Received: 24 July 2020 Revised: 12 December 2020 Accepted: 20 December 2020

DOI: 10.1002/ijfe.2435

RESEARCH ARTICLE

Day-of-the-week effect and market liquidity: A comparative

study from emerging stock markets of Asia†

Badal Khan1 | Muhammad Aqil1 | Syed Hasnain Alam Kazmi1,2 |

Syed Imran Zaman2,3

1
Department of Management Sciences,
Shaheed Zulfikar Ali Bhutto Institute of
Abstract
Science and Technology, Karachi, In the second half of the 20th century, it was observed that weekday-of-
Pakistan Monday generate significantly lower returns than other days of the week. Lit-
2
School of Economics and Management,
erature has shown that the day effect exists on weekdays of the week, which is
Southwest Jiaotong University, Chengdu,
China different from this day Monday. This study investigates day-of-the-week effect
3
Department of Business Administration, in returns and volume of stock in emerging Asian markets by using ordinary
Jinnah University for Women, Karachi, least square regression (OLS), generalized autoregressive conditional
Pakistan
heteroscedasticity (GARCH) (1, 1) and Kruskal-Wallis test. The data are
Correspondence obtained from Thomson Reuters from July 2013 to March 2019 for the market
Syed Imran Zaman, Department of indices and volume. The OLS and GARCH (1, 1) are employed through Evi-
Business Administration, Jinnah
University for Women, Karachi, Pakistan. ews, while Kruskal-Wallis test is used through SPSS. The results show a signifi-
Email: [email protected] cant day-of-the-week effect in returns in China, South Korea, Taiwan,
Thailand, Indonesia and Pakistan. On the other hand, no significant day-of-
Funding information
National Natural Science Foundation of the-week effect in returns is revealed in India and Malaysia. Furthermore, sig-
China nificant differences in traded volume on different days are also observed in all
markets except for Malaysia, where Monday is the least traded day in most
markets. No evidence for co-occurrence of day-of-the-week effect in returns
and volume is observed. The findings affirm inconsistency in the day-of-the-
week effect.

KEYWORDS
day-of-the-week effect, efficient market hypothesis, emerging markets, market liquidity,
GARCH (1, 1)

1 | INTRODUCTION many more. One of such abnormality in returns is the

day-of-the-week effect, where returns on a given day vary
The efficient market hypothesis of Fama (1970) assumes from other days of the week. From an early inquiry,
that prices of security reveal all available information. Cross (1973) observes that opening of markets after week-
Thus, return on security shall be consistent. However, ends are fairly unattractive in term of returns. It means
empirical inquiries show several anomalous patterns in that the first day-of-the-week that is, Monday generates
stock prices and returns that is, equity premium, size least returns. The differential in returns on different days
effect, overreaction, discount on close-end funds and of a week prompt more inquiries, which observe varia-
tion in days with maximum and minimum returns on a
†
Note: All Authors contributed equally and share the first equal given day of a week. Thus, the term Day-of-the-week
authorship rights1,2,3, effect, weekend effect, Monday effect and Tuesday effect

Int J Fin Econ. 2021;1–18. wileyonlinelibrary.com/journal/ijfe © 2021 John Wiley & Sons, Ltd. 1
2 KHAN ET AL.

are coined to test the differences in stock returns on the are used to capture the day-of-the-week effect. Con-
given days (Berument & Kiymaz, 2001; Dobis & versely, stock liquidity is an important indicator for
Louvet, 1996; Solnik & Bousquet, 1990). understanding investor behaviour that is not widely con-
Recent literature indicates that the day-of-the-week sidered for the testing day-of-the-week effect. Stock
effect has diminished or disappeared. Some evidence liquidity is a condition where stock is traded easily with-
from developed market is established in this regard. out affecting its price. It is a positive feature that investors
Olson, Mossman, and Chou (2015) claim that such effects seek when making an investment decision. The theory
have been declining since 1970s and it cannot be traced that emphasizes that markets and instruments should be
from long-term data. Similarly, Trick (2018) finds a mod- liquid, as liquidity in the market creates harmony and
erate day-of-the-week effect in few sectors in the United peace, and help prices to remain stable (Amihud, Men-
States, but claims that the overall market does not exhibit delson, & Pedersen, 2006). The draining of liquidity from
the day-of-the-week effect. Additionally, Abrahamsson markets has been associated with crises and chaos. In
and Creutz (2018) show that the OMXS30 index shows almost each stock market crisis (eg. 1987, 1998 and 2008)
no day of the effect. Furthermore, Rossi and flight of liquidity from the market has been observed.
Gunardi (2018) find no weekend effect in developed Building from the above premises that this study investi-
European markets that is, France, Germany, Spain and gates the day-of-the-week effect in Asian Emerging Mar-
Italy. However, Miss, Charifzadeh, and Herberger (2020) kets as recommended by (Akbalik & Ozkan, 2017;
show persistent Monday effect in Germany. And Rossi & Gunardi, 2018). Kim and Nofsinger (2008)
Birru (2018) finds a day-of-the-week effect on speculative emphasize that the uniqueness and size of Asian mar-
stocks but no such effects are revealed from non- kets, the cognitive differences of Asian cultures with the
speculative stocks. He also claims that the day-of-the- rest of the world make it important to inquire about
week effect does not exist after 1975. Similarly, Chiah financial behaviour in Asian markets. These emerging
and Zhong (2019) show that investment in speculative markets have similar political dynamics. Therefore,
stocks surge on Friday, which results in day of the week 6 years of data from elven emerging exchanges are used
effect in the markets. in the analysis. Considering the sensitivity of the day-of-
Interestingly, the day-of-the-week is more prominent the-week effect, three approaches that is, ordinary least
in frontier and emerging markets. Gbeda and square (OLS), generalized autoregressive conditional
Peprah (2018) explore the day-of-the-week effect in Nai- heteroscedasticity (GARCH) (1, 1) and Kruskal-Wallis
robi Stock Exchange, Adaramola and Adekanmbi (2020) test are used for analysis.
find the day-of-the-week effect in Nigeria. Akbalik and The findings reveal the day-of-the-week effect in
Ozkan (2017) find the day-of-the-week effect in Indone- return in China, South Korea, Taiwan, Thailand, Indone-
sia. Furthermore, Paital and Panda (2018) approve the sia and Pakistan, while no day-of-the-week effect in
existence of the day-of-the-week effect in India that is in return is observed in India and Malaysia. Moreover, the
contrast with the findings of Akbalik and Ozkan (2017). evidence for the day-of-the-week effect in returns is fairly
Likewise, Chancharat, Maporn, Phuensane, and week in the Philippines. Interestingly, the highest and
Chancharat (2018) discover anomalies in market returns the lowest returns observed in different markets vary. In
in Thailand. While, Islam and Sultana (2015) observe the Malaysia, Thailand, Indonesia and Pakistan, Mondays
day-of-the-week effect in Bangladesh. Recently, Singh, generate the lowest returns. In India and Philippines,
Bhattacharjee, and Kumar (2020) discover higher returns Tuesdays produce the lowest average returns. Thursday
in emerging markets of China, India and Brazil at the is the least profitable day in China, while in South Korea
turn of month that is significantly different than other and Taiwan Friday produce the least daily returns on
days of months. average. Contrary to the basic Monday effect, in China,
The theory associates days of the week effect with India, Taiwan and Philippines, Monday generates the
market inefficiency (Islam & Sultana, 2015; Birru, 2018; highest average daily returns. Therefore, the results do
Abrahamsson & Creutz, 2018; Paital & Panda, 2018; show the day-of-the-week anomaly that is different from
Kazmi, Wahab, Zaman, & Kou, 2018; Miss, Charifzadeh, the classical weekend effect where Mondays are least
and Herberger, 2020). Thus, the existence of day-of-the- profitable. The results are identical to the findings of the
week effect rejects the efficient market hypothesis. recent literature that show variation in the day-of-the-
Although some evidence from developed markets sug- week effect.
gests that it has disappeared. However, few studies affirm The significant differences in traded volume across
it exists specifically in less established markets. Interest- days are also observed in China, India, South Korea,
ingly day-of-the-week effect is not consistent across mar- Thailand, Indonesia, Philippines and Pakistan where
kets and time. Furthermore, returns from an exchange Monday is the least traded day in each market. However,
KHAN ET AL. 3

day-of-the-week effect in volume is not discovered in 2.1 | Conceptualization of the model

Malaysia only. Results for the day-of-the-week effect in
the volume are more consistent across markets than the The efficient market hypothesis explains that the previ-
results for day-of-the-week effect in returns. Subsequent ous information does not allow the investors to gain
paper follows with literature review in Section 2, Section 3 abnormal returns from the market, hence the market
discusses methods and data, results are detailed in Sec- exhibits weak form of efficiency. So the existence of day-
tion 4, and at last, the discussion and concluding remarks of-the-week effect that is examined using historical infor-
are presented in Section 5. mation indicate weak form of efficiency in a market
(Adaramola & Adekanmbi, 2020; Ahmed, Khan, Paul, &
Kazmi, 2018; Islam & Sultana, 2015; Rossi &
2 | LITERATURE R EVIEW Gunardi, 2018). The theorists that make conclusion about
market efficiency based on day-of-the-week effect check
This section highlights the related literature. Pioneer of if the results are consistent across different sample
the day-of-the-week effect (Cross, 1973) explores negative periods or not. So market is either efficient or inefficient.
returns on Monday at Standard and Poor index. A lot of Meanwhile the adoptive market hypothesis assumes that
researches on the day-of-the-week effect finds mixed market consists of participants with different capabilities
results. The market lacks consensus that causes it. Dobis and capacities and there are certain participants with
and Louvet (1996) consider market and exchange charac- superior information available for processing and deci-
teristics to be responsible for the day-of-the-week effect sion making. They are often followed by other investors
but, they are unable to explain this phenomenon for the in the market. So at times, it is not possible for the mar-
US and UK markets. Gayaker, Yalcin, and Ber- ket to be efficient forever, hence market efficiency varies
ument (2020) and Adaramola and Adekanmbi (2020) asso- over time (Urquhart & McGroarty, 2016). According to
ciate weekend effect with overnight interest rates. AMH, the day-of-the-week effect should not be consistent
Some studies focus on efficiency, for instance, Le over different sampling periods. Hence, day-of-the-week
Tran and Leirvik (2019) develop a model to measure mar- effect may change to different days of the week over time.
ket efficiency—adjusted market inefficiency magnitude This study explains day-of-the-week effect in different
(AMIM)—argue that market efficiency varies from mar- markets over a single sampling period. As a result, if the
ket to market and time to time, conclude that it supports results match with the findings of literature, it supports
the adoptive market efficiency theory. Likewise, the EMH, whereas any variation favours the AMH.
Ardalan (2018) and Jalees, Kazmi, and Zaman (2016) pro- Day-of-the-week effect in returns of a market can be
pose an analytical tool to predict the association of ratio- studied through obtaining the market returns. Market
nal decisions and market efficiency. It claims that returns are obtained through the indices by calculating
thinking is costly if properly used in market decision the percentage change in the value of the index from pre-
making the cost of thinking produces the best possible vious day as computed (Boubaker, Essaddam, Nguyen, &
decisions to reflect in prices. But more often, we lazily Saadi, 2017; Solnik & Bousquet, 1990). Moreover market
make decisions by consuming less time than required for liquidity has multiple measures for example, traded vol-
thinking, thus it creates market inefficiency. This theory ume, turnover ratio, trading value, illiquidity a high low
of physiological effect has often been associated with the ratio, sentiment index, illiquidity ratio of Amiud (ILR)
day-of-the-week effect. At a given time, a market consists and relative spread ratio (Apergis, Artikis, & Kyriazis,
of participants with varying strengths and decisional 2015). However, the traded volume has been the most
capacities. More powerful decisions push the market as common factor used as proxy for market liquidity and in
per their power of superior decisions and other partici- each cases, when it is tested for day-of-the-week effect
pants follow them. Hence, markets operate in a particular (Abbas & Javid, 2015; Huang, Hu, & Liao, 2010). To assess
inefficient manner; this effect is called the adoptive market the day-of-the-week effect, variant techniques of auto-
hypothesis (AMH). It argues that the market returns fol- regressive conditional heteroscedasticity (ARCH) and
low an adaptive pattern. Urquhart and McGroarty (2016) GARCH e.g., ARCH, GARCH (1, 1), E-GARCH, TGARCH
elaborate the adaptive market hypothesis (AMH) that the (Gbeda & Peprah, 2018; Richards & Willows, 2019;
market efficiency is not a condition to either exist or not to Trick, 2018) OLS (Birru, 2018; Zhang, Gu, & Zhou, 2019),
exist, it is a condition that evolves with time according to Analysis of variance (ANOVA) Hussain, Hamid, Akash,
market characteristics. It is just like the theory of evolution Shahid, and Imdad Khan (2011) and some nonparametric
where fittest survives. Hence, in a given time the decisions tests were employed. Whereas most common techniques
made by strong participants are followed, allowing predict- were different types of GARCH from family of GARCH,
ability of returns. OLS and Kruskal-Wallis rank test. Chen, Kwok, and
4 KHAN ET AL.

Rui (2001) claimed that day-of-the-week effect is effect is not forced by the volatility of returns. Hence,
influenced by econometric tools while Boubaker they concluded that the markets exhibit inefficiency. Li,
et al. (2017) argue that the GARCH(1,1) is the most effi- Zhang, and Zheng (2018) found evidence of seasonality
cient techniques of GARCH family, hence with GARCH more in advanced markets than emerging markets, and
(1,1), OLS and Krskal-wallis test are employed that are dis- associate this effect with regression effect, January effect
cussed in detail in Section 3.2. and Carhart type risk premium. Aggarwal and
Figure 1 shows the framework of the study through a Rivoli (1989) work on the market anomalies of emerging
thorough review of literature the hypotheses and markets of Hong Kong, Singapore, Malaysia and the Phil-
research designs are S. constructed where the economet- ippines, find a strong day-of-the-week effect with very
ric models for OLS, GARCH and Kurskal Wallis test are low returns on Monday while very high returns on
designed. Afterwards the data collected from Thomson Tuesday. The effect is assumed to be associated with the
Reuters are processed through Eviews and SPS. market's operational hours where US markets open
almost 13 hours from these markets. Berument and
Dogan (2012) reveal differences in returns and volatility
2.2 | Day-of-the-week effect in emerging on different days in the United States. Ţilica and
markets Oprea (2014) examined Romanian markets and find the
day-of-the-week effect, after incorporating international
Emerging markets are the markets with strong growth market risk, it is concluded that the day-of-the-week
and growth potential, technological development and effect has been transferred from international markets.
economic philosophy. Meanwhile, the emerging markets Specifically the results from emerging markets of Asia
have certain issues with political uncertainties, currency also support the existence of the day-of-the-week effect.
fluctuation and debt (Lhabitant, 2007). The bulk of litera- Kambal, Li, Letavina, and Mohamed (2018) investigate
ture is available to show the day-of-the-week effect in the day-of-the-week effect in China and find that market
emerging markets. Seif, Docherty, and Shamsuddin (2017) generates higher returns on Friday while the lowest
from an inquiry of Brazil, the Czech Republic, Hungary, returns on Thursday. Chen et al. (2001) have shown a
Malaysia, Mexico, Poland, South Africa, Taiwan and Tur- day-of-the-week effect in China after 1995 where Tuesday
key, find seasonality in daily returns where daily, weekly yield the least returns. After incorporating volatility and
and monthly effects are observed. It is claimed that the risk from the rest of the world, they conclude that the

F I G U R E 1 Framework of the
research.
Source: Self-Constructed
KHAN ET AL. 5

day-of-the-week effect is subject to spillover from the rest the week. Results are consistent with different models.
of the world. Lately, Ali and Ülkü (2020) show that in Therefore, they recommend a mixed model for future
Chinese market early days of the week yield higher studies. It is worthy that the effect shall be tested on a
returns, indicating contrary Monday effects. Aziz and larger sample of similar characteristics. Hence, using
Ansari (2015) study both Sensex and Nifty indexes of highly recommended GARCH and Kruskal-Wallis test,
India and find a day-of-the-week effect where Sensex following hypothesis is tested to approve the day-of-the-
yields abnormal positive returns on Monday while Nifty week phenomenon in the emerging markets of Asia.
generated abnormal positive returns on Wednesday. Hla,
Gunathilaka, and Isa (2015) find differences in market Proposition 1. The returns vary on different days of the
returns in four Malaysian indexes, results are consistent week in emerging markets of Asia.
across markets and models. Both parametric and non-
parametric tests show that Friday is the most profitable
day while Mondays yield significantly lower returns. Sim- 2.3 | Market liquidity
ilarly, Muhammad and Rahman (2010) find a day-of-the-
week effect in Malaysia with positive significant returns Market liquidity is associated with market efficiency, and
on Friday and negative returns on Monday. Furthermore, efficiency does not allow markets to behave abnormally.
Abdullah, Baharuddin, Shamsudin, Mahmood, and Chordia, Roll, and Subrahmanyam (2008) argue that
Sahudin (2011) find that Monday generates significantly liquidity improves market efficiency mainly through ava-
lower returns than other days while Friday generates iling the arbitrage opportunity. Few studies from
higher returns. Lim and Chia (2010) use non-parametric (Birru, 2018; Huang et al., 2010) have tried to find the
tests to find the day-of-the-week effect in Indonesia, association of DOW and market liquidity but find no or
Malaysia, Philippines, Singapore and Thailand where vague evidence.
different results are obtained for different tests. Market Various studies approve that market liquidity is
anomalies are observed in all markets except for associated with returns, efficiency and investors' psy-
Singapore. Boonkrong and Arjrith (2018) find the day- che that may result in a day-of-the-week effect. Li,
of-the-week effect in Thailand with Monday as the Zhang, and Li (2019) find that stock liquidity affects
least profitable day and Friday as the most profitable stock returns. Moreover, the flight to liquidity creates
day. Abbas and Javid (2015) investigate the day-of-the- psychological effects and herding. Chordia et al. (2008)
week effect and volume in SAARC. The results show a claim that liquidity creates market efficiency, where
day-of-the-week effect both in volume and returns in the existence of arbitrage is filled through orders and
Pakistan, India, Sri Lanka and Bangladesh. Mean- liquidity. Li et al. (2019) assume that liquidity affects
while, Shah and Abdullah (2015) using a similar returns positively. In the case of flight of liquidity and
approach on three subsamples of governing parties in stock market crisis, the liquid stocks experience less
Pakistan that is, The Pakistan Muslim League Quaid, fluctuation in prices. Furthermore, switching from less
Pakistan People's Party and Pakistan Muslim League liquid stocks to more liquid stocks effects the flight of
Nawaz, finds of day-of-the-week effect during the gov- liquidity. Indars, Savin, and Lublóy (2019) find that
ernment of Pakistan People's Party only. Lately, the concentrated buying behaviour depends on the
Anjum (2020) discover no end of the year effect and flow of information, uncertainty, liquidity and fluctua-
weekend effect in Pakistan, however, it reveals higher tion in oil prices. The market with high foreign traders
returns on Fridays than other days of the week, and and corporate mergers is associated with market
March and December yield higher returns than other liquidity. Hence, the market with greater capitalization
months of the year. and participants would be more efficient and liquid.
The techniques and models adopted to test the day- While studying the sentiment–liquidity relationship,
of-the-week effect are probed. Boubaker et al. (2017) Boubaker et al. (2017) and Liu, Chan, Alam Kazmi,
using various techniques Levene's test and GARCH claim and Fu (2015) reveal investor sentiments influence the
that day-of-the-week effect and volatility fluctuate with stock market liquidity. The effect is stronger for for-
choice of the model and period of the study, while during eign investors. Moreover, Huang et al. (2010) argued
the financial crisis the weekend effect declines but vola- stock liquidity influenced the returns, however,
tility in return surges. Lai, Bai, Chang, Wei, and Birru (2018) finds no association between the day-of-
Luo (2012) used non-parametric techniques to gauge the the-week effect and stock liquidity. The theory shows
day-of-the-week-effect in China and find a strong market liquidity plays a vital role in eliminating mar-
Thursday effect in the Chinese stock market; the returns ket inefficiency. Thus, the-day-of-the-week effect in
on Thursday are significantly lower than other days of liquidity is tested through the proposition;
6 KHAN ET AL.

Proposition 2. The traded volume differs on different 3 | M E T H O D O L O G Y AN D D A T A

days of the week in the emerging markets of Asia.
To assess the day-of-the-week effect, various techniques of
autoregressive conditional heteroscedasticity (ARCH) and
2.4 | Market size and structure generalized autoregressive conditional heteroscedasticity
(GARCH) i.e., ARCH, GARCH (1, 1), E-GARCH,
The vast literature covering the day-of-the-week effect TGARCH have been used (Gbeda & Peprah, 2018;
shows that the developed markets are more efficient Trick, 2018). Some studies also incorporate OLS
than developing and frontier markets, day-of-the- (Birru, 2018; Zhang et al., 2019) and ANOVA (Hussain
week is more visible in emerging and frontier mar- et al., 2011). Few nonparametric tests are also used due to
kets. A market is classified as developed based on non-normality of time series data. The most common tech-
financial development, market structure and partici- nique is the Kruskal-Wallis rank test (Hla et al., 2015;
pation, and political stability (Lhabitant, 2007). MSCI Lim & Chia, 2010). Chen et al. (2001) claim that day-of-
classifies and weights markets based on operations the-week effect is influenced by econometric tools while
and size of corporations. Hence, countries with larger Boubaker et al. (2017) argue that the GARCH(1,1) is the
cap corporations and operations have higher weights. most efficient technique of GARCH family, hence GARCH
MSCI has allocated 33% weights to China, 13.02% to (1,1), OLS and Kruskal-Wallis tests are used to test day-of-
South Korea, 11.35% to Taiwan, 9.16% to India, 7.23% the-week effect. For that, the daily returns from an index
to Brazil, and only 27.56% to other 19 countries classi- are calculated using the standardized formulae
fied in emerging markets (MSCI, 2020). Since the
larger markets have more participants and better r t = lnSt − lnSt − 1 ð1Þ
structure, they may exhibit efficiency in operations.
Hence, the effect is tested through the following where rt shows the daily logarithmic returns generated
proposition. by an index, St presents the closing value of an index on a
day, and St − 1 shows the closing value of the index on
Proposition 3. The larger markets are more efficient than the previous day. Moreover, stock liquidity is measured
smaller markets. through the traded volume.

TABLE 1 Daily distribution of returns

Market N Average returns Monday Tuesday Wednesday Thursday Friday

All 14,887 0.0262 0.0170 0.0312 0.0457 0.0040 0.03238
China
SSE 180 1,369 0.0429 0.1668 0.0951 −0.0205 −0.1303 0.1110
Shenzhen 1,381 0.0428 0.1528 0.1968 0.0323 −0.1723 0.0073
India
Sensex 30 1,394 0.0497 0.0919 −0.0271 0.0530 0.0739 0.0575
NSE Nifty 1,367 0.0511 0.1007 −0.0339 0.0572 0.0664 0.0665
South Korea 1,387 0.0076 −0.0205 0.0987 0.0007 −0.0060 −0.0348
Taiwan 1,368 0.0194 0.0431 0.0389 0.0097 0.0260 −0.0212
Malaysia 1,387 −0.0064 −0.0552 0.0234 −0.0461 0.0284 0.0156
Thailand 1,387 0.0083 −0.0890 −0.0139 0.1063 −0.0255 0.0569
Indonesia 1,366 0.0248 −0.0577 −0.0184 0.1018 0.0866 0.0110
Philippine 1,084 0.015058 0.0476 −0.0811 0.0470 0.0183 0.0427
Pakistan 1,397 0.0363 −0.1768 0.0384 0.1609 0.0839 0.0755

Note: Table provides the distribution of returns for each day of the week from July 30, 2013 to March 29, 2019. The non-trading days are not considered in
analysis. The returns are obtained from changes in the value of index—([Index at time t—index at time t–1]/index at time t–1). Where, the dividends are not
considered. It contains 14,887 observations from 11 indexes, obtained from Thomson Reuters. It shows most profitable and least profitable days in the sampling
period.
KHAN ET AL. 7

TABLE 2 Descriptive and diagnostic statistics

Std. Skew- Jarqua- Levene's P- ADF

Market N Mean Dev. ness Kurtosis Bera statistics Df2 Value Test
China
SSE180 1,369 0.043 1.529 −0.829 9.9889 2,943 3.716 1,364 0.005 −35.430
[0.000] [0.000]
Shenzhen 1,381 0.043 1.717 −0.984 7.0751 1,178 6.631 1,376 0.000 −34.139
[0.000] [0.000]
India
Sensex 1,394 0.047 0.873 −0.423 6.0335 576.1 1.933 1,389 0.103 −34.211
[0.000] [0.000]
Nifty 1,367 0.051 0.893 −0.417 5.9979 551.5 1.827 1,362 0.121 −33.568
[0.000] [0.000]
South 1,384 0.008 0.797 −0.373 4.8771 235.8 0.602 1,382 0.661 −37.265
Korea [0.000] [0.000]
Taiwan 1,368 0.019 0.819 −0.835 8.9534 2,179 1.771 1,363 0.132 −36.977
[0.000] [0.000]
Malaysia 1,386 −0.006 0.568 −0.458 5.6058 440.9 0.878 1,382 0.476 −34.509
[0.000] [0.000]
Thailand 1,387 0.008 0.909 −0.303 6.6245 780.5 3.056 1,382 0.016 −35.927
[0.000] [0.000]
Indonesia 1,366 0.024 0.945 −0.445 6.4877 737.4 2.172 1,361 0.070 −34.86
[0.000] [0.000]
Philippine 1,048 0.015 1.079 0.096 12.3353 3,937 0.810 1,079 0.516 −32.078
[0.000] [0.000]
Pakistan 1,397 0.036 0.956 −0.347 5.6972 451.6 5.338 1,392 0.000 −31.611
[0.000] [0.000]

Note: The table displays the distribution of data. The averages, variation, skew-ness and kurtosis for returns (whole sample) in each market obtained through
Eviews. Furthermore, it demonstrates the Levene's Statistics for homogeneity, the Jarqua-Bera statistics and Augmented Ducky Filler test of stationarity and
trends in data. The information in [] show P values for the given coefficients.

The standard model to test for the day-of-the-week dependency, moreover, the traded volume is controlled
effect is presented in Equation 2 as suggested (Solnik & to check its effect on returns.
Bousquet, 1990).
X
5

r t = δ0 + δ1 D1 + δ2 D2 + δ3 D3 + δ4 D4 + εt ð2Þ r t = β0 + β1 r t − 1 + δ j Dj + ε t ð5Þ
j=1

X
5
r t = δ0 + δj Dj + εt ð3Þ The rt − 1 represents the returns of the previous
j=1
day. The literature presents that the seasonal anoma-
lies are better estimated with such models that cap-
where Friday is the base period while, D1, D2, D3 and ture volatility in time series data (Abrahamsson &
D4 are the dummies for Tuesday, Wednesday, Creutz, 2018; Chancharat et al., 2018; Rossi &
Thursday and Monday. The equation assumes that Gunardi, 2018; Trick, 2018). Boubaker et al. (2017)
returns and volume traded are the same for each day- argue that the day-of-the-week effect with the latest
of-the-week. Where data and better techniques has either vanished or
transferred to different days as suggested by previous
δ1 = δ2 = δ3 = δ4 = δ5 ð4Þ studies. Therefore, simple OLS and MANOVA do not
account for autocorrelation and heteroscedasticity, so
Following Boubaker et al. (2017), a single lag for daily results could be spurious. Similarly, Berument and
returns is added in the model to capture linear Kiymaz (2001) claim that the time series data are
8 KHAN ET AL.

TABLE 3 Ordinary least square estimates for returns

Market Constant Monday Tuesday Wednesday Thursday Friday

All 0.0040 0.1301 0.0271 0.0417 0.0284
(0.2083) (0.4738) (0.9962) (1.5347) (1.0367)
[0.8350] [0.6356] [0.3192] [0.1249] [0.2999]
China
SSE 180 −0.1303 0.2971** 0.2254* 0.1098 0.2413**
(−1.540) (2.1311) (1.858) (0.897) (2.1027)
[0.1237] [0.0333] [0.0633] [0.369] (0.0357)
Shenzhen −0.1723* 0.3251** 0.3691*** 0.2046 0.1596
(−1.7941) (2.2650) (2.7958) (1.4491) (1.4185)
[0.073] [0.0237] [0.0052] [0.1475] [0.1563]
India
Sensex 30 −0.0271 0.1189 0.0801 0.1010 0.0846
(−0.516) (1.5180) (1.2308) (1.3425) (1.1005)
NSE Nifty [0.6057] [0.1293] [0.2186] [0.1796] (0.2713)
−0.033926 0.1359* 0.0911 0.1003 0.1004
(−0.6310) (1.6815) (1.3709) (1.2972) (1.2674)
[0.5282] [0.0929] [0.1706] [0.1948] (0.2052]
South Korea 0.0987** −0.1191* −0.0979 −0.1047* −0.1334**
(2.3798) (−1.9120) (−1.5683) (−1.6861) (−2.083)
[0.0175] (0.0561) [0.1170] [0.0920] (0.0374)
Taiwan −0.0212 0.0644 0.0602 0.0310 0.0473
(−0.4898) (1.0073) (0.9624) (0.4857) (0.7030)
[0.6243] [0.4140] [0.3360] [0.6272] [0.4821]
Malaysia 0.0365 −0.0895* −0.0182 −0.0866* −0.0223
(1.0880) (−1.8584) (−0.3811) (−1.8110) (−0.4665)
[0.2599] [0.0504] [0.6814] [0.0598] [0.6243]
Thailand −0.0887 0.7479 0.1950** 0.0632 0.1455**
(−1.3445) (0.8949) (2.3457) (0.7196) (2.0104)
[0.1790] [0.3710] [0.0191] [0.4719] [0.0446]
Indonesia −0.0577 0.0393 0.1595* 0.1442 0.0687
(−0.8836) (0.4887) (1.8551) (1.6100) (0.8622)
[0.3771] [0.6251] [0.0638] [0.1076) [0.3887]
Philippine −0.0811 0.1287 0.1281 0.0995 0.1238
(−1.2434) (1.3816) (1.5282) (0.9966) (1.1618)
[0.2140] [0.1674] [0.1268] [0.3191] [0.2456]
Pakistan −0.1768*** 0.215282*** 0.3378*** 0.2607*** 0.2523***
(−2.6343) (2.4503) (3.9503) (3.1880) (3.7904)
[0.0085] [0.0144] [0.0000] [0.0015) [0.0002]

Note: The table presents day-of-the-week effect in each market from ordinary least square regressions on Eviews. The base group is selected on the basis of
either most traded or least traded day. That is left blank in row of each market. Each row presents coefficients, t-statistics in (), and the significance of the test
in [], where, *** shows level of sig at 1%, ** presents level of sig at 5% and * shows level of sig at 10%.

subject to variation across time that makes simple OLS heteroscedasticity (GARCH) is employed. A GARCH
inefficient to predict results. Recently, Richards and process is estimated
Willows (2019), Miss, Charifzadeh, and Herberger
(2020), Anjum, 2020, and Ali and Ülkü (2020) incorpo- ht = α + βht − 1 + wεt − 1 ð6Þ
rated GARCH and variants of GARCH to test day-of-the-
week effect in time series data. Thus, to account for heter- Here, ht is the conditionality volatility, βht − 1 shows
oscedasticity generalized autoregressive conditional the effects of shocks, and wεt − 1 indicates the response of
KHAN ET AL. 9

TABLE 4 Estimates from GARCH (1, 1)

Market Constant Monday Tuesday Wednesday Thursday Friday

China
SSE 180 −0.0070 0.1931** 0.1532* −0.0363 0.0352
(−0.1094) (2.0675) (1.8056) (−0.4587) (0.4144)
[0.9129] [0.0387] [0.0710] [0.6465] [0.6786]
Shenzhen −0.141* 0.2868** 0.3319*** 0.1413 0.1285
(−1.791) (2.4620) (2.6965) (1.3534) (1.1330)
[0.0731] [0.0138] [0.0070] [0.1759] [0.2572]
India
Sensex 30 0.0285 0.0979 0.0368 0.053 0.035
(0.6011) (1.4682) (0.5759) (0.809) (0.518)
[0.5478] [0.1420] [0.5646] [0.419] [0.604]
NSE Nifty 0.0279 0.1110 0.037 0.040688 0.0448
(0.5830) (1.6020) (0.5672) (0.6186) (0.6608)
[0.5599] [0.1092] [0.5706] [0.5361] [0.5087]
South Korea 0.1221*** −0.1304* −0.1234** −0.1169* −0.159**
(2.7544) (−1.9324) [0.0533] (−1.997) (−1.7764) (−2.374)
[0.0059] [0.0458] [0.0757] [0.0176]
Taiwan −0.0185 0.1266** 0.1169* 0.0366 0.0743
(−0.4407) (1.9948) (1.848) (0.6193) (1.1599)
[0.6594] [0.0461] [0.0645] [0.5357] [0.2461]
Malaysia −0.0071 0.0078 0.0449 −0.035034 0.0295
−0.2510 (0.2076) (1.1219) (−0.8558) (0.7761)
[0.8019] [0.8355] [0.2619] [0.3921] [0.4377]
Thailand 0.1192*** −0.154** −0.1020* −0.1278** −0.0623
(2.8437) (−2.5249) (−1.7145) (−2.2438) (−0.995)
[0.0045] [0.0116] [0.0864] [0.0248] [0.3197]
Indonesia 0.1179** −0.152** −0.1083* −0.0436 −0.0874
(2.5642) −2.2433 (−1.6974) (−0.6469) (−1.294)
[0.0103] [0.0249] [0.0896] [0.5177] [0.1957]
Philippine −0.0356 0.110073 0.065499 0.2083** 0.0222
(−0.6253) (1.2381) (0.8392) (2.4363) (0.2424)
[0.5317] [0.2157] [0.4013] [0.0148] [0.8085]
Pakistan −0.0274 0.1063 0.1931*** 0.192*** 0.137**
(−0.5119) (1.6217) (2.8116) (2.6879) (2.3006)
[0.6087] [0.1049] [0.0049] [0.0072] [0.0214]

Note: The table shows statistics for GARCH (1, 1), for mean equation only. In each row coefficients, t-statistics in () and level of significance in [] are presented.
Where *** shows level of sig at 1%, ** presents level of sig at 5% and * shows level of sig at 10%.

volatility to the shocks. To make it more specific to this (Akbalik & Ozkan, 2017; Hla et al., 2015). It is com-
study, the GARCH is designed by incorporating dummies puted as
for days of the week.
12 X k
R2i
X
5 KW = −3ðn + 1Þ ð8Þ
nðn + 1Þ i = 1 ni
ht = α + βht − 1 + wεt − 1 + δ j Dj ð7Þ
j=1

where KW stands for Kruskal-Wallis test, n shows the num-

Apart from GARCH, the bulk of literature also ber of total observations in the sample, ni represents the num-
approves the use of the non-parametric test to gauge the ber of observation in the i trading day, k is the number of
day-of-the-week effect in returns. The most common trading days, while Ri is the rank sum of the i trading days.
technique employed is Kruskal-Wallis test that is applica- The kruskal-Wallis test is designed to rank days on the basis
ble on the data violating the assumption of normality of higher number of returns on given day compared with
10 KHAN ET AL.

TABLE 5 Kruskal-Wallis test for returns

Market Monday Tuesday Wednesday Thursday Friday P-Value

All Markets 8,254.68 8,205.37 8,084.22 7,948.96 8,143.37 0.082
China
SSE 180 744.54*** 713.36 642.71**m 614.71***m,f 713.46** 0.000***
Shenzhen 748.58*** 751.32*** 662.54 621.05***m,t 673.65 0.000***
India
Sensex 30 727.72 661.86 687.66 702.57 708.49 0.382
NSE Nifty 717.68 649.14 673.02 680.85 700.54 0.305
South Korea 687.82 740.01 681.63 689.35 671.13 0.295
Taiwan 709.07 696.98 676.39 685.12 655.01 0.572
Malaysia 671.62 723.50 666.63 716.01 691.39 0.341
Thailand 658.02 687.46 736.11 681.20 704.71 0.212
Indonesia 650.21 666.49 729.23 729.23 667.15 0.118
Philippine 570.28 514.38 550.39 546.27 533.01 0.441
Pakistan 632.55** 690.02 737.52**m 720.05 715.08 0.021**

Note: The table provides estimates of Kruskal-Wallis Rank test for returns. It ranks days on the basis higher returns generated on each day. The subscripts m, t,
w, th and f show the significance of difference associated with related day. Where, m = Monday, t = Tuesday, w = Wednesday, th = Thursday and f = Friday.
While, *** shows level of sig at 1%, ** presents level of sig at 5% and * shows level of sig at 10%.

TABLE 6 Daily distribution of volume

Market N Total Monday Tuesday Wednesday Thursday Friday

China
SSE 180 1,370 4.881 4.900 4.883 4.881 4.869 4.871
India
Sensex 30 1,395 4.083 4.064 4.073 4.084 4.081 4.113
NSE Nifty 1,368 4.277 4.236 4.268 4.276 4.310 4.297
South Korea 1,388 4.888 4.855 4.880 4.901 4.912 4.894
Taiwan 1,369 4.311 4.292 4.304 4.321 4.321 4.316
Malaysia 1,388 4.7075 4.7077 4.7072 4.7075 4.7077 4.7075
Thailand 1,388 4.317 4.285 4.323 4.325 4.343 4.306
Indonesia 1,367 3.862 3.859 3.890 3.855 3.863 3.844
Philippine 1,085 5.040 4.980 5.030 5.069 5.063 5.054
Pakistan 1,398 5.031 4.984 5.032 5.048 5.050 5.037

Note: The table presents distribution of traded volume for each day of the week from July 30, 2013 to March 29, 2019. The logged value of traded volume for
each day is used in analysis. The non-trading days are not considered in analysis. It contains 13,516 observations from 10 indexes, obtained from Thomson
Reuters.

other days of the week (Lim & Chia, 2010; Wong, Hui, & 3.1 | Data
Chan, 1992). For rejecting day-of-the-week effect in return, it
should have to result in equal number of higher returns on Data are collected from Thomson Reuters from July 2013
each day of the week for the period tested. If it yields signifi- to March 2019 from the stock indexes of emerging stock
cant high number of returns on any given day compared with markets of Asia. The SSE 180 index, Shenzhen composite
other days of the week, it would affirm existence of day-of- index (SZSE), Sensex 30 index, Nifty 50, KOSPI 200 Index,
the-week effect. TAIEX, KLCI, SET 100 index, JSX and KSE 100 index.
KHAN ET AL. 11

TABLE 7 Descriptive and diagnostic statistics for volume

ADF
Std. Skew- Jarqua- Levene's P- ADF First
Market N Mean Dev. ness Kurtosis Bera Statistics Df2 Value Test Order
China
SSE 180 1,370 4.88 0.271 0.880 3.225 179.7 3.716 1,364 0.005 −3.367
[0.000] [0.0123]
India
Sensex 1,395 4.08 0.192 1.952 15.63 10,165 1.933 1,389 0.103 −34.211
[0.000] [0.000]
Nifty 1,368 4.27 0.149 0.296 4.756 195.96 1.827 1,362 0.121 −4.670
[0.000] [0.000]
South 1,388 4.88 0.112 0.417 3.023 40.44 0.602 1,382 0.661 −5.902
Korea [0.000] [0.000]
Taiwan 1,369 4.31 0.094 0.082 3.688 28.61 1.771 1,363 0.132 −8.285
[0.000] [0.000]
Malaysia 1,388 4.70 0.018 −0.081 2.480 17.13 0.878 1,382 0.476 −2.807 2.871
[0.000] [0.0574] [0.004]
Thailand 1,388 4.31 0.147 −0.293 6.6245 31.06 3.056 1,382 0.016 −9.815
[0.000] [0.000]
Indonesia 1,367 3.86 0.859 −0.357 1.233 206.7 2.172 1,361 0.070 −1.282 −12.204
[0.000] [0.6395] [0.000]
Philippine 1,085 5.04 0.223 1.914 27.42 27,635 0.810 1,079 0.516 −4.537
[0.000] [0.000]
Pakistan 1,398 5.03 0.200 −0.127 2.727 8.13 5.338 1,392 0.000 −7.034
[0.017] [0.000]

Note: The table displays the distribution of data. The averages, variation, skew-ness and kurtosis for traded volume in each market are presented. Furthermore,
it demonstrates the Jarqua-Bera statistics and Augmented Ducky Filler test of normality and trends in data. It also demonstrates first order of ADF test to
detect the rectification of trends.

For analysis, the differences are studied on individual Wednesday is the most profitable day with average daily
samples. returns 0.0457%.
The least and most profitable days vary country
wise. In China, Thursday is the least profitable days
4 | R E SUL T S A N D F I NDI N GS with negative average returns, while Monday is most
profitable. In India, Wednesday generates least
This section provides findings for the day-of-the-week in returns, returns on Wednesday are negative. Similar to
returns and liquidity. It also presents descriptive informa- China, Monday is the most profitable in the selected
tion and diagnostic checks. period. In South Korea, Friday generates the least
returns while Tuesday yields highest daily returns.
Moreover, Friday also yields the lowest daily returns in
4.1 | The day-of-the-week effect in Taiwan while the most profitable day is Monday. Least
Returns profitable days in Malaysia are Monday while
Thursday yields highest daily returns. Monday also
Table 1 presents the descriptive information for returns generates the least daily returns in Thailand while
generated on each day. The total 14,887 observations are Friday produces the highest average daily returns.
obtained from 11 indexes show that markets generate Similarly, Monday generates the least returns in Indo-
0.0262% daily returns on average. Thursday is the least nesia too but the highest returns in Indonesia are pro-
profitable day for the selected sample with daily average duced on Thursday. In Philippines, Tuesday is least
returns 0.0040% only, followed by 0.0170% on Monday, profitable while Monday produces highest returns. In
0.0312% on Tuesday and 0.0323% on Friday, furthermore, Pakistan, Monday yields lowest returns but highest
12 KHAN ET AL.

TABLE 8 Ordinary least square estimates for volume

Market Constant Monday Tuesday Wednesday Thursday Friday

China
SSE 180 4.9000*** −0.0161** −0.0189** −0.0302*** −0.0285***
(161.69) (−2.336) (−2.401) (−3.161) (−3.386)
[0.0000] [0.0196] [0.0165] [0.0016] [0.0007]
India
Sensex 30 4.1113*** −0.0491*** −0.0407*** −0.0294*** −0.0327***
(322.78) (−3.6969) (−3.3574) (−2.3815) (−2.4373)
[0.0000] [0.0002] [0.0008] [0.0174] [0.0149]
NSE Nifty 4.3105*** −0.0739*** −0.0429* −0.0336*** −0.0126
(400.23) (−8.6020) (−4.8934) (−3.8958) (−1.4159)
[0.0000] [0.0000] [0.0000] [0.0001] [0.1570]
South Korea 4.9122*** −0.0572*** −0.0321*** −0.0111** −0.0178***
(585.16) (−9.3869) (−5.5981) (−1.9884) (−3.2552)
[0.0000] [0.0000] [0.0000] [0.0470] [0.0012]
Taiwan 4.2922*** 0.0124*** 0.0292*** 0.0290*** 0.0247***
(595.76) (3.3069) (6.4499) (6.3327) (5.0198)
[0.0000] [0.0010] [0.0000] [0.0000] [0.0000]
Malaysia 4.7072*** 0.0005 0.0002 0.0005 0.0002
(617.90) (1.1909) (0.6017) (0.9626) (0.7430)
[0.0000] [2339] [0.5475] [0.3359] [0.4576]
Thailand 4.2858*** 0.0373*** 0.0394*** 0.0574*** 0.0206***
(370.06) (4.7059) (4.3944) (6.4089) (2.5926)
[0.0000] [0.0000] [0.0000] [0.0000] [0.0096]
Indonesia 3.8447 0.0146 0.0453 0.0103 0.0190
(16.365) (0.6699) (2.2218) (0.5880) (0.9757)
[0.0000] [0.5030] [0.0265] [0.5566] [0.3294]
Philippine 4.9808*** 0.0492*** 0.0891*** 0.0826*** 0.0735***
(231.33) (2.8024) (4.4998) (4.0429) (3.6265)
[0.0000] [0.0052] [0.0000] [0.0001] [0.0003]
Pakistan 4.9847*** 0.0478** 0.0642*** 0.0658*** 0.0528**
(308.08) (5.7332) (6.1732) (5.9547) (5.2423)
[0.0000] [0.0000] [0.0000] [0.0000] [0.0000]

Note: Table shows day-of-the-week effect in each market with OLS statistics for volume. The base group is selected on the basis of either most traded or least
traded day. Each row presents coefficients, t-statistics and the significance of the test, where, *** shows level of sig at 1%, ** presents level of sig at 5% and *
shows level of sig at 10%.

returns are observed on Wednesdays. Therefore, there from the first lag. The abnormality and auto-correlation
is no consensus on the least and most profitable days suggest that the results from OLS may be inefficient.
across markets. Results from OLS estimates are presented in Table 3.
Table 2 presents diagnostic checks for the data. The It shows very strong evidence for the day-of-the-week
Jarqua-Bera shows the highly abnormal returns for all effect in Pakistan, where Monday yields lowest returns
samples. Levenes's statistics approve homogeneity of than rest of the days of the week, while moderate evi-
returns for day of the week in India (both Sensex 30 and dence for the day-of-the-week effect in returns is
NSE Nifty), and in South Korea, Taiwan, Malaysia, Indo- observed in China, South Korea and Thailand. Moreover,
nesia and Philippines. Furthermore, non-homogeneity is week evidence at 10% level of significance is confirmed in
revealed in China (SSE 180 and Shenzhen composite India (NSE Nifty only), Malaysia and Indonesia. No evi-
Index), Thailand and Pakistan. The ADF test applied at dence for the day-of-the-week effect is seen in Taiwan,
level with only intercept reveals stationary in the data. Philippines and NSE Nifty (India). Day-of-the-week effect
Furthermore, the correlogram from residual statics reveal varies across markets. In China, both SSE 180 index and
that the squared roots of residual are auto-correlated Shenzhen composite index show Thursday yields lowest
KHAN ET AL. 13

TABLE 9 GARCH (1, 1) for Volume

Market Constant Monday Tuesday Wednesday Thursday Friday

China
SSE 180 4.88*** 0.02*** 0.009 0.006718 −0.00254
(95.033) (3.356) (1.422) (1.169) (−0.405)
[0.0000] [0.000] [0.1548] [0.2423] [0.6851]
India
Sensex 30 4.06*** 0.022** 0.0081 0.0207* 0.05***
(253.61) (2.041) (0.678) (1.776) (3.910)
[0.0000] [0.0412] [0.4977] [0.0757] [0.0001]
NSE Nifty 5.23*** 0.03*** 0.035*** 0.069*** 0.06***
(344.09) (4.960) (4.782) (9.979) (8.883)
[0.0000] [0.0000] [0.0000] [0.0000] [0.0000]
South Korea 4.85*** 0.02*** 0.045*** 0.053*** 0.03***
(449.40) (5.640) (8.467) (9.392) (6.825)
[0.0000] [0.0000] [0.0000] [0.0000] [0.0000]
Taiwan 5.91*** 0.02*** 0.035*** 0.037*** 0.03***
(5.623) (3.444) (5.295) (4.812) (3.719)
[0.0000] [0.0006] [0.0000] [0.0000] [0.0002]
Malaysia 5.71*** 7.08E-05 −9.78E-05 1.84E-05 −0.00011
(933.19) (0.629) (−0.71) (0.134) (−1.106)
[0.0000] [0.5287] [0.4783] [0.8932] [0.2684]
Thailand 6.27*** 0.04*** 0.048*** 0.062*** 0.02***
(513.21) (5.455) (5.934) (7.657) (3.617)
[0.0000] [0.0000] [0.0000] [0.0000] [0.0003]
Indonesia 9.53*** 0.03*** 0.044*** 0.040*** 0.021***
(670.07) (4.489) (5.335) (5.158) (2.761)
[0.0000] [0.0000] [0.0000] [0.0000] [0.0058]
Philippine 4.999*** 0.057*** 0.087*** 0.082*** 0.071***
(151.6) (3.520) (4.811) (3.981) (3.309)
[0.0000] [0.0004] [0.0000] [0.0001] [0.0009]
Pakistan 4.990*** 0.052*** 0.065*** 0.066*** 0.053***
(274.59) (6.6502) (7.1447) (6.7956) (5.6006)
[0.0000] [0.0000] [0.0000] [0.0000] [0.0000]

Note: The table displays statistics for GARCH (1, 1), for mean equation only. In each row coefficients, t-statistics and level of significance are presented. Where,
*** demonstrates level of sig at 1%, ** presents level of sig at 5% and * shows level of sig at 10%.

returns. The returns in SSE 180 index for Thursday are than Wednesday and Thursday, respectively. Most signifi-
lower than Monday at 5% level of significance and lower cant result for day-of-the-week effect is observed in
than Tuesday and Friday at 10% level of significance. Pakistan where Monday produces significantly lower
Similarly, Shenzhen yields significantly low returns on returns then rest of the days of the week at 1% level of
Thursday at 5% level of significance then Monday and significance.
Tuesday. In India, only NSE Nifty shows weak evidence Table 4 provides results for GARCH (1, 1) where
of lower returns on Tuesday at 10% level of significance returns are regressed with the constant in mean equation.
than Monday. Tuesday generates higher return at 5% and It provides the day-of-the-week effect in conditional
10% level of significance than Friday and Monday in mean. The results are moderately different from the
South Korea. The returns on Thursday are greater than results of OLS estimates. In China, SSE 180 index shows
Wednesday and Monday at 10% level of significance in similar results but Shenzhen show improvement in
Malaysia. Monday yields significantly lower returns in strength of the effect on Tuesday. The weak day-of-the-
Thailand than Wednesday and Friday at 5% and 10% week effect in India, prevalent in OLS has vanished. The
level of significance. Similarly, in Indonesia, Monday day-of-the-week effect in South Korea has spread to all
generates lower return at 5% and 10% level of significance days of the week with Tuesday yielding higher average
14 KHAN ET AL.

TABLE 10 Kruskal-Wallis test for volume

Market Monday Tuesday Wednesday Thursday Friday P-Value

China
SSE 180 709.84 684.66 688.21 676.29 669.27 1.615
[0.805]
India
Sensex 30 632.02*** 680.92 703.66 707.9 766.81***m 16.152
[0.003]
NSE Nifty 717.68 649.14 673.02 680.85 700.54 4.834
[0.305]
South Korea 580.04*** 663.03**th 738.95***m 771.17***m 718.58***m 38.973
[0.000]
Taiwan 614.81*** 660.86 729.89**m 714.02***m 702.62 0.034
[0.007]
Malaysia 698.03 687.39 694.95 698.38 693.67 0.136
0.998
Thailand 604.51*** 704.3**m 719.04***m 773.89***m 664.25 27.223
[0.000]
Indonesia 655.16 704.11 696.38 699.23 664.92 3.479
[0.481]
Philippine 440.33*** 537.34***m 589.84***m 581***m 560.47***m 30.932
[0.000]
Pakistan 612.11*** 704.12 734.33***m 736.37***m 710.85**m 17.856
[0.001]

Note: The table provides estimates from Kruskal-Wallis Rank test for volume. It ranks days on the basis higher volume traded in each day. The subscripts m, t,
w, th and f show the significance of difference associated with related day. Where, m = Monday, t = Tuesday, w = Wednesday, th = Thursday and f = Friday.
While, *** shows level of sig at 1%, ** presents level of sig at 5% and * shows level of sig at 10%.

returns than Wednesday and Friday at 5% level of signifi- Wednesday too. Moreover, the returns on Thursday are
cance, it also generates higher returns than Thursday and also significantly lower than Tuesday and Monday at 1%
Mondays at 10% level of significance. OLS shows no evi- level of significance in Shenzhen. The results for Pakistan
dence for the day-of-the-week effect in Taiwan, but reveal that the returns on Wednesday are significantly
GARCH (1, 1) shows Friday generates significantly lower higher than Monday at 5% level of significance.
returns than Monday and Tuesday at 5% and 10% level of
significance, respectively. Moreover, day-of-the-week
effect in returns is not visible in Malaysia. In Thailand, 4.2 | Day-of-the-week effect in volume
Tuesdays generate significant higher returns than
Thursday and Monday at 5% and 10% level of signifi- Table 6 presents descriptive statics for the volume traded
cance. In Philippines, Tuesday generates lower returns in each market on a given day. In SSE 180, the highest
than Thursday at 5% level of significance. The results for volume traded on average on Monday while the lowest
Pakistan are slightly weaker as it shows strong evidence average volume is traded on Thursday. The highest turn-
for the day-of-the-week effect with Monday generating over is also recorded on Monday in Malaysia, but in
the lowest returns then other days of the week except for Malaysia, the lowest average volume traded on Tuesday.
Tuesday. In Sensex 30, the highest average volume is traded on
Table 5 displays the results for Kruskal-Wallis test. Wednesday and the lowest volume traded on Tuesday.
The findings indicate existence of day-of-the-week effect Wednesday also recorded, the highest average volume in
for returns only in China and Pakistan. The order of Philippines, but in Philippines, the lowest average vol-
ranks shows the higher ranks are more profitable that ume traded on Monday. In Taiwan, Friday results as the
lower ranks. It shows that the average returns on highest traded day and Thursday shows the lowest traded
Monday and Friday are higher than Thursday in SSE average. The data for NSE Nifty, KOSPI, SET 100, JKSE
180 index, while the returns on Monday are higher than and KSE 100 index show that the highest average volume
KHAN ET AL. 15

is traded on Thursday while the lowest volume is traded the days at 1% except Tuesday. In Taiwan, significant evi-
on Monday. dence show low traded volume on Monday than other
Table 7 presents descriptive and diagnostic tests for days of the week at 5% level of significance. The lowest
traded volume. The Jarqua-Bera shows non-normally dis- traded volume is observed on Monday in Thailand, Phil-
tributed data for all markets. Furthermore, the ADF test ippines and Pakistan, which differs from other days of
shows stationary series for all market except for Malaysia the week at 1% level of significance.
and Indonesia, respectively, that becomes stationary after
first change.
Table 8 presents day-of-the-week effect in volume 5 | CONCLUSION AND
with OLS statistics. It shows significant day-of-the-week DISCUSSIONS
effect in volume in China, India, South Korea, Taiwan,
Thailand, Philippines and Pakistan. Chinese trade the The results show day-of-the-week effect in return in
most on Mondays that reduces afterwards. However, China, South Korea, Taiwan, Thailand, Indonesia and
Indian trades the least with openings of the exchange on Pakistan and no day-of-the-week effect in return is rev-
Monday, but trading improves afterwards. That is visible ealed in India and Malaysia. The evidence for day-of-the-
both in Sensex 30 and NSE Nifty. The highest traded day week effect in returns is fairly week in Philippines. The
is Friday in Sensex and it is Thursday in Nifty, with a sig- highest and lowest returns observed on different markets
nificant difference in trading volume. Similarly Monday vary. In Malaysia, Thailand, Indonesia and Pakistan the
is the least traded day in South Korea, while Thursday is lowest returns are evident on Monday. In India and Phil-
the highest traded day, the difference is significant at 1% ippines, Tuesday yields the lowest average returns.
level of significance. In Taiwan, investors trade the least Thursday is least profitable in China, while in South
on Tuesday, compared with other days of the week at 1% Korea and Taiwan, Friday yields the least daily returns
level of significance. Finally, in Thailand, Philippines and on average. Interestingly in China, India, Taiwan and
Pakistan, the least trading takes place on Monday, that is Philippines Monday generates highest average daily
significant at 1% level of significant across days. However, returns. Therefore, the results do show day-of-the-week
the results do not show day-of-the-week effect in volume anomaly that is different from classical weekend effect
in Malaysia and Indonesia. Thus, the results from OLS where Mondays are the least profitable days. The results
approve the existence of day-of-the-week effect in vol- are identical to the findings of recent literature that
ume. It is very interesting that in almost all markets— shows variation in day-of-the-week effect.
except for China, where investor trade most on The significant differences in traded volume across
Mondays—the lowest volume is traded on Monday. days are also observed in China, India, South Korea, Tai-
Table 9 shows the results from GARCH (1, 1) for differ- wan, Thailand, Indonesia, Philippines and Pakistan
ences in average traded volume. The results affirm the esti- where Monday is the least traded day in each market—
mates of OLS. Both the techniques show strong evidence of except for China and NSE Nifty (India). In a recent
differences in trading volume at 1% level of significance in enquiry, Richards and Willows (2019) present that
China, India, South Korea, Taiwan, Thailand, Indonesia, Mondays in China are the least traded day in the week.
Philippines and Pakistan. The only market without day-of- Moreover, the day-of-the-week effect in volume is not
the-week effect in volume is Malaysia. Indonesia shows no discovered only in Malaysia. Malaysia is the only market
differences in trading volume in OLS but it shows day-of-the- with no day-of-the-week effect, both in returns and vol-
week effect in volume in GARCH (1, 1). ume. Results for day-of-the-week effect in volume are
The results of the Kruskal-Wallis test in Table 10 indi- more consistent across markets than the results for day-
cate existence of day-of-the-week effect in India (Sensex of-the-week effect in returns.
30 only), South Korea, Taiwan, Thailand, Philippines and Furthermore, the lowest and the highest average
Pakistan. It does not show day-of-the-week effect in returns in a day are exactly matched with lowest and
China, Malaysia and Indonesia. Monday is the day with highest average volume in China only. In India
lowest volume traded in all the markets where it shows (Sensex 30), Thailand, Indonesia and Pakistan, the
the evidence for the day-of-the-week effect. Only SSE lowest average return observed in a day is matched
180 and NSE Nifty are the markets where it shows with the lowest average volume. This could indicate
Monday with higher traded volume but it is not different that liquidity direct the returns but the evidence are
from the traded volume of other days. The results show visible only from 5/20 or 25% cases, hence it might
that the traded volume is higher in Friday than Monday have occurred due to sheer coincidence and it would
at 1% level of significance in Sensex 30. In South Korea, be flawed to conclude that the day-of-the-week effect
the traded volume is significantly lower than the rest of in return and traded volume co-occur.
16 KHAN ET AL.

Therefore, it is concluded that the emerging mar- DATA AVAILABILITY STATEMENT

kets of Asia that is, China, South Korea, Taiwan, Data is available on request.
Thailand, Indonesia and Pakistan exhibit day-of-the-
week effect in returns. Whereas, there is no day-of-the- ORCID
week anomaly in returns in India and Malaysia. More- Syed Hasnain Alam Kazmi https://orcid.org/0000-
over, day-of-the-week effect in volume is present in all 0002-5008-7365
emerging markets except for Malaysia. Since the pat- Syed Imran Zaman https://orcid.org/0000-0002-9107-
tern of day-of-the-week effect is slightly stronger in 1386
smaller markets, yet it shows very significant day-of-
the-week effect in China (Both Shenzhen and Shang- RE FER EN CES
hai), South Korea and Taiwan, hence it is concluded Abbas, S., & Javid, A. Y. (2015). The Day-of-the-Week Anomaly in
that the market size is not a determinant of day-of-the- Market Returns, Volume and Volatility in SAARC Countries (No.
week effect and efficiency. It may be explained by mar- 2015: 129), Pakistan: Pakistan Institute of Development Eco-
ket structure and characteristics as suggested by Ber- nomics. https://www.pide.org.pk/pideorgp/pdf/Working%
20Paper/WorkingPaper-129.pdf.
ument & Kiymaz, (2001).
Abdullah, R. N. J. R., Baharuddin, N., Shamsudin, N.,
The results show variation and sensitivity in the day-
Mahmood, W., & Sahudin, Z. (2011). The Day of the Week
of-the-week effect with econometric models. Similar Effect on Bursa (Bourse) Malaysia Shariah-Compliant Market.
results are obtained from multiple studies (Akbalik & Interdisciplinary Journal of Research in Business, 1(4), 29–36.
Ozkan, 2017; Huang et al., 2010; Hussain et al., 2011; Abrahamsson, A., & Creutz, S. (2018). Stock Market Anomalies: The
Kambal et al., 2018; Lai et al., 2012; Shah & Day-Of-The-Week-Effect: An empirical study on the Swedish
Abdullah, 2015). The variation in trading volume has also Stock Market: A GARCH Model Analysis, Sweden: (Master The-
been observed previously. Abbas and Javid (2015) find sis). Jönköping University, Jönköping International Business
School.
day-of-the-week effect in volume in SAARC countries
Adaramola, A. O., & Adekanmbi, K. O. (2020). Day-of-the-week
including Pakistan and India. Huang et al. (2010), find effect in Nigerian stock exchange: adaptive market hypothesis
day-of-the-week effect in volume in Taiwan. The day-of- approach. Investment Management & Financial Innovations, 17
the-week effect in returns does not show any pattern (1), 97–108.
market wise, but in volume, it is clearly visible that Aggarwal, R., & Rivoli, P. (1989). Seasonal and Day-of-the-Week
Mondays are least traded days in the emerging Asian Effects in Four Emerging Stock Markets. Financial Review, 24
markets. (4), 541–550.
Ahmed, S., Khan, A., Paul, S., & Kazmi, S. H. A. (2018). Role of
The evidence for the day-of-the-week effect in vol-
Green Information System and Information Cycle in Environ-
ume supports the theories that associate the day-of the
mental Performance. In International Conference on Manage-
week effect with psychological theories and the infor- ment Science and Engineering Management, 465–476. Cham:
mational absorption in the weekends. Moreover, the Springer. https://doi.org/10.1007/978-3-319-93351-1_37
significant evidence in very inefficient markets that is, Ali, F., & Ülkü, N. (2020). Weekday Seasonality of Stock Returns:
Pakistan where day-of-the-week effect in returns and The Contrary Case of China. Journal of Asian Economics, 68,
volume co-occur, hints that day-of-the-week effect in 101201. https://doi.org/10.1016/j.asieco.2020.101201
highly inefficient markets has particular investor senti- Amihud, Y., Mendelson, H., & Pedersen, L. H. (2006). Liquidity and
Asset Prices. Foundations and Trends® in Finance, 1(4), 269–364.
ments and information flow explained by theories but
Anjum, S. (2020). Impact of market anomalies on stock exchange: a
it is modified when markets evolve. Meanwhile the
comparative study of KSE and PSX. Future Business Journal, 6
day-of-the-week effect shows market inefficiency (1), 1–11.
hence it could be emerging through market mecha- Akbalik, M., & Ozkan, N. (2017). Day of the Week Effect in the
nism and informational flow. Moreover, the changing Stock Markets of Fragile Five Countries after 2008 Global
phenomenon of day-of-the-week affect indicates that Financial Crisis. In Global Financial Crisis and Its Ramifica-
the day-of-the-week effect is led by random informa- tions on Capital Markets (pp. 507–518). Cham: Springer.
tional flow where critical information is revealed on Apergis, N., Artikis, P. G., & Kyriazis, D. (2015). Does stock market
liquidity explain real economic activity? New evidence from two
specific days in certain periods. So studying the flow of
large European stock markets.. Journal of International Financial
information and its absorption with market structure Markets, Institutions and Money, 38, 42-64. https://www.
may reveal the causes of day-of-the-week effect in sciencedirect.com/science/article/abs/pii/S1042443115000554.
returns and volume. Ardalan, K. (2018). Neurofinance versus the efficient markets
hypothesis. Global Finance Journal, 35, 170–176.
A C K N O WL E D G E M E N T Aziz, T., & Ansari, V. A. (2015). The Day of the Week Effect: Evi-
Thank you to the reviewers to provide us the insightful dence from India. Afro-Asian Journal of Finance and Account-
ing, 5(2), 99–112.
feedback for this article.
KHAN ET AL. 17

Berument, M. H., & Dogan, N. (2012). Stock Market Return and Jalees, T., Kazmi, S. H. A., & Zaman, S. I. (2016). The Effect of
Volatility: Day-of-the-Week Effect. Journal of Economics and Visual Merchandising, Sensational Seeking and Collectivism on
Finance, 36(2), 282–302. Impulsive Buying Behavior. Journal of Systems Science and
Berument, H., & Kiymaz, H. (2001). The Day of the Week Effect on Information, 4(4), 321–333. https://doi.org/10.21078/JSSI-2016-
Stock Market Volatility. Journal of Economics and Finance, 25 321-13
(2), 181–193. Kambal, M. E. M., Li, H. Y., Letavina, A., & Mohamed, A. O. K.
Birru, J. (2018). Day of the Week and the Cross-Section of Returns. (2018). Does the Day of the Week form Still Persistent on the
Journal of Financial Economics, 130(1), 182–214. Chinese Stock Market? Asian Journal of Economics, Business
Boonkrong, P., & Arjrith, N. (2018). Impact of Weekdays on the and Accounting, 6(4), 1–10.
Return Rate of Stock Price Index: Evidence from the Stock Kazmi, S. H. A., Wahab, A., Zaman, S. I., Kou, Y., & (2018). Rise of
Exchange of Thailand. Journal of Finance and Accounting, 6(1), Digital Media to Triumph Brand Loyalty. In 2018 10th Interna-
35–41. tional Conference on Intelligent Human-Machine Systems and
Boubaker, S., Essaddam, N., Nguyen, D. K., & Saadi, S. (2017). On Cybernetics (IHMSC) (Vol. 1(1), pp. 297–300). China: IEEE.
the Robustness of Week-Day Effect to Error Distributional https://doi.org/10.1109/IHMSC.2018.00075. https://ieeexplore.
Assumption: International Evidence. Journal of International ieee.org/document/8530332/.
Financial Markets, Institutions and Money, 47, 114–130. Kim, K. A., & Nofsinger, J. R. (2008). Behavioral finance in Asia.
Chancharat, S., Maporn, S., Phuensane, P., & Chancharat, N. Pacific-Basin Finance Journal, 16(1–2), 1–7.
(2018). Volatility of Holiday Effects in Thai Stock Market. Lai, D., Bai, A., Chang, K. C., Wei, H., & Luo, L. (2012). Nonpara-
Kasetsart Journal of Social Sciences, 41(2), 401–406. metric Analysis of the Shenzhen Stock Market: The day of the
Chen, G., Kwok, C. C., & Rui, O. M. (2001). The Day-of-the-Week Week Effect. Mathematical and Computer Modelling, 55(3–4),
Regularity in the Stock Markets of China. Journal of Multina- 1186–1192.
tional Financial Management, 11(2), 139–163. Le Tran, V., & Leirvik, T. (2019). A simple but powerful measure of
Chiah, M., & Zhong, A. (2019). Day-of-the-week effect in market efficiency. Finance Research Letters, 29, 141–151.
anomaly returns: International evidence. Economics Let- Lhabitant, F. S. (2007). Handbook of hedge funds (Vol. 332),
ters, 182, 90–92. England: John Wiley and Sons.
Chordia, T., Roll, R., & Subrahmanyam, A. (2008). Liquidity and Lim, S. Y., & Chia, R. C. J. (2010). Stock Market Calendar Anoma-
Market Efficiency. Journal of Financial Economics, 87(2), lies: Evidence from ASEAN-5 Stock Markets. Economics Bulle-
249–268. tin, 30(2), 996–1005.
Cross, F. (1973). The behavior of stock prices on Fridays and Li, S., Zhang, T., & Li, Y. (2019). Flight-To-Liquidity: Evidence from
Mondays. Financial Analysts Journal, 29(6), 67–69. China's Stock Market. Emerging Markets Review, 38, 159–181.
Dobis, M., & Louvet, P. (1996). The Day-of-the-Week Effect: The Li, F., Zhang, H., & Zheng, D. (2018). Seasonality in the Cross
International Evidence. Journal of Banking and Finance, 20(9), Section of Stock Returns: Advanced Markets versus Emerging
1463–1484. Markets. Journal of Empirical Finance, 49, 263–281.
Fama, E. F. (1970). Efficient capital markets: A review of theory Liu, C., Chan, Y., Alam Kazmi, S. H., & Fu, H. (2015). Financial
and empirical work. The Journal of Finance, 25(2), 383–417. fraud detection model: Based on random forest. International
Gayaker, S., Yalcin, Y., & Berument, M. H. (2020). The day of the Journal of Economics and Finance, 7(7), 178–188. https://doi.
week effect and interest rates. Borsa Istanbul Review, 20(1), org/10.5539/ijef.v7n7p178
55–63. Miss, S., Charifzadeh, M., & Herberger, T. A. (2020). Revisiting the
Gbeda, J. M., & Peprah, J. A. (2018). Day-of-the-Week Effect and monday effect: a replication study for the German stock mar-
Stock Market Volatility in Ghana and Nairobi Stock Exchanges. ket. Management Review Quarterly, 70(2), 257–273.
Journal of Economics and Finance, 42(4), 727–745. MSCI Emerging Markets Index (2020). China: IEEE. Retrieved from
Hla, D. T., Gunathilaka, C., & Isa, A. H. M. (2015). Day-of-the- https://www.msci.com/emerging-markets
Week Effect: An Enduring Anomaly in Bursa Malaysia. IUP Muhammad, N. M. N., & Rahman, N. M. N. A. (2010). Efficient
Journal of Applied Finance, 21(4), 38. Market Hypothesis and Market Anomaly: Evidence from Day-
Huang, Z., Hu, O., & Liao, B. S. (2010). Short Sale, Stock Liquidity, of-the-Week Effect of Malaysian Exchange. International Jour-
and the Day-of-the-Week Effect: Evidence from the Taiwan nal of Economics and Finance, 2(2), 35–42.
Stock Market. Review of Pacific Basin Financial Markets and Olson, D., Mossman, C., & Chou, N. T. (2015). The Evolution of the
Policies, 13(01), 71–90. Weekend Effect in US Markets. The Quarterly Review of Eco-
Hussain, F., Hamid, K., Akash, I., Shahid, R., & Imdad Khan, M. nomics and Finance, 58, 56–63.
(2011). Day of the week Effect and Stock Returns: (Evidence Paital, R. R., & Panda, A. K. (2018). Day of the Week and Weekend
from Karachi stock exchange-Pakistan). Far East Journal of Effects in the Indian Stock Market. Theoretical Economics Let-
Psychology and Business, 3(1), 25–31. ters, 8(11), 2559–2588.
 (2019). Herding Behaviour in
Indars, E. R., Savin, A., & Lublóy, A. Richards, D. W., & Willows, G. D. (2019). Monday mornings: Indi-
an Emerging Market: Evidence from the Moscow Exchange. vidual investor trading on days of the week and times within a
Emerging Markets Review, 38, 468–487. day. Journal of Behavioral and Experimental Finance, 22,
Islam, R., & Sultana, N. (2015). Day of the Week Effect on Stock 105–115.
Return and Volatility: Evidence from Chittagong Stock Exchange. Rossi, M., & Gunardi, A. (2018). Efficient Market Hypothesis and
European Journal of Business and Management, 7(3), 165–172. Stock Market Anomalies: Empirical Evidence in Four
18 KHAN ET AL.

European Countries. Journal of Applied Business Research, 34 500 and Its Sectors”. (Masters Theses. 7786. USA: Missouri Uni-
(1), 183–192. versity of Science and Technology.
Seif, M., Docherty, P., & Shamsuddin, A. (2017). Seasonal Anoma- Urquhart, A., & McGroarty, F. (2016). Are Stock Markets Really
lies in Advanced Emerging Stock Markets. The Quarterly Efficient? Evidence of the Adaptive Market Hypothesis. Inter-
Review of Economics and Finance, 66, 169–181. national Review of Financial Analysis, 47, 39–49.
Shah, S. M. M., & Abdullah, F. (2015). A Study of Day of the Week Wong, K. A., Hui, T. H., & Chan, C. Y. (1992). Day-of-the-week
Effect in Karachi Stock Exchange during Different Political effects: evidence from developing stock markets. Applied Finan-
Regimes in Pakistan. Business and Economic Review, 7(1), 41–66. cial Economics, 2(1), 49–56.
Singh, G., Bhattacharjee, K., & Kumar, S. (2020). Turn-of-the- Zhang, T., Gu, G. F., & Zhou, W. X. (2019). Order Imbalances and
month effect in three major emerging countries. Managerial Market Efficiency: New Evidence from the Chinese Stock Mar-
Finance. https://doi.org/10.1108/MF-01-2020-0013 (ahead-of- ket. Emerging Markets Review, 38, 458–467. https://doi.org/10.
print), https://www.emerald.com/insight/content/doi/10.1108/ 1016/j.ememar.2018.12.003.
MF-01-2020-0013/full/html.
Solnik, B., & Bousquet, L. (1990). Day-of-the-week effect on the
Paris Bourse. Journal of Banking and Finance, 14(2–3),
How to cite this article: Khan B, Aqil M, Alam
461–468.
Kazmi SH, Zaman SI. Day-of-the-week effect and
Ţilica, E. V., & Oprea, D. (2014). Seasonality in the Romanian Stock
Market: the Day-of-the-Week Effect. Procedia Economics and market liquidity: A comparative study from
Finance, 15, 704–710. emerging stock markets of Asia. Int J Fin Econ.
Trick, M. A. (2018). An Investigation of the Influence of the 2021;1–18. https://doi.org/10.1002/ijfe.2435
2007–2009 Recessions on the Day of the Week Effect for the S&P