investment behavior during covid pandemic: an evidence

Business Review Business Review

Volume 16 Issue 1 January-June 2021 Article 4

8-16-2021

Investment behavior during COVID pandemic: An evidence from Investment behavior during COVID pandemic: An evidence from

emerging Asian Markets emerging Asian Markets

Sana Tauseef Institute of Business Administration (IBA)

Follow this and additional works at: https://ir.iba.edu.pk/businessreview

Part of the Finance and Financial Management Commons, Operations and Supply Chain Management

Commons, and the Portfolio and Security Analysis Commons

This work is licensed under a Creative Commons Attribution 4.0 International License.

iRepository Citation iRepository Citation Tauseef, S. (2021). Investment behavior during COVID pandemic: An evidence from emerging Asian Markets. Business Review, 16(1), 76-100. Retrieved from https://doi.org/10.54784/1990-6587.1265

This article is brought to you by iRepository for open access under the Creative Commons Attribution 4.0 License and is available at https://ir.iba.edu.pk/businessreview/vol16/iss1/4. For more information, please contact [email protected].

https://iba.edu.pk/

https://iba.edu.pk/

https://ir.iba.edu.pk/businessreview

https://ir.iba.edu.pk/businessreview/vol16

https://ir.iba.edu.pk/businessreview/vol16/iss1

https://ir.iba.edu.pk/businessreview/vol16/iss1/4

https://ir.iba.edu.pk/businessreview?utm_source=ir.iba.edu.pk%2Fbusinessreview%2Fvol16%2Fiss1%2F4&utm_medium=PDF&utm_campaign=PDFCoverPages

http://network.bepress.com/hgg/discipline/631?utm_source=ir.iba.edu.pk%2Fbusinessreview%2Fvol16%2Fiss1%2F4&utm_medium=PDF&utm_campaign=PDFCoverPages




https://creativecommons.org/licenses/by/4.0/



https://doi.org/10.54784/1990-6587.1265

https://ir.iba.edu.pk/


https://ir.iba.edu.pk/businessreview/vol16/iss1/4

mailto:[email protected]

Business Review: (2021) 16(1):76-100Original Paper

Investment behavior during the COVID pandemic:An evidence from emerging Asian markets

Sana Tauseef

Abstract This study investigates the investor behavior in Asian emerging mar-kets during the COVID pandemic period and assesses the asymmetric patternsand the influence of global and regional return dynamics on local investor be-havior. It uses daily firm-level data from January 2001 to December 2020 fornine Asian emerging countries: China, India, Indonesia, Korea, Malaysia, Pak-istan, Philippines, Taiwan and Thailand to estimate the impact of extreme pricemovements, market liquidity and US and Chinese market returns on local returndispersions using a non-linear specification. The findings indicate that rationalprice behavior prevailed in all markets over the COVID period. Though herdingmanifested itself occasionally during the low trading volume days in the pan-demic period; overall, herd formation was not observed conditioned on domesticor non-domestic factors. Infact, our comparison of pre-COVID and COVID pe-riods shows a significant shift in investment behavior away from herding andin favor of market efficiency for emerging markets of China, India and Korea.The findings on prevalence of rational price behavior are unique and reflectiveof improved informational environment in Asian emerging markets.

Keywords Rational price behavior · Herding · Return dispersions · Pandemic

1 Introduction

Market efficiency hypothesis assumes that investors form rational expectationsand make informed decisions based on their own independent analysis. As analternate, behavioral finance argues that investors operate with cognitive andemotional biases and their decisions are not always optimal. One such sys-tematic bias reflected in investors’ behavior is herding. Herding occurs wheninvestors imitate each other resulting in a large group of investors making sametrading decisions. Such behavior may result due to rational investing as well.For example, investors using the same source of information or using similar

Sana TauseefInstitute of Business Administration, Karachi-PakistanE-mail: [email protected]

https://ir.iba.edu.pk/businessreview/vol16/iss1/4DOI: https://doi.org/10.54784/1990-6587.1265

Published by iRepository, August 2021

Investment behavior during the Covid Pandemic...

methodologies reach to same investment decisions (Hirshleifer et al 1994). Therational herd behavior (also referred to as spurious herding) is more commonamong institutional investors who have access to homogenous information, usesame analytical approaches or have similar dispositions towards certain stockscharacteristics and their trading speeds up the price adjustment (Gompers andMetrick 2001). On the contrary, irrational herding (also referred to as true herd-ing) occurs when investors blindly copy the trading actions of other investorsrather than relying on their independent analysis. Herding backed by irrationalinvestor behavior is of particular concern in context of asset pricing theoriessince such behavior may aggravate volatility, slow down price adjustment, cre-ate further mispricing and hence destabilize financial markets. Despite cleartheoretical distinction between spurious and true herding, most empirical stud-ies gauge the clustering of investment decisions irrespective of the underlyingreasons for such behavior.

Recently, emerging markets have been an attraction for foreign investorsowing to their rapid economic growth (OECD 2019). These markets are char-acterized as having less sophisticated investors, low transparency levels andinformation asymmetry. As a result, it is argued that traders in these marketsmay base their investment actions on decisions of others who may be moreinformed by following market consensus (Gelos and Wei 2005). Consequently,herd behavior is expected to be relatively stronger for emerging markets thandeveloped markets (Chang et al 2000). However, existing literature from emerg-ing Asian markets is limited. Furthermore, these emerging markets are focusedon enhancing public equity, financial liberalization and economic integration.Indeed, Asian emerging markets of China, India and Japan have been rankedamong the top ten markets based on the number of public equity listings overthe period 2009-2018 (OECD 2019). As a matter of fact, market dynamics inthese markets are evolving rapidly over time rendering the findings based on olddata invalid and requiring reinvestigation of the investor behavior from thesemarkets in the most recent environment. In this paper, we contribute to theliterature by assessing the investor behavior in nine Asian emerging markets:China, India, Indonesia, Korea, Malaysia, Pakistan, Philippines, Taiwan andThailand over the period from 2001 to 2020. Available studies from these coun-tries focus mainly on investment behavior in individual countries. For instance,Tan et al (2008), Chiang et al (2010) and Chong et al (2017) investigated in-vestor behavior in China; Prosad et al (2012) and Chauhan et al (2020) checkedfor herding in Indian market; Brahmana et al (2012) and Dehghani and Sapian(2014) conducted studies on herding from Malaysia; Akbar et al (2019), Ra-jput and Bhutto (2019) and Jabeen and Rizavi (2019) reported evidence ofherding from Pakistan and Chang et al (2012) investigated herding in Taiwan.Findings of these studies from individual markets are based on different dataperiods making them incomparable. We include all Asian emerging countriesin our study to provide a comparative analysis of rationality or herding acrossemerging countries in Asian region.

More specifically, we investigate whether investment behavior in the Asianemerging countries varies in its significance prior to and during the COVID-19 outbreak. Literature show that herding becomes more apparent during the

Business Review: (2021) 16(1):76-100 77



S. Tauseef

period of crisis. For example, Goodfellow et al (2009) provided evidence thatinvestors deviate away from rational behavior towards herding during marketdownswings. In addition, Zhou and Lai (2009) observed that investors are morelikely to herd while selling stocks and less likely while buying. On similar lines,Chiang et al (2010) provided supportive evidence of herd formation in US andLatin American markets during the crisis periods. The recent COVID periodpresents a case to test the earlier reported findings. Hence, our study stresseson the presence of irrational investor behavior and change in investor behaviorover the recent COVID outbreak period. Furthermore, in line with existing lit-erature, we also investigate if investor behavior exhibits asymmetric propertiesconditional upon extreme market performance and extreme trading volume.

Finally, we test for the impact of two equity markets, US and China, onlocal investment behavior over the COVID period. We consider the impact ofthe US market since it is the most dominant international market. Further,the Chinese market is included because of its dominant position in the regionand its significant economic partnership with most Asian emerging countries.In an integrated global environment, the investment behavior in one market isexpected to be influenced by global factors and hence, findings on influence ofthese non-domestic factors would provide a hint about the level of integrationof these markets.

Our empirical tests indicate the presence of rational investor behavior inall markets over the COVID period. Upon conditioning to market liquidity, weobserve herding on low trading volume days for China, India, Korea and Pak-istan; however, herd formation was not observed on any other occasion based ondomestic or non-domestic factors. Furthermore, our comparison of the investorbehavior in COVID and pre-COVID periods shows a shift in price behavioraway from herding and in favor of market efficiency for China, India and Korea.Our findings, thus, support prevalence of efficient market conditions in Asianemerging markets and are of particular relevance to foreign portfolio investors.Disappearing herd behavior and a shift towards rational behavior in these mar-kets in recent period is indicative of improved informational efficiency and trans-parency in these markets making the market destabilizing incidents less likely.Given the impressive growth of Asian emerging markets in recent years (OECD2019) and the evidence of diversification opportunities that emerging marketsoffer (Conover et al 2002), prevalence of rational pricing in these markets is apositive news and should provide confidence to the global portfolio investors toincrease their portfolio allocations in these markets.

The remainder of the paper is structured as follows: In section 2, we pro-vide review of empirical studies on herding, section 3 presents the descriptionof data and the methodological details. In section 4, we provide the discussionof empirical results and in section 5, we provide concluding remarks.

2 Literature review

Empirical investigation of investment behavior in financial markets focuses onthe cross-sectional return dispersions. Christie and Huang (1995) argued that

78 Business Review: (2021) 16(1):76-100




investors are more likely to follow the market consensus during times of unusualmarket movements which would result in smaller cross-sectional dispersions.Chang et al (2000) extended the previous models by arguing that presence ofherding can be detected by employing non-linear relationship between returnsand dispersions. According to their framework, if herding exists, return disper-sions may decrease or increase at a decreasing rate with an increase in marketreturn.

Since the seminal works, a number of studies have been conducted to exam-ine the investors’ behavior in international equity markets. The literature on in-vestor behavior from developed stock markets presents mixed findings. Christieand Huang (1995) tests provided evidence against herding behavior and in favorof rational pricing for US equity market. Chiang et al (2010) reported evidenceof herding in Asian markets and selected developed markets. However, theirstudy did not find indication of herding for US and Latin American markets.Similarly, Mobarek et al (2014) did not find significant herd behavior in Euro-pean markets.

The evidence of herding behavior is stronger in emerging markets than inmature markets because more speculators having relatively short investmenthorizons are present in emerging markets (Chang et al 2000). Demirer andKutan (2006) examined investor behavior in the two Chinese equity markets:Shanghai and Shenzhen Stock Exchange. Taking into context daily stock returndata from 1999 to 2002 for 375 Chinese stocks, they found negligible evidenceon herding. Infact, the study reported equity return dispersions to be signifi-cantly higher during periods of large changes in the aggregate market index.Considering the differing characteristics of investors on two exchanges in China,Tan et al (2008) investigated investor behavior in dual-listed Chinese A-shareand B-share stocks and found evidence of herding within both the Shanghai andShenzhen markets. Moreover, they provided evidence of herding in both risingand falling market conditions for Chinese stock market and reported herding tobe more pronounced during timings of rising markets, high trading volume, andhigh volatility.

Evidence of herding is also available from other Asian emerging markets. Forexample, Prosad et al (2012) examined the stocks that make up Nifty 50 equityindex in India over the period from 2006 to 2011 and documented presence ofherding during the bull phase over periods of market stress. In a recent study,Chauhan et al (2020) also found evidence of herding from the Indian market.Their study observed that herding exists more significantly in large-cap stocksthan in small-cap stocks and reported it as a significant risk factor which ispriced in the Indian market.

Agarwal et al (2011) documented no evidence of herding found in investors’behavior across brokerage firms in Indonesia; however, herd behavior prevailedfor both domestic and foreign investors associated with particular brokeragefirms. Further, Putra, Rizkianto and Chalid (2017) confirmed presence of herd-ing in Indonesia and proved herding to be stronger in Indonesia than in Singa-pore.

Brahmana et al (2012) tested the role of herd behavior in determining theday-of-the week anomaly in Bursa Malaysia over the period from 1990 until

Business Review: (2021) 16(1):76-100 79



S. Tauseef

2010. They found evidence of Monday herding over complete sample period aswell as sub-periods. Further, herd behavior did not exist in market downswingsand market upswings through the week. Hence, the study reported herd behav-ior to be the determinant of Investor’s Monday irrationality in Malaysian stockmarket. Dehghani and Sapian (2014) examined investors’ herding behavior inthe IPO aftermarket from 2001 to 2011 in Bursa Malaysia and provided evidencethat the investors in the technology sector tend to follow the market movementduring down market conditions. Significant herding in the technology sector wasrelated to small firm size in the sector, which makes them considered to be therisky ones by the investors.

Akbar et al (2019) observed herding for five out of eleven industries in Pak-istan and reported that stock trading volume is a predictor of herding than stockreturns. Using time series data for the period from 2000 to 2016, the study notonly reported significant evidence of herding behavior in the Pakistan stockmarket over the entire sample period but also found it to be more pronouncedunder extreme market conditions, market volatility and financial crisis. Fur-ther, it was found that herding behavior before any concerning crisis increased,whereas it decreased at the time of crisis. Javaira and Hassan (2015) and Yousafet al (2018)) also confirmed prevalence of herding in Pakistan’s equity marketduring the crisis period.

Chang et al (2012) reported that herding prevails in Taiwan market. Further,the study concluded that significant gains can be made based on portfolio con-struction strategies based on individual investors’ herding. Moreover, Huang andWang (2017) investigated the impact of changes in the volatility index on theinvestment behavior of market participants in the Taiwan stock market. Theyvalidated that investors’ reaction to bad news is faster than to good news whentheir fear increases. In addition, they revealed that herding behavior tends toexist on days with a large trading volume. Recently, Jirasakuldech and Emekter(2020) investigated herding behavior of investors in Thailand under differentmarket conditions, crisis periods, and major market structural changes. Thestudy found that herding behavior occurs during an extreme market movementof negative or positive returns, in the declining market environment, duringtimeline when there is greater trading volume, and during economic downturnas measured by negative GDP output gap.

The evidence on presence of herding from Asian emerging markets is substan-tial; however, previous studies are focused on individual markets. We considerall nine Asian emerging markets in our sample and investigate investor behaviorover the recent COVID-19 pandemic. Also, we test for the impact of two equitymarkets, US and China, on local investment behavior to provide evidence onthe spillover effects of herding from the two dominant equity markets.

3 Data and methodology

We used daily stock price data for all stocks listed on following nine Asianemerging markets based on MSCI market classification: China, India, Indone-sia, Korea, Malaysia, Pakistan, Philippines, Taiwan and Thailand. We obtained

80 Business Review: (2021) 16(1):76-100




the data for the entire population over the period January 2001 till Decem-ber 2020 from Thomson-Reuters Data Stream. Along the lines of Christie andHuang (1995) and Chang et al (2000), our methodology relies on the idea thatif investors’ actions are based on collective actions of market, security returnswould not deviate far from the market returns, hence indicating presence of herdbehavior. For each country, we calculated the cross-sectional absolute deviation(CSAD) as measure of return dispersions:

CSADt =Σn

i=1|Ri,t −Rm,t|n

(1)

where Ri is individual stocks return and Rm is return on respective market’smajor equity index. For each market, we divide our sample period into two sub-periods: COVID period starts from the day first COVID case was reported inthe respective country and ends on December 31, 2020 and pre-COVID periodstarts from January 01, 2001 and ends one day before the first COVID casewas reported in the respective country. We report the equity index, date offirst reported COVID case and total COVID cases for sample countries in tableA1 in the appendix. We checked for the presence of herding or rational pricebehavior in the two periods and possibility of change in investor behavior overthe pandemic period, using the following non-linear specification:

CSADCOV IDt = α+ βCOV ID

1 |RCOV IDm,t |+ βCOV ID

2 (RCOV IDm,t )2 + εt

CSADpre−COV IDt = α+βpre−COV ID

1 |Rpre−COV IDm,t |+βpre−COV ID

2 (Rpre−COV IDm,t )2+εt

(2)

Given the rational asset pricing expectations, dispersions would increase withan increase in absolute market returns, reflected in the positive significant val-ues of βCOV ID

1 and βpre−COV ID1 in the above model. In presence of herding,

dispersions might decrease or increase at a decreasing rate, reflected in the neg-ative significant values of βCOV ID

2 and βpre−COV ID1 (Chang et al 2000) and a

significant difference in the two values would indicate a change in investor be-havior over the pandemic period.

We investigate the asymmetric patterns of herding behavior over the COVIDand pre-COVID periods with respect to extreme price movement and liquidity.In literature, one, five and ten percent in the upper and lower tails of marketreturn and market trading volume are used to define extreme price movementand liquidity respectively. Owing to the small number of observations in theCOVID period for each market, we consider ten percent of the highest and low-est return and trading volume within each period as extreme conditions. Theasymmetric behavior with respect to extreme price movement is estimated asfollows:

CSADCOV IDt = α+β1D

COV IDup |Rm,t|+β2DCOV ID

down |Rm,t|+β3DCOV IDup (Rm,t)

2

+ β4DCOV IDdown (Rm,t)

2 + εt

CSADpre−COV IDt = α+ β1D

pre−COV IDup |Rm,t|+ β2D

pre−COV IDdown |Rm,t|

+ β3Dpre−COV IDup (Rm,t)

2 + β4Dpre−COV IDdown (Rm,t)

2 + εt (3)

Business Review: (2021) 16(1):76-100 81



S. Tauseef

where Dup and Ddown are dummy variables which equal 1 for days with ex-treme positive and negative market returns, respectively. Significant negative(positive) values for of β3 and β4 would indicate presence of herd (rational)behavior on days of extreme price movement and β4 < β3 would indicate morepronounced herding on days with negative market returns. The asymmetricbehavior with respect to market liquidity is estimated as follows:

CSADCOV IDt = α+ β1|Rm,t|+ β2(Rm,t)

2 + β3DCOV IDhigh (Rm,t)

2

+ β4DCOV IDlow (Rm,t)

2 + εt

CSADpre−COV IDt = α+ β1|Rm,t|+ β2((Rm,t)

2 + β3Dpre−COV IDhigh (Rm,t)

2

+ β4Dpre−COV IDlow (Rm,t)

2 + εt

(4)

where Dhigh and Dlow are dummy variables which equal 1 for days with highand low market liquidity, respectively. Significant negative (positive) values forof β3 and β4 would indicate presence of herd (rational) behavior on days of highand low liquidity and β3 < β4 would indicate more pronounced herding on dayswith high trading volume.

Finally, we test for the effect of US market and Chinese market on local in-vestment behavior in sample markets over the COVID period using the followingmodel:

CSADCOV IDt = α+βCOV ID

1 |RCOV IDm,t |+βCOV ID

2 (RCOV IDm,t )2+βCOV ID

3 (RCOV IDUS,t )2+εt

CSADCOV IDt = α+βCOV ID

1 |RCOV IDm,t |+βCOV ID

2 (RCOV IDm,t )2+βCOV ID

3 (RCOV IDCh,t )2+εt

(5)

In the above model, RUS is US market return proxied through S&P 500 indexand RCh is Chinese market return proxied through Shanghai Composite Index.Significantly negative values of β2 would indicate presence of domestic herdingwhereas significantly negative values of β4 would indicate notable influence ofreturn dynamics of US and Chinese markets on local investment behavior ofrespective Asian market. Estimations of regression coefficients in all tests aredone using Newey and West (2017) approach to adjust for heteroscedasticityand autocorrelation.

4 Empirical results

4.1 Descriptive statistics

In table 1, we provide summary statistics for mean daily returns (Rm) andcross-sectional absolute deviations (CSAD) for sample countries. Mean dailyreturn of sample countries is higher in the COVID period with the exception ofPakistan and Philippines. Over the COVID period, mean daily returns rangebetween the lowest of -0.02 percent for Philippines and highest of 0.10 percent for

82 Business Review: (2021) 16(1):76-100




Korea. Volatility, as measured by standard deviation of returns, is also higher inthe COVID period for most countries. India and Philippines had most extremedaily price movements with minimum daily returns of -13.94 percent and -17.49and maximum daily returns of 8.40 percent and 6.88 percent, respectively.

Mean daily cross-sectional deviations reduced in COVID period for samplecountries except China, Korea and Malaysia. Mean daily dispersions for oursample range from lowest of 1.69 percent for China and highest of 3.69 percentfor India in the pandemic period. We also report test statistics for AugmentedDicky-Fuller (ADF) test on return and dispersion series. These tests indicatethat return and dispersion series for all sample countries exhibit stationarity.

4.2 Investor behavior over COVID and Pre-COVID periods

We begin our investigation of investor behavior in nine Asian markets by es-timating Equation 2. We present our estimated results in table 2. Our resultsshow positive coefficients on |Rm| across all countries for both COVID and pre-COVID periods. Furthermore, these values are significant with the exceptionof Korea and Malaysia in COVID period. In line with theoretical expectationsbased on rational asset pricing, these results confirm that return dispersionswiden with increase in market returns across all countries. The coefficient of(Rm)2 tests if the dispersions increase at a decreasing rate which would be in-dicative of herd behavior. We observe significant negative coefficient of (Rm,t)

2

for China in pre-COVID period only, whereas in the COVID period, none of thecountries show significant negative coefficient on (Rm)2. Infact, the positive val-ues on the coefficient are significant for China, India and Korea in the COVIDperiod and for Malaysia, Philippines and Thailand in the pre-COVID period.

In table 2, we also report the test statistic on the difference of coefficientson (Rm)2 between the COVID and pre-COVID period. We observe significantchange in investor behavior between the two periods for China, India, Koreaand Philippines. For Philippines, positive significant coefficient of (Rm)2 in pre-COVID period turned to negative for COVID period indicating a shift awayfrom rational pricing and towards herd formation; however, as noted above,herd behavior for Philippines in COVID period is not significant. In contrast,for China, India and Korea, negative coefficient of (Rm)2 in pre-COVID pe-riod turned to positive for COVID period indicating a shift in price behavioraway from herding. The significant difference in coefficient of (Rm)2 betweentwo periods along with significant positive coefficient in COVID period providesevidence of shift towards and prevalence of rational investor behavior in thesecountries.

Business Review: (2021) 16(1):76-100 83



S. Tauseef

Table

1:

Su

mm

ary

stati

stic

sof

mark

etre

turn

san

dcr

oss

-sec

tion

al

dev

iati

on

s

Pan

elA

:C

OV

IDp

erio

d

Ch

ina

Ind

iaIn

don

esia

Kore

aM

ala

ysi

aR

mC

SA

DR

mC

SA

DR

mC

SA

DR

mC

SA

DR

mC

SA

D277

229

200

235

230

Mea

n0.0

72

1.6

93

0.0

62

3.6

93

0.0

63.7

04

0.1

1.9

90.0

23.7

3S

td.

Dev

.1.2

54

0.4

14

2.0

82

0.7

65

1.8

21.1

56

1.8

20.6

51.2

81.0

3M

ax

5.5

54

4.4

12

8.4

9.7

59.7

10.6

8.2

57.2

83.9

29.7

2M

in-8

.039

0.9

24

-13.9

04

3.0

39

-6.8

12.3

4-8

.77

1.2

3-8

.05

2.4

9A

DF

-15.5

96**

-8.1

17**

-5.3

79**

-2.7

18**

-10.5

90**

-3.7

22**

-8.8

17**

-3.5

42**

-8.5

45**

-3.8

89**

Pakis

tan

Ph

ilip

pin

esT

aiw

an

Th

ail

an

dR

mC

SA

DR

mC

SA

DR

mC

SA

DR

mC

SA

D211

220

232

236

Mea

n0.0

62.8

7-0

.02

3.1

20.0

91.6

30.0

38

2.3

51

Std

.D

ev.

1.6

40.6

52.3

11.5

11.2

60.6

51.2

92

0.7

61

Max

5.5

6.8

46.8

815.7

33.7

5.6

89.6

39

13.5

7M

in-7

.21.8

-17.4

91.7

6-7

.21

0.8

5-1

2.5

92

1.2

96

AD

F-1

2.2

43**

-3.1

40**

-7.8

98**

-4.2

00**

-8.5

02**

-3.7

22**

-5.3

21**

-2.7

94**

Pan

elB

:P

re-C

OV

IDp

erio

d

Ch

ina

Ind

iaIn

don

esia

Kore

aM

ala

ysi

aR

mC

SA

DR

mC

SA

DR

mC

SA

DR

mC

SA

DR

mC

SA

D4573

4728

4666

4706

4682

Mea

n0.0

07

1.6

20.0

48

4.4

55

0.0

63.9

30.0

31.9

60.0

22.6

6S

td.

Dev

.1.5

73

0.6

99

1.3

92

1.8

96

1.3

1.5

31.3

60.5

60.7

70.6

7M

ax

9.4

01

7.5

95

16.3

34

34.1

76

7.6

217.6

111.2

89.3

26.2

88.9

6M

in-9

.256

0.5

3-1

3.0

54

1.8

77

-10.9

51.7

2-1

2.8

1.0

1-6

.81

1.4

AD

F-6

6.3

80**

-7.6

62**

-49.2

08**

-2.0

78**

-61.2

87**

-4.4

75**

-67.2

20**

-8.1

89**

-62.9

65**

-6.7

57**

Pakis

tan

Ph

ilip

pin

esT

aiw

an

Th

ail

an

dR

mC

SA

DR

mC

SA

DR

mC

SA

DR

mC

SA

D4682

4709

4687

4653

Mea

n0.0

73.4

20.0

33.6

60.0

21.7

8-0

.044

2.8

8S

td.

Dev

.1.3

61.1

11.1

71.5

31.3

62.3

41.9

09

1.1

14

Max

8.2

819.6

110.8

913.3

410.5

3108.2

6.1

78

9.6

71

Min

-8.9

40.8

8-1

2.0

21.4

3-9

.65

0.5

8-1

6.1

01

1.8

57

AD

F-6

3.2

15**

-7.4

02**

-47.0

06**

-5.8

62**

-71.6

70**

-17.7

40**

-68.3

83**

-6.6

84**

Note

:T

he

tab

lere

port

ssu

mm

ary

stati

stic

sof

mea

nd

aily

mark

etre

turn

sR

man

dcr

oss

-sec

tion

al

ab

solu

ted

evia

tion

s(C

SA

D)

for

sam

ple

cou

ntr

ies

inC

OV

IDan

dp

re-C

OV

IDp

erio

ds.

CO

VID

per

iod

start

sfr

om

the

day

firs

tC

OV

IDca

sew

as

rep

ort

edin

the

resp

ecti

ve

cou

ntr

yan

den

ds

on

Dec

emb

er31,

2020

an

dp

re-C

OV

IDp

erio

dst

art

sfr

om

Janu

ary

01,

2001

an

den

ds

on

ed

ay

bef

ore

the

firs

tC

OV

IDca

sew

as

rep

ort

ed.

All

figu

res

exce

pt

test

stati

stic

sare

inp

erce

nta

ge.

AD

Fre

fers

toth

ete

stst

ati

stic

sfr

om

Au

gm

ente

d-D

icky

Fu

ller

test

.

84 Business Review: (2021) 16(1):76-100




Table

2:

Un

con

dit

ion

al

her

din

ges

tim

ati

on

s

Pan

elA

:C

OV

IDp

erio

d

Ch

ina

Ind

iaIn

don

esia

Kore

aM

ala

ysi

aP

akis

tan

Ph

ilip

pin

esT

aiw

an

Th

ailan

dC

on

stant

0.0

16**

0.0

33**

0.0

31**

0.0

17**

0.0

35**

0.0

26**

0.0

23**

0.0

13**

0.0

31**

(30.9

7)

(55.6

99)

(26.1

98)

(31.7

95)

(26.7

26)

(36.7

66)

(18.8

98)

(20.8

2)

(26.1

98)

βCovid

10.1

13*

0.2

25**

0.4

77**

0.0

94

0.1

66

0.2

42*

0.5

88**

0.3

95**

0.4

77**

(2.3

38)

(2.6

97)

(4.2

13)

(1.3

01)

(0.7

06)

(2.3

43)

(3.3

96)

(3.2

99)

(4.2

13)

βCovid

21.5

58*

1.8

94**

1.4

13

4.1

86**

4.2

85

0.4

9-0

.143

-0.7

12

1.4

13

(2.2

31)

(2.8

37)

(0.7

73)

(2.9

01)

(1.3

72)

(0.2

67)

(-0.1

48)

(-0.3

59)

(0.7

73)

Ad

just

edR

20.1

70.7

12

0.4

16

0.5

40.1

11

0.2

65

0.4

95

0.2

49

0.3

68

Pan

elB

:P

re-C

OV

IDp

erio

d

Ch

ina

Ind

iaIn

don

esia

Kore

aM

ala

ysi

aP

akis

tan

Ph

ilip

pin

esT

aiw

an

Th

ailan

dC

on

stant

0.0

13**

0.0

42**

0.0

35**

0.0

16**

0.0

25**

0.0

32**

0.0

35**

0.0

12**

0.0

21**

(59.5

25)

(45.8

11)

(52.8

1)

(77.4

82)

(103.8

68)

(66.7

56)

(45.2

24)

(15.1

9)

(63.0

75)

βPre−Covid

10.3

72**

0.2

31**

0.4

49**

0.3

74**

0.2

40**

0.2

74**

0.1

11

0.5

97**

0.2

17**

(12.6

29)

(3.8

94)

(8.2

76)

(11.4

64)

(4.0

02)

(5.8

56)

(1.7

06)

(3.5

44)

(3.8

87)

βPre−Covid

2-1

.996**

0.1

77

1.9

17

-0.5

91

5.0

09**

-0.4

78

3.7

58**

0.2

02

3.6

22**

(-3.7

87)

(0.3

63)

(1.7

07)

(-0.9

06)

(3.1

27)

(-0.4

97)

(2.4

89)

(0.1

05)

(2.7

47)

Ad

just

edR

20.2

07

0.0

16

0.1

10.3

78

0.1

13

0.0

50.0

23

0.0

66

0.2

46

Tes

tS

tat(βCovid

1=βPreCovid

1)

-4.5

87**

-0.0

59

0.2

26

-3.5

38**

-0.3

06

-0.2

85

2.5

82*

-0.9

77

2.0

59*

Tes

tS

tat

(βCovid

2=βPreCovid

2)

4.0

62**

2.0

80*

-0.1

93.0

17**

-0.2

06

0.4

67

-2.1

75*

-0.3

31

-0.9

8

Note

:T

he

tab

lere

port

sth

ees

tim

ate

dco

effici

ents

for

Equ

ati

on

2.

Est

imati

on

sin

volv

eN

ewey

-Wes

tco

nsi

sten

tes

tim

ato

rs.

CO

VID

per

iod

start

sfr

om

the

day

firs

tC

OV

IDca

sew

as

rep

ort

edin

the

resp

ecti

ve

cou

ntr

yan

den

ds

on

Dec

emb

er31,

2020

an

dp

re-C

OV

IDp

erio

dst

art

sfr

om

Jan

-u

ary

01,

2001

an

den

ds

on

ed

ay

bef

ore

the

firs

tC

OV

IDca

sew

as

rep

ort

ed.

Fig

ure

sin

pare

nth

esis

are

t-st

ati

stic

s.*

an

d**

ind

icate

sign

ifica

nce

at

5p

erce

nt

an

d1

per

cent,

resp

ecti

vel

y.

Business Review: (2021) 16(1):76-100 85



S. Tauseef

4.3 Asymmetries in investor behavior

We now test for the asymmetric patterns in investor behavior based on extremeprice movements and market liquidity. Relationship between return deviationsand market returns distinguishing between up and down markets is estimatedusing Equation 3 and results are reported in table 3. For the COVID period,coefficient on the linear term |Rm| is significantly positive for Indonesia, Pak-istan and Thailand in up market, and significantly positive for China, Pakistan,Philippines and Taiwan in down market. For the pre-COVID period, coefficientsof |Rm| are positive for all countries and significant for most countries. Coeffi-cient on (Rm)2 in both periods is either significantly positive or insignificantlynegative for all countries except for China in pre-COVID period, providing anoverall evidence against herding and in support of rational asset pricing.

We now turn our attention to the only market in our sample for which wecapture evidence of herding in pre-COVID period. The coefficient of (Rm)2 issignificant negative for China in both up and down markets for pre-COVID pe-riod, suggesting a non-linear relationship between market return and dispersion.For example, based on the coefficients of |Rm| (0.20) and (Rm)2 (1.89) in upmarket, dispersion increases as market return increases till a maximum return of5.32 percent. Maximum return is calculated by substituting the estimated coef-ficients of |Rm| and (Rm,t)

2 into the quadratic relationship specified in Equation3. Beyond this threshold return, dispersion will decrease. The finding indicatespresence of herding in Chinese market on days with extreme price movements.The difference between coefficients of (Rm)2 for up and down markets is not sig-nificant indicating that market price movement is not a determinant of herdingasymmetry in China. Furthermore, our results in table 3 again provide evidencethat herd behavior disappeared and did not exist during the COVID period inChina.

Investor behavior on days with high and low liquidity are estimated by Equa-tion 4 and we present the results in table 4. When conditioned on liquidity, weobserve investors exhibiting herd behavior in some markets during both the peri-ods. Specifically, herd formation is notable for China, India, Korea and Pakistanon low liquidity days in the COVID period. Similarly, for the pre-COVID pe-riod, we observe herd behavior in China, Malaysia and Taiwan on low liquiditydays and in Korea on both low and high liquidity days. Furthermore, signifi-cant difference in the coefficients for high and low liquidity for some countriesin either or both periods indicates asymmetric pattern in herd behavior basedon market liquidity. However, our evidence of herd formation on days with lowliquidity is in contrast with the findings of previous studies (Tan et al 2008;Indars et al 2019) that reported herd formation in periods of high liquidity.

86 Business Review: (2021) 16(1):76-100




Table

3:

Her

din

ges

tim

ati

on

sco

nd

itio

ned

on

extr

eme

pri

cem

ovem

ents

Pan

elA

:C

OV

IDp

erio

d

Ch

ina

Ind

iaIn

don

esia

Kore

aM

ala

ysi

aP

akis

tan

Ph

ilip

pin

esT

aiw

an

Th

ailan

dC

on

stant

0.0

16**

0.0

35**

0.0

34**

0.0

18**

0.0

37**

0.0

27**

0.0

27**

0.0

15**

0.0

34**

(36.7

68)

(107.1

46)

(36.5

1)

(46.9

46)

(35.1

4)

(49.4

93)

(40.9

91)

(34.1

5)

(36.5

1)

βCovid

1-0

.073

0.1

71

0.6

30**

0.1

12

-0.6

05

0.4

24*

0.6

06

0.0

90.6

30**

(-1.7

47)

(1.4

8)

(5.2

35)

(1.7

97)

(-1.7

48)

(2.0

72)

(1.3

28)

(0.3

73)

(5.2

35)

βCovid

20.1

79**

0.0

76

0.3

72

0.0

06

-0.6

59

0.2

16*

0.3

59**

0.4

14**

0.3

72

(3.6

82)

(1.4

93)

(1.8

26)

(0.0

68)

(-1.1

15)

(2.2

85)

(3.4

04)

(2.7

7)

(1.8

26)

βCovid

34.5

54**

4.7

55*

1.0

82

2.1

08*

37.3

89*

-4.9

53

0.5

35

9.9

25

1.0

82

(6.4

32)

(2.3

38)

(0.8

73)

(2.3

31)

(2.1

69)

(-1.1

78)

(0.0

75)

(0.8

73)

(0.8

73)

βCovid

40.5

76

2.6

95**

-1.6

45

6.8

48**

44.6

36

0.7

01

1.0

41

-1.9

17

-1.6

45

(0.8

67)

(6.7

4)

(-0.4

21)

(6.3

46)

(0.8

78)

(0.3

73)

(1.7

19)

(-0.8

96)

(-0.4

21)

Ad

just

edR

20.1

78

0.7

69

0.4

54

0.5

81

0.0

96

0.2

88

0.6

06

0.2

59

0.4

4T

est

Sta

t(β

1=β2)

-3.9

29**

0.7

52

1.0

93

10.0

78

0.9

23

0.1

43

-1.1

41

1.0

93

Tes

tS

tat

(β3

=β4)

4.0

99**

0.9

94

0.6

66

-3.3

67**

-0.1

35

-1.2

27

0.4

09

1.0

24

0.6

66

Pan

elB

:P

re-C

OV

IDp

erio

d

Ch

ina

Ind

iaIn

don

esia

Kore

aM

ala

ysi

aP

akis

tan

Ph

ilip

pin

esT

aiw

an

Th

ailan

dC

on

stant

0.0

15**

0.0

44**

0.0

37**

0.0

18**

0.0

26**

0.0

33**

0.0

36**

0.0

15**

0.0

22**

(75.5

88)

(51.0

36)

(63.7

64)

(117.1

92)

(125.8

)(8

2.7

32)

(55.9

23)

(79.4

690

(102.1

56)

βPre−Covid

10.2

01**

0.1

99**

0.3

22**

0.2

56**

0.2

67**

0.2

67**

0.0

87

0.4

35**

0.0

95**

(6.9

16)

(3.4

9)

(2.6

71)

(4.8

21)

(4.1

81)

(5.3

3)

(1.6

28)

(2.5

08)

(2.4

32)

βPre−Covid

20.4

00**

0.0

30.1

49**

0.3

19**

0.1

37*

0.1

26

0.0

20.5

10**

0.5

10**

(12.1

88)

(0.4

58)

(3.0

09)

(8.8

86)

(2.3

22)

(1.9

88)

(0.3

57)

(2.9

2)

(4.0

62)

βPre−Covid

3-1

.887**

0.8

69

9.7

11*

0.7

55

4.8

43**

-0.0

96.2

23**

3.3

99

7.1

30**

(-3.6

49)

(1.2

06)

(2.2

84)

(0.5

13)

(2.7

02)

(-0.0

74)

(5.9

79)

(1.3

19)

(6.1

11)

βPre−Covid

4-1

.818**

1.8

41**

3.0

75**

-0.3

62

6.0

96**

1.1

37

3.3

00**

-0.4

35

3.1

05**

(-3.1

94)

(2.9

)(2

.793)

(-0.4

68)

(3.0

76)

(1.3

49)

(3.5

25)

(-0.1

76)

(2.7

92)

Ad

just

edR

20.2

21

0.0

13

0.1

28

0.3

47

0.1

16

0.0

47

0.0

25

0.0

63

0.2

49

Tes

tS

tat

(β1

=β2)

-4.5

55**

1.9

28

1.3

25

-7.7

70**

1.4

97

1.7

42

0.8

57

-0.3

08

-1.4

33

Tes

tS

tat

(β3

=β4)

-0.0

89

-1.0

12

1.5

11

0.6

72

-0.4

69

-0.8

35

2.0

89*

1.0

74

2.4

97**

Note

:T

he

tab

lere

port

sth

ees

tim

ate

dco

effici

ents

for

Equ

ati

on

3.

Est

imati

on

sin

volv

eN

ewey

-Wes

tco

nsi

sten

tes

tim

ato

rs.

CO

VID

pe-

riod

start

sfr

om

the

day

firs

tC

OV

IDca

sew

as

rep

ort

edin

the

resp

ecti

ve

cou

ntr

yan

den

ds

on

Dec

emb

er31,

2020

an

dp

re-C

OV

IDp

erio

dst

art

sfr

om

Janu

ary

01,

2001

an

den

ds

on

ed

ay

bef

ore

the

firs

tC

OV

IDca

sew

as

rep

ort

ed.

Fig

ure

sin

pare

nth

esis

are

t-st

ati

stic

s.*

an

d**

ind

icate

sign

ifica

nce

at

5p

erce

nt

an

d1

per

cent,

resp

ecti

vel

y.

Business Review: (2021) 16(1):76-100 87



S. Tauseef

Table

4:

Her

din

ges

tim

ati

on

sco

nd

itio

ned

on

extr

eme

mark

etli

qu

idit

y

Pan

elA

:C

OV

IDp

erio

d

Ch

ina

Ind

iaIn

don

esia

Kore

aM

ala

ysi

aP

akis

tan

Ph

ilip

pin

esT

aiw

an

Th

ailan

dC

on

stant

0.0

16**

0.0

34**

0.0

30**

0.0

17**

0.0

34**

0.0

26**

0.0

24**

0.0

13**

0.0

23**

(33.1

94)

(70.9

43)

(26.6

52)

(31.8

38)

(20.0

62)

(36.1

07)

(71.4

94)

(23.0

03)

(25.8

14)

βCovid

10.0

92

0.1

13

0.5

09**

0.1

11

0.3

81

0.2

45*

0.5

2**

0.3

17**

0.4

89**

(1.9

56)

(1.5

86)

(4.5

09)

(1.4

38)

(1.0

36)

(2.2

94)

(2.5

55)

(3.3

71)

(6.0

48)

βCovid

21.3

43*

5.9

07**

1.6

64

3.9

29**

-5.4

02

0.6

74

0.8

59**

0.1

71

-0.6

57

(2.2

39)

(3.2

79)

(1.1

89)

(2.5

95)

(-0.5

65)

(0.3

44)

(3.7

04)

(0.0

99)

(-1.0

29)

βCovid

31.5

29

0.2

05

0.3

82

0.7

11

8.1

53

-3.8

11

-1.7

05

25.4

28**

-4.5

38

(1.8

52)

(0.5

14)

(0.0

96)

(0.3

2)

(1.5

04)

(-1.5

94)

(1.8

22)

(2.9

41)

(-1.3

21)

βCovid

4-1

06.6

69**

-3.4

88*

-2.9

58

-15.9

33**

-10.8

95

-1.5

09*

-35.7

63

-18.6

86

9.4

05**

(-5.1

36)

(-2.3

82)

(-1.4

05)

(-3.9

74)

(-0.6

12)

(-2.2

98)

(0.7

32)

(-1.8

82)

(3.8

03)

Ad

just

edR

20.2

30.7

66

0.4

20.5

52

0.1

44

0.2

63

0.7

08

0.3

33

0.4

1T

est

Sta

t(β

3=β4)

29.8

77**

2.4

33**

0.7

39

3.6

31**

1.0

24

-0.9

29

0.5

34

3.3

51**

0.7

39

Pan

elB

:P

re-C

OV

IDp

erio

d

Ch

ina

Ind

iaIn

don

esia

Kore

aM

ala

ysi

aP

akis

tan

Ph

ilip

pin

esT

aiw

an

Th

ailan

dC

on

stant

0.0

13**

0.0

42**

0.0

35**

0.0

16**

0.0

25**

0.0

32**

0.0

24**

0.0

13**

0.0

21**

(58.6

99)

(93.4

9)

(51.2

71)

(85.5

26)

(95.1

22)

(66.2

49)

(107.6

81)

(59.0

98)

(134.5

9)

βPre−Covid

10.4

02**

0.3

54**

0.4

79**

0.3

48**

0.2

9**

0.2

56**

0.1

08**

0.3

73**

0.1

75**

(13.0

22)

(7.0

96)

(6.9

4)

(13.1

87)

(4.0

39)

(4.7

58)

(3.1

29)

(11.5

11)

(9.3

35)

βPre−Covid

2-3

.163**

-3.3

23**

0.7

55

0.4

23

2.7

37

-0.8

61

4.0

27**

3.1

09**

3.9

08**

(-5.5

62)

(-3.4

19)

(0.5

01)

(0.7

77)

(1.3

34)

(-0.8

48)

(5.0

89)

(3.9

88)

(12.2

66)

βPre−Covid

33.0

59**

0.5

59

-4.6

52

-1.5

59**

5.6

75

-0.5

49

4.2

86**

2.3

63*

2.8

11*

(4.5

53)

(0.5

2)

(-1.2

04)

(-2.5

26)

(1.7

5)

(-0.5

52)

(3.0

19)

(2.0

65)

(2.2

35)

βPre−Covid

4-7

.872**

5.6

55**

5.2

55

-1.4

66**

-12.9

94*

24.9

29*

2.2

3-5

.409**

10.6

84**

(-4.0

06)

(5.7

58)

(1.3

06)

(-2.6

3)

(-2.3

94)

(2.3

23)

(0.4

87)

(-2.7

68)

(10.6

6)

Ad

just

edR

20.2

37

0.0

23

0.1

14

0.3

83

0.1

27

0.0

94

0.1

10.4

21

0.2

64

Tes

tS

tat

(β3

=β4)

5.2

63**

-3.4

97

-1.7

76

-0.1

12

2.9

53

-2.3

64*

0.4

29

3.4

32**

-1.3

62

Note

:T

he

tab

lere

port

sth

ees

tim

ate

dco

effici

ents

for

Equ

ati

on

4.

Est

imati

on

sin

volv

eN

ewey

-Wes

tco

nsi

sten

tes

tim

ato

rs.

CO

VID

per

iod

start

sfr

om

the

day

firs

tC

OV

IDca

sew

as

rep

ort

edin

the

resp

ecti

ve

cou

ntr

yan

den

ds

on

Dec

emb

er31,

2020

an

dp

re-C

OV

IDp

erio

dst

art

sfr

om

Janu

ary

01,

2001

an

den

ds

on

ed

ay

bef

ore

the

firs

tC

OV

IDca

sew

as

rep

ort

ed.

Fig

ure

sin

pare

nth

esis

are

t-st

ati

stic

s.*

an

d**

ind

icate

sign

ifica

nce

at

5p

erce

nt

an

d1

per

cent,

resp

ecti

vel

y.

88 Business Review: (2021) 16(1):76-100




Overall, our empirical tests indicate the presence of rational investor behav-ior in all markets over the COVID period. We observe herding on low trad-ing volume days for few markets; however, herd formation was not observedfor most markets when conditioned to domestic market factors (price move-ments and liquidity). Our findings on prevalence of efficient market conditionsin Asian emerging markets contradict with the findings about significant herd-ing reported in previous studies; for example, Chiang et al (2010) documentedevidence of herd behavior in Asian markets. Nonetheless, our results based onthe most recent data period hint about the improved information efficiency andtransparency in Asian emerging markets.

4.4 Role of US and Chinese Markets

We now check the impact of return dynamics in foreign markets on local investorbehavior over the COVID period. Panels A and B in Table 5 present the estima-tions of Equation 5 with respect to US and Chinese market returns, respectively.None of the coefficient for (RUS)2 and (RCh)2 is significantly negative suggest-ing that the local herds are not formed by the return dynamics of the dominantglobal and regional markets. Infact, we report significant positive coefficientsof RUS for China, Malaysia, Pakistan, Philippines and Taiwan, indicating anincrease in return dispersions in these markets with an absolute change in USreturn. With respect to regional influence, coefficient of RCh is significant onlyfor Taiwan. Most importantly, the findings regarding local investment behaviorin all markets after controlling for non-domestic factors remain the same andmatch the results from panel B of table 2. Our findings on insignificant impactof non-domestic factors is in line with low levels of integration of Asian emergingmarkets within global financial environment.

Business Review: (2021) 16(1):76-100 89



S. Tauseef

Table

5:

Her

din

ges

tim

ati

on

sco

ntr

ollin

gfo

rU

San

dC

hin

ese

mark

etre

turn

sover

Covid

per

iod

Pan

elA

:Im

pact

of

US

mark

et

Ch

ina

Ind

iaIn

don

esia

Kore

aM

ala

ysi

aP

akis

tan

Ph

ilip

pin

esT

aiw

an

Th

ailan

dC

on

stant

0.0

16**

0.0

33**

0.0

31**

0.0

18**

0.0

35**

0.0

26**

0.0

23**

0.0

22**

0.0

13**

(32.9

31)

(60.2

7)

(23.1

76)

(33.7

31)

(28.5

94)

(36.9

72)

(17.0

59)

(21.9

98)

(21.9

72)

βCovid

10.0

94

0.1

96**

0.4

62**

0.0

61

0.2

29

0.2

45**

0.5

20**

0.5

04**

0.4

18**

(1.8

35)

(2.6

74)

(3.5

09)

(0.8

28)

(1.2

37)

(2.4

8)

(3.1

99)

(3.3

42)

(3.3

88)

βCovid

21.8

04**

1.9

73**

0.4

75

4.3

36**

-5.0

11

-0.2

9-0

.485

-1.6

31

-3.5

19

(2.3

69)

(3.5

62)

(0.1

95)

(2.6

6)

(-1.5

72)

(-0.1

56)

(-0.5

02)

(-1.6

85)

(-1.0

15)

βCovid

30.1

85*

0.4

89

1.2

0.4

55

3.0

74**

0.7

21**

3.8

11**

2.4

51**

0.9

41

(2.0

36)

(1.4

75)

(0.8

82)

(1.5

05)

(3.8

18)

(3.1

67)

(4.0

32)

(5.2

98)

(1.0

61)

Ad

just

edR

20.1

75

0.7

20.4

35

0.5

58

0.1

73

0.2

91

0.5

68

0.4

61

0.2

75

Pan

elB

:Im

pact

of

Ch

ines

eM

ark

et

Ch

ina

Ind

iaIn

don

esia

Kore

aM

ala

ysi

aP

akis

tan

Ph

ilip

pin

esT

aiw

an

Th

ailan

dC

on

stant

-0.0

33**

0.0

30**

0.0

18**

0.0

35**

0.0

26**

0.0

23**

0.0

13**

0.0

22**

(53.9

54)

(23.6

59)

(31.4

94)

(24.4

87)

(34.0

71)

(16.9

07)

(18.2

89)

(17.0

17)

βCovid

1-

0.2

31**

0.4

78**

0.0

88

0.1

64

0.2

35*

0.6

16**

0.3

80**

0.5

85**

(2.6

76)

(3.6

82)

(1.1

5)

(0.6

69)

(2.1

88)

(3.4

24)

(2.9

74)

(3.3

02)

βCovid

2-

1.8

36**

1.3

44

4.2

64**

4.1

28

0.5

29

-0.3

11

-1.0

15

-1.2

36

(2.6

49)

(0.6

52)

(2.9

0)

(1.2

54)

(0.2

77)

(-0.3

03)

(-0.5

79)

(-1.0

46)

βCovid

3-

0.2

38

1.7

48

-0.2

50.6

96

1.0

80.8

11

2.4

34**

2.8

24

(0.4

3)

(0.7

05)

(-0.5

50)

(0.6

22)

(1.3

88)

(0.7

85)

(4.1

41)

(1.1

13)

Ad

just

edR

2-

0.7

13

0.4

23

0.5

48

0.1

05

0.2

62

0.5

03

0.2

64

0.3

84

Note

:T

he

tab

lere

port

sth

ees

tim

ate

dco

effici

ents

for

Equ

ati

on

5.

US

mark

etre

turn

isp

roxie

dth

rou

gh

S&

P500

ind

exan

dC

hin

ese

mark

etre

turn

isp

roxie

dth

rou

gh

Sh

an

gh

ai

Com

posi

teIn

dex

..E

stim

ati

on

sin

volv

eN

ewey

-Wes

tco

nsi

sten

tes

tim

ato

rs.

CO

VID

per

iod

start

sfr

om

the

day

firs

tC

OV

IDca

sew

as

rep

ort

edin

the

resp

ecti

ve

cou

ntr

yan

den

ds

on

Dec

emb

er31,

2020.

Fig

ure

sin

pare

nth

esis

are

t-st

ati

stic

s.*

an

d**

ind

icate

sign

ifica

nce

at

5p

erce

nt

an

d1

per

cent,

resp

ecti

vel

y.

90 Business Review: (2021) 16(1):76-100




4.5 Robustness of results

We perform a number of supplementary calculations to verify the robustnessof our results. Though the dispersion measure we used is preferred due to itslow sensitivity to the outliers, we employed cross section standard deviation(CSSD) as an alternate measure to confirm that our conclusions are not af-fected by the choice of dispersion measure. Secondly, we ran all models usingquantile regression which does not require the conditions of linear regression.Also, we rechecked our results of conditional herding over pre-covid period us-ing 5 percent and 1 percent criteria for defining extreme price movements andliquidity. We present the results in tables A2-A5 in the appendix. Our resultsremain qualitatively same with the use of alternate dispersion measure, regres-sion model and defining criteria for extreme conditions.

We consider COVID period from the day first COVID case was reportedin the respective country till December 31, 2020. To check for the possibilityof irrational price behavior in the short-term period following COVID shock,we performed our estimations on a sub-period of initial 30 days during COVIDoutbreak and found similar results (presented in table A6 in appendix). Finally,our pre-COVID period comprises the period of financial crisis 2007-2008 whichwas characterized by low returns and high volatility in most equity markets.Since the investment behavior might be significantly different over the crisisperiod, we report the results of the crisis period separately in table A7 in theappendix. Over the financial crisis period, we observe some evidence of herding,conditional or unconditional, for China, India, Korea, Pakistan and Thailand.For four of these markets, investor herd behavior was not observed in overallpre-COVID period. In addition, we do not find any market that showed herdingin overall pre-COVID period but not in financial crisis period. The comparisonssuggest that our findings of investment behavior over the pre-COVID period isnot impacted by the financial crisis period.

5 Conclusion

We investigate the investor behavior in the fast-growing Asian emerging marketsbefore and after the COVID outbreak using the data for the period 2001 to2020. Our findings indicate that rational price behavior prevailed in all marketsover the COVID period and herding manifested itself only in few countriesduring the low trading volume days. Interestingly enough, our comparison ofthe pre-COVID and COVID periods shows a significant shift in price behavioraway from herding and in favor of market efficiency for the emerging markets ofChina, India and Korea. Given the evidence of herding during crisis in literature,our findings of rational investor behavior over the COVID period are unique.We attribute the findings supporting rationality hypothesis in Asian emergingmarkets to the improved information quality, various financial liberalizationmeasures, and a substantial rise in available equity base in these markets. Ourfindings will be of particular interest to the investors, particularly those witha global outlook. Evidence in support of rational price behavior suggests that

Business Review: (2021) 16(1):76-100 91



S. Tauseef

occurrence of destabilizing outcomes is less likely and, hence, could provideconfidence to the global portfolio investors to increase their portfolio allocationsin these markets.

References

Agarwal S, Chiu IM, Liu C, Rhee SG (2011) The brokerage firm effect in herding: Evidencefrom indonesia. Journal of Financial Research 34(3):461–479

Akbar US, Oad Rajput SK, Bhutto NA (2019) Do investors herd with industries or markets?evidence from pakistan stock exchange. Cogent Economics & Finance 7(1):1698,089

Brahmana R, Hooy CW, Ahmad Z (2012) The role of herd behaviour in determining theinvestor’s monday irrationality. Asian Academy of Management Journal of Accounting &Finance 8(2)

Chang CY, Chen HL, Jiang ZR (2012) Portfolio performance in relation to herding behaviorin the taiwan stock market. Emerging Markets Finance and Trade 48(sup2):82–104

Chang EC, Cheng JW, Khorana A (2000) An examination of herd behavior in equity markets:An international perspective. Journal of Banking & Finance 24(10):1651–1679

Chauhan Y, Ahmad N, Aggarwal V, Chandra A (2020) Herd behaviour and asset pricing inthe indian stock market. IIMB Management Review 32(2):143–152

Chiang TC, Li J, Tan L (2010) Empirical investigation of herding behavior in chinese stockmarkets: Evidence from quantile regression analysis. Global Finance Journal 21(1):111–124

Chong TTL, Liu X, Zhu C (2017) What explains herd behavior in the chinese stock market?Journal of Behavioral Finance 18(4):448–456

Christie WG, Huang RD (1995) Following the pied piper: Do individual returns herd aroundthe market? Financial Analysts Journal 51(4):31–37

Conover CM, Jensen GR, Johnson RR (2002) Emerging markets: when are they worth it?Financial Analysts Journal 58(2):86–95

Dehghani P, Sapian RZZ (2014) Sectoral herding behavior in the aftermarket of malaysianipos. Venture Capital 16(3):227–246

Demirer R, Kutan AM (2006) Does herding behavior exist in chinese stock markets? Journalof international Financial markets, institutions and money 16(2):123–142

Gelos RG, Wei SJ (2005) Transparency and international portfolio holdings. The Journal ofFinance 60(6):2987–3020

Gompers PA, Metrick A (2001) Institutional investors and equity prices. The quarterly journalof Economics 116(1):229–259

Goodfellow C, Bohl MT, Gebka B (2009) Together we invest? individual and institu-tional investors’ trading behaviour in poland. International Review of Financial Analysis18(4):212–221

Hirshleifer D, Subrahmanyam A, Titman S (1994) Security analysis and trading patterns whensome investors receive information before others. The Journal of finance 49(5):1665–1698

Huang TC, Wang KY (2017) Investors fear and herding behavior: evidence from the taiwanstock market. Emerging Markets Finance and Trade 53(10):2259–2278

Indars ER, Savin A, Lubloy A (2019) Herding behaviour in an emerging market: Evidencefrom the moscow exchange. Emerging Markets Review 38:468–487

Jabeen S, Rizavi SS (2019) Herd behaviour, short-lived phenomenon: Evidence from pakistanstock exchange. The Lahore Journal of Business 8(2)

Javaira Z, Hassan A (2015) An examination of herding behavior in pakistani stock market.International journal of emerging markets

Mobarek A, Mollah S, Keasey K (2014) A cross-country analysis of herd behavior in europe.Journal of International Financial Markets, Institutions and Money 32:107–127

Newey WK, West KD (2017) A simple, positive semi-definite, heteroskedasticity and auto-correlation consistent covariance matrix. No1 (33) 2014 p 125

OECD (2019) Equity market review of asia 2019. OECD Capital Market Series, ParisProsad JM, Kapoor S, Sengupta J (2012) An examination of herd behavior: An empirical evi-

dence from indian equity market. International Journal of Trade, Economics and Finance3(2):154

92 Business Review: (2021) 16(1):76-100




Tan L, Chiang TC, Mason JR, Nelling E (2008) Herding behavior in chinese stock markets:An examination of a and b shares. Pacific-Basin finance journal 16(1-2):61–77

Yousaf I, Ali S, Shah SZA (2018) Herding behavior in ramadan and financial crises: the caseof the pakistani stock market. Financial Innovation 4(1):1–14

Zhou RT, Lai RN (2009) Herding and information based trading. Journal of Empirical Finance16(3):388–393

Appendix

Table A1: Sample countries

Country Equity Index Date of firstCOVID case

Total COVIDcases (31-12-20)

China Shanghai Composite 17-Nov-19 87052India Nifty 50 30-Jan-20 10286329Indonesia IDX Composite 2-Mar-20 743198Korea KOSPI 20-Jan-20 60740Malaysia KLCI 25-Jan-20 113010Pakistan KSE 100 26-Feb-20 479715Philippines PSEI Composite 30-Jan-20 474055Taiwan TAIEX 21-Jan-20 799Thailand SET 13-Jan-20 6884

Note: The table reports the major equity market index, date of first COVIDcase and number of total COVID cases for nine Asian emerging countries.

Business Review: (2021) 16(1):76-100 93



S. Tauseef

Table

A2

:Un

con

dit

ion

al

her

din

ges

tim

ati

on

su

sin

gC

SS

Das

dis

per

sion

mea

sure

Pan

elA

:C

OV

IDp

erio

d

Ch

ina

Ind

iaIn

don

esia

Kore

aM

ala

ysi

aP

akis

tan

Ph

ilip

pin

esT

aiw

an

Th

ailan

dC

on

stant

0.0

16**

0.0

33**

0.0

31**

0.0

17**

0.0

35**

0.0

26**

0.0

23**

0.0

13**

0.0

31**

(30.9

7)

(55.6

99)

(26.1

98)

(31.7

95)

(26.7

26)

(36.7

66)

(18.8

98)

(20.8

2)

(26.1

98)

βCovid

10.1

13*

0.2

25**

0.4

77**

0.0

94

0.1

66

0.2

42*

0.5

88**

0.3

95**

0.4

77**

(2.3

38)

(2.6

97)

(4.2

13)

(1.3

01)

(0.7

06)

(2.3

43)

(3.3

96)

(3.2

99)

(4.2

13)

βCovid

21.5

58*

1.8

94**

1.4

13

4.1

86**

4.2

85

0.4

9-0

.143

-0.7

12

1.4

13

(2.2

31)

(2.8

37)

(0.7

73)

(2.9

01)

(1.3

72)

(0.2

67)

(-0.1

48)

(-0.3

59)

(0.7

73)

Ad

just

edR

20.1

70.7

12

0.4

16

0.5

40.1

11

0.2

65

0.4

95

0.2

49

0.3

68

Pan

elB

:P

re-C

OV

IDp

erio

d

Ch

ina

Ind

iaIn

don

esia

Kore

aM

ala

ysi

aP

akis

tan

Ph

ilip

pin

esT

aiw

an

Th

ailan

dC

on

stant

0.0

13**

0.0

42**

0.0

35**

0.0

16**

0.0

25**

0.0

32**

0.0

35**

0.0

12**

0.0

21**

(59.5

25)

(45.8

11)

(52.8

1)

(77.4

82)

(103.8

68)

(66.7

56)

(45.2

24)

(15.1

9)

(63.0

75)

βPre−Covid

10.3

72**

0.2

31**

0.4

49**

0.3

74**

0.2

40**

0.2

74**

0.1

11

0.5

97**

0.2

17**

(12.6

29)

(3.8

94)

(8.2

76)

(11.4

64)

(4.0

02)

(5.8

56)

(1.7

06)

(3.5

44)

(3.8

87)

βPre−Covid

2-1

.996**

0.1

77

1.9

17

-0.5

91

5.0

09**

-0.4

78

3.7

58**

0.2

02

3.6

22**

(-3.7

87)

(0.3

63)

(1.7

07)

(-0.9

06)

(3.1

27)

(-0.4

97)

(2.4

89)

(0.1

05)

(2.7

47)

Ad

just

edR

20.2

07

0.0

16

0.1

10.3

78

0.1

13

0.0

50.0

23

0.0

66

0.2

46

Tes

tS

tat

(βCovid

1=βPreCovid

1)

-4.5

87**

-0.0

59

0.2

26

-3.5

38**

-0.3

06

-0.2

85

2.5

82*

-0.9

77

2.0

59*

Tes

tS

tat

(βCovid

2=βPreCovid

2)

4.0

62**

2.0

80*

-0.1

93.0

17**

-0.2

06

0.4

67

-2.1

75*

-0.3

31

-0.9

8

Note

:T

he

tab

lere

port

sth

ees

tim

ate

dco

effici

ents

for

Equ

ati

on

2.

Cro

ss-s

ecti

on

al

stan

dard

dev

iati

on

(CS

SD

)is

use

das

mea

sure

of

retu

rnd

isp

er-

sion

.E

stim

ati

on

sin

volv

eN

ewey

-Wes

tco

nsi

sten

tes

tim

ato

rs.

CO

VID

per

iod

start

sfr

om

the

day

firs

tC

OV

IDca

sew

as

rep

ort

edin

the

resp

ecti

ve

cou

ntr

yan

den

ds

on

Dec

emb

er31,

2020

an

dp

re-C

OV

IDp

erio

dst

art

sfr

om

Janu

ary

01,

2001

an

den

ds

on

ed

ay

bef

ore

the

firs

tC

OV

IDca

sew

as

rep

ort

ed.

Fig

ure

sin

pare

nth

esis

are

t-st

ati

stic

s.*

an

d**

ind

icate

sign

ifica

nce

at

5p

erce

nt

an

d1

per

cent,

resp

ecti

vel

y.

94 Business Review: (2021) 16(1):76-100




Table

A3

:U

nco

nd

itio

nal

her

din

ges

tim

ati

on

su

sin

gqu

anti

lere

gre

ssio

n

Pan

elA

:C

OV

IDp

erio

d

Ch

ina

Ind

iaIn

don

esia

Kore

aM

ala

ysi

aP

akis

tan

Ph

ilip

pin

esT

aiw

an

Th

ailan

dC

on

stant

0.0

15**

0.0

34**

0.0

30**

0.0

18**

0.0

35**

0.0

26**

0.0

28**

0.0

13**

0.0

23**

(37.2

98)

(96.3

54)

(41.7

57)

(29.7

05)

(35.7

41)

(35.8

45)

(5.4

98)

(28.9

38)

(41.6

3)

βCovid

10.1

58**

0.1

21**

0.3

21**

0.0

21

-0.1

47

0.2

31*

0.4

78

0.1

70.3

306**

(3.2

08)

(2.8

84)

(3.5

68)

(0.2

83)

(-1.0

42)

(2.4

75)

(0.6

01)

(1.9

44)

(5.6

97)

βCovid

21.0

41

2.4

39**

4.7

67**

4.9

12**

7.5

45**

-0.2

83

-9.0

87

3.9

26

0.0

61

(1.4

93)

(7.6

66)

(5.2

15)

(3.4

19)

(4.3

89)

(-0.1

89)

(-0.3

36)

(1.7

42)

(0.1

7)

Ad

just

edR

20.0

90.2

86

0.2

38

0.2

07

0.0

36

0.1

26

0.0

52

0.1

29

0.1

59

Pan

elB

:P

re-C

OV

IDp

erio

d

Ch

ina

Ind

iaIn

don

esia

Kore

aM

ala

ysi

aP

akis

tan

Ph

ilip

pin

esT

aiw

an

Th

ailan

dC

on

stant

0.0

21**

0.0

35**

0.0

30**

0.0

16**

0.0

24**

0.0

30**

0.0

30**

0.0

12**

0.0

21**

(40.6

69)

(209.9

9)

(99.1

95)

(138.2

9)

(166.3

45)

(114.2

59)

(29.9

95)

(98.2

56)

(162.0

85)

βPre−Covid

10.6

91**

0.1

44**

0.5

28**

0.2

90**

0.0

79*

0.2

36**

0.2

25

0.4

16**

0.0

27

(8.4

78)

(6.9

37)

(11.7

87)

(14.8

87)

(2.2

81)

(7.6

78)

(0.9

72)

(15.8

82)

(1.3

51)

βPre−Covid

2-5

.285**

1.8

62**

0.5

72

1.5

00**

7.3

14**

0.1

11

2.0

49

1.6

46

7.4

34**

(-4.8

72)

(5.8

17)

(0.8

85)

(3.0

51)

(11.6

1)

(0.2

52)

(0.2

4)

(1.9

27)

(18.3

06)

Ad

just

edR

20.1

62

0.0

19

0.0

60.2

09

0.0

29

0.0

28

0.0

14

0.2

19

0.0

94

Tes

tS

tat

(βCovid

1=βPreCovid

1)

-0.4

91

-2.0

55

-0.9

25

-1.3

54

-0.0

18

0.4

79

-0.5

48

0.9

15

Tes

tS

tat

(βCovid

2=βPreCovid

2)

4.9

06**

1.2

77

3.7

48**

2.2

47*

0.1

27

-0.2

53

-0.3

93

0.9

46

-13.6

5**

Note

:T

he

tab

lere

port

sth

ees

tim

ate

dco

effici

ents

for

Equati

on

2.

Est

imati

on

sare

don

eu

sin

gqu

anti

lere

gre

ssio

ns.

CO

VID

per

iod

start

sfr

om

the

day

firs

tC

OV

IDca

sew

as

rep

ort

edin

the

resp

ecti

ve

cou

ntr

yan

den

ds

on

Dec

emb

er31,

2020

an

dp

re-C

OV

IDp

erio

dst

art

sfr

om

Janu

ary

01,

2001

an

den

ds

on

ed

ay

bef

ore

the

firs

tC

OV

IDca

sew

as

rep

ort

ed.

Fig

ure

sin

pare

nth

esis

are

t-st

ati

stic

s.*

an

d**

ind

icate

sign

ifica

nce

at

5p

erce

nt

an

d1

per

cent,

resp

ecti

vel

y.

Business Review: (2021) 16(1):76-100 95



S. Tauseef

Table

A4

:Con

dit

ion

al

her

din

ges

tim

ati

on

sover

pre

-Covid

per

iod

(usi

ng

5p

erce

nt

crit

eria

)

Pan

elA

:C

on

dit

ion

edon

Extr

eme

Pri

ceM

ovem

ents

Ch

ina

Ind

iaIn

don

esia

Kore

aM

ala

ysi

aP

akis

tan

Ph

ilip

pin

esT

aiw

an

Th

ailan

dC

on

stant

0.0

15**

0.0

44**

0.0

38**

0.0

19**

0.0

26**

0.0

33**

0.0

36**

0.0

16**

0.0

23**

(73.0

03)

(51.8

46)

(65.0

73)

(116.2

33)

(120.9

07)

(85.0

25)

(57.7

95)

(78.5

05)

(101.0

91)

βPre−Covid

10.2

17**

0.1

94**

0.2

67

0.3

07**

0.3

03**

0.3

30**

0.0

57

0.5

10.0

97*

(6.7

88)

(3.2

11)

(1.5

81)

(5.5

83)

(3.7

18)

(4.5

38)

(0.9

25)

(1.7

52)

(1.9

45)

βPre−Covid

20.4

96**

-0.0

14

0.0

80.3

43**

0.2

08**

0.1

82

0.0

60.5

79**

0.2

25**

(10.9

49)

(-0.1

84)

(1.2

17)

(7.9

98)

(3.1

31)

(1.6

69)

(1.0

43)

(1.9

53)

(4.6

49)

βPre−Covid

3-2

.304**

0.8

64*

10.5

01

-0.2

49

3.7

4-1

.484

6.6

31**

1.7

25

6.9

48**

(-4.0

61)

(1.2

06)

(2.0

44)

(-0.1

90)

(1.8

81)

(-0.9

16)

(6.3

58)

(0.3

86)

(5.6

21)

βPre−Covid

4-3

.465**

2.3

39*

3.8

60*

-0.7

78

4.3

73*

-0.0

57

2.6

54**

-1.8

78

2.4

80*

(-5.0

73)

(3.5

11)

(3.0

54)

(-0.9

68)

(2.2

88)

(-0.0

50)

(3.3

06)

(-0.4

13)

(2.4

08)

Ad

just

edR

20.1

95

0.0

11

0.1

12

0.3

12

0.1

13

0.0

43

0.0

24

0.0

57

0.2

4T

est

Sta

t(β

1=β2)

-5.0

44**

2.1

34*

1.0

32

-0.5

16

0.9

01

-1.1

31

-0.4

6-0

.165

-1.8

33

Tes

tS

tat

(β3

=β4)

1.3

08

-1.5

08

1.2

55

0.3

46

-0.2

3-0

.72

3.0

22**

0.5

66

3.6

12

Pan

elB

:C

on

dit

ion

edon

Extr

eme

Mark

etL

iqu

idit

y

Ch

ina

Ind

iaIn

don

esia

Kore

aM

ala

ysi

aP

akis

tan

Ph

ilip

pin

esT

aiw

an

Th

ailan

dC

on

stant

0.0

13**

0.0

42**

0.0

35**

0.0

16**

0.0

25**

0.0

32**

0.0

24**

0.0

13**

0.0

21**

(58.8

34)

(45.5

73)

(52.5

66)

(82.3

98)

(90.2

81)

(63.9

26)

(72.3

57)

(58.6

1)

(63.3

48)

βPre−Covid

10.4

03**

0.2

36**

0.4

72**

0.3

71**

0.3

62**

0.3

01**

0.1

15**

0.3

81**

0.2

05**

(13.0

86)

(3.9

27

)(7

.941)

(12.8

84)

(4.4

01)

(5.1

53)

(2.8

58)

(11.9

52)

(3.6

84)

βPre−Covid

2-3

.153**

0.3

92**

1.1

41

-0.6

89

-2.5

13

-1.8

39

4.0

00**

3.1

33**

3.6

83*

(-5.5

28)

(0.8

05)

(0.8

31)

(-1.1

55)

(-0.8

27)

(-1.6

52)

(3.7

37)

(4.8

61)

(2.8

1)

βPre−Covid

32.9

94*

-1.5

43

-1.2

62.2

86*

8.9

28**

0.9

56

3.2

45

2.2

53*

3.1

21

(4.0

39)

(-1.0

82)

(-0.2

30)

(3.0

59)

(3.2

63)

(0.8

9)

(1.5

39)

(2.1

28)

(0.7

26)

βPre−Covid

4-1

5.1

66**

-96.2

42**

3.3

71

-10.7

35**

-24.3

39**

8.5

66

3.2

34

-18.8

52**

6.8

57*

(-4.1

71)

(-7.4

78)

(0.9

84)

(-5.2

60)

(-3.9

90)

(1.0

74)

(0.7

14)

(4.6

62)

(2.3

77)

Ad

just

edR

20.2

31

0.0

23

0.1

11

0.3

86

0.1

33

0.0

60.1

08

0.4

18

0.2

51

Tes

tS

tat

(β3

=β4)

-6.5

72**

7.3

13**

0.7

16

5.9

91**

4.9

76**

-0.9

46

0.0

02

5.0

49**

-0.7

21

Note

:T

he

tab

lere

port

sth

ees

tim

ate

dco

effici

ents

for

Equ

ati

on

s3

an

d4.

Extr

eme

mark

etre

turn

san

dex

trem

em

ark

etliqu

idit

yare

de-

fin

edas

five

per

cent

of

the

hig

hes

tan

dlo

wes

tm

ark

etre

turn

san

dtr

ad

ing

volu

me,

resp

ecti

vel

y.E

stim

ati

on

sin

volv

eN

ewey

-Wes

tco

nsi

sten

tes

tim

ato

rs.

Pre

-CO

VID

per

iod

start

sfr

om

Janu

ary

01,

2001

an

den

ds

on

ed

ay

bef

ore

the

firs

tC

OV

IDca

sew

as

rep

ort

ed.

Fig

ure

sin

pare

n-

thes

isare

t-st

ati

stic

s.*

an

d**

ind

icate

sign

ifica

nce

at

5p

erce

nt

an

d1

per

cent,

resp

ecti

vel

y.

96 Business Review: (2021) 16(1):76-100




Table

A5

:C

on

dit

ion

al

her

din

ges

tim

ati

ons

over

pre

-Covid

per

iod

(usi

ng

1p

erce

nt

crit

eria

)

Pan

elA

:C

on

dit

ion

edon

Extr

eme

Pri

ceM

ovem

ents

Ch

ina

Ind

iaIn

don

esia

Kore

aM

ala

ysi

aP

akis

tan

Ph

ilip

pin

esT

aiw

an

Th

ailan

dC

on

stant

0.0

16**

0.0

44**

0.0

39**

0.0

20**

0.0

26**

0.0

34**

0.0

36**

0.0

17

0.0

23**

(65.6

66)

(52.5

93)

(65.3

76)

(92.6

15)

(111.0

03)

(80.5

9)

(59.8

15)

(32.5

78)

(97.1

48)

βPre−Covid

10.3

89**

0.2

68**

0.1

2-0

.763**

0.5

78**

0.3

05**

0.1

32

0.1

09

0.3

89*

(4.9

75

(3.2

58)

(0.2

96)

(-7.1

09)

(3.8

96)

(1.9

95)

(1.6

34)

(0.9

19)

(2.1

91)

βPre−Covid

20.5

93**

-0.0

16

-0.0

63

-0.7

16**

0.4

92**

-0.0

52

0.1

02

0.1

31

0.3

37**

(8.1

79)

(-0.1

14)

(-0.5

42)

(-4.6

10)

(3.1

41)

(-0.2

52)

(0.9

75)

(1.2

85)

(4.8

65)

βPre−Covid

3-4

.844**

0.1

85**

13.0

26

42.0

29**

-2.3

89

-1.2

29

5.5

12**

7.0

35**

2.3

9(-

4.9

00)

(0.3

01)

(1.4

79)

(8.8

68)

(-0.8

17)

(-0.4

50)

(6.3

02)

(3.3

2)

(0.8

96)

βPre−Covid

4-4

.954**

2.2

65

5.4

01**

42.9

74**

-1.3

31

2.9

20*

2.1

28

4.4

31**

1.1

67

(-5.2

38)

(1.9

21)

(3.3

24)

(5.0

74)

(-0.5

07)

(1.0

06)

(1.7

83)

(2.5

91)

(1.2

45)

Ad

just

edR

20.0

68

0.0

08

0.0

80.0

05

0.0

96

0.0

13

0.0

20.0

20.2

Tes

tS

tat

(β1

=β2)

-1.9

12

1.7

28

1.0

32

0.2

46

0.3

99

1.3

95

3.3

82**

-0.1

39

-0.5

28

Tes

tS

tat

(β3

=β4)

0.0

15

-1.5

64

1.2

55

-0.0

97

-0.2

69

-1.0

41

2.2

87*

0.9

56

0.4

33

Pan

elB

:C

on

dit

ion

edon

Extr

eme

Mark

etL

iqu

idit

y

Ch

ina

Ind

iaIn

don

esia

Kore

aM

ala

ysi

aP

akis

tan

Ph

ilip

pin

esT

aiw

an

Th

ailan

dC

on

stant

0.0

13**

0.0

42**

0.0

35**

0.0

16**

0.0

25**

0.0

32**

0.0

24**

0.0

13**

0.0

21**

(61.6

57)

(45.7

61)

(52.7

73)

(77.3

97)

(106.5

68)

(66.7

39)

(71.8

37)

(58.6

96)

(63.1

46)

βPre−Covid

10.3

71**

0.2

35**

0.4

31**

0.3

76**

0.2

51**

0.2

66**

0.1

07**

0.3

77**

0.2

11**

(13.4

09)

(3.9

2)

(8.0

37)

(11.4

68)

(4.4

19)

(5.6

26)

(2.6

66)

(11.8

51)

(3.7

77)

βPre−Covid

2-2

.255**

0.1

78

2.0

25*

-0.6

17

4.0

05**

-0.2

35

4.4

19**

3.3

62**

3.6

58**

(-4.4

29)

(0.3

68)

(1.8

13)

(-0.9

47)

(3.0

97)

(-0.2

20)

(4.3

51)

(5.4

89)

(2.7

81)

βPre−Covid

36.5

22**

-3.4

99*

27.6

38

-1.3

76

11.0

29**

1.8

94*

82.9

72**

4.3

07

7.4

69**

(4.3

96)

(-2.0

75)

(8.3

15)

(-1.7

90)

(2.6

92)

(1.4

61)

(2.8

76)

(1.1

43)

(4.2

8)

βPre−Covid

4-2

4.1

32**

-154.1

45**

15.4

24

-23.2

65**

-19.3

59**

-1.6

87

16.6

61

-14.2

37*

32.0

74**

(-4.1

81)

(-6.2

23)

(2.0

26)

(-3.4

18)

(-2.7

61)

(-1.3

40)

(1.1

78)

(-1.9

70)

(6.4

18)

Ad

just

edR

20.2

25

0.0

18

0.1

14

0.3

79

0.1

21

0.0

50.1

25

0.4

15

0.2

53

Tes

tS

tat

(β3

=β4)

5.1

44**

6.0

81**

0.7

16

3.1

95**

3.7

42**

-1.9

82

2.0

64*

2.2

75**

-4.6

48

Note

:T

he

tab

lere

port

sth

ees

tim

ate

dco

effici

ents

for

Equ

ati

on

s3

an

d4.

.E

xtr

eme

mark

etre

turn

san

dex

trem

em

ark

etliqu

idit

yare

defi

ned

as

on

ep

erce

nt

of

the

hig

hes

tan

dlo

wes

tm

ark

etre

turn

san

dtr

ad

ing

volu

me,

resp

ecti

vel

y.E

stim

ati

on

sin

volv

eN

ewey

-Wes

tco

nsi

s-te

nt

esti

mato

rs.

Pre

-CO

VID

per

iod

start

sfr

om

Janu

ary

01,

2001

an

den

ds

on

ed

ay

bef

ore

the

firs

tC

OV

IDca

sew

as

rep

ort

ed.

Fig

ure

sin

pare

nth

esis

are

t-st

ati

stic

s.*

an

d**

ind

icate

sign

ifica

nce

at

5p

erce

nt

an

d1

per

cent,

resp

ecti

vel

y.

Business Review: (2021) 16(1):76-100 97



S. Tauseef

Table

A6

:Un

con

dit

ion

al

her

din

ges

tim

ati

on

sfo

r30-d

ay

CO

VID

sub

-per

iod

Ch

ina

Ind

iaIn

don

esia

Kore

aM

ala

ysi

aP

akis

tan

Ph

ilip

pin

esT

aiw

an

Th

ailan

dC

on

stant

0.0

12**

0.0

33**

0.0

58**

0.0

15**

0.0

36**

0.0

35**

0.0

24**

0.0

10**

0.0

22**

(29.1

19)

(55.1

)(9

.506)

(14.2

31)

(19.2

4)

(12.6

74)

(10.4

78)

(4.8

72)

(46.9

32)

βCovid

1-0

.06

0.2

24**

-0.6

02

0.0

14

-1.2

53**

0.1

44

0.6

05

0.5

14**

-0.1

2(-

0.2

66)

(2.6

75)

(-1.7

24)

(0.1

3)

(-2.7

23)

(0.8

23)

(1.5

19)

(2.4

7)

(-1.2

12)

βCovid

210.9

18

1.9

00**

10.4

09**

5.6

02

80.6

44**

-0.2

51

-12.7

43

-2.9

97

10.4

27**

(0.7

67)

(2.8

29)

(3.3

9)

(2.0

01)

(5.0

85)

(-0.0

80)

(-1.0

77)

(-1.2

13)

(2.7

58)

Ad

just

edR

20.0

11

0.7

11

0.2

70.5

45

0.6

78

0.0

34

0.1

12

0.3

95

0.1

45

Note

:T

he

tab

lere

port

sth

ees

tim

ate

dco

effici

ents

for

Equ

ati

on

2.

Est

imati

on

sin

volv

eN

ewey

-Wes

tco

nsi

sten

tes

tim

ato

rs.

CO

VID

sub

-per

iod

per

iod

con

sist

sof

the

init

ial

30

days

sin

ceth

ed

ay

of

firs

tC

OV

IDca

sefo

rre

spec

tive

cou

ntr

y.F

igu

res

inp

are

nth

esis

are

t-st

ati

stic

s.*

an

d**

ind

icate

sign

ifica

nce

at

5p

erce

nt

an

d1

per

cent,

resp

ecti

vel

y.

98 Business Review: (2021) 16(1):76-100




Table

A7

:H

erd

ing

esti

mati

on

sover

the

fin

an

cial

cris

isp

erio

d

Pan

elA

:U

nco

nd

itio

nal

her

din

g

Ch

ina

Ind

iaIn

don

esia

Kore

aM

ala

ysi

aP

akis

tan

Ph

ilip

pin

esT

aiw

an

Th

ailan

dC

on

stant

0.0

20**

0.0

36**

0.0

41**

0.0

19**

0.0

32**

0.0

30**

0.0

38**

0.0

19**

0.0

24**

(27.8

48)

(41.9

42)

(31.1

62)

(32.6

48)

(27.9

78)

(10.9

22)

(35.0

86)

(18.5

52)

(23.1

17)

βFin

cris

is1

0.2

74**

0.0

22

0.0

67

0.3

63**

0.2

38

0.4

77

0.1

13

0.2

74**

0.1

85

(4.8

45)

(0.3

68)

(0.5

83)

(5.0

98)

(1.7

16)

(1.4

77)

(0.9

69)

(2.6

37)

(1.7

47)

βFin

cris

is2

-2.6

51***

3.3

72**

6.2

50**

-0.8

34

2.7

72

-0.6

67

2.3

59

3.8

70**

5.8

45**

(-3.0

61)

(4.0

39)

(3.7

22)

(-0.8

34)

(1.2

56)

(-0.1

55)

(1.5

66)

(2.7

61)

(4.1

32)

Ad

just

edR

20.0

88

0.3

08

0.2

35

0.4

06

0.1

02

0.0

97

0.1

02

0.3

70.4

93

Pan

elB

:H

erd

ing

Est

imati

on

sC

on

dit

ion

edon

Extr

eme

Pri

ceM

ovem

ents

Ch

ina

Ind

iaIn

don

esia

Kore

aM

ala

ysi

aP

akis

tan

Ph

ilip

pin

esT

aiw

an

Th

ailan

dC

on

stant

0.0

23**

0.0

37**

0.0

42**

0.0

22**

0.0

34**

0.0

33**

0.0

39**

0.0

23**

0.0

26**

(44.4

65)

(64.3

55)

(32.7

58)

(44.3

49)

(33.2

26)

(23.8

56)

(48.4

77)

(31.1

89)

(31.5

84)

βFin

cris

is1

0.0

02

-0.2

99*

-0.0

82

0.3

59**

0.4

77

0.3

70*

0.2

39

0.2

19

0.2

26

(0.0

45)

(-2.0

99)

(-0.2

51)

(5.6

68)

(2.1

11)

(2.0

55)

(1.1

33)

(1.3

71)

(1.6

76)

βFin

cris

is2

0.3

36**

-0.0

99

-0.3

00**

0.2

95**

0.0

73

1.1

38

0.0

40.0

20.1

54

(4.0

16)

(-1.5

31)

(-2.4

34)

(4.8

94)

(0.5

97)

(1.8

01)

(0.3

54)

(0.1

8)

(1.1

12)

βFin

cris

is3

-0.8

65

12.5

03**

14.5

69

-1.9

79**

-3.2

85

-5.7

72

2.8

59

5.7

63**

5.6

05**

(-1.3

88)

(3.9

61)

(1.8

33)

(-2.3

89)

(-0.7

54)

(-1.8

34)

(0.6

68)

(2.6

79)

(3.2

16)

βFin

cris

is4

-2.9

85*

3.5

21**

7.9

52**

-0.1

27

5.6

85**

-11.4

51

2.4

65*

5.0

68*

5.3

65*

(-2.3

26)

(7.0

85)

(4.8

1)

(-0.1

64)

(3.1

26)

(-1.3

6)

(2.1

9)

(2.2

55)

(2.1

79)

Ad

just

edR

20.1

66

0.4

49

0.3

39

0.4

12

0.1

14

0.1

62

0.1

39

0.3

62

0.4

97

Business Review: (2021) 16(1):76-100 99



S. Tauseef

Table

A7

:C

onti

nu

ed

Pan

elB

:H

erd

ing

Est

imati

on

sC

on

dit

ion

edon

Extr

eme

Mark

etL

iqu

idit

y

Ch

ina

India

Ind

on

esia

Kore

aM

ala

ysi

aP

akis

tan

Ph

ilip

pin

esT

aiw

an

Th

ailan

dC

on

stant

0.0

20**

0.0

36**

0.0

41**

0.0

19**

0.0

31**

0.0

31**

-0.0

20**

0.0

24**

(27.1

14)

(41.1

87)

(27.8

01)

(32.2

7)

(23.2

11)

(11.9

08)

-(1

7.9

48)

(21.1

85)

βFin

cris

is1

0.2

81**

0.0

65

0.0

83

0.3

77**

0.5

87*

0.3

19

-0.2

72**

0.2

50*

(4.5

69)

(0.8

4)

(0.6

39)

(5.3

5)

(2.1

56)

(1.1

89)

-(2

.497)

(2.0

16)

βFin

cris

is2

-2.7

12**

2.4

96

4.4

02*

-1.1

19

-14.2

58

0.0

93

-3.8

69**

4.5

87*

(-2.6

48)

(1.5

84)

(2.2

52)

(-1.2

61)

(-1.4

13)

(0.0

23)

-(2

.767)

(2.1

76)

βFin

cris

is3

-0.1

20.6

65

15.1

97**

0.2

96

12.5

46

-14.4

11*

-0.7

82

1.0

75

(-0.1

47)

(0.6

06)

(6.5

24)

(0.5

68)

(1.9

)(-

2.1

49)

-(0

.466)

(0.8

26)

βFin

cris

is4

-2.4

-10.5

89**

2.4

54

-8.2

69**

1.7

59

43.0

71

--0

.045

-14.4

59**

(-1.9

69)

(-4.8

36)

(0.7

98)

(-3.2

51)

(0.1

9)

(6.0

66)

-(-

0.0

13)

(-3.2

59)

Ad

just

edR

20.0

86

0.3

17

0.3

0.4

11

0.1

26

0.2

55

-0.3

66

0.5

03

Note

:T

he

tab

lere

port

sth

ees

tim

ate

dco

effici

ents

for

Equ

ati

on

2-4

over

the

fin

an

cial

cris

isp

erio

d.

We

con

sid

er2008-2

009

as

the

fin

an

cial

cris

isp

erio

d.

Req

uir

edd

ata

over

the

per

iod

isn

ot

availab

lefo

rP

hilip

pin

es.

Fig

ure

sin

pare

nth

esis

are

t-st

ati

stic

s.*

an

d**

ind

icate

sign

ifica

nce

at

5p

erce

nt

an

d1

per

cent,

resp

ecti

vel

y.

100 Business Review: (2021) 16(1):76-100



investment behavior during covid pandemic: an evidence

Documents