412 Afro-Asian J. Finance and Accounting, Vol. 14, No.

3, 2024

Structural change, information asymmetry and

volatility in Indian stock market: evidence from
pre-and-post-COVID-19 outbreak

T. Mohanasundaram* and M. Rizwana

Department of Management Studies,
M.S. Ramaiah Institute of Technology,
Bangalore, 560054, India
Email: tmohansun@[Link]
Email: rizumehar@[Link]
*Corresponding author

S. Sathyanarayana
M.P. Birla Institute of Management,
Bangalore, 560001, India
Email: sathya4u.s@[Link]

Padmalini Singh
R.V. Institute of Management,
Bangalore, 560041, India
Email: padmalinisingh@[Link]

Abstract: This paper examines the Indian stock market’s interconnection with
the stock markets of the top four economies before and after the COVID-19
outbreak. The log-returns of daily data for Sensex, S&P 500, SSE Composite,
Nikkei 225 and DAX were used in the study. The log-return series of all stock
indices were found to be stationary. The exponential-GARCH model is applied
to assess the information asymmetry and to model the volatility spillover on the
Indian stock market. The ARCH and GARCH terms were positive and
significant during both the pre-COVID and post-COVID outbreak periods
representing that market news and previous period variances were significantly
increasing the volatility in the market. The ensemble of events during the
pre-COVID period confirms the negative significant volatility spillover of
bourses on the Indian markets, and continues to be so in the post-COVID
period, except in case of the US market where there is a positive significant
return and volatility spillover. The portfolio managers, regulators, policymakers
and other market participants may consider the change in information
transmission during the pre-COVID and post-COVID-19 outbreak phases from
these foreign markets to India while making investment-related decisions.

Keywords: COVID-19; market returns; volatility; investment; asymmetry;

spillover; EGARCH.

Copyright © 2024 Inderscience Enterprises Ltd.

 Structural change, information asymmetry and volatility 413

Reference to this paper should be made as follows: Mohanasundaram, T.,

Rizwana, M., Sathyanarayana, S. and Singh, P. (2024) ‘Structural change,
information asymmetry and volatility in Indian stock market: evidence from
pre-and-post-COVID-19 outbreak’, Afro-Asian J. Finance and Accounting,
Vol. 14, No. 3, pp.412–431.

Biographical notes: T. Mohanasundaram received his Doctoral in Finance

from the Anna University, India. He is currently working as an Associate
Professor at the M.S. Ramaiah Institute of Technology, Bangalore, India. He
has more than 18 years of collective experience in industry and academics. He
has published several articles in reputed journals and is actively involved in
consultancy works for corporates.

M. Rizwana is a graduate in Applied Science (IT) and a Post-graduate in MBA,

and currently working as an Associate Professor in the Department of
Management Studies and Research Centre, M.S. Ramaiah Institute of
Technology, Bangalore. She obtained her Doctoral degree from the Bharathiar
University, Coimbatore. She has contributed several papers and articles to
various conferences and journals both at National and International levels. Her
areas of interest include entrepreneurship development, industry institute
interaction, environmental management, services marketing and consumer
behaviour.

S. Sathyanarayana has over 21 years of teaching experience at various premier

management institutes. He published many research articles in the field of
finance, marketing, HR and general management. He presented many papers at
national and international conferences and won many ‘best paper’ awards
besides chairing conference sessions at various conferences.

Padmalini Singh is a Marketing Professor whose interests span the fields of

consumer behaviour, digital marketing and entrepreneurship that she developed
while working in the ISB, Hyderabad and IIM Ranchi. She teaches courses on
marketing management, consumer behaviour, advertising, digital marketing
and international marketing. A pure educationist by choice, she has been
involved with teaching and research in marketing field all her professional life.

1 Introduction

Many empirical research studies have explored how information transmission and shocks
influence financial market volatility. These empirical works have identified the channels
through which such shocks are transmitted. The structural change in the relationship
between the financial markets is evident in many studies. In terms of the financial market,
structural changes denote a persistent and long-term change in the composition of
information flows to one market from other markets. In the recent past, liberalisation and
deregulation of several stock markets have led to a surge in the inflow of direct and
portfolio investments from advanced economies to other economies. The factors of
information transmission from one market to other markets have been due to regulations,
necessary reforms, global ease of trade and investment policies, macro-economic
similarities, capital inflows, foreign assets and liabilities, free workforce mobility,
advances in information technology, rapid broadcasting speed of global news, etc. These
factors have improved the information transmission between the markets (Booth et al.,
414 T. Mohanasundaram et al.

1997; Dornbusch et al., 2000; Gallo and Otranto, 2007). These transmissions led to
profound growth in the financial market and economy. The economic and financial
market integration across the globe has been enlarged and reinforced (Singh et al., 2010).
The stock market in particular is vulnerable to social, economic and financial shocks.
There are numerous instances where the stock markets have witnessed such a crisis.
Some of the notable economic and financial shocks are: Mexican Peso Crisis (1994), the
Asian Financial Crisis (1997), the Subprime Crisis in the USA (2007–2008), the
European Debt Crisis (2010), Brexit Referendum (2016), etc. There are adequate
empirical studies that have found the association between the global markets to have
strengthened and improved significantly because of the crisis (Cha and Oh, 2000; Hoque,
2007). Whenever the crisis has occurred, the literature has given greater emphasis on the
financial market volatility and spillover among the major developed and emerging
markets. In the present study, an attempt is thus made to address the key objectives of
understanding the Indian stock markets’ interconnectedness with the top four economies
of the world. The study also investigates the asymmetric responses and the spillover
effect before and after the COVID-19 outbreak from these economies to the Indian stock
markets.
The novel coronavirus (COVID-19) cases were first observed in China, before
spreading to other parts of the world. The growth rate of COVID-19 cases was
exponential and in no time had affected millions of people causing social, economic and
global meltdown. The fast spread of this virus compelled the World Health Organization
(WHO) to declare this as a pandemic and forced almost all the countries to announce a
lockdown to fight against the virus. The sudden and unexpected nationwide lockdown
brought lot of chaos in society and severely hit the livelihood of many people. The
governments’ financial assistances and relief measures helped the citizens and businesses
in overcoming the economic strain to some extent, but this led to a dampening effect in
the governments’ budget. The critical effect of this epidemic was found to be much more
than the apprehensions of public and on public health apprehensions (Bakhshi and
Chaudhary, 2020). The pandemic was found to have significant negative impact on the
markets throughout the world. Financial integration and well connectedness of the global
markets ensured that, the impact had severe ramifications across the global financial
markets within few months. Therefore, it was felt very essential to analyse the volatility
of financial assets and the effects of information spillover from one economy to another
economy to this pandemic from the perspective of policymakers and portfolio managers
as it created uncertainty and thus adversely affecting growth prospects (Poshakwal and
Murinde, 2001). The volatility in the stock market hampers economic performance
through consumer spending (Garner, 1988) and it may also affect business investment
spending (Gertler and Hubbard, 1988). Higher volatility in the stock market is thus
perceived as an increased risk in the equity market and may lead to the movement of
funds from risky investment to less-risky investment avenues. As the volatility spillover
generally results from the interdependence among the nations, it is essential to understand
the underlying drivers of cross-country stock market correlation and volatility from the
perspective of market participants and policymakers (Baele, 2005).
The variations in return and volatility of the stock market and the succeeding
contagion because of information transmission across the worldwide financial markets
during the crisis period grasped extensive attention from researchers, policy-makers,
investors and other stakeholders because, the current coronavirus crisis is different from
the past ones in its velocity and intensity. According to Bloomberg data, before
 Structural change, information asymmetry and volatility 415

COVID-19, the total market capitalisation in India was about $2.16 trillion but the
COVID-19 outbreak caused market capitalisation to reduce to $1.3 trillion on 23rd March
2020. The Indian capital markets were observed to loom under a lot of anxiety despite the
string of financial reforms introduced by RBI and the Government of India. The
pandemic has wedged the premise of the corporate sector and the uncertainty prevails
(Ravi, 2020). Among the various measures undertaken, the nation-wide lockdown
announced by the Indian Government help check the contagion and revive the economy
later on. The subsequent periods after the lockdown helped the country’s stock market
capitalisation to cross the $2.5 trillion mark, thereby attaining eighth position among the
global stock markets and fourth in terms of stock market capitalisation growth rate of
17.4% next to China, the USA and Hong Kong. Given this context, the present study
attempts to understand and measure how Indian markets are interconnected with the stock
markets of top-four foreign economies of the globe. The study is different from the
previous studies by investigating the response of Indian stock markets to the information
flow from the stock markets of top four economies and by examining the possible change
in information transmission and volatility spillover before and after the COVID-19
outbreak. Earlier studies have not examined the information transmission and volatility
spillover from the top four economies considered in this study with Indian stock markets
specifically during the pre-and-post-COVID-19 outbreak. This study enables to
understand whether COVID-19 pandemic crisis unlike any other financial crisis
observed, induced any significant volatility spillover effects on Indian stock markets from
the stock markets of four major economies in the world which is not found in the
literature. The results of the study can be considered by the small investors and portfolio
managers while following portfolio allocation methods in case of volatility spillover. The
results further enable the regulators to understand how the global markets will have the
impact on the Indian economy represented by the broader indices like BSE Sensex in
investment related decision making.
The current empirical study proceeds as follows: Section 2 briefs out the review of
available literature on the planned topic. Section 3 highlights the theoretical framework
and research design adopted for the current study to realise the stated objectives. Section
4 deals with the analysis of data and discussion on the outcome. In Section 5, the
summary of the major outcomes is given and an expressive conclusion has been drawn.
The important outcomes are compared with other similar existing studies.

2 Literature review

The most imperative task in estimating the returns and volatility is to understand the
essence and amount of spillover in financial information and its transmission across
international stock markets. Numerous modelling techniques are applied to capture
volatility clustering in financial data. Engle (1982) proposed ARCH, Bollerslev (1986)
proposed GARCH and Nelson (1991) discovered exponential GARCH (EGARCH).
Later, various ARCH extension models have been developed to capture the volatility and
spillover effects. Several researchers attempted to discover the economic integration of
global stock markets (Bae and Karolyi, 1994; Worthington and Higgs, 2004; Li, 2007;
Yepes-Rios et al., 2015; Paramati et al., 2016). Sabri (2004) examined the increasing
stock return volatility of emerging economies such as Mexico, Korea, South Africa,
416 T. Mohanasundaram et al.

Turkey and Malaysia. The study found that the volume of stock trading and exchange
rates were common predicting factors for the volatility of these emerging markets.
Mishra et al. (2007) explored the long-run relationship and bidirectional volatility
spillover between the stock and forex markets in India using GARCH, EGARCH and
cointegration approach. Bhargava et al. (2012) investigated if the volatility in the US
dollar interest rate swap market spillover to the Indian swap market. The result shows
that there is a unidirectional volatility spillover from US swap market to the Indian swap
market. Moreover, the volatility spillover impact was found to be asymmetric for
one-year swaps. Kumar (2013) found the presence of return and volatility spillover
effects between forex rates and stock prices within the India, Brazil and South Africa
(IBSA) countries. Li and Giles (2014) investigated the stock market interconnections and
volatility spillover between the USA, Japan and six emerging countries in Asia for the
period of 20 years from 1993 to 2012 using asymmetrical MGARCH model. The study
found a unidirectional volatility spillover from the USA to the Japan and the Asian
emerging markets. Further, the study exposed the bidirectional volatility spillover
between the US market and the Asian markets during the Asian financial crisis.
Researchers across the world have shown a keen interest in modelling the effect of
the COVID-19 virus outbreak on stock market movements and its consequent influence
on spillover effects. Anh and Gan (2020) discovered the impact of COVID-19 spread and
the following lockdown on daily stock returns in Vietnam. The panel-data analysis was
used to estimate the effect of the daily upsurge in the COVID-19 infected cases during
pre-lockdown and lockdown on Vietnam’s 723 listed firms. The study found a contrary
behaviour of the Vietnam stock market before and during the lockdown phase. The study
has established that Vietnam’s stock market was hit hard during the COVID-19 outbreak.
Thakur (2020) analysed the US stock market movements during the COVID-19 epidemic
by employing the VAR model and using time series data from 23rd Jan. 2020 to
19th June 2020. The result showed that the Standard and Poor (S&P) index has exhibited
an adverse causality effect with the upsurge in the number of new cases at the global
level. Yan et al. (2020) analysed the consequence of COVID-19 on the stock market and
probable investing strategies by selecting various industries like the entertainment sector,
travel and tourism and gold investment. The study established that the pandemic has an
adverse effect in the short-run but in long-run, it will make a course correction and
increase.
Baker et al. (2020) opined that compared to any other infectious disease which had hit
the world, the COVID-19 pandemic has left a strong trace on the US stock market. The
study used text-based methods to assess volatility in the stock market and assessed the
potential explanations for the incomparable stock market response to the COVID-19
epidemic. Narayan et al. (2020a, 2020b) investigated the association between the
Japanese yen and the country’s stock returns and found that the decline in the value of the
Japanese yen against the US dollar has resulted in gains in the returns of Japanese stock
and the association was strong during the COVID-19 phase – January 2020 to August
2020 – in comparison to the pre-COVID period. Baek et al. (2020) have conducted an
industry-level analysis on the impact of COVID-19 on US stock market volatility. The
features viz., machine learning methods and economic indicators were used to identify
the regime change in volatility. The study found that volatility is largely determined by
certain macroeconomic indicators and is sensitive to COVID-19 that too negative news is
more impactful than positive news, suggesting a negativity bias. Adnan et al. (2020)
analysed the response of capital markets to COVID-19 by using everyday individual
 Structural change, information asymmetry and volatility 417

stock’s return of 311 registered firms. The study revealed that the domestic stock market
displays an astonishing market response to the announcement.
Capelle-Blancard and Desroziers (2020) assessed the stock markets’ integration on
account of COVID-19 news, the successive lockdowns and the policy announcements.
The study used a panel of 74 countries for the period of four months from January 2020
to April 2020 and analysed the effect of the pandemic on the stock market. The outcome
of the study revealed that during the earlier stage of the pandemic, the stock markets have
overlooked the virus epidemic, before reacting severely to the mounting number of
diseased people. The stock market volatility increased and apprehensions about the
pandemic raised. Later, share prices recovered across the world. The characters that are
specific to the concerned nation had no impact on the behaviour of the stock market but
the credit schemes, government assurances and low-interest rates reduce the downfall of
stock markets. Overall, the stock prices had been lesser sensitive to the particular
country’s economic variables before the crisis compared to its short-term response at the
time of crisis. Cao et al. (2020) using panel data analysis found that the stock price index
inclined to move with COVID-19’s domestic and global spreads. Sharif et al. (2020)
examined the association between the COVID-19 spread, oil price instability, stock
market price volatility, geopolitical risk and economic policy ambiguity in the USA. The
analysis revealed that the impact of the COVID-19 on the geopolitical risk is
considerably greater than the economic uncertainty of the USA. The risk of COVID-19
spread is observed differently over the short-run and long-run.
Gherghina et al. (2020) examined the influence of COVID-19 on the financial market
by considering daily stock market returns for seven countries viz., the USA, Germany,
the UK, Spain, Italy, France and Romania. The study applied an autoregressive
distributed lag (ARDL) model and the impact of the pandemic outbreak on the Romanian
stock market was analysed. The empirical study has provided affirmation that the
ten-year old government bond of Romania is more highly influenced by news related to
COVID-19 than the stock index of the Bucharest Stock Exchange. Yong and Laing
(2020) investigated the reaction of the US stock market to WHO’s COVID-19 Global
Emergency news with the main focus on firms having global exposure. The study found
that international acquaintance through imports and exports and foreign assets were
important and adversely related with standardised cumulative abnormal returns in the
short-run, the effect reverses in the long-run. He et al. (2020) have conducted an event
study to understand the performance of the stock market and behavioural trends of
Chinese industries to the COVID-19 pandemic. The study exposed that the predominant
sectors like electricity, transportation, environment and mining industries that were
unfavourably hit by the pandemic.
To sum up, despite many studies being available on the COVID-19 pandemic and its
impact on the stock market, there are hardly any studies that are intended to capture the
changing dynamism of information spillover and volatility in the Indian stock market due
to COVID-19 spread. Therefore, the current study tries to understand the response of the
Indian stock markets to the information flow by examining the volatility spillover from
the stock markets of the top-four economies namely the USA, Japan, China and
Germany.
418 T. Mohanasundaram et al.

3 Objectives, methodology and framework

Although many research studies focused on capturing the spillover effects between the
financial markets and its effect on market volatility, there was no specific study found on
examining the change in spillover effect between the chosen stock markets due to the
COVID-19 pandemic. Thus, this study is carried out to:
1 Explore the possible structural change in Indian stock market movements due to
COVID-19 pandemic.
2 Examine the information transmission from other advanced stock markets to the
Indian stock market during pre-and-post-COVID outbreak.
3 Investigate asymmetries in the Indian stock market in response to positive and
negative news.
This study empirically assesses the possible asymmetry in information transmission and
volatility spillover from stock markets of the top four economies to the Indian stock
market. As of 2019, the top five economies in terms of nominal GDP are: the USA,
China, Japan, Germany and India ([Link]
economies-in-the-world-and-their-growth-in-2020-2020-01-22). Table 1 describes the
secondary data sources for the stock market data used in the study.
Table 1 Sources of data

Stock index Frequency Source

Sensex (India) Daily [Link]
S&P 500 (USA) [Link]
SSE Composite Index (China)
Nikkei 225 (Japan)
DAX (Germany)

The daily adjusted closing stock price data from January 2019 to February 2021 are used.
The stock price returns were computed using the formula given in equation (1).
Pt
Rt = log (1)
Pt −1

Pt is the adjusted closing price of the stock index on day t, and Pt–1 is the adjusted closing
price of the stock index for the previous day.

3.1 Descriptive statistic and trend analysis

Descriptive statistics numerically summarises the dataset to provide greater insight into
data. It includes measures of central tendency and measures of dispersion. The
descriptive statistic will form a base for advanced inquiry. Jarque-Bera (JB) test helps to
identify whether the underlying time series data is following normal distribution or not.
The null hypothesis (H0) of the JB test is ‘Data series is normally distributed’. The JB test
statistic identifies whether the underlying time series have skewness and kurtosis
matching normality assumption. Trend analysis is used to recognise the pattern in data
 Structural change, information asymmetry and volatility 419

and used to predict future movements based on the observed pattern. The trend analysis
offers valuable input to make better decisions.

3.2 Test for stationarity

‘Unit-root with break test’ is employed to test the structural change and stationarity of
data. Structural change refers to a paradigm shift in the way the market operates. When
structural breaks are present, the Dickey-Fuller (DF) tests are biased toward the
non-rejection of a unit root. Regressing one random walk time series data on another
random walk time series data may result in spurious regression. To avoid this,
it is necessary to ensure the underlying time series is stationary. The augmented
Dickey-Fuller (ADF) unit-root test is useful to find out whether the data series is
stationary or not. ADF test can control for autocorrelation in error terms by adding the
lagged values of the dependent variable as an independent variable in the regression
equation.

3.3 Test for data distribution

The normal distribution of data is tested through JB test. It is a goodness-of-fit test to
determine whether the collected time series data (sample) have the kurtosis and skewness
matching normality assumption. The JB test equation is:
n  2 1 
JB =  S + ( K − 3) 2   (2)
6  4 
S = skewness, K = kurtosis and n = sample size.

3.4 Ordinary least squares regression analysis

It is a popular statistical method to analyse the connection between the predictor
variable(s) and predicted variable. Ordinary least squares (OLS) is considered to be one
of the most powerful estimation techniques used in statistical analysis which relies on
certain assumptions. The estimation results of OLS are considered to be consistent when
the regressors are exogenous, having unbiased estimators, and when the error terms are
homoscedastic and serially uncorrelated.

3.5 Diagnostic tests

The two important diagnostic tests are the autocorrelation test and the heteroscedasticity
test. The residuals of the estimated model should be free from autocorrelation and
heteroscedasticity. In simple words, autocorrelation in the econometric model refers to
the correlation of error terms with its past values. Heteroscedasticity refers to varying
variance. An ideal statistical model should have equal variance in its error terms.

3.6 EGARCH model

Nelson (1991) has offered an EGARCH model on a logarithmic expression. It is an
improvement over the GARCH model and it is useful in recognising the presence or
420 T. Mohanasundaram et al.

absence of leverage (asymmetry) effect. In a traditional GARCH model, both positive

and negative error terms are presumed to have a symmetric effect on the volatility, i.e.,
good and bad news have a similar effect on the volatility. In reality, this assumption in
stock market returns is mostly violated. In the stock market, volatility increases more for
bad news than for good news.
In order to examine the volatility spillover, we follow the stepwise procedure
proposed by Jebran and Iqbal (2016). First, we compute the returns of stock indices to
examine the conditional variance of stochastic segments of returns. Secondly, in order to
examine the ARCH effect, we apply the ARCH test by considering the first lag of all
variables and checking for the significance of the chi-square statistic. If significant, it
provides confirmation for using EGARCH model on variables with problem of
autocorrelation and heteroscedasticity. As a third step, we examine the cross-market
volatility spillover between the indices by first generating the volatility residual series
from a specific EGARCH model for each index separately which acts a proxy for shock
emanating to other markets and then examining the volatility spillover using
EGARCH (1, 1) from foreign markets to Indian markets by using the volatility residual
series of foreign market indices as shock emanating to Indian markets and vice versa.
The EGARCH model to examine volatility spillover from foreign markets to Indian
markets is given below:
Rt ( Indian ) = α 0 + α1 Rt −1 + α 2 Rt −1( foreign ) + εt (3)

εt − 1 εt − 1
ht ( Indian ) = β 0 + β1ht −1 + β 2 +φ + δresid ( Foreign ) (4)
ht −1 ht −1

Equation (3) is the conditional mean and equation (4) is the conditional variance
equation. In the conditional mean equation, Rt(Indian) is the return of Indian indices and α2
measures the foreign market changes on Indian indices. In the conditional variance
equation, ht(Indian) represents log of conditional variance of Indian indices, β0 represents
constant of volatility, β1ht–1 represents the consistency which is function of volatility,
ε −1 ε −1
β2 t captures the impact of changes in news on volatility, φ t measures
ht −1 ht −1
asymmetric effect of volatility and finally, δresid(Foreign) examines the volatility spillover
emanating from foreign market to Indian market.

4 Analysis and discussion

4.1 Time plot and descriptive statistic

The daily closing value of stock market indices of India (Sensex), the USA (S&P 500),
China (SSE), Germany (DAX) and Japan (Nikkei) are considered in this study. The
period of the study is from January 2019 to February 2021. The daily closing values are
converted into log returns for analysis purpose. The primary step in a time series analysis
is to plot the variable’s data series against time and observe its movement. This is a
simple and easier method to understand how the variable performed over time. The
presence of trend in either mean or variance implies that the data is non-stationary.
Figure 1 shows the time plot of return series of Sensex, S&P 500, SSE, DAX and Nikkei.
 Structural change, information asymmetry and volatility 421

Figure 1 Time plot of stock market indices log returns (see online version for colours)

The stock market returns of all the economies considered in the study show massive
volatility in their returns during the end of the first quarter in 2020 on the account of
widespread COVID-19 across the globe. Most of the economies in the world
implemented a stringent lockdown which completely restricts the movement of people
and halted all economic activities. This sudden uncertainty caused by the pandemic
created distress in the financial markets, especially in the stock market. It is evident from
Figure 1 that volatility has increased in all the stock market returns from the 1st quarter of
2020, i.e., since the COVID-19 outbreak compared to prior periods. The stock market
returns time series plot exhibits continuous mean reversion. The volatility clustering
(large shocks cluster together and small shocks cluster together) seems to be present in
the stock returns time plot.
422 T. Mohanasundaram et al.

The descriptive statistic of the return series of all stock market indices summarises
how the data series are distributed. Table 2 displays the result of descriptive statistics.
The standard deviation is higher than the mean and median for all index returns. The
maximum return for Sensex is 8.59% and the minimum is –8.53%. The ideal values of
skewness and kurtosis should be 0 and 3 respectively for a data series following a normal
distribution. The negative skewness values of Sensex, S&P 500, SSE and DAX indicate
that data are negatively skewed, i.e., data series has a longer left tail. Higher kurtosis
values represent a leptokurtic characteristic of the distribution. The probability value less
than 5% indicate that data are not normally distributed in all the five stock market return
series.
Table 2 Descriptive statistics

Indices Sensex S&P 500 SSE DAX Nikkei

Mean 0.00071 0.000911 0.000742 0.00055 0.000874
Median 0.001499 0.001842 0.000699 0.001068 0.000757
Max. 0.085947 0.089683 0.075482 0.104143 0.077314
Min. –0.085316 –0.127652 –0.080392 –0.130549 –0.062736
SD 0.016486 0.017201 0.013048 0.016954 0.014014
Skewness –0.591564 –1.353641 –0.293741 –0.849565 0.227813
Kurtosis 10.8521 18.18224 9.724949 16.72973 8.435602
Jarque-Bera 1,179.658 4,449.396 852.5401 3,580.633 556.6359
Prob. 0.0000 0.0000 0.0000 0.0000 0.0000

4.2 Unit root with breakpoint test

The underlying time series data are tested for the presence of unit-root. The stationary
time series data is desirable for modelling and estimating the relationship between the
variables because most statistical tests and techniques depend on the assumption that
statistical properties remain constant over time. As the purpose of the study is to find out
the possible change in information transmission and volatility in the Indian stock market
between the pre-and-post-COVID-19 outbreak, ‘unit root with break test’ is carried out to
investigate the presence of a structural break in Sensex return series. The test statistic
used is the ADF test. The ADF test includes the lagged difference terms of the dependent
variable in the equation, to have a serially uncorrelated error term. Table 3 displays the
results of the Breakpoint unit root test for Sensex and ADF test results of other stock
indices return series.
The null hypothesis ‘Data series has the unit root’ is rejected for all the log-returns of
stock indices. This reflects that all the underlying variables are stationary at the level. The
unit root test with breakpoints indicates Sensex returns have a structural break on
03/23/2020. Considering the breakpoint date as a base, the information transmission and
volatility before and after the structural break date is analysed in the study.
 Structural change, information asymmetry and volatility 423

Table 3 Unit root with breakpoint test

Null hypothesis: data series has unit root

Break date: 03/23/2020
Break selection: minimise Dickey-Fuller t-statistic
ADF test statistic t-statistic Prob.
Sensex –22.93429 0.0000
S&P 500 –12.50045 0.0000
SSE –18.8393 0.0000
DAX –12.38875 0.0000
Nikkei –19.92737 0.0000

4.3 OLS regression for pre-and-post-COVID-19 outbreak

A regression model having Sensex returns as the dependent variable and the returns of
S&P 500, SSE, DAX and Nikkei as independent variables has been estimated during the
pre-COVID-19 outbreak period (1st January 2019 to 22nd March 2020) to assess the
return spillover from the top four economies in the world to Indian stock market.
Similarly, to assessment was also done to understand the return spillover after the
structural break date too. Table 4 provides OLS regression estimation result during the
pre-COVID-19 period and post-COVID-19 period.
Table 4 OLS regression during pre-and-post-COVID-19 phase

Dependent variable: Sensex Method: least squares

Null hypothesis: no significant relationship between IV and DV
Period: pre-COVID-19 Obs.: 248
Variables Coeff. Std. error t-statistic Prob.
S&P 500_R 0.223994 0.061096 3.666238 0.0003
SSE_R 0.106319 0.057897 1.836363 0.0675
DAX_R 0.245322 0.075956 3.229777 0.0014
NIKKEI_R 0.120479 0.074465 1.617924 0.107
C -0.000621 0.000681 -0.913004 0.3621
R-squared: 0.401856 Adj. R-squared: 0.39201
F-statistic: 40.81425 Prob. (F-stat): 0.0000 Durban-Watson stat: 1.925785
Period: post-COVID-19 Obs.: 201
Variables Coeff. Std. error t-statistic Prob.
S&P 500_R 0.364841 0.089295 4.085797 0.0001
SSE_R 0.184296 0.100484 1.834091 0.0682
DAX_R 0.082977 0.089386 0.928299 0.3544
NIKKEI_R 0.313701 0.083222 3.769428 0.0002
C 0.000604 0.00113 0.534143 0.5938
R-squared: 0.34552 Adj. R-squared: 0.332163
F-statistic: 25.86859 Prob. (F-stat): 0.0000 Durban-Watson stat: 2.138002
424 T. Mohanasundaram et al.

During the pre-COVID-19 phase, the null hypothesis of ‘No significant relationship
between the independent variable and dependent variable’ is rejected for all variables
except for Nikkie as the probability value of t-statistic is less than 5% level. Similarly,
during the post-COVID-19 period, all the independent variables viz., S&P 500, SSE and
Nikkei are significant at the 5% level. Considering the pre-COVID period, the adjusted
R2 is 39.20% which shows the explanatory power of the model. The overall model is
significant as the probability value of the F-statistic is significant. The Durban-Watson
test statistic is closer to 2 indicating there is no first-order autocorrelation in error terms.
In the post-COVID-19 phases, the adjusted R2 is 33.21% which exhibits that the model’s
explanatory power is lower than the pre-COVID period. The Durban-Watson value is
2.13 which is higher than the desired value of 2.

Figure 2 Residual time plot of OLS regression, (a) pre-COVID-19 residual plot
(b) post-COVID-19 residual plot (see online version for colours)

(a)

(b)

The pre-and-post-COVID-19 OLS model residuals are displayed in Figure 2. The

residuals in the plot are clustered based on the size of movements, i.e., there is a volatility
 Structural change, information asymmetry and volatility 425

clustering, as the big movements in residuals are clustered together and small movements
in residuals are clustered together. Such clustering in residuals is violating the constant
variance assumption of OLS. The OLS estimation is reliable only if the model has
constant variance and free from autocorrelation issue. Thus, to verify whether or not the
residuals of the model is abiding with the OLS assumption of constant variance and no
autocorrelation, the ARCH heteroscedasticity test and Breusch-Godfrey serial correlation
LM test are carried out for both pre-and-post-COVID-19 periods. The results are shown
in Table 5.
Table 5 Diagnostic tests for OLS regression analysis

ARCH heteroscedastity test NH: no ARCH effect

Pre-COVID-19 outbreak
F-statistic 61.75259 Prob. F (1, 245) 0.0000
Obs. ∗ R-squared 49.72375 Prob. chi-square (1) 0.0000
Post-COVID-19 outbreak
F-statistic 19.27414 Prob. F (1, 198) 0.0000
Obs. ∗ R-squared 17.74177 Prob. chi-square (1) 0.0000
Breush-Godfrey serial correlation test NH: no serial correlation
Pre-COVID-19 outbreak
F-statistic 0.766687 Prob. F (2, 241) 0.4657
Obs. ∗ R-squared 1.567937 Prob. chi-square (2) 0.4566
Post-COVID-19 outbreak
F-statistic 4.666298 Prob. F (2, 194) 0.0105
Obs. ∗ R-squared 9.225534 Prob. chi-square (2) 0.0099

The null hypothesis of the ARCH test is rejected both in pre-and-post-COVID-19

phases as the p-value is less than 5% level indicating there is an ARCH effect
(heteroscedasticity) in the models’ residual. The p-value of the Breush-Godfrey serial
correlation test for the pre-COVID phase is higher than the 5% level representing there is
no autocorrelation issue. However, the p-value is less than 5% level for post-COVID-19
outbreak, signifying that the model suffers from the autocorrelation issue. The presence
of the ARCH effect obliges to use ARCH framework to address the heteroscedasticity
issue.

4.4 Volatility spillover and leverage effect

To assess the information asymmetry and to model the volatility spillover in the Indian
stock market, the EGARCH model is used. The traditional GARCH model is symmetry
to all the information and therefore fails to capture the market’s asymmetric response to
the good news and bad news. In this context, the traditional GARCH model’s conditional
variance fails to capture the Indian stock market’s asymmetric response to positive and
negative shocks received from other stock markets. This possible leverage effect
(asymmetric response) is more valuable in predicting market volatility. Table 6 to Table 7
represents the outcomes of the EGARCH model during both pre-and-post-COVID-19
outbreak.
 426

Table 6
model

SENSEX_R = α0 + α1 ∗ SENSEX_R(–1) + α2 ∗ FOREIGN MARKET INDICES_R(–1)

LOG(GARCH) = β0 + β1 ∗ ABS(RESID(–1) / @SQRT(GARCH(–1))) + β1 ∗ RESID(–1) / @SQRT(GARCH(–1)) + φ ∗ LOG(GARCH(–1)) + δ ∗ RES_FOREIGN MARKET
INDICES
Mean equation coefficients Variance equation coefficients
Indices
φ δ
T. Mohanasundaram et al.

α0 α1 α2 β0 β1 β2
Pre-COVID-19 outbreak phase Obs.: 248
S&P 500 (USA) → Sensex 0.000** –0.030** 0.171** –0.931** 0.062 0.221** 0.906** –17.504**
SSE Composite (China) → Sensex 0.000 0.003 0.055 –0.805** 0.106* 0.263** 0.921** –8.319**
Nikkei 225 (Japan) → Sensex 0.000 –0.013 0.113 –0.882** 0.119** 0.226** 0.915** –12.957**
DAX (Germany) → Sensex 0.001 –0.032 0.067 –5.195** 0.426** 0.262** 0.487** –38.708**
Post-COVID-19 outbreak phase Obs: 201
S&P 500 (USA) → Sensex 0.001 –0.033 0.326** –0.281** 0.156** 0.137** 0.954** 3.940**
SSE Composite (China) → Sensex 0.002 0.021 –0.098 –0.179** 0.125** 0.130** 0.968** –1.361
Nikkei 225 (Japan) → Sensex 0.002 –0.013 –0.013 –0.279** 0.005 0.163** 0.968** –0.892
DAX (Germany) → Sensex 0.002 –0.095 0.157 –0.250 0.020 0.156** 0.973** –2.326
Note: **Significant at 1% and *significant at 5%.
Spillover effects from foreign stock markets to Indian stock market using EGARCH
 Structural change, information asymmetry and volatility 427

The first part of Table 6 reveals the coefficients of return and variance equations during
the pre-COVID-19 outbreak phase. The α2 value is positive and significant only for S&P
500 to Sensex, indicating positive return spillover from the USA to India. In the variance
equation, β1 (GARCH term) represents the impact of prior period volatility on current
volatility. It is not significant for S&P 500 to Sensex but is positive and significant for
SSE composite to Sensex, Nikkei 225 to Sensex and DAX to Sensex. β2 (ARCH term)
captures the change in volatility news and it is found to be significant in all four
scenarios. The value of φ is positive and significant for all four foreign market indices to
Sensex. It measures the asymmetric effect of volatility. The positive and significant
leverage effect coefficient (φ) indicates that shocks are asymmetric wherein the positive
shocks have stronger impact than the negative shocks. In the context of the volatility
spillover from foreign markets (S&P 500, SSE Composite, Nikkei 225 and DAX) to
Indian market (Sensex) we find the volatility spillover parameter (δ) is negative and
significant. The estimate of AR(1) – EGARCH (1, 1) model exhibits volatility spillover
from the these foreign markets to Indian market. The estimation of the model is also
validated by checking the autocorrelation and ARCH effect in the residuals and squared
residuals.
Table 7 Heteroscedasticity test after EGARCH model for the post-COVID period

ARCH heteroscedastity test NH: No ARCH effect

Post-COVID-19 outbreak S&P 500 on Sensex
F-statistic 0.449715 Prob. F(1,198) 0.5033
Obs. ∗ R-squared 0.453228 Prob. chi-square (1) 0.5008
Post-COVID-19 outbreak SSE on Sensex
F-statistic 0.001441 Prob. F(1,198) 0.9698
Obs. ∗ R-squared 0.001455 Prob. chi-square (1) 0.9696
Post-COVID-19 outbreak Nikkei on Sensex
F-statistic 0.009203 Prob. F(1,198) 0.9237
Obs. ∗ R-squared 0.009296 Prob. chi-square (1) 0.9232
Post-COVID-19 outbreak DAX on Sensex
F-statistic 0.020275 Prob. F(1,198) 0.8869
Obs. ∗ R-squared 0.020477 Prob. chi-square (1) 0.8862

In the second part of Table 6, information pertaining to post-COVID outbreak phase are
given. There is a positive return spillover from S&P 500 to Sensex. GARCH term (β1) is
positive and significant for S&P 500 to Sensex and SSE Composite to Sensex signifying
that volatility is consistent from these markets. The ARCH (β2) and EGARCH (φ) terms
are significant for all four foreign markets to Sensex. The significance of β2 represents
that news on change in volatility affects the conditional volatility in the Indian market.
The positive φ values indicate the leverage effect wherein the positive shocks are having
greater effect. The volatility spillover parameter (δ) is positive and significant for the
USA to India market. This specifies that there is a positive volatility spillover effect from
USA to India. The spillover effects are not observed from other foreign markets to India
during the post-COVID phase. The results show that during the post-COVID-19
outbreak, Indian stock market volatility is triggered by positive spillover effect from the
428 T. Mohanasundaram et al.

USA. The EGARCH model is tested for its reliability and robustness by checking the
presence of heteroscedasticity and autocorrelation. The valid model should be free from
the issues of heteroscedasticity and autocorrelation. Table 7 and Figure 3 show the results
of the diagnostic checking of the EGARCH model.

Figure 3 Serial correlation LM test: post-COVID outbreak period, (a) S&P 500 on Sensex
(b) SSE composite on Sensex (c) Nikkei on Sensex (d) DAX on Sensex

(a) (b)

(c) (d)

The ARCH heteroscedasticity test shows that the p-value is greater than a 5% level of
significance which indicates that the null hypothesis of ‘no heteroscedasticity’ is not
rejected. Similarly, the serial correlation test displays that the null hypothesis of ‘no
 Structural change, information asymmetry and volatility 429

autocorrelation’ is not rejected. Thus, the EGARCH model in Table 6 is free from both
heteroscedasticity and autocorrelation.

5 Conclusions and policy implications

The drastic change in financial markets as a result of COVID-19 outbreak has brought a
structural change in the stock market movements. This study has investigated the
structural change, information spillover and volatility in the Indian stock market. The
study was undertaken to assess the return and volatility spillover from the stock market
indices of the top four economies –the USA (S&P 500), China (SSE), Germany (DAX)
and Japan (Nikkei 225) to the Indian stock market (Sensex). The breakpoint unit root test
identifies the structural change date as 23rd March 2020. The augmented unit root test
revealed that all the stock market return series are stationary. An OLS regression analysis
has been modelled to estimate the long-run relationship between Sensex returns and the
foreign stock market returns during the pre-COVID and the post-COVID outbreak
phases. During the pre-COVID outbreak phase, S&P 500 and DAX are found
to be a significant determinant of Sensex returns. On the other hand, S&P 500 and
Nikkei 225 are found to be a significant determinant of Sensex returns during the
post-COVID outbreak phase. Thus, India’s stock market integration with these
economies varied during the pre-and-post-COVID-19 outbreak. The diagnostic test for
the OLS regression analysis revealed that the models’ residual has heteroscedasticity. The
post-COVID outbreak regression model also suffers from autocorrelation issue.
The presence of heteroscedasticity in the residual demands the use of the ARCH
model. The EGARCH model confirmed that the information on volatility and past
variances are observed to be significant during the pre-COVID and post-COVID phases.
The study was mainly conducted to examine the contrasting changes appearing in the
global markets due to pandemic situation observed. The revision strategies adopted by
the portfolio managers and investors have been towards the ray of hope for better news,
better macroeconomic factors and better than normal growth rates. The shocks from the
foreign markets are found to have major bearing on the Indian markets. The results of the
volatility spillover suggest that positive and negative news of pandemic definitely has
been seen to provide momentum. These results suggest the unwavering impact of
pandemic in top four economies viz., the USA, China, Germany and Japan. This would
have resulted in positive inflow in Indian markets and thus seen as an opportunity for
portfolio reallocation and investment in the Indian markets during the pre-COVID
period which is evident from the significant negative spillover coefficients in the
EGARCH(1, 1) model. The post-COVID outbreak phase shows that the volatility
spillover to India is not significant as was found in the pre-COVID period from China,
Germany and Japan. The positive and significant volatility spillover from the USA during
the post-COVID phase suggest that long-term cointegration has resumed and signifies
strong integration of Indian market with US markets and reallocation might not provide
the return and diversification benefit as during the pre-COVID period.
430 T. Mohanasundaram et al.

References
Adnan, A.T.M., Hasan, M.M. and Ahmed, E. (2020) ‘Capital market reactions to the arrival of
COVID-19: a developing market perspective’, Economic Research Guardian, Vol. 10, No. 2,
pp.97–121.
Anh, D.L.T. and Gan, C. (2020) ‘The impact of the COVID-19 lockdown on stock market
performance: evidence from Vietnam’, Journal of Economic Studies, Vol. 48, No. 4,
pp.836–851.
Bae, K. and Karolyi, G. (1994) ‘Good news, bad news and international spillovers of stock return
volatility between Japan and the U.S.’, Pacific-Basin Finance Journal, Vol. 2, No. 4,
pp.405–438.
Baek, S., Mohanty, S.K. and Glambosky, M. (2020) ‘COVID-19 and stock market volatility:
an industry level analysis’, Finance Research Letters, Vol. 37, DOI: 10.1016/[Link].2020.
101748.
Baele, L. (2005) ‘Volatility spillover effects in European equity markets’, Journal of Financial and
Quantitative Analysis, Vol. 40, No. 2, pp.373–401.
Baker, S.R., Bloom, N., Davis, S.J., Kost, K., Sammon, M. and Viratyosin, T. (2020)
‘The unprecedented stock market reaction to COVID-19’, The Review of Asset Pricing
Studies, Vol. 10, No. 4, pp.742–758.
Bakhshi, P. and Chaudhary, R. (2020) ‘Responsible business conduct for the sustainable
development goals: lessons from COVID-19’, International Journal of Disaster Recovery and
Business Continuity, Vol. 11, No. 1, pp.2835–2841.
Bhargava, V., Malhotra, D.K. and Russel, P. (2012) ‘An empirical examination of volatility
spillover between the Indian and US swap markets’, International Journal of Emerging
Markets, Vol. 7, No. 3, pp.289–304.
Bollerslev, T. (1986) ‘Generalized autoregressive conditional hetroscedasticity’, Journal of
Econometrics, Vol. 31, No. 3, pp.307–327.
Booth, G., Chowdhury, M., Martikanian, T. and Tse, Y. (1997) ‘Intraday volatility in international
stock index futures markets: meteor showers or heat waves?’, Management Science, Vol. 43,
No. 11, pp.1564–1576.
Cao, K.H., Li, Q., Liu, Y. and Woo, C.K. (2020) ‘COVID-19’s adverse effects on a stock market
index’, Applied Economics Letters, Vol. 28, No. 14, pp.1157–1161, DOI: 10.1080/13504851.
2020.1803481.
Capelle-Blancard, G. and Desroziers, A. (2020) The Stock Market is Not the Economy? Insights
from the COVID-19 Crisis, 16 June, CEPR COVID Economics.
Cha, B. and Oh, S. (2000) ‘The relationship between developed equity markets and the Pacific
Basin’s emerging equity markets’, International Review of Economics & Finance, Vol. 9,
No. 4, pp.299–322.
Dornbusch, R., Park, Y.C. and Claessens, S. (2000) ‘Contagion: how it spreads and how it can be
stopped’, World Bank Research Observer, Vol. 15, No. 2, pp.177–197.
Engle, R.F. (1982) ‘Autoregressive conditional heteroscedasticity with estimates of the variance of
United Kingdom inflation’, Econometrica, Vol. 50, No. 4, pp.987–1007.
Gallo, G.M. and Otranto E. (2007) ‘Volatility transmission across markets: a multi-chain Markov
switching model’, Applied Financial Economics, Vol. 17, No. 8, pp.659–670.
Garner, A.C. (1988) Has Stock Market Crash Reduced Customer Spending? Economic Review,
April, pp.3–16, Federal Reserve Bank of Kanas City.
Gertler, M. and Hubbard, R.G. (1988) ‘Financial factors in business fluctuations’, Proceedings –
Economic Policy Symposium – Jackson Hole, Federal Reserve Bank of Kanas City, pp.33–72.
Gherghina, S.C., Armeanu, D.S. and Joldeș, C.C. (2020) ‘Stock market reactions to COVID-19
pandemic outbreak: quantitative evidence from ARDL bounds tests and Granger causality
analysis’, International Journal of Environmental Research and Public Health, Vol. 17,
No. 18, DOI: 10.3390/ijerph17186729.
He, P., Sun, Y., Zhang, Y. and Li, T. (2020) ‘COVID–19’s impact on stock prices across different
sectors – an event study based on the Chinese stock market’, Emerging Markets Finance and
Trade, Vol. 56, No. 10, pp.2198–2212.
 Structural change, information asymmetry and volatility 431

Hoque, H. (2007) ‘Co-movement of Bangladesh stock market with other markets: cointegration and
error correction approach’, Managerial Finance, Vol. 33, No. 10, pp.810–820.
Jebran, K. and Iqbal, A. (2016) ‘Examining volatility spillover between Asian countries’ stock
markets’, China Finance and Economic Review, Vol. 4, No. 6, DOI: 10.1186/s40589-016-
0031-1.
Kumar, M. (2013) ‘Returns and volatility spillover between stock prices and exchange rates’,
International Journal of Emerging Markets, Vol. 8, No. 2, pp.108–128.
Li, H. (2007) ‘International linkages of the Chinese stock exchanges: a multivariate GARCH
analysis’, Applied Financial Economics, Vol. 17, No. 4, pp.285–297.
Li, Y. and Giles, D.E. (2014) ‘Modelling volatility spillover effects between developed stock
markets and Asian emerging stock markets’, International Journal of Finance & Economics,
Vol. 20, No. 2, pp.155–177.
Mishra, A.K., Swain, N. and Malhotra, D.K. (2007) ‘Volatility spillover between stock and foreign
exchange markets: Indian evidence’, International Journal of Business, Vol. 12, No. 3,
pp.343–358.
Narayan, P.K., Devpura, N. and Hua, W. (2020a) ‘Japanese currency and stock market –
what happened during the COVID-19 pandemic?’, Economic Analysis and Policy, Vol. 68,
pp.191–198.
Narayan, P.K., Phan, D.H.B. and Liu, G. (2020b) ‘COVID-19 lockdowns, stimulus packages, travel
bans, and stock returns’, Finance Research Letters, Vol. 38, DOI: 10.1016/[Link].2020.101732.
Nelson, D.B. (1991) ‘Conditional heteroscedasticity in asset returns: a new approach’,
Econometrica, Vol. 59, No. 2, pp.347–700.
Paramati, S.R., Roca, E. and Gupta, R. (2016) ‘Economic integration and stock market dynamic
linkages: evidence in the context of Australia and Asia’, Applied Economics, Vol. 48, No. 44,
pp.4210–4226.
Poshakwal, S. and Murinde, V. (2001) ‘Modelling the volatility in East European emerging stock
markets: evidence on Hungary and Poland’, Applied Financial Economics, Vol. 11, No. 4,
pp.445–456.
Ravi, S. (2020) Impact of COVID-19 on the Indian Stock Markets, May 2020,
BW Businessworld [online] [Link]
The-Indian-Stock-Markets/11-05-2020-191755 (accessed 20 January 2021).
Sabri, N.R. (2004) ‘Stock return volatility and market crisis in emerging economies’, Review of
Accounting and Finance, Vol. 3, No. 3, pp.59–78.
Sharif, A., Aloui, C. and Yarovaya, L. (2020) ‘COVID-19 pandemic, oil prices, stock market,
geopolitical risk and policy uncertainty nexus in the US economy: fresh evidence from
the wavelet-based approach’, International Review of Financial Analysis, Vol. 101496,
DOI: 10.1016/[Link].2020.101496.
Singh, P., Kumar, B. and Pandey A. (2010) ‘Price and volatility spillovers across North American,
European and Asian stock markets’, International Review of Financial Analysis, Vol. 19,
No. 1, pp.55–64.
Thakur, S. (2020) ‘Effect of COVID-19 on capital market with reference to S&P 500’,
International Journal of Advanced Research, Vol. 8, No. 6, pp.1180–1188.
Worthington, A. and Higgs, H. (2004) ‘Transmission of equity returns and volatility in Asian
developed and emerging markets: a multivariate GARCH analysis’, International Journal of
Finance & Economics, Vol. 9, No. 1, pp.71–80.
Yan, B., Stuart, L., Tu, A. and Zhang, T. (2020) Analysis of the Effect of COVID-19 on the Stock
Market and Investing Strategies, DOI: 10.2139/ssrn.3563380.
Yepes-Rios, B., Gonzalez-Tapia, K. and Gonzalez-Perez, M.A. (2015) ‘The integration of stock
exchanges: the case of the Latin American Integrated Market (MILA) and its impact on
ownership and internationalization status in Colombian brokerage firms’, Journal of
Economics, Finance and Administrative Science, Vol. 20, No. 39, pp.84–93.
Yong, H.H.A. and Laing, E.S. (2020) ‘Stock market reaction to COVID-19: evidence from US
firms’ international exposure’, International Review of Financial Analysis, Vol. 76, Nos. 1–4,
DOI: 10.1016/[Link].2020.101656.