Emerging Markets Finance and Trade

ISSN: (Print) (Online) Journal homepage: [Link]

Does the Indian Financial Market Nosedive

because of the COVID-19 Outbreak, in Comparison
to after Demonetisation and the GST?

Alok Kumar Mishra , Badri Narayan Rath & Aruna Kumar Dash

To cite this article: Alok Kumar Mishra , Badri Narayan Rath & Aruna Kumar Dash (2020) Does
the Indian Financial Market Nosedive because of the COVID-19 Outbreak, in Comparison to after
Demonetisation and the GST?, Emerging Markets Finance and Trade, 56:10, 2162-2180, DOI:
10.1080/1540496X.2020.1785425

To link to this article: [Link]

Published online: 25 Jul 2020.

Submit your article to this journal

Article views: 1146

View related articles

View Crossmark data

Citing articles: 3 View citing articles

Full Terms & Conditions of access and use can be found at

[Link]
EMERGING MARKETS FINANCE AND TRADE
2020, VOL. 56, NO. 10, 2162–2180
[Link]

Does the Indian Financial Market Nosedive because of the

COVID-19 Outbreak, in Comparison to after
Demonetisation and the GST?
Alok Kumar Mishraa, Badri Narayan Rathb, and Aruna Kumar Dashc
a
School of Economics, University of Hyderabad, Hyderabad, India; bDepartment of Liberal Arts, Indian
Inistitute of Technology Hyderabad, India; cDepartment of Economics, The ICFAI Foundation for Higher
Education, India

ABSTRACT KEYWORDS
We investigate the impact of COVID-19 on the Indian financial Coronavirus; stock returns;
market and compare it with the outcomes of two recent struc- exchange rate; FII; systemic
tural changes of the Indian economy: demonetization and risks; India
implementation of the Goods and Services Tax (GST). Using JEL CLASSIFICATION
daily stock return, net foreign institutional investment, and G15; G18
exchange rate data from January 3, 2003 to April 20, 2020, we
find negative stock returns for all the indices during the COVID-
19 outbreak, unlike during the post-demonetization and GST
phases. Markov switching vector autoregression shows the
impact of COVID-19 on stock returns is severe in comparison
to that of demonetization and the GST.

1. Introduction
As of June 9, 2020, the COVID-19 outbreak had spread to 213 countries, with
nearly 7 million confirmed cases and 404,396 deaths worldwide (World Health
Organization website, [Link] This pandemic has impacted
the world’s economies, and India, the focus of our study, is not immune. One
opinion is that India is facing a prolonged period of economic slowdown,
although the exact magnitude of the economic loss cannot be predicted (Dev
and Sengupta 2020).
This paper aims to examine the impact of the COVID-19 outbreak on the
Indian financial market, proxied by stock returns, the exchange rate between
the Indian rupee (INR) and the US dollar (USD), and foreign institutional
investor (FII) net inflows. The aim is to compare the outcomes of the effects of
COVID-19 with those of two other major economic policy events in India,
namely, demonetization (November 8, 2016) and implementation of the
Goods and Services Tax (GST, July 1, 2017). Both policies had implications
for India’s equity and foreign exchange markets. As India was adjusting to the
aftermath of the demonetization and the GST, the unprecedented outbreak of

CONTACT Badri Narayan Rath badrirath@[Link] Department of Liberal Arts, Indian Institute of
Technology Hyderabad, Sangareddy, Telangana 502285
© 2020 Taylor & Francis Group, LLC
 EMERGING MARKETS FINANCE AND TRADE 2163

the COVID-19 pandemic halted all progress and posed a major challenge. The
government’s response to COVID-19 was a lockdown, restricting people’s
movement and business activities. An immediate effect of COVID-19 was
also noted in the Indian financial markets,1 which further motivated us to
examine the dynamic impact of the COVID-19 pandemic on Indian equity
and foreign exchange markets.
A research strand related to our paper examines, for instance, the dynamic
relation between stock returns, exchange rates, mutual funds, and oil and gold
prices (Dhingra, Gandhi, and Bulsara 2016; Edelen and Warner 2001; Garg
and Dua 2014; Jain and Biswal 2016; Katechos 2011; Mishra 2004; Narayan,
Narayan, and Prabheesh 2014; Sensoy and Tabak 2015). These studies show
that India’s stock market is influenced by mutual funds (Edelen and Warner
2001; Narayan, Narayan, and Prabheesh 2014), gold and crude oil prices (Jain
and Biswal 2016), currency risk (Garg and Dua 2014), exchange rates and
foreign equity flows (Dhingra, Gandhi, and Bulsara 2016; Mishra 2004), and
long-range dependence (Sensoy and Tabak 2015).
A second strand of research to which our study relates examines the effect of
pandemics on financial markets (Al-Awadhi et al. 2020; Ali, Alam, and Rizvi
2020; Apergis and Apergis 2020; Barro, Ursúa, and Weng 2020; Donadelli,
Kizys, and Riedel 2017; Gil-Alana and Monge 2020; Gormsen and Koijen
2020; Granger, Huang, and Yang 2000; Haroon and Rizvi 2020; Ichev and
Baker Marinc 2018; Lagoarde-Segot and Leoni 2013; Narayan 2020; Phan and
Narayan 2020; Zhang, Hu, and Ji 2020; Fu and Shen 2020; Liu, Wang and Lee
2020; Qin, Zhang and Su 2020). These studies show how financial markets and
their volatility and the energy markets have been impacted by COVID-19. Of
these, the works closest to ours are those of Ali, Alam, and Rizvi (2020); Al-
Awadhi et al. (2020); Katechos (2011); Phan and Narayan (2020); Sensoy and
Tabak (2015). These studies show that global financial market risks and
uncertainty have increased substantially in response to the pandemic, with
significant negative effects on stock returns.
We add to the understanding of COVID-19’s effect on India’s stock market
by proposing a dynamic model that relates stock returns, the INR–USD
exchange rate, and FII flows. To examine the impact of COVID-19, we
consider the equity indices of the BSE (Bombay Stock Exchange), such as
Standard & Poor’s (S&P) BSE Sensex, BSE mid-caps and small-caps, and the
BSE 100, along with three sectoral indices (BSE-Auto, BSE-Bankex, and BSE-
Realty). The foreign exchange market is proxied by the nominal bilateral
exchange rate between the INR and the USD. Since FIIs are an important
component of the Indian financial market, we consider net foreign investment.
Our empirical approach is motivated by the portfolio balance model
(Branson 1983; Frankel 1983). According to this model, an exogenous increase
in stock prices will increase the demand for money in anticipation of future
expected returns. The short-term interest rate will therefore rise, which will
2164 A. K. MISHRA ET AL.

attract more foreign investment in the domestic economy. Our study uses
daily data spanning from April 8, 2003 to April 20, 2020. The logical rationale
behind the use of this data period is to consider various structural reforms,
such as the implementations of demonetization in 2016 and of the GST in
2017, in addition to the COVID-19 pandemic in 2020. Demonetization and
implementation of GST are two recent major economic reforms that India has
undertaken, where the market has responded differently. In the case of
COVID-19 pandemic outbreak, the financial market responded negatively.
The impact is also found severe. As most of our data is in time series in nature,
the trend break is also found to be significant. This has motivated us to
understand whether the outcome of the COVID-19 impact in comparison to
two major economic reforms is tremendous. To comparatively assess the
impacts of these three shocks (demonetization, GST, and COVID-19), first,
we estimate the growth rates, returns, and volatility of these key indicators and
compare the results. In the second stage, we employ a Markov switching vector
autoregressive (MS-VAR) model to examine the nonlinear dynamics among
these variables.
The major findings of the study are as follows. First, the stock returns of all
the indices are negative during the COVID-19 outbreak, which was not the
case during the post-demonetization and GST phases. Second, the volatility of
the benchmark stock price is 2.77, compared to 0.51 and 0.59 in the post-
demonetization and GST phases, respectively. Third, the consequences of the
IRF (impulse response function) of the MS-VAR model indicate that shocks
(innovation) in exchange rate returns have asymmetric impacts on stock
returns in the COVID-19 phase. However, shocks in stock returns negatively
affect exchange rate returns in both the short and long run. Fourth, the
transition matrix shows that the impact of COVID-19 is determined to be
severe in the context of India’s stock liquidity index (Sensex), in comparison to
the impact of demonetization and GST implementation.
The remainder of the paper is organized as follows. Section 2 presents the
methodology and data. Section 3 discusses the statistical analysis and empirical
results. Finally, Section 4 highlights the main conclusions and policy
implications.

2. Methodology and Data

2.1. Markov-switching VAR (MS-VAR) Model
In the empirical literature, the impact of the event on the stock market is
generally analyzed by the event study methods. However, in the event analysis,
the t-test or other nonparametric tests are used to test the null hypothesis
(example, no abnormal returns in case of stock market) at the time of the
event. This may lead to misleading results because of the kurtosis and
 EMERGING MARKETS FINANCE AND TRADE 2165

volatility-clustering characterization of the financial time series, especially the

data on securities which are continuously traded in the market. To avoid this
limitation of the event study approach, we proposed to model the returns in
the event of window through a Markov Switching model with two regimes:
Regime 1: Normal market conditions; Regime 2: Abnormal market conditions.
The general idea behind the class of Markov switching models is that the
parameters and the variance of an autoregressive process depend upon an
unobservable regime variable which represents the probability of being in
a particular state of the world. A major advantage of the MS-VAR model is
its flexibility in modeling time series subject to nonlinearity and regime shifts.
The model is estimated based on the Gausssian Maximum Likelihood estima-
tion (for detailed discussion see Castellano and Scaccia 2010). The MS-VAR
model is presented as follows.
The portfolio balance model (Branson 1983; Frankel 1983), as highlighted
in the introduction, is modeled by employing a three-variable MS-VAR model.
These variables include the return on the liquidity index Sensex (RSEN), net
FIIs (NETFII), and the return on the exchange rate of the INR versus the USD
(RUSD). �
Let zt ¼ RSENt ; NETFIIt; RUSDt ’ be a 3 × 1vector of observations at time t.
�

Specifically, the structural representation of the M-state MS-VAR can be

expressed as

β0;st zt ¼ μst þ β1;st zt 1 þ . . . þ βp;st zt p

~ ð0; ωst Þ
þ εt ; εt N (1)

where εt is the structural error term, which follows a Gaussian distribution,

with a diagonal covariance matrix ωst at time t. The reduced form of the model
can be obtained by premultiplying β−10,st on both sides of Equation (1),
yielding

zt ¼ αst þ γ1;st zt 1 þ . . . þ γp;st zt p

~ ð0; �st Þ
þ φt; φt N (2)

where αst = β−10,stμst represents the 3 × 1 time-varying intercepts and γi,st

= β−10,st βi,st,,i = 1, 2, . . ., p,denotes the 3 × 3 VAR coefficient matrices at time t.
The covariance matrix for the reduced-form error φt = β−10,st εt can be
decomposed as Σst = β−10,st ω st β−10,st
The regime indicator variable St is assumed to follow an M-state Markov process
with transition probabilities PrðSt ¼ jjSt 1 ¼ iÞ ¼ Pij; i; j ¼ 1; . . . ; M: The transi-
tion probabilities of the Markov Switching process determine the probability of the
volatility switching to another regime, and thus the expected duration of each
regime (Chevallier, 2011a). Transition probabilities can be constant or a time-
varying function of exogenous variables (Cai 1994; Gray 1996; Hamilton and
Susmel 1994). The model is estimated based on a Gaussian maximum likelihood
procedure. The calculation of the covariance matrix uses using the second partial
2166 A. K. MISHRA ET AL.

derivatives of the log-likelihood function, and the transition matrix P controls the
probability of a switch from state 1 to 2:
� �
p11 p21
P¼
p12 p22

where the sum of each column in p is equal to one.

In this paper, the Markov switching model is estimated by assuming two
states. The choice of the two-state process is motivated by the fact that this
model is intuitively appealing for tracking the bull–bear cycle of the Indian
equity market for the three phases of post-demonetization, GST implementa-
tion, and the COVID-19 outbreak.
Since the MS-VAR model is nonlinear, a generalized impulse response
function is defined as the difference between a shock trajectory (i.e. incorpor-
ating the initial identified shock) and a baseline trajectory (i.e. without the
initial identified shock). All exogenous random variables, except the initial
identified shock, are identical between these two trajectories (Karame 2010,
2015; Koop, Pesaran, and Potter 1996).

2.2. Data and Measurement of Variables

The empirical analysis is based on daily data from January 3, 2003 to April 20,
2020. The selection of the time span is based purely on the availability of data.
The study uses three key variables: stock returns, exchange rate returns, and
FII net investment. We use the daily stock closing prices of the Sensex as the
benchmark and other indices (BSE 100, mid-caps and small-caps, and sectoral
indices) and estimate the first difference of the natural logarithmic stock prices
series to obtain stock returns. However, the growth rate of the stock indices
and the exchange rates are computed as the ratio of the difference between the
current and previous prices upon previous price. The volatility series is
calculated using a five-day rolling standard deviation method. This study
uses nominal bilateral exchange rate data between the INR and the USD and
calculates the return series. The FII net investment is estimated by taking the
difference between FII gross purchases and FII gross sales. A positive sign for
this net investment series indicates inflows, and negative figures indicate
outflows. Since both the stock and foreign exchange markets do not trade on
weekends or holidays, we adjust the COVID-19 series accordingly by mapping
the exact dates between these four series. In addition to the three key indica-
tors, we use data on the number of confirmed COVID-19 cases from
January 30, 2020 to April 20, 2020.
We use three binary dummy variables to denote the periods after the
demonetization and the GST and the period of the COVID-19 outbreak;
these are, respectively, equal to one for the post-demonetization or GST period
 EMERGING MARKETS FINANCE AND TRADE 2167

and after the Ministry of Health and Family Welfare of the Government of
India reported the first COVID-19 case, and zero otherwise. Specifically, from
November 8, 2016 to April 20, 2020 denotes the post-demonetization period,
from July 1, 2017 to April 20, 2020 the GST period, and from January 30, 2020
to April 20, 2020 the COVID-19 outbreak period. All the data were collected
from the CEIC database, which provides the most accurate and in-depth data
available on the Indian economy and more than 200 other countries.

3. Empirical Results
3.1. Preliminary Analysis

This section starts with preliminary evidence on the relation between stock
returns, exchange rate returns, FII net investment, and the COVID-19 out-
break by presenting a range of descriptive statistics, returns, growth rates,
volatilities, and correlations among the key indicators. Figure 1 shows the plots
of the four data series. The FII net flows and the number of new COVID-19
cases are shown in the first row of the figure and return series of the exchange
rate and stock price are presented in the second row of Figure 1. We notice
that all four series show diverse shape.
We present descriptive statistics in Panel A of Table 1. We note that the
average positive daily return of the INR versus the USD is minimal (0.01) in
comparison to the average daily return of the Sensex and the broad stock index
(BSE 100) during the sample period. The kurtosis coefficient, a measure of the
thickness of the tail of the distribution, is quite high for all the variables, which
follow a platykurtic distribution, with a fatter shorttail. This result implies that
a Gaussian distribution for the respective variables cannot be assumed. This
finding is further strengthened by the results of a Jarque–Bera test. The
Jarque–Bera test indicates that the returns on both assets are not normally
distributed (unconditionally), since the time series appear to be slightly
skewed (negatively skewed for both stock indices, such as RSEN and RBSE
100, and positively skewed for the return series of the INR versus the USD
exchange rate). Hence, this result rejects the null hypothesis of normality at
any conventional confidence levels.
Panel B, of Table 1 reports statistically significant correlation coefficients
between stock returns, exchange rate returns, and net FIIs. A significant
negative correlation is found between the returns on the exchange rate and
stock indices and net foreign investments. However, the correlation coefficient
between net FII investment and the exchange rate return is extremely low
(−0.05), in comparison to the correlation coefficient between the exchange rate
and stock price returns. This result supports the stock-oriented model of
a negative relation between stock prices and exchange rates.
2168 A. K. MISHRA ET AL.

Number of new COVID-19 cases in India

FII net investment
200000 2000
150000 1500
100000
1000
50000
500
0
04-04-2003
04-04-2005
04-04-2007
04-04-2009
04-04-2011
04-04-2013
04-04-2015
04-04-2017
04-04-2019
-50000 0

16-03-2020

23-03-2020

30-03-2020

06-04-2020

13-04-2020

20-04-2020
-100000
-150000
-200000

Exchange rate returns Stock returns

5 20
4
15
3
10
2
5
1
0 0
04-04-2003

04-04-2005

04-04-2007

04-04-2009

04-04-2011

04-04-2013

04-04-2015

04-04-2017

04-04-2019
04-04-2003
04-04-2004
04-04-2005
04-04-2006
04-04-2007
04-04-2008
04-04-2009
04-04-2010
04-04-2011
04-04-2012
04-04-2013
04-04-2014
04-04-2015
04-04-2016
04-04-2017
04-04-2018
04-04-2019
04-04-2020

-1 -5
-2 -10
-3 -15
-4

Figure 1. Time series plots of data series.

This graph is plotted by computing the returns of Sensex and exchange rate (INR vs. US$). Both the
return series exhibit volatility clustering and their trend is stationary. The tendency of positive
COVID-19 is kept on mounting at an alarming rate from April 2020 onwards.

After presenting the descriptive statistics, we analyze the growth rates,

returns, and volatilities of the stock prices, the exchange rate, and FII net
flows by comparing three events. The post-demonetization phase is consid-
ered to be from November 9, 2016 to June 30, 2017 and the GST phase from
July 1, 2017 to January 29, 2020. Similarly, January 30, 2020 to April 20, 2020
covers the COVID-19 phase. We arrive at the following insights. First, the
Sensex stock prices show negative growth (−22.65%) during the COVID-19
pandemic phase, compared to positive growth rates of 12.07% and 31.9%
during the post-demonetization and GST phases, respectively.
Second, other stock indices, such as mid-caps and small-caps, and all
sectors, except healthcare, show drastic negative growth rates during the
COVID-19 phase, compared to the post-demonetization and GST phases.
Third, both the returns and volatility series of the stock indices also reveal
a precarious scenario during the COVID-19 phase, in comparison to the
other events. All the stock indices, except healthcare, show negative returns
and high volatility during this ongoing COVID-19 pandemic. Fourth, among
 EMERGING MARKETS FINANCE AND TRADE 2169

Table 1. Descriptive statistics.

RSEN RBSE100 NETFII RUSD
Panel A: Summary Statistics
Mean 0.05 0.05 2808.44 0.01
Median 0.09 0.11 1792.00 0.00
Maximum 15.98 15.49 163775.7 4.01
Minimum −14.10 −13.88 −161203.9 −3.00
Std. Dev. 1.45 1.46 16203.60 0.45
Skewness −0.21 −0.40 0.74 0.15
Kurtosis 15.03 14.26 21.78 9.15
Jarque-Bera 24793.59 21796.74 60671.97 6496.65
Probability 0.00 0.00 0.00 0.00
Observations 4101 4101 4101 4101
RSEN RBSE100 NETFII RUSD
Panel B: Correlation Matrix
RSEN 1
RBSE100 0.98 1
(0.00)
FIITNET 0.057 0.06 1
(0.00) (0.00)
RUSD −0.30 −0.30 −0.05 1
(0.00) (0.00) (0.00)
Figures in parentheses are probability values.
This table reports the summary statistics of the variables (Panel A) and the pair wise correlation matrix (Panel B).

the three major stock indices, the growth and returns of small-cap compa-
nies perform worse than the mid-caps and BSE Sensex, whereas the BSE
Sensex stock price is more volatile compared to the mid-cap and small-cap
stock indices. Fifth, among the sectors, Realty, Metal, Bankex, and Auto
perform the worse in terms of stock returns and stock price growth. The
important demand for health-care products and services during the COVID-
19 pandemics likely reflected in their positive stock price returns with low
volatility. After healthcare, fast-moving consumer goods companies are
the second-best performers during the ongoing pandemic, which also show
a relatively less affected sector as compared to other sectors. Sixth, by
accounting for nominal exchange rate movement between the INR and the
USD, our results reveal that the INR depreciated by 7.2% during the
COVID-19 phase, as opposed to its 2.95% appreciation during the demone-
tization phase; however, the INR depreciated by 9.94% during the GST
phase, more than during the COVID-19 phase. When we compare the
exchange rate volatilities, however we again find that the exchange rate series
is more volatile during the COVID-19 phase than in the post-
demonetization and GST phases.
Finally, the net FII flow figures show negative growth during all three
phases, but this indicator is less volatile in the COVID-19 phase compared
to during the other two phases. This result implies that, although the average
gross sales of FIIs are higher than their average gross purchase, the growth of
FII outflows from the Indian financial market is still lower in the COVID-19
2170 A. K. MISHRA ET AL.

phase compared to the post-demonetization and GST phases. This finding is

not surprising, because the uncertainty due to COVID-19 pandemic is higher
in developed countries and other emerging countries such as China, Russia,
and Brazil. Therefore, though foreign investors are pulling out their funds,
particularly from short-term debt funds, during the ongoing COVID-19 out-
break, its growth as well as volatility figures are still much lower in comparison
to the post-demonetization and GST phases.

3.2. MS-VAR Results

In this section, after discussing the preliminary analysis, we present the MS-
VAR results. Prior to this, however, we need to address the stationarity of the
data series. To do so, we employ both augmented Dickey–Fuller (ADF) and
Phillips–Perronunit root tests. The unit root test results reveal that all four
variables—Sensex returns (RSEN), BSE 100 returns (RBSE 100), exchange rate
returns (RUSD), and net FII flows (NETFII)—are stationary in levels, that is,
I(0).2
We further investigate the distributional pattern of the stock returns,
exchange rate returns, and net FII flows by applying the BDS test of Broock
et al. (1996). This is a well-established technique to check the spatial depen-
dence and nonlinearity of a time series. In particular, this test identifies if
a time series is an independent and identically distributed process. Our results
reject the null hypothesis of independent and identical distributions for all
epsilon (ε) values of close points and m embedding dimensions at the 1%
significance level. Epsilon (ε) represents the distance between a selected pair of
points for measuring the independent and identical distribution of residuals,
whereas the embedding dimensions represent the number of consecutive
points used in the set. Thus, the results of the BDS test suggest a likely non-
linear structure in the stock and exchange rate returns and net FII flows.
Table 2 reports the maximum likelihood estimates of the Markov switching
model, an intercept switching VAR model with regime-dependent variances
but constant autoregressive coefficients. The coefficients are obtained with the
expected maximization algorithm. The MSI(2)-VAR(2)-EVENT(3) model
refers to the average return of series of the daily stock indices—both liquidity
(Sensex) and broad based (BSE 100)—in both stages 1 and 2, respectively, for
three different events, such as the most immediate effects of the demonetiza-
tion, the implementation of the GST, and the impact of the COVID-19 out-
break in the Indian equity market.
In this model, the intercept is assumed to be state dependent. The order of is
set to P = 2 (and P = 1 in some cases) by minimizing the Akaike information
criterion (AIC). In the model, the first regime (St = 1) represents a bear
(depression) state, while the second regime (St = 2) represents a bull (growth)
state, since regime 1 coefficients are smaller than regime 2 coefficients. The
Table 2. Markov Switching intercepts VAR based on MLE results.
S&P BSE SENSEX as S&P BSE 100 as a
Dependent Variable Demonetization Phase GST Phase COVID-19 Phase Dependent Variable Demonetization Phase GST Phase COVID-19 Phase
Regime1—Dependent Intercepts
1 −3.975 0.146 −1.520 C1 −0.142 −0.175 −0.893
(−14.699) (6.152) (−7.898) (−0.372) (−0.570) (−4.024)
Dummy1(−1) 4.291 −0.103 −6.641 Dummy1(−1) 0.229 0.444 −6.695
(11.304) (−1.763) (−10.156) (0.475) (0.919) (−10.522)
Regime2—Dependent Intercepts
C2 0.180 −5.662 0.136 C2 0.064 0.056 0.089
(6.477) (−18.049) (5.662) (2.398) (2.246) (3.745)
Dummy2(−1) −0.155 5.991 0.298 Dummy2(−1) −0.047 −0.036 0.315
(−3.036) (12.533) (1.298) (−0.869) (−0.571) (1.412)
Autoregressive Coefficients
RSEN(−1) −0.005 0.017 0.026 RBSE 100(−1) 0.054 0.055 0.056
(−0.288) (1.137) (1.660) (3.329) (3.412) (3.509)
RSEN(−2) −0.076 −0.073 RBSE 100(−2) −0.046
(−4.663) (−4.551) (−2.786)

EMERGING MARKETS FINANCE AND TRADE

NETFII(−1) 2.57E-07 −3.68E-06 −2.28E-06 NETFII(−1) 1.14E-06 1.30E-06 −2.78E-06
(0.166) (−2.831) (−1.489) (0.733) (0.933) (1.949)
NETFII(−2) 1.06E-07 −1.41E-06 NETFII(−2) 8.34E-07
(0.069) (−0.938) (0.546)
RUSD(−1) −0.083 −0.061 −0.006 RUSD(−1) −0.076 −0.049 −0.037
(−1.619) (−1.268) (−0.122) (1.436) (−0.952) (−0.721)
RUSD(−2) −0.110 −0.059 RUSD(−2) −0.114
(−2.172) (−1.130) (2.167
Variances
σ12 RSEN 1.789 1.757 1.933 σ12 RETBSE 2.131 2.136 2.036
(39.222) (42.142) (42.145) (45.268) (45.277) (44.638)
σ12 NETFII 978.292 −44.384 362.993 σ12 NETFII 1076.505 1055.121 137.007
(3.273) (−0.178) (1.174) (3.872) (3.642) (0.437)
σ12 RUSD −0.201 −0.182 −0.133 σ12 RUSD −0.198 −0.198 −0.169
(−16.695) (−16.979) (−12.514) (−18.643) (−18.567) (−15.200)
Transition Matrix Coefficients
P11-C −1.135 4.389 −0.060 P11-C −2.65 −3.564 −0.960
(−3.473) (24.482) (−0.282) (−3.757) (−3.312) (−2.739)

2171
(Continued)
 2172
A. K. MISHRA ET AL.
Table 2. (Continued).
S&P BSE SENSEX as S&P BSE 100 as a
Dependent Variable Demonetization Phase GST Phase COVID-19 Phase Dependent Variable Demonetization Phase GST Phase COVID-19 Phase
P21-C −3.845 1.483 −3.860 P21-C −4.668 −4.647 −4.152
(−24.778) (4.190) (−23.610) (−25.926) (−26.668) (−20.672)
Diagnostic Statistics
Resid Covariance 79256500 84705921 80332916 Resid Covariance 79037700 84247094 84282924
Log Likelihood −54306.360 −54492.190 −54397.930 Log Likelihood −54252.130 −54442.770 −54559.280
AIC/SC 26.515/26.574 26.595/26.640 26.560/26.619 AIC/SC 26.489/26.548 26.571/26.616 26.628/26.673
Number of Coefficients 38 29 38 Number of Coefficients 38 29 29
Figures indicated in parenthesis are Z-statistics
This table presents MSI (2)-VAR (2)-EVENT (3) with regime-dependent variances and with constant autoregressive coefficients. The parameters are estimated by using the Expectations-
Maximization (EM) algorithm. The estimated coefficients indicate that the nonlinear model fit the data better in COVID-19 period in the context of the return Sensex and the phase of
demonetization in the case of return on BSE 100.
 EMERGING MARKETS FINANCE AND TRADE 2173

COVID-19 dummy is highly statistically significant in regime 1 in the context

of both Sensex and BSE 100 returns. Examination of the coefficients of the two
intercepts (c1 and c2), which are both statistically significant, shows the pre-
sence of switches in growth (in terms of returns) between the two regimes. In
regime 1 (bear), the average daily return of the Sensex is negative in both the
demonetization and COVID-19 phases (−3.97% and −1.52%, respectively) and
positive (0.14%) in the GST phase. However, the daily average return of the
BSE 100 is negative (−0.14%, −0.17%, −0.89%, respectively) in the demoneti-
zation, the GST and COVID-19 phases. However, in regime 2 (bull), the
average returns of the Sensex (except in the GST phase) and the BSE 100 are
positive across all phases. In regime 1, the demonetization and COVID-19
dummies are statistically significant in the case of Sensex returns with a one-
period lag. It is important to note that the COVID-19 dummy negatively
affects Sensex returns for a one-day lag, but the demonetization dummy has
a positive impact. Similar findings are found in the case of the BSE 100,
although the demonetization and GST dummy coefficients are not statistically
significant. However, their impact is positive in regime 2 bull, except for the
demonetization dummy in the case of Sensex returns and the demonetization
and GST dummies in the case of the BSE 100. The autoregressive coefficients
are negative, implying negative serial correlation in the return series of stock
indices. Interestingly, we find the exchange rate (INR vs. USD) returns and
stock returns to have a negative impact in the context of both regimes and
stock indices. This empirical finding strengthens our theoretical argument of
the negative relation between stocks and the foreign exchange rate, as noted by
the portfolio balance (stock-oriented) approach.
The estimated coefficients presented in Table 2 indicate that, overall, the
nonlinear models fit the data better in the COVID-19 period in the context of
Sensex returns, and in the demonetization phase in the context of BSE 100
returns, in comparison to other phases. Similarly, the nonlinear models with
regime-dependent variance fit the data better, as indicated by the higher
estimated log-likelihood values. The volatility of Sensex returns is very high
in the COVID-19 period, and the volatility of the BSE 100 is very high in the
demonetization phase, as indicated by their corresponding variances. The AIC
favors the demonetization phase over the GST and COVID-19 phases in the
context of both Sensex and BSE 100 returns, although the difference is
extremely small among three phases. The intercept regime-dependent MS(2)-
VAR(2)-EVENT(3) model passes all the misspecifications tests and demon-
strates robustness by negative log-likelihood and positive AIC values.
The transition matrix with constant expected durations is reported in Table 3.
In this table, P11 is the probability of staying in regime 1 and P22is the probability
of being in regime 2, and P12 is the probability of transition from regime 1 to
regime 2and P21 is the probability of transition from regime 2 to regime 1. In the
demonetization period in the context of Sensex returns, during the bear phase,
2174 A. K. MISHRA ET AL.

Table 3. Transition probabilities and expected duration matrix.

Demonitisation: RSEN Demonitisation: RBSE 100
Regime 1 Regime 2 Duration Regime 1 Regime 2 Duration
Regime 1 0.317 0.682 (P12) 1.464 Regime 1 0.981 (P11) 0.018 53.864
(P11) (P12)
Regime 2 0.014 0.985 (P22) 67.959 Regime 2 0.981 (P21) 0.018 (P22) 1.018
(P21)
GST:RSEN GST:RBSE 100
Regime 1 0.251 0.748 1.336 Regime 1 0.299 (P11) 0.700 (P12) 1.426
(P11) (P12)
Regime 2 0.013 (P21) 0.986 73.281 Regime 2 0.018 (P21) 0.981 P(22) 53.487
(P22)
COVID 19: RSEN COVID 19: RBSE 100
Regime 1 0.985 (P11) 0.014 70.668 Regime 1 0.018 (P11) 0.981 (P12) 1.018
(P12)
Regime 2 0.704 (P21) 0.295 1.418 Regime 2 0.018 (P21) 0.981 (P22) 55.479
(P22)
P11 . . . P22 are transition probabilities. The average expected duration of being in regime one is computed using
formula suggested by Hamilton (1989).
This table presents the transition matrix with constant expected durations in both the regimes. The estimated
probabilities associated with both the regimes suggest that the impact of COVID-19 is severe in the context of
Sensex in comparison to the impact of demonetization and GST implementation. The average expected duration of
return Sensex being in the Regime 1 (bear phase) is nearly 70 days, while that of Regime 2 (Bull phase) is only
2 days.

the estimated probability of being in regime 1 is 31.73%. On the other hand, the
probability of the series switching from regime 1 to regime 2 is 68.26%.
However, once the economy finds itself in the bull phase (regime 2), the
probability that it will be in a bull phase is estimated to be 98.52%. The
probability that the series will switch from regime 2 to regime 1 is lower
(1.47%).In the GST phase, the estimated probability that the Sensex return will
remain in regime 1 is 25.15%; however, the probability that they will continue to
remain there during the bull phase (regime 2) is 98.63%. Interestingly, the results
show that, in the COVID-19 phase, the estimated probability of the Sensex
return continuing to remain in regime 1 (bear phase) is 98.58%. On the other
hand, the probability that the Sensex return will switch from regime 1 to regime
2 is 1.41%. The probability that the Sensex return series will switch from regime
2 to regime 1 is higher (70.49%). However, the estimated probability associated
with the Sensex return remaining in the bull phase (regime 2) is only 29.50%.
These results suggest that the impact of COVID-19 is severe in the context
of the Indian liquidity index, Sensex, in comparison to the impact of demo-
netization and the implementation of the GST. The average expected duration
of the Sensex return in regime 1 (bear) is nearly 70 days, whereas that for
regime 2 (bull) is only 2 days. Similarly, in the demonetization period, the
average expected duration of the Sensex return in regime 1 is nearly 2 days,
whereas that for regime 2 is almost 68 days. In the GST phase, the average
expected duration of the Sensex return in regime 1 is nearly 2 days,whereas
that for regime 2 is almost 74 days.
 EMERGING MARKETS FINANCE AND TRADE 2175

However, in the context of the broad-based stock index, the BSE 100, the
impact of demonetization is severe in comparison to the impact of the
implementation of the GST and the COVID-19 outbreak. The probability of
BSE 100 returns remaining in the bear phase (regime 1) is 98.14% in the
demonetization phase; however, the probability of them remaining there in
regime 2 is lower (1.85%). The transition probability from regime 1 to regime
2 is 1.85%, whereas the probability of switching from regime 2 to regime 1 is
98.14%. The average expected duration for BSE 100 returns in regime 1 is
52 days, whereas that for regime 2 is 1 day. However, during the COVID-19
phase, the average expected duration of return BSE 100 in regime 1 is only
1 day, whereas that in regime 2 is almost 55 days.

3.2.1. MS-VAR Impulse Response Function

The results from the comparative return and volatility analysis for the demo-
netization, GST, and COVID-19 phases, as well as the findings for the transi-
tion probabilities and expected durations in the previous section motivated us
to examine the potential dynamic nonlinear response of all the variables in the
system to a shock or innovation in each variable in both the short and long
run. The regime-dependent impulse response functions are derived from the
MSI(2)-VAR(2)-EVENT(3) model (Koop, Pesaran, and Potter 1996).
Although we generated the impulse response function results for all the
phases, only important results inconsistent with previous findings are pre-
sented here. The generalized Markov switching impulse response function of
Sensex (liquidity index) returns in the COVID-19, GST, and demonetization
phases are presented in Figures 2 and 3, respectively. In Figure 2, during the
COVID-19 phase, a one-standard-deviation shock (innovation) to net foreign
investment (NETFII) has no noticeable impact on the Sensex return in periods
1 and 2. From the second period onwards, the response gradually declines
until the third period, when it hits its steady-state value. Beyond this period,
the Sensex return remains in the positive region. Similarly, a one-standard-
deviation shock to the exchange rate (INR vs. USD) return has a negative
impact on the Sensex return, and this negative effect increases sharply up to
the second period, when it hits its steady-state value and, after that, decays
toward zero.
Thus, it can be concluded that innovation in the exchange rate return has
asymmetric impacts on stock returns. However, a one-standard-deviation
shock to the Sensex return initially increases NETFII up to the second period.
This positive response is flattened up to the third period, after which it sharply
declines until the sixth period, when it hits its steady-state value. Therefore, it
can be concluded that shocks to the Sensex return will have asymmetric
impacts on net foreign investment in both the short and long run. Similarly,
a one-standard-deviation shock in the Sensex return hurts the exchange rate.
2176 A. K. MISHRA ET AL.

Response to Genera lized One S.D. Innova !ons

Response of RSEN to RSEN Response of RSEN to FIIT_NET Response of RSEN to RUSD

1.2 1.2 1.2

0.8 0.8 0.8

0.4 0.4 0.4

0.0 0.0 0.0

1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10

Response of FIIT_NET to RSEN Response of FIIT_NET to FIIT_NET Response of FIIT_NET to RUSD

12,000 12,000 12,000

8,000 8,000 8,000

4,000 4,000 4,000

0 0 0

1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10

Response of RUSD to RSEN Response of RUSD to FIIT_NET Response of RUSD to RUSD

.3 .3 .3

.2 .2 .2

.1 .1 .1

.0 .0 .0

1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10

Figure 2. GIRFS return Sensex-COVID-19.

This graph is the dynamic responses of return Sensex, net foreign investment and exchange rate
due to the generalized one standard deviation shock to each of the variables in 10 days’ time
ahead in the COVID-19 phase.

In Figure 3, the generalized impulse response function of the BSE 100

(broad-based index) return generated by Markov switching VAR with two
regimes is presented for the demonetization phase. In line with the Sensex
return, a one-standard-deviation shock to the net FII initially decreased the
BSE 100 return until the second period, after which the effect dissipates.
However, a one-standard-deviation shock to the exchange rate (INR vs.
USD) return has an initially negative impact on the BSE 100 return. This
negative effect increases sharply up to the second period and then flattens, up
to the fourth period, after which it hits its steady-state value. Similarly, a one-
standard-deviation innovation to the BSE 100 return has initially increases the
net FII up to the second period. The response gradually declines from the third
period onwards, up to the ninth period; after the10th period, the effect decays
toward zero. Thus, shocks to the BSE 100 return will harm NETFII in both the
short and long run. Similarly, a shock in the exchange rate return will have
asymmetric impacts on NETFII. Finally, a one-standard-deviation innovation
to the BSE 100 return has a negative impact on the exchange rate return. This
result is in line with the portfolio balance approach (Branson and Frankel
1983) to determining exchange rates.
 EMERGING MARKETS FINANCE AND TRADE 2177

Response to Generalized One S.D. Innova!ons

Response of RETBSE to RETBSE Response of RETBSE to FIIT_NET Response of RETBSE to RUSD

1.0 1.0 1.0

0.5 0.5 0.5

0.0 0.0 0.0

1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10

Response of FIIT_NET to RETBSE Response of FIIT_NET to FIIT_NET Response of FIIT_NET to RUSD

12,000 12,000 12,000

8,000 8,000 8,000

4,000 4,000 4,000

0 0 0

1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10

Response of RUSD to RETBSE Response of RUSD to FIIT_NET Response of RUSD to RUSD

.4 .4 .4
.3 .3 .3
.2 .2 .2
.1 .1 .1
.0 .0 .0
-.1 -.1 -.1
1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10

Figure 3. GIRFS return BSE 100-Demonetization.

This graph is the dynamic responses of return BSE 100, net foreign investment and exchange rate
due to the generalized one standard deviation shock to each of the variables in 10 days’ time
ahead in the demonetization phase.

4. Conclusions
This paper examines the impact of the COVID-19 outbreak on Indian finan-
cial markets, emphasizing the dynamic interlinkages between stock returns,
net FII flows, and exchange rate volatility. The outcomes of the COVID-19
outbreak are being compared with those of two recent major structural
changes of the Indian economy, namely, demonetization and implementation
of the GST. The MS-VAR nonlinear dynamic model—MSI(2)-VAR(2)-
EVENT(3)—is employed to estimate regime-dependent intercepts, autore-
gressive coefficients, variances, transition probabilities, and expected dura-
tions between the regimes by using an expectation maximization algorithm.
The main findings of the study are as follows. First, benchmark BSE Sensex
stock prices experience a negative growth of −22.6% during the COVID-19
phase, compared to positive growth during the demonetization and GST
phases. Second, the stock returns of all the indices are negative during the
COVID-19 phase, unlike during the demonetization and GST phases. Third,
the volatility of benchmark stock prices is 2.77 in the COVID-19 phase,
compared to 0.51 and 0.59 in the demonetization and GST phases, respec-
tively. Fourth, realty has been the worst-affected sector by the COVID-19
pandemic, with stock price declines of −41.67%, unlike in the other sectors.
However, of all the sectors, Bankex’s stock returns are the most volatile. Fifth,
2178 A. K. MISHRA ET AL.

the INR has depreciated by 7.2% against the USD and is also more volatile
during the COVID-19 phase, compared to the other two phases. However, net
FII flows are more volatile in the GST phase compared to the COVID-19
phase. Sixth, the impact of COVID-19 is severe in the context of the Indian
stock liquidity index, the BSE Sensex, in comparison to the impacts of demo-
netization and implementation of the GST. The average expected duration of
the Sensex return in regime 1 (bear) is nearly 70 days, whereas that for regime
2 (bull) is only 2 days. However, we find that, in the context of a broad stock
index, that is, the S&P BSE 100, the impact of demonetization is significant in
comparison to the impact of the GST and COVID-19. Seventh, the conse-
quences of the impulse response function of the MSI(2)-VAR(2) model
assumes that the shocks (innovation) to the exchange rate return have asym-
metric impacts on stock returns in the COVID-19 phase. Even so, the return
on the exchange rate is negatively impacted due to the shock in stock returns
in both the short and long run. A shock in net FII has a positive impact on the
stock return and a negative impact on the exchange rate return in both the
short and long run during the post-COVID-19 period.
Our findings have important policy implications for the Reserve Bank of
India (RBI) in this crisis period. They suggest that RBI should be extremely
cautious in its intervention during the period of INR depreciation. We
propose that RBI intervene in forward contract policy, along with determin-
ing the trading limits for commercial banks to arrest speculative forces and
stop the precipitous fall of the INR against the USD. This approach will
stabilize the Indian financial market and especially the equity market by
minimizing the outflow of FIIs. The research in this paper could be extended
by considering switching between multiple regimes based on the examina-
tion of time-varying causality, rather than our prespecified regimes in the
MS-VAR model.

Notes
1. India’s stock market capitalization declined by 27.4% in March 2020; however, it jumped
back up by 14% in April 2020. The FII net investment, which was INR 89.702 billion in
February 2020, decreased drastically and was negative in both March and April 2020. In
the foreign exchange market, we note that the nominal bilateral exchange rate between
the INR and the USD has been depreciating during this ongoing COVID-19 phase,
dropping to 76.06 on 23 April 2020, a depreciation of 6.4%.
2. The unit root tests results are not presented here due to space constraints, but they are
available upon request.

References
Al-Awadhi, A. M., K. Al-Saifi, A. Al-Awadhi, and S. Alhamadi. 2020. Death and contagious
infectious diseases: Impact of the COVID-19 virus on stock market returns. Journal of
 EMERGING MARKETS FINANCE AND TRADE 2179

Behavioural and Experimental Finance 27 (September):100326. doi:10.1016/j.

jbef.2020.100326.
Ali, M., N. Alam, and S. A. R. Rizvi. 2020. Coronavirus (COVID-19)–An epidemic or
pandemic for financial markets. Journal of Behavioral and Experimental Finance
27:100341. doi:10.1016/[Link].2020.100341.
Apergis, N., and E. Apergis. 2020. Can the COVID-19 pandemic and oil prices drive the US
Partisan Conflict Index?. Energy Research Letters 1 (1): 13144. doi:10.46557/001c.13144.
Barro, R. J., J. F. Ursúa, and J. Weng 2020. The coronavirus and the great influenza pandemic:
Lessons from the “Spanish flu” for the coronavirus’s potential effects on mortality and
economic activity. National Bureau of Economic Research Working Papers No. 26866.
[Link]
Branson, W. H. 1983. Macroeconomic determinants of real exchange risk. In Managing foreign
exchange risk, ed. R. J. Herring, 33–74. Cambridge: Cambridge University Press.
Broock, W. A., J. A. Scheinkman, W. D. Dechert, and B. LeBaron. 1996. A test for indepen-
dence based on the correlation dimension. Econometric Reviews 15 (3):197–235.
doi:10.1080/07474939608800353.
Cai, J. 1994. A Markov model of switching-regime ARCH. Journal of Business & Economic
Statistics 12 (3):309–16. doi:10.1080/07350015.1994.10524546.
Castellano, R., and L. Scaccia. 2010. Bayesian hidden Markov models for financial data. In Data
analysis and classification: From exploration to confirmation, ed. F. Palumbo, C. N. Lauro,
and M. J. Greenacre, 453–61. Berlin-Heidelberg: Springer.
Chevallier, J. 2011a. Macroeconomics, finance, commodities: Interactions with carbon markets
in a data-rich model. Economic Modelling 28 (1–2):557–67. doi:10.1016/j.
econmod.2010.06.016.
Chevallier, J. 2011b. Detecting instability in the volatility of carbon prices. Energy Economics 33
(1):99–110. doi:10.1016/[Link].2010.09.006.
Dev, S. M., and R. Sengupta 2020. Covid-19: Impact on the Indian economy. Indira Gandhi
Institute of Development Research Working PaperNo.2020-013. [Link]
publication/[Link]
Dhingra, V. S., S. Gandhi, and H. P. Bulsara. 2016. Foreign institutional investments in India:
An empirical analysis of dynamic interactions with stock market return and volatility. IIMB
Management Review 28 (4):212–24. doi:10.1016/[Link].2016.10.001.
Donadelli, M., R. Kizys, and M. Riedel. 2017. Dangerous infectious diseases: Bad news for main
street, good news for wall street? Journal of Financial Markets 35 (C):84–103. doi:10.1016/j.
finmar.2016.12.003.
Edelen, R. M., and J. B. Warner. 2001. Aggregate price effects of institutional trading: A study of
mutual fund flow and market returns. Journal of Financial Economics 59 (2):195–220.
doi:10.1016/S0304-405X(00)00085-4.
Frankel, J. A. 1983. Monetary and portfolio-balance models of exchange rate determination. In
Economic interdependence and flexible exchange rates, ed. J. S. Bhandari, B. H. Putnam, and
J. H. Levin, 84–115. Cambridge: MIT Press.
Fu, M., and H. Shen. 2020. COVID-19 and corporate performance in the energy industry.
Energy Research Letters 1 (1): 12967. doi:10.46557/001c.12967.
Garg, R., and P. Dua. 2014. Foreign portfolio investment flows to India: Determinants and
analysis. World Development 59:16–28. doi:10.1016/[Link].2014.01.030.
Gil-Alana, L. A., and M. Monge. 2020. Crude oil prices and COVID-19: Persistence of the
shock. Energy Research Letters 1 (1): 13200. doi:10.46557/001c.13200.
Gormsen, N. J., and R. S. J. Koijen 2020. Coronavirus: Impact on stock prices and growth
expectations. University of Chicago, Becker Friedman Institute for Economics Working Paper
No. 2020-22. doi:10.2139/ssrn.3555917.
2180 A. K. MISHRA ET AL.

Granger, C. W. J., B. N. Huang, and C. W. Yang. 2000. A bivariate causality between stock
prices and exchange rates: Evidence from recent Asian flu. The Quarterly Review of
Economics and Finance 40 (3):337–54. doi:10.1016/S1062-9769(00)00042-9.
Gray, S. F. 1996. Modelling the conditional distribution of interest rates as a regime switching
process. Journal of Financial Economics 42 (1):27–62. doi:10.1016/0304-405X(96)00875-6.
Hamilton, J. D. 1989. A new approach to the economic analysis of nonstationary time series
and the business cycle. Econometrica 57 (2):357–84. doi:10.2307/1912559.
Hamilton, J. D., and R. Susmel. 1994. Autoregressive conditional heteroskedasticity and
changes in regime. Journal of Econometrics 64 (1–2):307–33. doi:10.1016/0304-4076(94)
90067-1.
Haroon, O., and S. A. R. Rizvi. 2020. COVID-19: Media coverage and financial market
behaviour-A sectoral inquir. Journal of Behavioural and Experimental Finance 2020.
doi:10.1016/[Link].2020.100343.
Ichev, R., and M. Baker Marinc. 2018. Stock prices and geographic proximity of information:
Evidence from the Ebola outbreak. International Review of Financial Analysis 56:153–66.
doi:10.1016/[Link].2017.12.004.
Jain, A., and P. C. Biswal. 2016. Dynamic linkages among oil price, gold price, exchange rate,
and stock market in India. Resources Policy 49:179–85. doi:10.1016/[Link].2016.06.001.
Karame, F. 2010. Impulse-response functions in Markov-switching structural vector autoregres-
sions: A step further. Economics Letters 106 (3):162–65. doi:10.1016/[Link].2009.11.009.
Karame, F. 2015. Asymmetries and Markov-switching structural VAR. Journal of Economic
Dynamics & Control 53:85–102. doi:10.1016/[Link].2015.01.007.
Katechos, G. 2011. On the relationship between exchange rates and equity returns: A new
approach. Journal of International Financial Markets, Institutions and Money 21 (4):550–59.
doi:10.1016/[Link].2011.03.001.
Koop, G., M. H. Pesaran, and S. M. Potter. 1996. Impulse response analysis in nonlinear multi-
variate models. Journal of Econometrics 74 (1):119–47. doi:10.1016/0304-4076(95)01753-4.
Lagoarde-Segot, T., and P. L. Leoni. 2013. Pandemics of the poor and banking stability. Journal
of Banking & Finance 37 (11):4574–83. doi:10.1016/[Link].2013.04.004.
Liu, L., E. Z. Wang, and C. C. Lee. 2020. Impact of the COVID-19 pandemic on the crude oil
and stock markets in the US: A time-varying analysis. Energy Research Letters 1 (1): 13154.
doi:10.46557/001c.13154
Mishra, A. K. 2004. Stock market and foreign exchange market in India: Are they related? South
Asia Economic Journal 5 (2):209–32. doi:10.1177/139156140400500202.
Narayan, P. K. 2020. Oil price news and COVID-19—Is there any connection?. Energy Research
Letters 1 (1): 13176. doi:10.46557/001c.13176
Narayan, P. K., S. Narayan, and K. P. Prabheesh. 2014. Stock returns, mutual fund flows and
spillover shocks. Pacific-Basin Finance Journal 29:146–62. doi:10.1016/[Link].2014.03.007.
Phan, D. H. B., and P. K. Narayan. 2020. Country responses and the reaction of the stock
market to COVID-19 - A preliminary exposition. Emerging Markets Finance and Trade 56
(10): 2138–2150. doi:10.1080/1540496X.2020.1784719.
Qin, M., Y. C. Zhang, and C. W. Su. 2020. The Essential Role of Pandemics: A Fresh Insight
into the Oil Market. Energy Research Letters 1 (1): 13166. doi:10.46557/001c.13166
Sensoy, A., and B. M. Tabak. 2015. Time-varying long term memory in the European Union
stock markets. Physica A: Statistical Mechanics and Its Applications 436:147–58. doi:10.1016/
[Link].2015.05.034.
Zhang, D., M. Hu, and Q. Ji. 2020. Financial markets under the global pandemic of COVID-19.
Finance Research Letters 101528. doi:10.1016/[Link].2020.101528.