
Generalized Autoregressive Conditional Heteroscedasticity
II. Methodology
(GARCH) model to estimate the volatility of COVID-19 new
To test the null hypothesis of non-Granger causality us- cases. The prices of oil, natural gas and coal come from the
ing the BFGC-Q approach, we design a two-step procedure. Wind Financial Terminal, which is recognized as the leading
In the first step, to control for smooth breaks of determinis- firm in providing accurate and real-time data in China. Fig-
tic terms, the Fourier expansions are included in the follow- ure 1 plots the data. Panel A has the daily global new con-
ing Granger causality equation: firmed cases. The data show fluctuating upward trend. Pan-
el B shows oil price volatility: a significant drop from US$68
to US$20 can be observed. Panel C displays natural gas price
and shows a downward trend. Panel D describes coal price:
the fluctuation amplitude is obviously less than oil and nat-
ural gas. With respect to the results of the Jarque-Bera test
in Table 1 (Panel A), the variables except OIL are signifi-
where Y is energy price, X is volatility of COVID-19, T is cantly non-normally distributed. COVID-19 and OIL show
sample size, t is time, and are coefficients, i and j are a positive skewness. COVID-19 and GAS follow leptokurtic
subscript to count, p is the lag length, h is the maximum in- distribution, meaning that it contains heavy-tailed or pro-
tegration degree of the variables, and m is the number of fusion of outliers, and others follow platykurtic distribu-
covariates. To estimate Equation (1), we select the optimal tion. The standard unit root tests were conducted and all
value of k, say k* and optimal value of lags, say p*. To this variables were found to be stationary. The results are not re-
end, we do a grid search and for each , we se- ported here but are available upon request.
lect k = k* as optimum frequency when the sum of squared Table 1 (Panel B) points out that the asymmetric link be-
residuals are minimized and then select the optimal value tween COVID-19 and energy price exists in oil and natur-
of lags, p*, using the Akaike Information Criteria. We apply al gas market in quantiles [0.6, 0.8] and [0.8, 1]. This means
the standard restricted F test statistic to test the null hy- the high volatility of COVID-19 cases influences oil and nat-
pothesis of . By choosing the k* and p*, we esti- ural gas prices. According to Figure 1, we notice that high
mate Equation (1) by quantile regression rather than the or- volatility appears in January, February, and March of 2020,
dinary least squares estimator: which corresponds to the beginning of the COVID-19 out-
break. During this period, COVID-19 was declared a glob-
al pandemic. However, we find that there is no causal link
between COVID-19 and coal market. Coal has not been the
major fossil energy in the world, except China and other de-
veloping countries, and therefore it is less affected by glob-
al energy demand changes. Meanwhile, it does not have an
active future market and its trading volume is less than oil
where Z is a matrix of all covariates in the regression model and natural gas. Figure 1 (Panel B) points out that the coal
(2). Estimating regression model (2) by quantile regression price presents mild volatility pattern when compared to the
approach allows for testing the null hypothesis of Granger other two markets.
non-causality from X to Y at different quantiles as In summary, relying on BFGC-Q test, we find that the hy-
follows: pothesis of no causal link is rejected in oil and gas markets
in quantiles [0.6, 0.8] and [0.8, 1]. That means COVID-19
influences oil and gas prices when COVID-19 is highly
volatile. It is also noticed that there is no significant causal
The null hypothesis of Granger non-causality is tested by link in the coal market.
following Wald test:

Asian Economics Letters 2

 Asymmetric Link Between COVID-19 and Fossil Energy Prices

Table 1: Descriptive statistics and main results

Panel A: Descriptive statistics

Max. Min. Mean Median Skewness Kurtosis Jarque-Bera
COVID-19 0.681 0.093 0.202 0.119 1.689 4.553 183.689***
OIL 65.2 19.33 41.824 42.44 0.121 2.175 0.855
GAS 3.582 0.282 1.548 1.526 -0.459 6.386 114.366***
COAL 643.4 468.2 559.6 557.2 -0.331 2.525 5.899**
Panel B: Quantile Granger causality
Quantile Granger Causality Wald test Critical values
10% 5% 1%
0.2 COVID-19 ⇏ OIL 3.661 5.477 6.651 8.176
0.4 COVID-19 ⇏ OIL 3.113 3.211 3.768 4.622
0.6 COVID-19 ⇏ OIL 4.137** 3.229 3.992 5.333
0.8 COVID-19 ⇏ OIL 5.544* 5.012 5.732 7.339
0.2 COVID-19 ⇏ GAS 2.729 3.546 5.121 9.665
0.4 COVID-19 ⇏ GAS 3.012 3.711 4.274 6.236
0.6 COVID-19 ⇏ GAS 4.777** 3.516 3.943 6.151
0.8 COVID-19 ⇏ GAS 6.993** 4.437 5.191 9.094
0.2 COVID-19 ⇏ COAL 0.421 4.155 5.399 9.011
0.4 COVID-19 ⇏ COAL 0.163 3.339 4.202 5.654
0.6 COVID-19 ⇏ COAL 0.572 3.396 4.073 5.122
0.8 COVID-19 ⇏ COAL 2.616 4.911 6.064 8.337

The table has two parts. Panel A has descriptive statistics. It reports the maximum (max.), minimum (min.), mean, median, Skewness, Kurtosis and Jarque-Bera statistics for
COVID-19, oil, natural gas and coal, respectively. Panel B has the quantile causality results. The different quantiles are stated in column 1, column 2 indicates the direction of causali-
ty; the Wald test is reported in column 3; and the final three columns report the critical values. The subscripts ***, ** and * denote statistical significance at the 1%, 5% and 10% levels,
respectively.

IV. Concluding remarks

This paper investigates the link between COVID-19 and

fossil energy prices. The results indicate that the asym-
metric link exists in quantiles [0.6, 0.8] and [0.8, 1], im-
plying that only high volatility of COVID-19 cases influ-
ences oil and natural gas prices. However, there is no sig-
nificant causal link in the coal market due largely to global
economies less reliance on it.

Submitted: November 14, 2020 AEDT, Accepted: January 15,

2021 AEDT

Asian Economics Letters 3

 Asymmetric Link Between COVID-19 and Fossil Energy Prices

Figure 1: COVID-19 cases and prices data

This figure plots four data series: panel A contains a plot of the volatility of daily COVID-19 cases; and panels B, C, and D have time-series data for oil price, natural gas and
coal price, respectively.

This is an open-access article distributed under the terms of the Creative Commons Attribution 4.0 International License (CC-
BY-SA-4.0). View this license’s legal deed at https://creativecommons.org/licenses/by-sa/4.0 and legal code at https://cre-
ativecommons.org/licenses/by-sa/4.0/legalcode for more information.

Asian Economics Letters 4

 Asymmetric Link Between COVID-19 and Fossil Energy Prices

REFERENCES

Ali, M., Alam, N., & Rizvi, S. A. R. (2020). Coronavirus Liu, L., Wang, E.-Z., & Lee, C.-C. (2020). Impact of the
(COVID-19) - An epidemic or pandemic for financial COVID-19 pandemic on the crude oil and stock
markets. Journal of Behavioral and Experimental markets in the US: A time-varying analysis. Energy
Finance, 27, 100341. https://doi.org/10.1016/j.jbef.20 Research Letters, 1(1). https://doi.org/10.46557/001
20.100341 c.13154

Devpura, N., & Narayan, P. K. (2020). Hourly Oil Price Mensi, W., Sensoy, A., Vo, X. V., & Kang, S. H. (2020).
Volatility: The Role of COVID-19. Energy Research Impact of COVID-19 outbreak on asymmetric
Letters, 1(2). https://doi.org/10.46557/001c.13683 multifractality of gold and oil prices. Resources
Policy, 69, 101829. https://doi.org/10.1016/j.resourpo
Ertuğrul, H. M., Güngör, B. O., & Soytaş, U. (2020). l.2020.101829
The Effect of the COVID-19 Outbreak on the Turkish
Diesel Consumption Volatility Dynamics. Energy Narayan, P. K. (2020). Oil price news and
Research Letters, 1(3). https://doi.org/10.46557/001 COVID-19—Is there any connection? Energy Research
c.17496 Letters, 1(1). https://doi.org/10.46557/001c.13176

Fu, M., & Shen, H. (2020). COVID-19 and Corporate Phan, D. H. B., & Narayan, P. K. (2020). Country
Performance in the Energy Industry. Energy Research responses and the reaction of the stock market to
Letters, 1(1). https://doi.org/10.46557/001c.12967 COVID-19—a Preliminary Exposition. Emerging
Markets Finance and Trade, 56(10), 2138–2150. http
Gil-Alana, L. A., & Monge, M. (2020). Crude Oil Prices s://doi.org/10.1080/1540496x.2020.1784719
and COVID-19: Persistence of the Shock. Energy
Research Letters, 1(1). https://doi.org/10.46557/001 Polemis, M., & Soursou, S. (2020). Assessing the
c.13200 Impact of the COVID-19 Pandemic on the Greek
Energy Firms: An Event Study Analysis. Energy
Huang, W., & Zheng, Y. (2020). COVID-19: Structural Research Letters, 1(3). https://doi.org/10.46557/001
changes in the relationship between investor c.17238
sentiment and crude oil futures price. Energy
Research Letters, 1(2). https://doi.org/10.46557/001 Prabheesh, K. P., Padhan, R., & Garg, B. (2020).
c.13685 COVID-19 and the Oil Price – Stock Market Nexus:
Evidence From Net Oil-Importing Countries. Energy
IEA (International Energy Agency). (2020). The Research Letters, 1(2). https://doi.org/10.46557/001
impacts of the Covid-19 crisis on global energy c.13745
demand and CO2 emissions. Global Energy Review
2020. https://www.iea.org/fuels-and-technologies/ga Salisu, A. A., & Adediran, I. (2020). Uncertainty Due
s to Infectious Diseases and Energy Market Volatility.
Energy Research Letters, 1(2). https://doi.org/10.4655
Iyke, B. N. (2020a). COVID-19: The reaction of US oil 7/001c.14185
and gas producers to the pandemic. Energy Research
Letters, 1(2). https://doi.org/10.46557/001c.13912 Salisu, A. A., Ebuh, G. U., & Usman, N. (2020).
Revisiting oil-stock nexus during COVID-19
Iyke, B. N. (2020b). Economic Policy Uncertainty in pandemic: Some preliminary results. International
Times of COVID-19 Pandemic. Asian Economics Review of Economics & Finance, 69, 280–294. http
Letters, 1(2). https://doi.org/10.46557/001c.17665 s://doi.org/10.1016/j.iref.2020.06.023

Jefferson, M. (2020). A crude future? COVID-19s Sharma, S. S., & Sha, Y. (2020). Part A: Special
challenges for oil demand, supply and prices. Energy Section on COVID-19 Research. Emerging Markets
Research & Social Science, 68, 101669. https://doi.or Finance and Trade, 56(15), 3551–3553. https://doi.or
g/10.1016/j.erss.2020.101669 g/10.1080/1540496x.2020.1858617

Asian Economics Letters 5