




































 

 

 
1 

© 2021 Conscientia Beam. All Rights Reserved. 

MODELLING STOCK RETURNS VOLATILITY AND ASYMMETRIC NEWS EFFECT: A 
GLOBAL PERSPECTIVE   

 

 Kingsley 
Onyekachi Onyele1+ 

 Emmanuel 
Chijioke Nwadike2 

 

1Michael Okpara University of Agriculture, Umudike, Umuahia, Abia State, 
Nigeria.  

 
2Federal University of Technology, Owerri, Owerri, Imo State, Nigeria. 

 
 

 
(+ Corresponding author) 

 ABSTRACT 
 
Article History 
Received: 2 November 2020 
Revised: 4 December 2020 
Accepted: 21 December 2020 
Published: 11 January 2021 
 

Keywords 
Stock 
Returns 
Volatility 
Stock market 
Asymmetry  
GARCH. 

 
JEL Classification: 
C58; G10; L16. 

 
This paper modelled stock returns volatility using daily S&P Global 1200 index from 
1st September, 2010 to 30th September, 2020. The S&P 1200 represents a free-float 
weighted stock market index of global equities covering seven (7) regional stock 
market indices and approximately 70% of the global market capitalization, hence was 
used to compute global stock returns. The data analysis was carried out with 
Generalized Autoregressive Conditional Heteroskedasticity (GARCH) techniques. Of 
the variant GARCH models specified in this study, the symmetric GARCH-M (1,1) and 
the asymmetric TGARCH (1,1) models were found suitable for the estimation. The 
findings from the GARCH-M and TGARCH models revealed explosive volatility 
persistence and strong asymmetric news effect in the global stock market, respectively. 
The implication of volatility persistence is that current volatility shocks influenced 
expected returns over a long period. The asymmetric news effect showed that negative 
news (bad news) spurred stock returns volatility than positive news (good news) 
especially in 2020 which was due to the COVID-19 crisis as shown by the plot of the 
conditional variance. These results were consistent with the empirical findings of a 
number of studies in emerging markets. Hence, the study concludes that the global 
stock market exhibited high volatility persistence and leverage effect during the 
sampled period.  
 

Contribution/Originality: This study contributes to the literature by modelling global stock returns volatility 

and asymmetric news effect using a new stock index (S&P 1200 global index). The paper contributes the first 

logical analysis that volatility of S&P 1200 returns is explosive and largely influenced by news available in the 

global markets.  

 

1. INTRODUCTION 

In both developed and developing economies, the stock market is an integral component of the financial system 

that contributes immensely to capital formation, wealth creation and economic growth. Although it plays a 

prominent role in economic prosperity by deepening the financial system, problems occasioned by volatility of stock 

returns has immensely influenced the effective functioning of the global market. According to literature allied to 

stock markets, volatility is the level of uncertainty or risk associated with the value of financial assets (Engle & 

Patton, 2001). Periods of higher volatility connotes significant variation in the value of financial assets while lower 

volatility suggests that the value of financial assets does not dramatically change overtime (Banumathy & 

Azhagaiah, 2015). This volatility risk could cause financial shocks to investors, thus creating challenges of low 

capital investments in financial assets, vulnerability in market-making, loss of investors’ confidence and fickle stock 

Financial Risk and Management Reviews 
2021 Vol. 7, No. 1, pp. 1-15. 
ISSN(e): 2411-6408 
ISSN(p): 2412-3404 
DOI: 10.18488/journal.89.2021.71.1.15 
© 2021 Conscientia Beam. All Rights Reserved. 

 
 
 

 
 
 

 

 
 
 
 

https://orcid.org/0000-0002-4731-6139
https://orcid.org/0000-0001-5177-0230
https://www.doi.org/10.18488/journal.89.2021.71.1.15


Financial Risk and Management Reviews, 2021, 7(1): 1-15 

 

 
2 

© 2021 Conscientia Beam. All Rights Reserved. 

prices and returns (Bello, 2020; Chiang & Doong, 2014; Wang & Yang, 2017). Consequently, in a highly volatile 

stock market, it is difficult for quoted companies to raise sufficient funds as rational investors prefer stocks with less 

volatile returns/prices unlike the risk takers (Onyele, Opara, & Ikwuagwu, 2017). Though, returns volatility in the 

stock market may not necessarily be destructive all the time, but volatility persistence in market returns, especially 

in developed markets will likely lead to a crash in the global financial market due to increasing financial integration 

(Onyele. & Ikwuagwu, 2020).  

Volatility of stock returns is majorly triggered by investors’ expectations and perceptions of daily information 

(news effect) in the market. When there is upsurge in returns volatility as a result of news effect, efficiency and 

liquidity is altered as market participants receive the news with different mindset (Ho & Hung, 2012). In reality, 

however, bad news will accelerate returns volatility more than good news which may be interpreted by investors as 

higher risk-return tradeoff (Jegageevan, 2015). Since returns on financial assets is a function of the market risk, risk 

takers are expected to receive a rate of return that will compensate for the risk taken in making such long-term 

funds (such as, debentures, common share, bond and mortgage loan) available to economic units (Edem & Ogbonna, 

2020). This explains the long age maxim of Efficient Market Hypothesis (EMH) that all available information is 

correctly reflected in stock prices and thus stock prices rapidly react to any novel information at the moment it 

reaches the market participants (Brealey & Meyers, 2003). This informational fundamental comprise changes to 

firms’ operations, modifications in macroeconomic policies, twist in the level of investors’ risk-return preference, 

financial integration, natural disaster, etc. (Onyele, Ikwuagwu, & Onyekachi-Onyele, 2020; Sansa, 2020).  

Myriad of studies has indicated that stock markets at different time period exhibits volatility persistence, risk-

return tradeoff or asymmetric news effects; hence, it can be said that no conclusive model can fit every stock market 

all the time. Notwithstanding, enormous studies has been done for developed stock markets such as the United 

States, United Kingdom, Japan, etc. while some other studies focused on emerging stock markets like China, India, 

etc. (see, (Banumathy & Azhagaiah, 2015; Caporale, Karanasos, Yfanti, & Kartsaklas, 2019; Khedhiri, 2008; Lai, 

Cheong, & Lee, 2019; Wei, 2009)). In the developing markets of Africa, there have been research efforts towards 

modelling volatility of stock market returns and asymmetries in Nigeria, South Africa, Kenya, Morocco, Egypt, etc. 

(Bello, 2020; Jebari & Hakmaoui, 2017; Kuhe, 2018; Ndei, Muchina, & Wawure, 2019). These empirical studies, 

though with varying findings, used country specific market indices (such as, the Dow Jones, S&P 500, FTSE 100, 

Nikkei index, NIFTY index, NSE index, etc.) but the current study confirmed the stylized facts using the S&P 

global 1200 index that represents a free-float weighted stock market index of global equities covering seven (7) 

regional stock market indices and approximately 70% of the global market capitalization. In view of this research 

gap, the main goal of this paper is to model stock returns volatility and asymmetric news effects in the global stock 

market using the S&P Global 1200 index.  

The rest of the paper is organized as follows; Section 2 presents the literature review on issues concerning the 

modelling of stock market returns volatility. Sections 3 captures the model, data and methodology used for the 

estimation while section 4 presents the results and discussions. The conclusions of the study are documented in 

section 5.  

 

2. LITERATURE REVIEW 

2.1. Stylized Facts 

There are quite a number of stylized facts regarding volatility of stock market returns that has been confirmed 

by prior studies. Consequently, a good volatility model should reflect these stylized facts to a large extent. Features 

of stock returns series such as volatility clustering, volatility persistence, risk-return tradeoff and asymmetric news 

effect (leverage effect) has been confirmed in several empirical works. These stylized facts have been discussed 

below as follows:  

 



Financial Risk and Management Reviews, 2021, 7(1): 1-15 

 

 
3 

© 2021 Conscientia Beam. All Rights Reserved. 

a) Non-Normal Distribution: 

Distribution of stock returns as well as other financial time series are not normal or exhibits fatter tails also 

referred to as excess kurtosis (Fama, 1995; Mandelbrot, 1963). Hence, stock returns series are usually leptokurtic 

with fourth moment above 3. This stylized fact is common and it has been confirmed by many empirical studies.  

 

b) Volatility Clustering 

Volatility clustering is a situation where small and large values (of either signs) in the return series is likely to 

occur in clusters, that is, small moves being accompanied by large changes (Fama, 1995; Mandelbrot, 1963). This 

implies that volatility could be time-varying, that is, excessive volatility comes and goes over a period of time 

(Arouri, Lahiani, Lévy, & Nguyen, 2012; Cong, 2017).  

 

c) Volatility Persistence: 

Stock returns volatility is highly persistent or has long memory if it is characterized by insignificant 

autocorrelations of absolute or squared returns (Owidi & Mugo-Waweru, 2016). Persistence in volatility of stock 

returns series majorly affect future market volatility under influence of shocks. The implication of such volatility 

persistence is that today’s volatility shocks will affect the expectation of volatility over many periods in the future. 

Hence, volatility persistence is sequentially beneficial in forecasting future stock returns. 

 

d) Returns Are Mean Reverting:  

A mean-reverting volatility is interpreted to imply a level of normal volatility to which volatility will 

eventually return. Long-run predictions of volatility would often converge to this same level of normal volatility, 

not minding when the predictions were made (Engle & Patton, 2001). Though, many studies opined that mean-

reversion is a feature of volatility, they might be differences on the level of normal volatility and whether it is 

constant over all the time (Bello, 2020; Owidi & Mugo-Waweru, 2016).  

 

e) Asymmetric News Effect: 

One of the assumptions of volatility models is that the conditional variance of financial assets is influenced 

symmetrically by negative and positive innovation. For example, the GARCH (1,1) model permits the variance to 

be influenced by the square of the lagged innovation only, totally ignoring the effect of positive or negative 

innovation. Regarding stock returns, it is particularly said that positive and negative shocks/news would affect 

volatility (Wei, 2009). Sometimes, asymmetry is likened to a leverage effect and a risk premium (risk-return 

tradeoff) at other times (Engle & Patton, 2001). Here, news of higher volatility lowers demand for a particular stock 

due to risk aversion, which is accompanied by the increased volatility as predicted by the news.  

 

2.2. Theoretical Underpinning 

Theoretically, studies on stock returns volatility are often anchored on the Efficient Market Hypothesis (EMH) 

which was developed by Fama (1970). The EMH explains why stock prices is seen to follow a random walk. 

According to Fama (1970) an efficient market is one in which all available information are reflected in the stock 

prices. According to the EMH, the intrinsic value of shares and other financial assets is defined by the future 

discounted value of cash flows accruing to investors (Fauzel & Fauzel, 2016). Hence, if the stock market is efficient, 

all available information must be reflected in stock prices. This is needful for the assessment of a firm’s performance 

in the future, therefore the intrinsic and market value of a share should be equal (Dukes, Bowlin, & MacDonald, 

1987; Lo & MacKinlay, 1988). Hence, an information that may alter firm’s profitability in the future must be 

reflected in the share price immediately, else any delay in information diffusion to price would lead to irrationality 

as availability of some information could be exploited to predict or forecast profitability (Bohl & Henke, 2003; 



Financial Risk and Management Reviews, 2021, 7(1): 1-15 

 

 
4 

© 2021 Conscientia Beam. All Rights Reserved. 

Fama.. 1991). As such, in an efficient market, it is assumed that changes in share prices are unpredictable since there 

is random arrival of information. Using Equation 1 the random walk model is specified as follows:  

     (1) 

Where,   

 = share price at time . 

 = share price at time . 

random error with zero mean and finite variance. 

Equation 1 shows that share price at time  is equivalent to its price at time  in addition to a specific value 

that depends on arrival of unpredictable new information between  and . In other words, 

 does not depend on previous price changes.  

There are three levels of efficient markets among which is the weak-form efficiency whereby the information 

content of interest is historical prices (Fama, 1970). The weak form efficiency suggests that current stock prices 

reflect all information of previous prices and that investors cannot apply technical analysis of any form in their 

investment decisions (to determined undervalued or overvalued stocks) but can research firm’s financial statements 

to boost their chances of gaining returns higher than that of the market. On the other hand, the semi-strong form is 

hinged on the notion that investors cannot use either fundamental or technical analysis to obtain higher returns in 

the market since all publicly available information is used in the computation of current stock prices and that only 

information that is not available to the public (private information) can aid investors boost their returns above that 

of the market. The advocates of the strong form version states that all available information (both public and non-

public) is completed reflected in the current stock prices, that is, there is no type of information that can make an 

investor make returns higher than the market, not even insider knowledge give investors a predictive edge over the 

entire market. The building block of this study is the weak form of market efficiency.  

 

2.3. Empirical Review 

On the empirical sphere, studies on stock returns volatility dates back to the 1980s but the empirical studies 

have improved in recent times. The first reason adduced to this development is the fact that different data on stock 

market indices has been computed globally. Availability of these data has empowered researchers to conduct studies 

on stock returns volatility in less developed, emerging and developed countries across the world. The second reason 

for the current development in the literature is associated with advancement in econometric estimation methods as 

captured by various GARCH models applied in the literature. The econometric estimation models are built on the 

weak-form EMH, showing evidence of volatility persistence and asymmetric volatility/news or leverage effect. 

Notable among earlier studies are Bollerslev, Chou, and Kroner (1992); Bollerslev (1986); Ding, Granger, and Engle 

(1993); Engle (1982) who confirmed presence of volatility in financial times series. On the other hand, recent 

empirical studies on the subject emerged with different results due to time period, methodology and geography 

covered by the various studies. The gap identified in the empirical literature is that none of the prior empirical 

works had investigated stock returns volatility and asymmetric news effect using return series from S&P Global 

1200 index (see components of S&P Global Index in Table 1 in section 3).  



Financial Risk and Management Reviews, 2021, 7(1): 1-15 

 

 
5 

© 2021 Conscientia Beam. All Rights Reserved. 

Most recently, Bello (2020) used GARCH(1,1) and daily data from 2008-2018 to analyze stock returns 

volatility and found high volatility persistence in the Nigerian Stock Market. In another study, Musa, Adamu, and 

Dauran (2020) applied PGARCH and revealed that volatility persistence in the Nigerian Stock Market reduced 

significantly after unexpected shocks between 1987 and 2019. In Southern Asia, Iqbal, Saeed, and Shah (2020) used 

GARCH(1,1) model to analyze daily returns series from 2007 to 2019 and found explosive volatility of stock 

returns. Also, using daily data from 1999 to 2016, Edem and Ogbonna (2020) showed evidence of returns volatility 

persistence and asymmetric news effect in the Nigerian Stock Market. In Malaysia, Lai et al. (2019) used daily time 

series data spanning from 1996 to 2016 with GARCH (1,1) model to show that stock returns of oil & gas sector 

were most volatile. In Kenya, Ndei et al. (2019) analyzed the Nairobi Stock Market from 2010 to 2017 using 

GARCH(1,1) and TGARCH(1,1) and revealed persistent returns volatility, leverage effects, and absence of risk-

return trade-off. Using daily time series from 1997 to 2018, Caporale et al. (2019) showed that volatility persistence 

in the Korean stock market was driven by buy and sell trades depending on the type of investor trading and phase 

of business cycle. In a comparative analysis of America, Europe, Far East, BRICS stock markets from 1997 to 2008, 

Tsuji (2018) found evidence of persistent asymmetric volatility in all the markets. On the other hand, using the 

GARCH (1,1) model and daily return series from 1997 to 2008, De Gaetano (2018) found that returns in the BRICS 

market was time varying. In a study of emerging markets, Abdennadher and Hallara (2018) applied GARCH(1,1) 

on daily time series spanning from 2005 to 2015 and found that returns volatility varied with structural changes. 

Kuhe (2018) observed high volatility persistence in stock returns in the Nigeria Stock Market from 1999 to 2017.  

Using Frictionally Integrated EGARCH model, Jebari and Hakmaoui (2017) reported strong volatility persistence 

in the Moroccan stock market from 1993 to 2017.  

In other studies, Wang and Yang (2017) observed that long-term returns volatility in the Shanghai Stock 

Exchange, China was driven by negative returns.  Using TGARCH, Aguda (2016); Owidi and Mugo-Waweru 

(2016); Ndwiga and Muriu (2016) showed that stock returns volatility decreased with asymmetric effects in Nigeria 

and Kenya respectively. Ahmad, Ahmed, Vveinhardt, and Streimikiene (2016); Babikir, Gupta, and Owusu-Sekyere 

(2010) and Babikir et al. (2010) found that the Asian and South African markets exhibited volatility persistence of 

returns. Again, using the Symmetric GARCH (1,1) model, Adewale, Olufemi, and Oseko (2016) found high 

volatility persistence with no leverage effect in Nigeria. Banumathy and Azhagaiah (2015); Jegageevan (2015); 

Sethapramote and Prukumpai (2012) and Khedhiri (2008) found that stock returns volatility was driven by bad 

news in India, Sri Lanka, Thailand and UAE, respectively. Using Bivariate GARCH, Wang, Huang, and 

Padmanabhan (2015) found volatility persistence of stock returns in the United States. Bentes and Da Cruz (2010) 

analyzed the G7 markets using GARCH, IGARCH and FIGARCH and found persistent stock returns volatility in 

Germany, Italy and France, but less returns volatility in Japan. Comparing African markets, Alagidede and 

Panagiotidis (2009) found evidence of leverage effects. Chiang and Doong (2001) revealed that bad news was 

responsible for volatility persistence in most Asian markets. Berument and Kiymaz (2001) found that highest and 

lowest stock returns volatility in the United States was observed on Wednesday and Monday, respectively. In 

Japan, Bekaert and Wu (2000) indicated that returns volatility persistence and feedback at firm level is driven by 

strong asymmetries in the conditional covariances. 

 

3. METHODOLOGY AND DATA 

3.1. Methodology 

The degree of volatility which is also termed “conditional variance of a financial asset” must be estimated in a 

model that best show its time varying conditional variance (Engle, 1982; Tsay, 2010). Financial time series depend 

on three basic factors, viz; their own previous values (that is, autoregressive), past information (that is, conditional) 

and exhibit non-constant variance (that is, heteroscedasticity) which forms the bedrock of the popular 

Autoregressive Conditional Heteroscedasticity (ARCH) model. Hence, the presence of these fundamental features 



Financial Risk and Management Reviews, 2021, 7(1): 1-15 

 

 
6 

© 2021 Conscientia Beam. All Rights Reserved. 

should be well captured in the proposed volatility model(s) to be adopted in a research study of this nature (Cont, 

2005). Recent econometric techniques within the scope of GARCH family models provide the tool for solving this 

research problem.  

In this paper, as originated by Bollerslev (1986) and applied in recent empirical studies Bello (2020); Ndei et al. 

(2019); Banumathy and Azhagaiah (2015); Jegageevan (2015) the GARCH model was adopted to unravel stock 

returns volatility from a global perspective. In the GARCH model, the conditional variance is a function of its 

previous own lags. Stock returns volatility is determined by the magnitude of coefficients  and . If the addition of 

both parameters is equal to or approximately one (1), then volatility of the return series is said to be persistent and 

vice versa. In its simplest form, the symmetrical GARCH models (mean and variance equations) is specified as 

displayed in Equations 2 and 3: 

Mean equation:  

Variance equation:  

Where, 

   

  represents the stock market return at time . 

 denote the average return of the market. 

 indicate the residual return. 

In a GARCH model, the conditional variance equation is directly fitted into the mean equation, leading to the 

GARCH-M model (mean-reversion GARCH). Regarding the GARCH-M model, the coefficient  in the mean 

equation represents the risk premium. If  is positive, it indicates that there is positive relationship between stock 

return and its volatility, that is, an increase in mean return is determined by a rise in conditional variance as a proxy 

of higher risk. As a matter of fact, a positive and significant  indicates presence of risk-return tradeoff or risk 

premium.  In the GARCH-M model, stock return is dependent on its own volatility and as such a simple GARCH-

M (1,1) model is specified as shown in Equations 4 and 5:  

Mean equation:  

Variance equation:  

The shortcoming of symmetric GARCH models is that the conditional variance does not react asymmetrically 

to fluctuations in returns.  As a result, a number of models known as asymmetric models such as EGARCH and 

TGARCH, amongst others, have been developed to deal with this issue. In consonance with Banumathy and 

Azhagaiah (2015) to ascertain the relationship between asymmetric volatility and stock returns, the Exponential 

GARCH (EGARCH) and the Threshold GARCH (TGARCH) models were applied. The EGARCH model is hinged 

on logarithmic expression of the conditional variability. With the EGARCH model, the presence of leverage effect 

can be tested (Nelson, 1991). The presence of leverage effect or asymmetry is tested based on the hypothesis that 



Financial Risk and Management Reviews, 2021, 7(1): 1-15 

 

 
7 

© 2021 Conscientia Beam. All Rights Reserved. 

. The impact is said to be symmetric if . The EGARCH model is expressed as shown in Equation 6 

below:  

 
Where,  

 denote the log of the conditional variance. 

 represents the leverage term or asymmetry 

With regards to the Threshold GARCH (TGARCH), in tandem with Zakoian (1994) the following 

Equation 7 was applied:  

 

Where, 

 is the asymmetry or leverage parameter. In the TGARCH model, good news  and bad 

news  on the conditional variance. Here, the impact of good news is  while the impact of bad news is 

. As such, supposing  is positive and significant, negative shocks would have a greater impact on   

 Prior to estimating the GARCH models, the distributional properties of the return series were considered. 

To specify the distributional properties of the daily return series of S&P Global 1200 index, the descriptive statistic 

was carried out. Also, to ensure that stock returns are stationary as one of the conditions required for the 

application of GARCH models, unit root tests were conducted using the Augmented Dickey-Fuller Test (ADF) and 

Philips-Perron (PP) tests (Dickey & Fuller, 1979; Phillips & Perron, 1988). 

 

3.2. Description of Data 

The daily time series data of S&P Global (SPG) 1200 index used for this study spanned from the trading days 

between 1st Sept., 2010 and 30th Sept., 2020. The SPG daily data was sourced from 

(https://www.spglobal.com/spdji/en/indices/equity/sp-global-1200/#overview). The SPG rate of returns series 

(R_SPG) within the sampled period was calculated as natural logarithm (ln) of the first difference associated with 

daily closing stock prices. The formula used for computing the R_SPG is as shown in Equation 8: 

 

Where, 

  represent the logarithmic SPG daily returns for time t. 

  denote the closing price at time t. 

  shows the corresponding price in the period at time . 

 

https://www.spglobal.com/spdji/en/indices/equity/sp-global-1200/#overview


Financial Risk and Management Reviews, 2021, 7(1): 1-15 

 

 
8 

© 2021 Conscientia Beam. All Rights Reserved. 

Table-1. Components of S&P Global 1200 index. 

Country/Region Index 

United States S&P 500 
Hong Kong, Singapore, South Korea and Taiwan S&P Asia 50 
Australia  S&P/ASX 50 
Eurozone, Denmark, Norway, Sweden, Switzerland and UK S&P Europe 350 
Brazil, Chile, Colombia, Mexico and Peru S&P Latin America 40 
Japan S&P/TOPIX 150 
Canada S&P/TSX 60 

 

 

As detailed in Table 1, the S&P Global (SPG) 1200 represents a weighted stock index of global equities 

covering thirty-one (31) countries, seven (7) regional stock markets and about 70% of global market capitalization, 

including all ten (10) Global Industry Classification Standard (GICS) sectors. 

 

4. RESULTS AND DISCUSSIONS 

The logarithmic of the S&P Global 1200 index from 1st Sept., 2010 to 30th Sept., 2020 have been plotted in 

Figure 1: 

 

 
Figure-1. Trend of S&P Global 1200 index. 

 

Having computed the daily SPG returns using Equation 8, Figure 2, shows evidence of volatility clustering for 

R-SPG over the sampled period (1st Sept., 2010 to 30th Sept., 2020) which is one of the stylized facts of financial 

time series. There was higher volatility clustering during the trading days in 2011 and 2020 due to the European 

and US debt crisis as well as the recent COVID-19 pandemic. On the other hand, the histogram captured by Figure 

3 shows that the return series are not normally distributed based on Kurtosis (19.35118 > 3) and p-value (0.0000) of 

the Jarque-Bera test which rejects the normality distribution of the R_SPG series. The non-normal distribution of 

the R_SPG confirmed the stylized fact that distribution of financial returns series is largely leptokurtic.  

 

-1.5

-1.0

-0.5

0.0

0.5

1.0

1.5

2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020

R_SPG

 
Figure-2. Time plot of daily R_SPG. 



Financial Risk and Management Reviews, 2021, 7(1): 1-15 

 

 
9 

© 2021 Conscientia Beam. All Rights Reserved. 

0

200

400

600

800

1,000

1,200

-1.0 -0.5 0.0 0.5 1.0

Series: R_SPG

Sample 9/01/2010 9/30/2020

Observations 2624

Mean       0.003703

Median   0.007905

Maximum  1.105218

Minimum -1.313233

Std. Dev.   0.125186

Skewness  -1.120008

Kurtosis   19.35118

Jarque-Bera  29780.08

Probability  0.000000

 

Figure-3. R_SPG normality testing. 
 

Next, the study proceeded with the Augmented Dickey Fuller (ADF) and Philip-Perron (PP) unit root tests in 

order to investigate the level of stationarity of the R_SPG series. On the other hand, the presence of ARCH effect in 

the return series was confirmed by the ARCH-LM test. Details of the unit root and the ARCH-LM test results. As 

presented in Table 2 below, the R_SPG series has no unit root from the ADF and PP test approaches. The 

probability values of the ADF and PP t-Statistics are less than 0.05 which led to the conclusion that the R_SPG 

series for the sampled period is stationary, hence both the ADF and PP tests reject the hypothesis of non-

stationarity of the R_SPG series at all levels of significance. On the other hand, the ARCH-LM test which was used 

to investigate the presence of ARCH effect on the returns series is highly significant since the p-value (p < 0.05), 

leading to the rejection of the null hypothesis of “no ARCH effect” in the residuals. Based on the outcome of these 

preliminary tests, the appropriateness of the GARCH family models was justified.  

 
Table-2. Unit root test results. 

  ADF @ Level PP @ Level 
  t-Statistic Prob.* t-Statistic Prob.* 

Test statistic -19.00810 0.0000 -50.50163 0.0001 
Test critical values: 1% level -3.432657  -3.432657  
 5% level -2.862445  -2.862445  
 10% level -2.567297  -2.567297  
ARCH-LM Test Statistics: 
F-statistic 290.1747  Prob. F(1,2620) 0.0000 
Obs*R-squared 261.4407  Prob. Chi-Square(1) 0.0000 

 

 

4.1. Estimation of GARCH Models 

The best fit GARCH models are those with the highest Adjusted R-squared, lowest AIC and SIC, no serial 

correlation and heteroscedasticity in the residuals (Engle, 1982). Evidence of no serial correlation and 

heteroscedasticity is when the p-values of the Q-statistics and F-statistic are statistically insignificant. However, all 

the models passed the serial correlation and heteroscedasticity tests but varied slightly in other selection criteria 

such as the AIC and SIC, Log likelihood and coefficient of determination. Having, x-rayed the various GARCH 

models, the symmetric GARCH-M (1,1) and asymmetric TGARCH (1,1) models were selected.  

The symmetric GARCH(1,1) and GARCH-M(1,1) are reported in Table 3: 

 

 



Financial Risk and Management Reviews, 2021, 7(1): 1-15 

 

 
10 

© 2021 Conscientia Beam. All Rights Reserved. 

Table-3. Estimated result of GARCH (1,1) and GARCH-M (1,1) models. 

Parameters GARCH-M (1,1) GARCH-M (1,1) 

Mean Equation: 

 (mean return) 
0.007780{0.0000}*** 0.006100 {0.0028}*** 

⋌ (risk premium) -- 0.245237 {0.2066} 

Variance Equation: 

 (constant) 
0.000367{0.0000}*** 0.000368{0.0000}*** 

(ARCH effect) 
0.187895{0.0000}*** 0.187645{0.0000}*** 

(GARCH effect) 
0.796539{0.0000}*** 0.796632{0.0000}*** 

 (persistence coefficient) 
0.984434 0.984277 

   
Log likelihood 2429.474 2430.389 
AIC -1.848627 -1.848562 
SIC -1.837433 -1.835130 
Adjusted R-squared -0.008763 -0.016122 
   

Residual Diagnostics:    

Serial correlation  
(Q-statistic probabilities) 

Q > 0.05 
(No autocorrelation) 

Q > 0.05 
(No autocorrelation) 

   
Heteroscedasticity: 
F-statistic 0.612562 {0.4339} 0.534893 {0.4646} 

Note: ** and *** indicate rejection of the null hypothesis @ 5%, and 1 % levels of significance, respectively. Figures in parenthesis { ) 
are the probability values. 

 

From Table 3, though the SIC and AIC associated with the GARCH (1,1) and GARCH-M (1,1) varied slightly, 

the Adjusted R-squared (-0.016122) of the latter is greater than that of the former in absolute value. Hence the 

GARCH-M was accepted as the best fit model for the symmetric GARCH. It can be seen that the sum of ARCH and 

GARCH coefficients  of the GARCH-M (1,1) model is 0.984442. This indicates that volatility of the 

R_SPG series was highly persistent. It then suggests that the R_SPG for the sampled period is mean reverting. 

This finding is in line with those of Bello (2020); Jebari and Hakmaoui (2017); Wang and Yang (2017); Ahmad et al. 

(2016); Wang et al. (2015); Banumathy and Azhagaiah (2015) that stock markets of Nigeria, Morocco, China, Asian 

countries, United States and India exhibited high volatility persistence. Also, Ndei et al. (2019); Caporale et al. 

(2019); Aguda (2016); Berument and Kiymaz (2001) are of the view that stock returns volatility is persistent but 

varied with time, business cycle, etc. On the other hand, the estimated coefficient of the risk premium  in the 

mean equation is positive and statistically insignificant which indicates that volatility has no significant impact on 

expected returns of the global stock market which indicates lack of risk-return trade-off. This means that higher 

market risk arising from the conditional variance (volatility) did not necessarily trigger higher returns in the global 

market. In consonance with Ndei et al. (2019); Banumathy and Azhagaiah (2015) this implies that investors were 

not compensated for taking additional investment risks, but contrary to Ndwiga and Muriu (2016); Alagidede and 

Panagiotidis (2009) who found significant and positive risk premium in Kenya and selected African countries 

respectively.  

The plot of the conditional variance presented in Figure 4 below depicts that volatility of R_SPG was high 

within the trading days in late-2011 and early 2012 probably due to the US and European debt crisis. There was 

also increased volatility in the trading days within the third quarter of 2015 and decreased towards the end of the 

fourth quarter of the same year. Volatility increased and dropped within the trading days in the second quarter of 

2016. R_SPG volatility was also evident in the trading days of the first and fourth quarters of 2018 probably due to 



Financial Risk and Management Reviews, 2021, 7(1): 1-15 

 

 
11 

© 2021 Conscientia Beam. All Rights Reserved. 

the US/China trade war (see, Wang, Yao, and Bonne (2020)). Again, the R_SPG experienced explosive volatility 

during the trading days in the first and second quarters of 2020 which later trended downwards towards the 

trading days in the third quarter due to the COVID-19 pandemic and the lockdown restrictions. However, the 

downward trending variance curve during the trading days in the third quarter of 2020 is due to the easing of 

COVID-19 lockdown restrictions across the world. 

.0

.1

.2

.3

.4

.5

.6

.7

.8

.9

2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020

Conditional variance
 

Figure-4. Conditional variance of the R_SPG series. 

                 

In a bid to capture the asymmetries in the R_SPG series, the EGARCH and TGARCH models were estimated 

as reported in Table 4.  

 
Table-4. Estimated result of EGARCH (1,1) and TGARCH (1,1) Models. 

Parameters EGARCH (1,1) TGARCH (1,1) 

Mean Equation 

 (mean return) 
0.003613 {0.0180}** 0.004534 {0.0056}*** 

Variance Equation 

 (constant) 
-0.347425 {0.0000}*** 0.000353 {0.0000}*** 

(ARCH effect) 
0.247916 {0.0000}*** 0.075499 {0.0000}*** 

 (asymmetric effect) 
-0.129350 {0.0000}*** 0.192600 {0.0000}*** 

(GARCH effect) 
0.965908 {0.0000}*** 0.810012 {0.0000}*** 

 (persistence coefficient) 
1.213824 0.985511 

Log likelihood 2474.041 2459.418 
AIC -1.881846 -1.870696 
SIC -1.868414 -1.857264 
Adjusted R-squared -0.005529 -0.006363 
   

Residual Diagnostics:   

Autocorrelation 
(Q-statistic probabilities) 

Q > 0.05 (No 
autocorrelation) 

Q > 0.05 (No 
autocorrelation) 

Heteroscedasticity: 
F-statistic 0.037119 {0.8472} 0.831677 {0.3619} 

Note: ** and *** indicate rejection of the null hypothesis @ 5%, and 1 % levels of significance, respectively. Figures in parenthesis { 
} are the probability values. 

 



Financial Risk and Management Reviews, 2021, 7(1): 1-15 

 

 
12 

© 2021 Conscientia Beam. All Rights Reserved. 

From Table 4 above, it can be seen that the TGARCH emerged with the lowest AIC and SIC as well as the 

highest Adjusted R-squared in absolute value. Hence, the TGARCH was chosen as the best fit model. Looking at 

the TGARCH estimates, the coefficients of α and β are statistically significant at 1% level. The sum of the ARCH 

(α) and GARCH (β) parameters is 0.985511 which is approximately unity (1), implying that conditional variance 

(volatility) was explosive. The leverage effect coefficient is positive and statistically significant at 1% level, 

which indicates that bad news or negative shocks exert greater influence on volatility of R_SPG than good news 

which provides evidence of leverage effect. This implies that the global stock market exhibited persistent returns 

volatility with leverage effects (asymmetric news effects). Studies such as Edem and Ogbonna (2020); Ndei et al. 

(2019); Tsuji (2018); Banumathy and Azhagaiah (2015); Jegageevan (2015) lend credence to the existence of 

volatility persistence and leverage effects in various stock markets, but Adewale et al. (2016) found no significant 

leverage effect in the Nigerian stock market.  

 

5. CONCLUSION AND RECOMMENDATIONS 

This study modelled stock returns volatility and asymmetric news effect in the global stock market over the 

period 1st Sept., 2010 to 30th Sept., 2020. The global stock market index was measured by the S&P Global 1200 (see 

description in Table 1). Generalized ARCH models such as GARCH (1,1), GARCH-M (1,1), EGARCH (1,1) and 

TGARCH (1,1) were estimated for the empirical investigation. However, the symmetric GARCH-M (1,1) and 

asymmetric TGARCH (1,1) best fit the estimation when compared to other variants of GARCH model. From the 

estimation results, it was found that volatility of R_SPG was highly persistent. Though, the estimation results of 

GARCH-M (1,1) showed that higher volatility did not result to higher R_SPG, implying lack of risk-return trade-

off. On the other hand, the TGARCH (1,1) confirmed evidence of asymmetric volatility process in the global stock 

market, implying presence of leverage effect where bad news influenced R_SPG volatility more than good news. In 

summary, the empirical results significantly suggest that stock returns volatility in the global market persisted over 

a long period with no significant risk-return trade off, and that negative shocks or bad news exerted greater effect 

on global stock returns volatility than positive shocks or good news, especially in the trading days in 2020 which 

could be due to the COVID-19 crisis.  

Based on the findings of this study, market regulators across the globe need to ensure market stability so as to 

accommodate diverse risk-classes of international investors by modelling and aligning trading rules and regulations 

of both developed and emerging stock markets since these markets are largely integrated. As the empirical results 

showed evidence of leverage effect in the global stock market, it is recommended that regulators avail reliable 

platforms for information flow through software application and other possible means to facilitate the ease of 

accessing market information which in turn drives investors investment decisions. Hence, market stability and 

better information dissemination will reduce the magnitude of stock returns volatility and improve transparency in 

the global stock market.   

 

Funding: This study received no specific financial support.    
Competing Interests: The authors declare that they have no competing interests.  
Acknowledgement: Both authors contributed equally to the conception and design of the 
study. 

 

REFERENCES 

Abdennadher, E., & Hallara, S. (2018). Structural breaks and stock market volatility in emerging countries. International Journal 

of Business and Risk Management, 1(1), 9-16.Available at: https://doi:10.12691/ijbrm-1-1-2. 



Financial Risk and Management Reviews, 2021, 7(1): 1-15 

 

 
13 

© 2021 Conscientia Beam. All Rights Reserved. 

Adewale, A. O., Olufemi, A. P., & Oseko, M. S. (2016). Modelling volatility persistence and asymmetry with structural break: 

Evidence from the Nigerian stock market. Journal of Economics and Behavioral Studies, 8(6 (J)), 153-160.Available at: 

https://doi.org/10.22610/jebs.v8i6(j).1489. 

Aguda, N. A. (2016). A test of asymmetric volatility in the Nigerian stock exchange. International Journal of Economics, Finance 

and Management Sciences, 4(5), 263-268.Available at: https://doi:10.11648/j.ijefm.20160405.15. 

Ahmad, M., Ahmed, R. R., Vveinhardt, J., & Streimikiene, D. (2016). Empirical analysis of stock returns and volatility: Evidence 

from Asian stock markets. Technological and Economic Development of Economy, 22(6), 808–829.Available at: 

https://doi:10.3846/20294913.2016.1213204. 

Alagidede, P., & Panagiotidis, T. (2009). Modelling stock returns in Africa's emerging equity markets. International Review of 

Financial Analysis, 18(2), 1-11.Available at: https://https://doi.org/10.1016/j.irfa.2009.02.001. 

Arouri, M. E. H., Lahiani, A., Lévy, A., & Nguyen, D. K. (2012). Forecasting the conditional volatility of oil spot and futures 

prices with structural breaks and long memory models. Energy Economics, 34(1), 283-293.Available at: 

https://doi.org/10.1016/j.eneco.2011.10.015. 

Babikir, A., Gupta, R., & Owusu-Sekyere, E. (2010). Structural breaks and GARCH models of stock return volatility: The case of 

South Africa. Economic Modelling, 29(6), 2435-2443.Available at: 10.1016/j.econmod.2012.06.038. 

Banumathy, K., & Azhagaiah, R. (2015). Modelling stock market volatility: Evidence from India. Managing Global Transitions, 

13(1), 27-42. 

Bekaert, G., & Wu, G. (2000). Asymmetric volatility and risk in equity markets. The Review of Financial Studies, 13(1), 1-

42.Available at: https://doi.org/10.1093/rfs/13.1.1. 

Bello, A. I. (2020). Re-examining the effect of volatility persistence on Nigerian stock market returns: Mean-revert GARCH 

approach. Journal of Accounting and Management, 10(2), 34-40. 

Bentes, S. R., & Da Cruz, M. M. (2010). Is stock market volatility persistent? A fractionally integrated approach. Physica A, 32(5), 

3826-3830. 

Berument, H., & Kiymaz, H. (2001). The day of the week effect on stock market volatility. Journal of Economics and Finance, 25(2), 

181-193. 

Bohl, M. T., & Henke, H. (2003). Trading volume and stock market volatility: The Polish case. International Review of Financial 

Analysis, 12(5), 513-525.Available at: https://doi.org/10.1016/s1057-5219(03)00066-8. 

Bollerslev, T., Chou, R. Y., & Kroner, K. F. (1992). ARCH modeling in finance: A review of the theory and empirical evidence.  

Journal of Econometrics, 52(1-2), 5-59. 

Bollerslev, T. (1986). Generalized autoregressive conditional heteroskedasticity. Journal of Econometrics, 31(3), 307-327.Available 

at: https://doi.org/10.1016/0304-4076(86)90063-1. 

Brealey, R. A., & Meyers, S. C. (2003). Principles of corporate finance. New York: McGraw-Hill. 

Caporale, G. M., Karanasos, M., Yfanti, S., & Kartsaklas, M. (2019). Investors’ trading behaviour and stock market volatility  

during crisis periods: A dual long memory model for the Korean stock exchange. Economics and Finance Working 

Paper Series, No. 1921. 

Chiang, T. C., & Doong, S. (2014). Empirical analysis of stock returns and volatility: Evidence from seven Asian stock markets 

based on TAR-GARCH model. Review of Quantitative Finance and Accounting, 17(1), 301–318.Available at: 

10.1023/A:1012296727217. 

Chiang, T. C., & Doong, S.-C. (2001). Empirical analysis of stock returns and volatility: Evidence from seven Asian stock 

markets based on TAR-GARCH model. Review of Quantitative Finance and Accounting, 17(3), 301-318. 

Cong, H. (2017). Stock market volatility persistence performance of 2008’ crash: Evidence from the BRIC markets. Master’s Thesis 

Finance, Tilburg University, Netherlands.    

Cont, R. (2005). Volatility clustering in financial markets: Empirical facts and agent–based models (pp. 1-21). Ecole 

Polytechnique: Center for Applied Mathematics. 



Financial Risk and Management Reviews, 2021, 7(1): 1-15 

 

 
14 

© 2021 Conscientia Beam. All Rights Reserved. 

De Gaetano, D. (2018). Forecast combinations for structural breaks in volatility: Evidence from BRICS countries. Journal of Risk 

and Financial Management, 11(64), 1-13.Available at: https://doi.org/10.3390/jrfm11040064. 

Dickey, D. A., & Fuller, W. A. (1979). Distribution of the estimators for autoregressive time series with a unit root. Journal of 

American Statistical Association, 74(366), 427–431.Available at: https://doi.org/10.1080/01621459.1979.10482531. 

Ding, Z., Granger, C. W., & Engle, R. F. (1993). A long memory property of stock market returns and a new model. Journal of 

Empirical Finance, 1(1), 83-106. 

Dukes, W. P., Bowlin, O. D., & MacDonald, S. S. (1987). The performance of beta in forecasting portfolio returns in bull and 

bear markets using alternative market proxies. Quarterly Journal of Business and Economics, 11(4), 89-103. 

Edem, C. A., & Ogbonna, U. G. (2020). Evaluation of volatility persistence and asymmetric effect in Nigerian capital market. 

IOSR Journal of Economics and Finance, 11(1), 19-26.Available at: 10.9790/5933-1101061926. 

Engle, R. F., & Patton, A. J. (2001). What good is a volatility model? Quantitative Finance, 1(1), 237-245.Available at: 

https://doi.org/10.1088/1469-7688/1/2/305. 

Engle, R. F. (1982). Autoregressive conditional heteroskedasticity with estimates of the variance of UK inflation. Econometrica, 

50(4), 987–1007.Available at: https://doi.org/10.2307/1912773. 

Fama, E. F. (1995). Random walks in stock market prices. Financial Analysts Journal, 51(1), 75-80.Available at: 

https://doi.org/10.2469/faj.v51.n1.1861. 

Fama, E. F. (1970). Efficient capital markets: A review of theory and empirical work. Journal of Finance, 25(7), 383–417.Available 

at: https://doi.org/10.2307/2325486. 

Fama, E. F. (1991). Efficient capital markets. Journal of Finance, 1575-1617.Available at: https://doi.org/10.1111/j.1540-

6261.1991.tb04636.x. 

Fauzel, N., & Fauzel, S. (2016). Testing weak form efficiency: An international market perspective. International Review of 

Business Research Papers, 12(1), 171-192. 

Ho, J., & Hung, C.-H. (2012). Predicting stock market returns and volatility with investor sentiment: Evidence from eight 

developed countries. Journal of Accounting and Finance, 12(4), 49-65. 

Iqbal, H. R., Saeed, S. K., & Shah, S. Z. A. (2020). Structural breaks and volatility spillover in Southern Asian economies. 

SEISENSE Journal of Management, 3(1), 64-77.Available at: https://doi.org/10.33215/sjom.v3i1.260. 

Jebari, O., & Hakmaoui, A. (2017). Modeling persistence of volatility in the Moroccan exchange market using a fractionally 

integrated EGARCH. Turkish Economic Review, 4(4), 389-399. 

Jegageevan, S. (2015). Return volatility and asymmetric news effect in Sri Lankan stock market. Economic Research 

Department: Central Bank of Sri Lanka, Working Paper Series 166, 37-57. 

Khedhiri, S. (2008). Empirical analysis of the UAE stock market volatility. International Research Journal of Finance and Economics, 

15(2), 250-260. 

Kuhe, D. A. (2018). Modeling volatility persistence and asymmetry with exogenous breaks in the Nigerian stock returns. CBN 

Journal of Applied Statistics, 9(1), 167-196. 

Lai, N. G., Cheong, C. W., & Lee, C. L. (2019). Modelling volatility in the presence of abrupt jumps: Empirical evidence from 

Islamic stock markets. International Journal of Economics and Management, 13(1), 93-109. 

Lo, A. W., & MacKinlay, A. C. (1988). Stock market prices do not follow random walks: Evidence from a simple specification 

test. The Review of Financial Studies, 1(1), 41-66. 

Mandelbrot, B. (1963). The variation of certain speculative prices. Journal of Business, 36(5), 394–419. 

Musa, Y., Adamu, I., & Dauran, M. S. (2020). Forecasting of the Nigeria stock returns volatility using GARCH models with 

structural breaks. Asian Research Journal of Mathematics, 16(2), 39-50.Available at: 

https://doi:10.9734/ARJOM/2020/v16i230174. 

Ndei, C., Muchina, S., & Wawure, K. (2019). Modeling stock market return volatility in the presence of structural breaks: 

Evidence from Nairobi securities exchange, Kenya. International Journal of Research in Business and Social Science, 8(5), 

156-171.Available at: https://doi.org/10.20525/ijrbs.v8i5.308. 



Financial Risk and Management Reviews, 2021, 7(1): 1-15 

 

 
15 

© 2021 Conscientia Beam. All Rights Reserved. 

Ndwiga, D., & Muriu, P. W. (2016). Stock returns and volatility in an emerging equity market: Evidence from Kenya. European 

Scientific Journal, 12(4), 79-98.Available at: http://dx.doi.org/10.19044/esj.2016.v12n4p79. 

Nelson, D. B. (1991). Conditional heteroskedasticity in asset returns: A new approach. Econometrica, 59(2), 347-370.Available at: 

https://doi.org/10.2307/2938260. 

Onyele, K. O., Opara, C. C., & Ikwuagwu, E. B. (2017). Does stock market performance influence capital flight from Nigeria? 

International Journal of Economics and Business Management, 3(7), 22-37. 

Onyele, K., & Ikwuagwu, B. (2020). The globalization wave and stock market return in Africa. International Journal of Economics, 

Business and Management Studies, 7(2), 80-94.Available at: https://doi:10.20448/802.71.80.94. 

Onyele, K. O., Ikwuagwu, E. B., & Onyekachi-Onyele, C. (2020). Macroeconomic policies and stock market liquidity: Evidence 

from Nigeria. Economy, 7(1), 25-35.Available at: https://doi:10.20448/journal.502.2020.71.25.35. 

Owidi, O. H., & Mugo-Waweru, F. (2016). Analysis of asymmetric and persistence in stock return volatility in the Nairobi 

securities exchange market phases. Journal of Finance and Economics, 4(3), 63-73.Available at: https://doi:10.12691/jfe-

4-3-1. 

Phillips, P. C., & Perron, P. (1988). Testing for a unit root in time series regression. Biometrika, 75(2), 335-346.Available at: 

https://doi.org/10.1093/biomet/75.2.335. 

Sansa, N. A. (2020). The impact of the COVID - 19 on the financial markets: Evidence from China and USA. Electronic Research 

Journal of Social Sciences and Humanities, 2(2), 29-39. 

Sethapramote, Y., & Prukumpai, S. (2012). Structural breaks in stock returns volatility: Evidence from the Stock Exchange of 

Thailand. The Empirical Econometrics and Quantitative Economics Letters, 1(3), 113-130. 

Tsay, R. S. (2010). Analysis of financial time series (3rd ed.). Hoboken: John Wiley & Sons. 

Tsuji, C. (2018). Return transmission and asymmetric volatility spillovers between oil futures and oil  equities: New DCC-

MEGARCH analyses. Economic Modelling, 74(2), 167-185.Available at: 

https://doi.org/10.1016/j.econmod.2018.05.007. 

Wang, J., & Yang, M. (2017). Conditional volatility persistence. European Financial Management, 19(4), 1-60.Available at: 

http://dx.doi.org/10.2139/ssrn.3080693. 

Wang, J., Yao, J., & Bonne, G. (2020). The coronavirus market impact spreads globally. Retrieved from 

https://www.msci.com/www/blog-posts/the-coronavirus-market-impact/01732620365. 

Wang, C., Huang, C., & Padmanabhan, P. (2015). Transitory and permanent volatility components in the main US ETF markets: 

An empirical analysis. Frontiers in Finance and Economics, 13(2), 19-40. 

Wei, C.-C. (2009). An empirical analysis of the Taiwan institutional trading volume volatility spillover on stock market index 

return. Economics Bulletin, 29(2), 1264-1275. 

Zakoian, J.-M. (1994). Threshold heteroskedastic models. Journal of Economic Dynamics and Control, 18(5), 931-955.Available at: 

https://doi.org/10.1016/0165-1889(94)90039-6. 

 

 

 

 

 

 

Views and opinions expressed in this article are the views and opinions of the author(s), Financial Risk and Management Reviews shall not be responsible or 
answerable for any loss, damage or liability etc. caused in relation to/arising out of the use of the content. 

 

http://dx.doi.org/10.19044/esj.2016.v12n4p79
http://dx.doi.org/10.2139/ssrn.3080693
http://www.msci.com/www/blog-posts/the-coronavirus-market-impact/01732620365

