Modifications Committee Mod_18_10: Intra-Day Trading 25 November 2010.
Predicting Intra-day and Day of the Week Anomalies in ... · Predicting Intra-day and Day of the...
-
Upload
duongthuan -
Category
Documents
-
view
216 -
download
0
Transcript of Predicting Intra-day and Day of the Week Anomalies in ... · Predicting Intra-day and Day of the...
Predicting Intra-day and Day of the Week Anomalies in Turkish Stock Market
Kemal Eyuboglu (a)
Sinem Eyuboglu (b)
Rahmi Yamak (c)
According to Efficient Market Hypothesis which is presented by Fama (1965) in the finance
literature, any investor cannot gain abnormal returns. But various anomalies such as day or intra-day
effect which are frequently observed at the stock markets provide some abnormal returns to investors.
In the related literature, many studies found various anomalies for the different national and
international stock markets. But most of them used aggregate data in their econometric analysis. The
question is whether the same anomalies exist in the sub-indexes such as communication,
transportation, banking, mining etc. The purpose of this study is to investigate whether there are the
same anomalies such as intra-day effect and day of the week effect for 24 Borsa Istanbul (BIST) sub-
indexes. The data used in this study are daily and cover the period of 03.01.2005-11.02.2015 for
Turkey. Statistical results show that there is an evidence for intra-day effect in all 24 sub-indexes
(except communication) and day of the week effect in 3 sub-indexes for this period. Accordingly
temporal predictability of returns in the BIST indexes is under a strong intra-day effect and weak day
of the week effect. Moreover the existence of anomalies in the stock market show that investors are
not rational, in other words these anomaly patterns weak the validity of Efficient Market Hypothesis in
the context of Borsa Istanbul.
Key Words: Intra-day effect, day of the week effect, Borsa Istanbul, Least squares method
JEL Classification: G11, G12, G14
1. Introduction
Efficient Market Hypothesis presented by Fama (1965) assumes that stock prices reflect the public
disclosure of information therefore no investor can gain any abnormal returns. This hypothesis is
based on the assumption that investors behave rationally, different kinds of information related to
stocks could be gained by investors; thus the price of stocks is determined in term of this information.
However it has been reached some findings conflict with Efficient Market Hypothesis in the literature
and it is termed as anomalies. A significant part of these anomalies consist of the calendar anomalies.
Calendar anomalies arises hourly, daily, weekly, monthly, yearly or a specific pre or post-time of
period.
An inefficient market will allow investors to gain disproportionately abnormal returns with their
degree of risk, in other words, calendar anomalies allow them to obtain lower or higher returns at
certain times. Therefore the determination of the calendar anomalies for investors composes an
important part of decision-making process.
In the securities markets anomalies take an important place in terms of investors’ gains and they
have an extensive place in the finance literature. Especially intra-day and day-of-week effects have
(a)
Ph.D. Research Assistant, Karadeniz Technical University, Trabzon, Turkey, email: [email protected] (b)
Research Assistant, Karadeniz Technical University, Trabzon, Turkey, email: [email protected] (c)
Professor, Karadeniz Technical University, Trabzon, Turkey, email: [email protected]
often tested in empirical studies on a variety of the world’s stock markets. However, almost all
existing studies both in the world and Turkey literature used aggregated data in their econometric
analysis. The question is whether the same anomalies exist in the sub-indexes such as communication,
transportation, banking, mining, etc. In this study 23 sub-indexes of Borsa Istanbul (BIST) and also an
aggregated index (BIST-100) are separately used to determine whether there are intra-day and day of
the week anomalies. Firstly, the current literature is investigated for the intra-day and day of the week
anomalies. Then the data and methods are presented and finally findings are evaluated.
2. Literature Review
There are many studies searching for the existence of intra-day and day of the week anomalies for
different countries, different indexes and different periods in literature. Among these studies, Wood et
al. (1985) determined whether there was an intra-day effect in NYSE by considering minute-by-
minute price changes and found that the returns realized in the first 30 minutes and the last 1 minute of
the trading day were more than the rest of the day. McInish and Wood (1990) reached the similar
results for USA.
Another study on the intra-day effect was carried out by Harris (1986) for the period of 1981-
1983 for USA. By analyzing the trading day into 15 minutes of periods, Harris (1986) concluded that
intra-day effect in terms of returns existed significantly in USA. The similar results were separately
found by Ho et al. (1993) and Cheung (1995) for Hong Kong Stock Exchange. However, Cheung et al.
(1994) found out that there was not any significant difference between the morning and afternoon
session returns in Hong Kong Stock Exchange for 1990. The same finding was obtained by Smirlock
and Starks (1986) for USA and by Aitken et al. (1994) for Australian Securities Exchange. Jain and
Joh (1988) examined the existence of intra-day effect by using hourly returns in S&P 500 Index for the
period of 1979-1983. The results showed that Monday was the only day of the week on which
negative return was achieved.
Lockwood and Linn (1990) detected that returns in NASDAQ followed a decreasing trend in an
hour after the trading start; later on the other hand it followed a rising trend. For the period of 1992-
1993, Camino (1996) investigated the intra-day effect by dividing IBEX-35 Index into 15 minutes of
periods and found that the returns were statistically different in the first 4 hours following the opening
of trade. Niarchos and Alexakis (2003) repeated the same research on intra-day effect for Atina Stock
Market by using 15 minutes of frequencies in 1998 and found out that statistically negative and
significant returns showed up at 11:30 am. Ozenbas (2006) studied on intra-day effect on New York,
London, Germany and Paris Stock Markets for 2000 and found out that the opening times of Mondays
were more volatile compared to opening of other days and the closing times of Fridays were more
volatile compared to closing times of other days.
Tian and Guo (2007) studied the existence of intra-day effect in Shanghai Stock Market for the
period of 2000-2002 by dividing the sessions into 5 minutes. The findings showed that the volatility in
the morning session was more than the afternoon session. The same results were also found by
Strawinski and Slepaczuk (2008) for Poland. Tooma (2007) examined whether there was an intra-day
effect on Cairo and Alexandria Stock Market for 2005 and found that there was an intra-day effect.
Deev and Linnertová (2012) used 5 minutes data in Czech Stock Market and obtained that positive
returns could be achieved in the opening of trade on Monday and Thursday.
The intra-day effect studies carried out in Turkey on the other hand focused mainly on Istanbul
Stock Exchange Market (BIST 100 Index). Among these studies, Ozmen (1997) found that the lowest
return was gained on Monday in the afternoon session for the period of 1988-1996. Similarly, Bildik
(2000) investigated the existence of intra-day effect by using 15 minutes data for the period of 1996-
1999. He found that the returns were quite high and positive towards the opening and closing hours of
the day. Gokce and Sarıoglu (2004) studied the intra-day effect for the period of 1995-2003. The
findings showed that there was an intra-day effect and the highest returns realized in the morning
session of Tuesday and in the afternoon session of Friday. Abdioglu and Degirmenci (2013) examined
the intra-day anomaly for 2012 and obtained that there was an effect in this period. On the other hand
Kucukkocaoglu (2008) investigated whether there was an intra-day effect for the period of 2000-2002,
by using 15 minutes data for 8 different stocks and for different indexes. The findings indicated that
the volatility was maximum in the mornings until 2001, and then the volatility decreased significantly.
In the related literature day of the week effect was also studied for different countries,
different periods. Among these studies, Cross (1973) examined whether there was a day of the week
anomaly in S&P Index for the period of 1953-1970. The findings showed that the returns were
negative on Monday and positive on Friday. The similar results were also found by French (1980) for
USA, Poshakwale (1996) and, Nath and Dalvi (2004) for India, Chai et al. (2008) for Taiwan,
Singapore and Hong Kong. Jaffe and Westerfield (1985) investigated the day of the week effect for
Japan Stock Exchange Market for the period of 1970-1983, and obtained that the lowest return was
obtained on Tuesdays. The similar results were found by Solnik and Bousquet (1990) for France.
Condoyanni et al. (1987) selected USA, Australia, Canada, France, England, Japan and Singapore and
obtained that rate of returns were all statistically different in all countries except Australia.
Chen et al. (2000) studied the existence of day of the week anomaly in China stock market
and their findings showed that negative returns were gained on Tuesday. The similar results were
found by Lyroudi and Subeniotis (2002) for Athens and by Raj and Kumari (2006) for India Stock
Exchange. Berument and Kıymaz (2001) investigated the day of the week anomaly for S&P Index for
the period of 1973-1997 and concluded that the highest return was gained on Wednesday and the
lowest on Monday. Chukwuogor-Ndu (2007) scrutinized for 10 stock market and observed that
negative returns were on Mondays in the 7 out of 10 stock markets. On the other hand Kenourgios and
Samitas (2008) investigated the existence of the day of the week effect for the period of 1995-2000 in
Athens Stock Exchange. Their findings implied that there was a significant day of the week effect on
both returns and volume of Athens Stock Exchange. Worthington (2010) studied the effect for the
period of 1958-2005 in Australia Stock Exchange and concluded that the returns were negative on
Wednesday.
Nageswari et al. (2011), tested the existence of the day of the week anomaly in S&P CNX
Nifty and S&P CNX 500 Indexes for the period of 2002 to 2010. Their findings showed that the
highest returns were on Monday and the lowest were on Friday. Rodriguez (2012) examined the day of
the week effect for Argentina, Brazil, Chile, Colombia, Mexico and Peru and found that all countries
except Mexico, showed Monday and Friday effect. But Al-Jafari (2012) could not get any day of the
week effect for Oman Stock Market.
Mitra and Khan (2014) searched the existence of the effect for India Stock Market for the
period of 2001-2012. Their findings showed that the lowest returns were gained on Mondays however;
the lowest returns were not statistically significant.
Eken and Uner (1997) studied the calendar effects in Istanbul Stock Market Exchange (ISE)
for the period of 1988-2007. Their findings showed that there was a day of the week effect in the ISE.
The same results were also found by Guneysu and Yamak (2011), Abdioglu and Degirmenci (2013)
for ISE. Kıvılcım et al. (1997) repeated the research for the period of 1988-1996 and obtained that
Monday and Friday influenced the returns and for this reason the stock market in Turkey was not
effective in its weak form. Oguzsoy and Guven (2003) examined the day of the week effect in ISE-100
Index for the period of 1988-1999 and found that the returns were lowest on Tuesday and highest on
Friday. Kıyılar and Karakas (2005) investigated whether seasonal anomalies could be observed for
ISE Index for the period of 1988-2003. It was observed that the returns were highest on Friday and
Thursday and lowest on Monday. The same results were found by Atakan (2008) for ISE. However,
using GARCH model in their econometric analysis, Atakan and Kozanoglu (2007) could not find any
significant difference between Thursday and Friday.
Dicle and Hassan (2007) examined the existence of day of the week effect for all the indexes of
ISE for the period of 1987-2005. Their Findings showed that the return was negative on Monday and
positive on Thursday and Friday. The same results were also found by Cinko and Avcı (2009) for ISE
100 Index. Hamarat and Tufan (2008) pointed out that although there was a day of the week anomaly
in Tourism Index for the period of 1997-2005 there was not any month effect.
Ergul et al. (2009) studied whether the day of the week effect was valid in Second National
Market Index for the period of 1997-2007. Their findings showed that the highest return was on Friday
and the lowest on Wednesday. Cicek (2013) examined whether there was a day of the week effect in
BIST 100, Financial, Services, Industry and Technology Indexes for the period of 2008-2012. The
findings showed that the returns, except for Financial Index, were positive and highest on Monday. On
the other hand Konak and Kenderli (2014) found negative Monday effect in BIST 100 for the period
of 2005-2012.
Table 1
Literature Review
Intra-day Effect
Study Period Country Method Effect
Wood et al. (1985) 1971-1972 USA Statistical Tests Yes
Harris (1986) 1981-1983 USA Statistical Tests Yes
Smirlock and Starks (1986) 1963-1983 USA Statistical Tests No Jain and Joh (1988) 1979-1983 USA Statistical Tests Yes
McInish and Wood (1990) 1980-1984 USA OLS Method Yes Lockwood and Linn (1990) 1964-1989 USA Statistical Tests Yes
Ho et al. (1993) - Hong Kong - Yes Cheung et al. (1994) 1986-1990 Hong Kong - No Aitken et al. (1994) 1991-1992 Australia Statistical Tests No
Cheung (1995) 1986-1990 Hong Kong Statistical Tests Yes Camino (1996) 1992-1993 Spain Statistical Tests Yes Özmen (1997) 1988-1996 Turkey Statistical Tests Yes Bildik (2000) 1996-1999 Turkey Statistical Tests Yes
Niarchos ve Alexakis (2003) 1998 Greece OLS Method Yes Gökce and Sarıoglu (2004) 1995-2003 Turkey Statistical Tests Yes
Özenbaş (2006) 2000 USA, UK,
Germany, France Statistical Tests Yes
Tian and Guo (2007) 2000-2002 China Statistical Tests Yes Tooma (2007) 2005 Egypt Statistical Tests Yes
Kücükkocaoglu (2008) 2000-2002 Turkey OLS Method Yes Strawinski and Slepaczuk
(2008)
1998-2008 Poland OLS Method Yes Deev and Linnertová (2012) 2011-2012 Czech Republic GARCH Yes Abdioglu and Degirmenci
(2013)
2003-2012 Turkey OLS Method Yes
Day of the Week Effect Cross (1973) 1953-1970 USA OLS Method Yes French (1980) 1953-1977 USA OLS Method Yes
Jaffe and Westerfield (1985) 1970 -1983 Japan OLS Method Yes Condoyanni et al. (1987) 1969-1984 7 Countries OLS Method Yes
Solnik and Bousquet (1990) 1978-1987 France OLS Method Yes Poshakwale (1996) 1987-1994 India OLS Method Yes
Chen et al. (2000) 1992-1997 China ARCH-GARCH Yes Berument and Kıymaz (2001) 1973-1997 USA GARCH Yes Lyroudi and Subeniotis (2002) 1994-1999 Greece Statistical Tests Yes
Nath and Dalvi (2004) 1999-2003 India OLS Method Yes
Chukwuogor-Ndu (2005) 1998-2003 10 Eastern Asia
Countries Statistical Tests Yes in 7
Raj and Kumari (2006) 1987-1998 India OLS Method Yes
Kenourgios and Samitas (2008) 1995-2000 Greece OLS Method and
M-GARCH Yes
Chia et al. (2008) 2000-2006
Taiwan,
Singapore, Hong
Kong and
S. Korea
EGARCH-M Yes
Worthington (2010) 1958-2005 Australia Statistical Tests Yes
Nageswari et al. (2011) 2002-2010 India OLS Method No
Rodriguez (2012) 1993-2007
Argentina, Brazil,
Chile, Colombia,
Mexico and Peru
GARCH Yes (Except
Mexico)
Al-Jafari (2012) 2005-2011 Oman
GARCH (1,1),
TGARCH (1,1),
EGARCH (1,1)
No
Mitra and Khan (2014) 2001-2012 India OLS Method No
Eken and Uner (1997) 1988-2007 Turkey Statistical Tests Yes Kıvılcım et al. (1997) 1988-1996 Turkey OLS Method Yes
Oguzsoy and Guven (2003) 1988-1999 Turkey OLS Method Yes Kıyılar and Karakas (2005) 1988-2003 Turkey Statistical Tests Yes
Aktas and Kozanoglu (2007) 2001-2007 Turkey GARCH Yes Dicle and Hassan (2007) 1987-2005 Turkey ARCH-GARCH Yes
Hamarat and Tufan (2008) 1997-2005 Turkey Statistical Tests,
Probit Model
Yes Atakan (2008) 1987-2008 Turkey ARCH-GARCH Yes
Cinko and Avcı (2009) 1995-2008 Turkey OLS Method Yes Ergul et al. (2009) 1997-2007 Turkey OLS Method Yes
Guneysu and Yamak (2011) 1990-2010 Turkey OLS Method Yes Abdioglu and Degirmenci
(2013) 2003-2012 Turkey OLS Method Yes
Cicek (2013) 2008-2012 Turkey EGARCH Yes Konak and Kenderli (2014) 2005-2012 Turkey GARCH Yes
3. Data and Methodology
This study investigates whether there are intra-day and day of the week anomalies for Turkish
Stock Market. Disaggregated price indexes used in the study were collected from the official site of
Borsa Istanbul. Indexes which are used in the empirical analysis are shown in Table 2.
Table 2
Indexes used in the study
Index Name Index Name
BIST 100 Tourism Industrials Wholesale and Retail Trade
Food Beverage Telecommunication
Textile Leather Sports
Wood Paper Printing Financials
Chemical Petroleum Plastic Banks
Nonmetal Mineral Products Insurance
Basic Metal Leasing Factoring
Metal Products Machinery Holding and Investment
Services Real Estate İnvestment Trusts
Electricity Technology
Transportation Information Technology
3.1. Testing of Intra-day Effect
Closing prices of the morning and afternoon sessions of each index for testing of intra-day effect
were used for computation of session returns in the equation (1) as follows:
1
ln( )tt
t
PR
P
(1)
In the equation Rt is the session return for each index, Pt is the closing price of the each index on
session t, Pt-1 is the closing price of the each index on session t-1 and “ln” is naturel logarithm.
In order to examine intra-day anomalies, dummy variables were created and then the significance
level of each dummy variable was determined by t-test statistics under ordinary least squares
equations. The existence of intra-day effect for each index is estimated with the following regression.
1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10t tR D D D D D D D D D D (2)
In the equation, D1…D10 represent dummy variables. D1 is a dummy variable which takes the
value of 1 if session is a Monday morning session and 0 otherwise, D2 is a dummy variable which
takes the value of 1 if session is Monday afternoon session, and 0 otherwise; and so on. Rt is the
session return of the each index, the OLS coefficients β1 to β10 are the mean returns for morning
session of Monday through afternoon session of Friday, respectively. The stochastic term is shown by
t . The null hypothesis is 0 1 2 3 4 5 6 7 8 9 10: 0 H and the
alternative is H1: All β’s are not equal. If the null hypothesis is rejected then the returns must exhibit
some intra-day effect.
3.2. Testing of Day of the Week Effect
Using the closed prices of each index in the study, returns of each index are separately computed
using by the following equation.
1
ln( )tt
t
PR
P
(1)
In the equation, Rt is the daily return for day t of each index, Pt is the price of the each index on day t,
Pt-1 is the price of the index on day t-1 and “ln” is naturel logarithm.
In order to investigate day of the week anomalies, five dummy variables are created and then
tested by using t-statistics of coefficients of dummy variables included equation 3.
In this equation, Rt represents the daily return on the each index, D1 is a dummy variable which takes
the value of 1 if day is Monday and 0 otherwise, D2 is a dummy variable which takes the value of 1 if
day is Tuesday and 0 otherwise; and so on.
1 2 3 4 5 1 2 3 4 5 t tR D D D D D (3)
The OLS coefficients β1 to β5 are the daily mean returns from Monday to Friday, respectively. The
random error term is represented byt . The null hypothesis is 0 1 2 3 4 5: 0H and
the alternative is H1: All β’s are not equal.
If the null hypothesis is rejected, there must be a statistically significant difference among the
coefficients. This means that there is a day of the week effect in Borsa Istanbul.
4. Findings
Descriptive statistics of average returns for each index are reported in Table 3. As can be seen, the
average returns for each index are positive. In addition, the highest and lowest average returns appear
in Wholesale and Retail Trade and Electricity sectors, respectively. According to measures of the
standard deviations, the highest volatility appears in tourism sector and the lowest appears in nonmetal
mineral product sector.
Table 3
Descriptive Statistics of Index Returns Index Mean Maximum Minimum Std.
Deviation
Skewness Kurtosis
BIST 100 0.00023 0.08534 -0.07811 0.01207 -0.35022 7.22680 Industrials 0.00025 0.05996 -0.07257 0.01019 -0.85664 9.28654 Food Beverage 0.00028 0.08220 -0.09633 0.01295 -0.30910 7.20951 Textile Leather 0.00022 0.08869 -0.09591 0.01143 -1.12812 12.5763 Wood Paper Printing 0.00009 0.08506 -0.09014 0.01313 -0.65211 7.58976 Chemical Petroleum Plastic 0.00027 0.06638 -0.07497 0.01247 -0.52571 7.36215 Nonmetal Mineral Product 0.00023 0.06012 -0.07128 0.01004 -0.85491 9.93571 Basic Metal 0.00032 0.08876 -0.09732 0.01477 -0.26154 7.97760 Metal Products Machinery 0.00026 0.07362 -0.08722 0.01211 -0.65265 8.38066 Services 0.00029 0.07703 -0.07478 0.01059 -0.26921 7.85827 Electricity 0.00004 0.09369 -0.10560 0.01512 -0.19921 9.71731 Transportation 0.00039 0.08407 -0.08938 0.01563 -0.14734 6.30009 Tourism 0.00006 0.08979 -0.10501 0.01600 -0.49865 7.95507 Wholesale and Retail Trade 0.00043 0.09802 -0.08614 0.01231 -0.11492 8.96522 Telecommunication
Telecominic
Telecommunication
Telecommunication
0.00016 0.10779 -0.13248 0.01486 -0.04730 8.43253 Sports 0.00014 0.10510 -0.11231 0.01471
4
-0.09944 11.4345
5 Financials 0.00022 0.09880 -0.08172 0.01404 -0.20231 6.66519 Banks 0.00024 0.10330 -0.08739 0.01536 -0.08450 6.01927 Insurance 0.00026 0.08782 -0.09477 0.01446 -0.40355 8.08333 Leasing Factoring 0.00018 0.09101 -0.09859 0.01501 -0.03857 9.16233 Holding and Investment 0.00017 0.09342 -0.08301 0.01304 -0.44691 7.73649 Real Estate Investment Trusts 0.00009 0.06827 -0.09000 0.01267 -0.73163 8.01090 Technology 0.00032 0.07158 -0.09641 0.01297 -0.71365 8.92364 Information Technology 0.00018 0.07394 -0.09848 0.01362 -0.62982 9.23717
The regression results for intra-day effect are shown in Table 4. Whether there is a significant
difference between session returns is tested by using F statistics. The values of F statistics in all sub-
indexes (except Telecommunication) show that there is a significant difference between morning and
afternoon session. So, in 12 of the 24 sub-indexes (BIST 100, industrials, textile leather, wood paper
printing, nonmetal mineral products, electricity, transportation, tourism, leasing factoring, real estate
investment trust, technology and information technology) average return is positive for morning
session and negative for afternoon session.
Table 4
Regression results of intra-day effect
BIST 100 Industrials Food Beverage Textile
Leather
Wood Paper
printing
Chemical
Petroleum
Plastic
β β β β β β
Monday1 1.89E-05 0.0006 -0.0001 0.0015a
0.0021a 0.0001
Monday2 0.0004 0.0001 0.0005 -0.0007 -0.0012b 0.0005
Tuesday1 0.0006 0.0011a 0.0007 0.0017
a 0.0023
a 0.0006 Tuesday2 -0.0001 -0.0001 6.93E-06 -0.0010
b -0.0012
b 0.0002 Wednesday
1 0.0012
b 0.0015
a 0.0016
a 0.0016
a 0.0026
a 0.0012
b
Wednesday
2 -0.0012
b -0.0017
a -0.0006 -0.0027a
-0.0035a
-0.0014b
Thursday1 0.0019a
0.0015a
0.0016a
0.0027a
0.0031a
0.0010c
Thursday2 -0.0011b
-0.0012a -0.0007 -0.0014
a -0.0036
a -0.0007 Friday1 0.0007 0.0010
b 0.0002 0.0015a
0.0026a
0.0011b
Friday2 -0.0001 -0.0005 -0.0005 -0.0011b
-0.0022a -0.0001
F 3.170771a
6.307258a
2.268120b
12.22678a
20.78387a
2.411931a
Nonmetal
mineral
pro.
Basic Metal Metal Products
Machinery Services Electricity Transportation
β β β β β β
Monday1 0.0012a
0.0011c
0.0009c 0.0001 0.0021
a 0.0017
b
Monday2 -0.0007c 0.0001 0.0004 0.0002 -0.0014
b -0.0002 Tuesday1 0.0015
a 0.0013
b 0.0015
a 0.0005 0.0010 0.0024a
Tuesday2 -0.0002 0.0001 -0.0006 -0.0003 -0.0013b -0.0002
Wednesday
1 0.0022
a 0.0014
b 0.0014
a 0.0007 0.0018a
0.0024a
Wednesday
2 -0.0023
a -0.0022
a -0.0018
a -0.0010
a -0.0028
a -0.0025
a
Thursday1 0.0018a
0.0016b
0.0016a
0.0014b
0.0031a
0.0019a
Thursday2 -0.0017a -0.0009 -0.0019
a -0.0009
b -0.0025
a -0.0017
b
Friday1 0.0012a 0.0010 0.0013
b 0.0016
a 0.0012
b 0.0016
b
Friday2 -0.0007 -0.0004 -0.0005 0.0004 -0.0008 -0.0016b
F 12.32637a
3.550763a
6.287966a
3.650653a
9.012781a
7.177932a
Tourism
Wholesale
and Retail
Trade
Telecommunication Sports Financials Banks
β β β β β β
Monday1 0.0018a 0.0008 -0.0012
c 0.0007
-6.05E06 0.0001 Monday2 -0.0017
b 9.65E-05 0.0006 -0.0001 0.0004 0.0003 Tuesday1 0.0017
b 0.0001 7.29E-05 0.0003 0.0005 0.0007 Tuesday2 -0.0013
c -0.0001 -0.0003 3.71E-06 -0.0002 -0.0002 Wednesday
1 0.0020
a 0.0009
c -0.0001 0.0012b
0.0014b
0.0016b
Wednesday
2 -0.0034
a -0.0011
b -0.0001 -0.0026a
-0.0012b
-0.0013b
Thursday1 0.0035a
0.0015a 0.0008 0.0019
a 0.0022
a 0.0025
a
Thursday2 -0.0026a -0.0007 -0.0003 -0.0011
c -0.0011
c -0.0012
c
Friday1 0.0025a
0.0020a
0.0015b
0.0018a 0.0004 0.0005
Friday2 -0.0019a 0.0006 0.0007 -0.0008 -0.0003 -0.0005
F 11.63440a
3.526460a 1.365799 4.507841
a 2.893274
a 2.973785
a
Insurance
Leasing
Factoring
Holding and
Investment
Real
Estate
Investment
Trust
Technology Information
Technology
β β β β β β
Monday1 -0.0006 0.0010 -0.0006 0.0015a
0.0028a
0.0024a
Monday2 1.72E-05 -0.0009 0.0009c -3.52E-05 -0.0008 -0.0011
b
Tuesday1 0.0011c
0.0025a -7.07E05 0.0011
b 0.0008 0.0008 Tuesday2 -0.0001 -0.0008 0.0001 -0.0013
b -0.0005 -0.0006 Wednesday
1 0.0018
a 0.0022
a 0.0005 0.0015a
0.0018a
0.0019a
Wednesday
2 -0.0019
a -0.0025
a -0.0009 -0.0024a
-0.0028a
-0.0030a
Thursday1 0.0022a
0.0028a 0.0016 0.0019
a 0.0026
a 0.0025
a
Thursday2 -0.0006 -0.0026b
-0.0007a
-0.0023a
-0.0019a
-0.0020a
Friday1 0.0012b
0.0015b 9.82E-05 0.0021
a 0.0018
a 0.0017
a
Friday2 -0.0006 -0.0013b 0.0006 -0.0012
b -0.0007 -0.0010c
F 4.040646a
9.180555a
1.941683b
9.408071a
11.20642a
10.06372a
a denotes significance at the 1% level. b denotes significance at the 5% level. c denotes significance at the 10% level
Also, in 17 of the 24 indexes (BIST 100, industrials, textile leather, wood paper printing, nonmetal
mineral products, basic metal, metal products machinery, electricity, transportation, tourism,
wholesale and retail trade, telecommunication, sports, leasing factoring, real estate investment trust,
technology, information technology) morning sessions have higher return than the afternoon sessions.
The difference between 2 session returns can be interpreted as the validity of Efficient Market
Hypothesis in BIST.
In addition, both session returns of all days are found to be different for wood paper printing sector
and tourism sector. All other session returns except afternoon session of Monday are different from
each other for textile leather and real estate investment trust sectors. It is also determined that there is a
difference among all other session returns except afternoon session of Tuesday and Friday for
nonmetal mineral products, morning session of Tuesday and afternoon session of Friday for electricity,
afternoon session of Monday and Tuesday for transportation, morning and afternoon sessions of
Tuesday for information technology. It is also found that all session returns except Monday are
statistically different from each other, afternoon sessions of Tuesday and Friday for metal products
machinery sector, morning and afternoon sessions of Monday and afternoon session of Tuesday for
leasing factoring sector returns. Other session returns are statistically different except for morning
session of Monday and afternoon sessions of Monday, Tuesday, Friday returns for industrials sector.
All other session returns except afternoon sessions of Monday, Friday and morning and afternoon
sessions of Tuesday are statistically different for technology sector. Returns for five sessions in basic
metal, sports and insurance sectors, four sessions in BIST 100, chemical petroleum plastic, services,
wholesale and retail trade, financials and banks sectors are different from other sessions returns.
Morning sessions of Wednesday and Thursday in food beverage sector, morning session of Monday
and Friday in telecommunication sector, afternoon session of Monday and Thursday in holding and
investment sector returns are also different from other session returns.
Table 5 shows the results of the regression model in equation 3 for testing the day of the week
effect. Findings show that Monday’s returns are significantly positive for the metal products
machinery, real estate investment trust, technology; Tuesday’s returns are significantly positive for the
nonmetal mineral products, transportation, leasing factoring; Thursday’s returns are significantly
positive for the textile leather, insurance and Friday’s returns are significantly positive for the services,
wholesale and retail trade and telecommunication sectors. Further, for the textile leather, nonmetal
mineral products, metal products machinery, transportation, telecommunication, insurance, leasing
factoring and real estate investment trust sectors, someday returns are individually significant but the F
statistics are not statistically significant so this leads to conclude that there is no evidence of day of the
week effect. Also significant F-statistics for the services, wholesale and retail trade and technology
index indicate that day of the week effect is valid in these three sectors.
Table 5
Regression results of the day of week effect
BIST 100 Industrials Food Beverage
Textile
Leather
Wood Paper
Printing
Chemical
Petroleum
Plastic
β β β β β β
Monday 0.0004 0.0007 0.0003 0.0008 0.0009 0.0007
Tuesday 0.0004 0.0010 0.0007 0.0006 0.0010 0.0008 Wednesday 4.37E-05 -5.33E05 0.0010 -0.0010 -0.0008 -2.27E05
Thursday 0.0007 0.0003 0.0008 0.0013c -0.0004 0.0002
Friday 0.0005 0.0004 -0.0002 0.0004 0.0003 0.0009
F 0.436 0.925 0.791 1.691 0.474 0.676
Nonmetal
Mineral
Products
Basic Metal Metal Products
Machinery Services Electricity Transportation
β β β β β β
Monday 0.0005 0.0011 0.0014c 0.0003 0.0005 0.0015
Tuesday 0.0013b 0.0014 0.0008 0.0001 -0.0003 0.0021
b
Wednesday -0.0001 -0.0007 -0.0002 -0.0002 -0.0008 4.11E-05
Thursday 0.0001 0.0007 -0.0002 0.0004 0.0006 0.0001 Friday 0.0005 0.0005 0.0007 0.0021
a 0.0004 -6.35E05
F 1.19 1.45 1.211 2.194c 0.400 1.416
Tourism
Wholesale
and Retail
Trade
Telecommunication Sports Financials Banks
β β β β β β
Monday 0.0002 0.0008 -0.0005 0.0004 0.0004 0.0004
Tuesday 0.0003 8.30E-05 -0.0003 0.0004 0.0003 0.0004 Wednesday -0.0012 -0.0002 -0.0003 -0.0013 0.0001 0.0003
Thursday 0.0009 0.0008 0.0003 0.0007 0.0011 0.0012
Friday 0.0005 0.0026a
0.0023b 0.0010 0.0001 -4.18E05
F 0.584 2.793b 1.461 0.876 0.379 0.379
Insurance
Leasing
Factoring
Holding and
Investment
Real Estate
Investment
Trust
Technology Information
Technology
β β β β β β
Monday -0.0005 0.0001 0.0003 0.0014b
0.0019b 0.0012
Tuesday 0.0010 0.0016c -2.47E-05 -0.0001 0.0002 0.0001
Wednesday -6.72E05 -0.0002 -0.0003 -0.0008 -0.0008 -0.0009
Thursday 0.0016c 0.0002 0.0009 -0.0003 0.0006 0.0005
Friday 0.0005 0.0002 0.0007 0.0008 0.0010 0.0007
F 1.037 0.632 0.464 1.190 1.890c 0.970
a denotes significance at the 1% level. b denotes significance at the 5% level. c denotes significance at the 10% level
5. Conclusion
According to the Efficient Market Hypothesis, it is reflected to the stock prices instantly when the
information is disclosed. Thus any investor cannot gain abnormal returns. But anomalies such as day
and intra-day effect which are frequently observed at the stock markets provide some abnormal returns
to investors. In this study, it is tested whether there are intra-day effect and day of the week effect for
24 sub-indexes of Borsa Istanbul for the period of 2005-2015. For this purpose, first, each index
returns are computed, and then the dummy variables are created within the framework of anomalies.
Dummy variables are included to the right side of regression equation as explanatory variables and
tested whether there is a difference among the returns. Findings show that there is an evidence for
intra-day effect for all 24 sub-indexes (except telecommunication). The lowest returns occur on
afternoon session of Wednesday for BIST 100, industrials, textile leather, chemical petroleum plastic,
nonmetal mineral products, basic metal, services, electricity, transportation, tourism, wholesale and
retail trade, sports, financials, banks, insurance, holding and investment, real estate investment trust,
technology and information technology sectors, for wood paper printing, metal products machinery,
leasing factoring sectors on afternoon sessions of Thursday, for telecommunication sector on morning
session of Monday. The highest returns occur; on morning sessions of Thursday for BIST 100,
industrials, food beverage, textile leather, wood paper printing, basic metal, metal products machinery,
electricity, tourism, sports, financials, banks, insurance, leasing factoring, holding and investment and
information technology sectors, on morning sessions of Friday for services, wholesale and retail trade,
telecommunication and real estate investment trust sectors, on morning sessions of Wednesday for
chemical petroleum plastic, nonmetal mineral products and transportation sectors, on morning sessions
of Monday for technology sector.
In terms of the day of the week effect, returns on Mondays in metal products machinery, real
estate investment trust and technology sectors, returns on Tuesdays in nonmetal mineral products,
transportation and leasing factoring sectors, returns on Thursdays in textile leather and insurance
sectors, returns on Fridays in services, wholesale and retail trade and telecommunication sectors are
different compared to the other days of the week. So this situation presents some evidences for the
gaining abnormal returns with the timing of the trading decision in Borsa Istanbul. However someday
returns are statistically and individually significant, but the F statistics are statistically significant only
for the services, wholesale and retail trade and technology indexes. It means that there is a day of the
week effect only for these three indexes; metal products machinery, real estate investment trust and
technology sectors.
Moreover the existence of anomalies in the stock market in Turkey shows that investors are not
rational. In other words, these anomaly patterns weak the validity of Efficient Market Hypothesis in
the case of Borsa Istanbul.
References
Abdioglu, Z. & Degirmenci, N. (2013). Seasonal Anomalies in Istanbul Stock Exchange
(in Turkish). Business and Economics Research Journal, 4(3), 55-74.
Aitken, M., Frino, A. & Sayers, S. (1994). The intra-day impact of block trades on the
Australian Stock Exchange, Asia Pacific Journal of Management, 11(2), 237-253.
Aktas, H. & Kozanoglu, M. (2007). The GARCH model-based test of the day of the week
Effect at Istanbul Stock Exchange (in Turkish). Finansal Politik & Ekonomik Yorumlar
Dergisi, 44(514), 37-45.
Al-Jafari, M. K. (2012). An empirical investigation of the day-of- the-week effect on stock
returns and volatility: evidence from Muscat Securities Market. International Journal of
Economics and Finance, 4(7), 141-149.
Atakan, T. (2008). Testing the day-of-the-week effect and January effect anomalies at
Istanbul Stock Exchange with ARCH-GARCH models (in Turkish). İstanbul Üniversitesi
İşletme Fakültesi Dergisi, 37(2), 98-110.
Berument, H. & Kıymaz, H. (2001). The day of the week effect on stock market volatility.
Journal of Economics and Finance, 25(2), 181-193.
Bildik, R. (2000). Intra-day seasonalities on stock returns: evidence from the Turkish
Stock Market, http://papers.ssrn.com/sol3/papers.cfm?abstract_id=251503.
Camino, D., (1996). The role of information and trading volume on intradaily and weekly
returns pattern in the Spanish Stock market. Business Economics Series 01, Working Paper,
96-10.
Chen, G., Kwok, C. C. Y. & Rui, O.M. (2000). The day-of-the-week regularity in the
stock markets of China. Journal of Multinational Financial Management, 11, 139-163.
Cheung, Y., (1995). Intraday returns and the day-end effect: evidence from the Hong
Kong Equity Market. Journal of Business Finance & Accounting, 1023-1034.
Cheung, Y.; Ho, R., Yan-Ki; Pope, P. & Draper, P., (1994). Intraday stock return
volatility: The Hong Kong evidence, Pasific-Basin Finance Journal, 261-276.
Chia, R. C., Liew, V. K., Syed K. W., & Syed A. W. (2008). Day-of-the-week effects on
selected East Asian Stock Markets, Economics Bulletin, Accessecon, 7(5): 1-8.
Chukwuogor-Ndu, C. (2005). Day-of-the-week effect and volatility in stock returns:
evidence from East Asian financial markets. International Journal of Banking and Finance,
5(1), 153-163.
Condoyanni, L., O’Hanlon, J., & Ward, C. (1987). Day of the week effects on stock
returns: international evidence. Journal of Business Finance and Accounting, 159-174.
Cross F. (1973). The behavior of stock prices on Fridays and Mondays, Financial
Analysts Journal, 29, 67-69.
Cicek, M. (2013). The day-of-the-week effect on return and volatility in the Turkish Stock
Markets”, Journal of Applied Finance & Banking, 3(4), 143-167.
Cinko, M. & Avcı, E. (2009). Examining the day of the week effect in Istanbul Stock
Exchange (ISE), International Business & Economics Research Journal, 8(11), 1-5.
Deev, D. & Linnertová O. (2012). Intraday and intraweek trade anomalies on the Czech
Stock Market, Acta Universitatis Agriculturae et Silviculturae Mendelianae Brunensis, 60,
79-87.
Dicle, M. F. & Hassan, M. K. (2007). Day of the week effect in Istanbul Stock Exchange.
Scientific Journal of Administrative Development, 5, 1-27.
Eken, M.H. & Üner, T.Ö. (1997). Calender effects in stock exchanges: case of Istanbul
Stock Exchange . İMKB Dergisi, 12(45), 61-119.
Ergül, N., Akel, V. & Dumanoğlu, S. (2009). Is the day of the week effect valid in ISE
Second National?”, http://www.-finanskulup.org.tr-/assets/-
maliyefinans/82/Nuray_ErgulVeli_AkelSezai_Dumanoglu_Haftanin_Gunu_Etkisi_Ikinci_Ul
usal_Pazarda_Gecerlimidir.pdf, 1.2.2015.
French, K.R. (1980). Stock returns and the weekend effect, Journal of Financial
Economist, 8, 55-70.
Gokce A. G. & Sarıoglu S. E. (2004). Trading session effect: the evidence from Istanbul
Stock Exchange, 11th
Annual Conference of Multinational Finance Society, Istanbul.
Güneysu, F. & Yamak, N. (2011). Investigation of Day of the Week Effect on ISE for the
Periods of Crisis (in Turkish). Finans Politik & Ekonomik Yorumlar, 48(560), 33-44.
Hamarat, B. & Tufan, E. (2008). Is the tourism sector index efficient?, Anadolu
Üniversitesi Sosyal Bilimler Dergisi, 8(2), 169-184.
Harris, L. (1986). A transaction’s data study of weekly and intradaily patterns in stock
returns. Journal of Financial Economics, 1, 99-117.
Ho Y., Richard, C., Yan Leung, & Cheung, D. W. W. (1993). Intraday prices and trading
volume relationship in an emerging Asian Market-Hong Kong. Pasific-Basin Finance Journal
2, 203-214.
Jaffe, J. & Westerfield, R. (1985), Patterns in Japanese common stock returns: day of the
week and turn of the year effects, Journal of Financial Quantitative Analysis, 20, 261-272.
Jain, P. C. & Joh, C. H. (1988). The dependence between hourly prices and trading
volume, Journal of Financial and Quantitative Analysis, 23(3), 269-284.
Kenourgios, D. & Samitas, A. (2008). The day of the week effect patterns on stock market
return and volatility: Evidence for the Athens stock exchange. International Research Journal
of Finance and Economics, 15, 78-89.
Kıyılar, M. & Karakaş, C. (2005). Calendar-Based Market Anomalies in Istanbul Stock
Exchange (in Turkish), Yönetim Dergisi, İ.Ü. İşletme İktisadi Enstitüsü, 16-52.
Konak, F. & Kenderli, S. (2014). Analysis of the day-of-the-week Effect During the
Global Financial Crisis (in Turkish), Süleyman Demirel Üniversitesi İktisadi ve İdari Bilimler
Fakültesi Dergisi, 19(2), 275-286.
Kücükkocaoglu, G. (2008). Intra-day stock returns and close-end price manipulation In
the Istanbul Stock Exchange, Frontiers in Finance and Economics, 5(1), 46-84.
Lockwood, J. & Linn, C. (1990). An examination of stock market return volatility during
overnight and intraday periods 1964-1989, Journal of Finance, 2, 591-601.
Lyroudi, K. & Subeniotis, D. (2002). Market anomalies in the A.S.E: the day of the week
effect. SSRN Electronic Library ID-314394.
Kıvılcım, M., Muradoğlu, G. & Yazıcı, B. (1997). “An Analysis of the Day of the Week
Effect on the Istanbul Stock Exchange”, (in Turkish), İMKB Dergisi, 1(4), 15-25.
McInish, T.H., & Wood, R.A. (1990). A transaction data analysis variability of common
stock returns during 1980-1984. Journal of Banking and Finance, 1 (March), 99-112.
Mitra P. & Khan, G. S. (2014). An analysis of day of the week and intraday effects in the
Indian Stock Market: evidence from National Stock Exchange, Journal of Contemporary
Issues in Business Research, 3(3), 115-127.
Nageswari, P., Selvan, M. & Gayathri, J. (2011). Analysis of Monday effect in Indian
stock market, Research Journal of Business Management, 5(4): 170-177.
Nath, G. & Dalvi, M. (2004). Day of the week effect and market efficiency-evidence from
Indian equity market using high frequency data of National Stock
Exchange,http://isb.edu/caf/htmls/GolakaNathManojDalviDayoftheweekeffectandMarketEffi
ciency.pdf, 9.1.2015.
Niarchos, N. A., & Alexakis, C. A. (2003). Intraday stock price patterns in the Greek
Stock Exchange. Applied Financial Economics, 1, 13-22.
Oguzsoy C. B. & Guven S. (2003). Stock Returns and the day-of-the-week effect in
Istanbul Stock Exchange. Applied Economics, 35, 959-971.
Ozenbas, D. (2006). Day of the week effects in intra-day volatility patterns of equity
markets: A study of US and European Stock Markets. International Business & Economics
Research Journal, 5(6), 45-57.
Ozmen, T. (1997). Anomalies in world stock exchanges and empirical evidence in
Istanbul Stock Exchange, (in Turkish), Ankara: Sermaye Piyasası Kurulu Yayını.
Poshakwale, S. (1996). Evidence on weak form efficiency and day of the week effect in
the Indian Stock Market. Finance India, 10(3), 605-616.
Raj, M. & Kumari, D. (2006). Day-of-the-week and other market anomalies in the Indian
stock market. International Journal of Emerging Markets, 1(3), 235-246.
Rodriguez, W.K. J. (2012). Day of the week effect in Latin American Stock Market.
Revista de Analisis Economics, 27(1). 71-89.
Smirlock, M. & Starks, L. (1986). Day-of-the-week and intraday effects in stock returns.
Journal of Financial Economics, 1 (September), 197-210.
Solnik, B. & Bousquet, L. (1990). Day of the week effect on the Paris Bourse. Journal of
Banking & Finance, 14, 461-468.
Strawinski, P. & Slepaczuk, R. (2008). Analysis of high frequency data on the Warsaw
Stock Exchange in the context of efficient market hypothesis. Journal of Applied Economic
Sciences, 3(5), 306-319.
Tooma, E. A. (2007). Does the Egyptian Stock Exchange still have a day-end effect?.
International Journal of Business, 12(4), 461-470.
Tian, G.G. & Guo, M. (2007). Interday and intraday volatility: additional evidence from
Shanghai Stock Exchange. Review of Quantitative Finance & Accounting, 28, 287-306.
Wood R. A., McInish T. H. & Ord J. K. (1985). An investigation of transactions data for
NYSE stocks. The Journal of Finance, 40(3), 723-739.
Worthington, A. C. (2010). The decline of calendar seasonality in the Australian stock
exchange 1958–2005. Annals of Finance, 6(3), 421–433.