Amity Journal of Management Amity Business School

Vol. VI, No. 2, July - December 2018 Amity University, Madhya Pradesh (ISSN 2347 – 1832)

THE EFFECT OF SEASONALITY OVER STOCK EXCHANGES IN INDIA

Dr. Silky Vigg Kushwah1 and Ms. Sulekha Munshi2

ABSTRACT
Seasonal variations in production and sales are a well-known fact in business. Seasonality refers to
regular and repetitive fluctuation in a time series which occurs periodically over a span of less than a
year. The main cause of seasonal variations in time series data is the change in climate. Besides customs
and tradition, economic variables also affect stock market returns. Similarly, stock returns exhibits
systematic patterns at certain times of the day, week or month. The most common of these are monthly
patterns; certain months provide better returns as compared to others i.e. the month of the year effect. The
objective of this paper is to analyse whether the study of seasonality can help the investors to formulate
such strategies to outperform during specific events and earn high returns.

Keywords: Seasonality, Market Anomalies, Repetitive fluctuation

INTRODUCTION The existence of seasonality in stock returns

however violates an important hypothesis
Seasonal variations in production and sales
in finance that is efficient market
are a well-known fact in business.
hypothesis. The Efficient Market
Seasonality refers to regular and repetitive
Hypothesis (EMH) states that all stocks are
fluctuation in a time series which occurs
properly priced, and that abnormal returns
periodically over a span of less than a year.
cannot be earned by searching for
The main cause of seasonal variations in
mispriced stocks. Furthermore, future stock
time series data is the change in climate.
prices follow a random walk pattern, they
Besides customs and tradition, economic cannot be predicted. The presence of
variables also affect stock market returns. seasonality in stock returns violates market
Similarly, stock returns exhibits systematic efficiency principle because equity prices
patterns at certain times of the day, week or are no longer random and can be predicted
month. The most common of these are based on past pattern. This facilitates
monthly patterns; certain months provide market participants to devise trading
better returns as compared to others i.e. the strategy which could fetch abnormal profits
month of the year effect. Similarly, some on the basis of past pattern. Various studies
days of the week provides lower returns as took place and seasonal component is being
compared to other trading days i.e. days of recorded. They are called calendar
the week effect. anomalies (effects).
Market Anomalies are market patterns that In study of anomalies, 3 different types of
do seem to lead to abnormal returns more anomalies are found such as:
often than not, and since some of these
1. Calendar based 2. Announcement
patterns are based on information in Anomalies based
financial reports, therefore market  End-of-the-Day-effect Anomalies
anomalies pose an opportunity to study  Holiday effect  Earning-
them and form such a trading strategy to  Intra-Day effect Surprise effect
earn better returns.  January effect  Information
 Monday/Week-End Releasing
effect Hypothesis
1 Associate Professor, New Delhi Institute of  IPO‟s, Seasonal
 Monthly/Turn-of-the-
Management, New Delhi, India Equity
Month effect
2 Assistant Professor, Aditya Institute of Offerings and
 Tax-Year effect
Management, Science & Research, Mumbai, Buy-Backs
 Week-of-the-Month
India. effect  Pay-Out effect
 Day-of-the-Week effect  P/E Ratio effect

46
 is considered as a significant event.
3. Announcement based 4. Other The date 1st April is treated as event
Anomalies Anomalies date. Returns a week before and
 Earning-Surprise effect  Book-to-Market after the event are considered.
 Information Releasing effect
Hypothesis  Low-Beta-Firm 3. Change in calendar year
 IPO‟s, Seasonal Equity effect A new calendar year starts from 1st
Offerings and Buy-  Low Price Stock
January. The last week of calendar
Backs effect
year i.e. 25 December to 31st January
 Pay-Out effect  Momentum
 P/E Ratio effect effect
have 2 big effects like Christmas,
 Reversion to New Year eve. i.e. to obtain the
the Mean effect good sentiments of market 1tJanuary
 SEO is taken as event date. Returns a
Underperforma week before and after the event are
nce effect considered.
 Size effect
4. Diwali week
 Weather effect
One of the most auspicious day for
In this paper we have conducted Indians is Diwali. As every business
researchover returns of Nifty 50 for a encounter higher number of sales in
period of 10years i.e. 2007-2017. We have this week. Therefore, returns
considered 4 events to study seasonality. occurring in market before and after
They are: the event are considered for study.
The date on which Diwali occur
 Budget week
(different in every year) is
 Change in financial year considered as event day.
 Change in calendar year LITERATURE REVIEW
 Diwali week Several studies have been conducted to
examine the effect of seasonality over nifty
50 returns. Researchers used the concept of
1. Budget week EMH (Efficient Market Hypothesis). This
One of most awaited week of whole means efficient market hypothesis is a
year is budget week. On this day central paradigm in finance. The EMH
government announces its fiscal relates to how quickly and accurately the
policy for the next one year. Here market reacts to new information. New
government presents their proposed data are constantly entering the market
revenue and expenses for the place via economic reports, company
coming financial year. A significant announcements, political statements, or
portion of market sentiments are public surveys. If the market is
attached with this event. i.e. it is informational efficient then security prices
being considered for study. The date adjust rapidly and accurately to new
when budget was presented is information. According to this hypothesis,
considered as event date. Returns a security prices reflect fully all the
week before and after the event are information that is available in the market.
considered. Since all the information is already
incorporated in prices, a trader is not able
2. Change in financial year
to make any excess returns. Thus, EMH
A new financial year starts from 1st proposes that it is not possible to
April. Prominent change in books of outperform the market through market
accounts of every company has a timing or stock selection. However, in the
significant effect on every financial context of financial markets and
transaction took place. Therefore, it particularly in the case of equity market

47
seasonal component have been recorded. 30 index is used as a representative of stock
They are called calendar anomalies (effects) market. Tools like descriptive statistics are
in literature. Presence of seasonal anomalies used to understand relationship between
contradicts the concept of EMH. stock exchange and seasonal anomaly. It
has found that there is no significant
On this one of the studies carried out by
presence of anomaly returns over exchange
Ash Narayan Sah () illustrates that Nifty
returns. Khanna (2014) observes the day of
returns do get effected by seasonal
the week effect over stock markets. Sensex
abnormalities. Here he had found the effect
index is taken and returns of it are studied.
of days of week, weekend effect and
The daily stock price data of the Sensex has
seasonality returns effect over S&P Nifty 50.
been taken for the period of January 31,
Nifty 50 is being taken as a representative
2006 upto December, 31 2010. For returns of
of stock market. The monthly data on S&P
working days, logathim returns are used.
Nifty for the period from April 1997 to
For further analysis ARIMA
March2009 is collected from NSE Website.
(Autoregressive Integrated Moving
He has used statistical tools i.e. descriptive
Average) model has been constructed,
statistics and Annova to examine its effect.
descriptive statistics are used. Study
The study concluded daily and monthly
observed that there is an existence of
seasonality over Nifty 50. The presence of
relationship between days of week returns
seasonality effect over nifty is statistically
effect over stock exchanges.
significant. The results established that the
Indian stock market is not efficient and Sarma(2004) studies the effect of seasonality
investors can improve their returns by on stock exchanges of emerging markets.
timing their investment. Three indexes are used in the study
SENSEX, NATEX, BSE 200. Tools like
Kumar & Ravidarshini (2010) conducted a
descriptive statistics are used. In order to
study to examine April Anomaly And
check test hypothesis are used. Data is
return predictability in stock market. The
collected for a period, January 1st 1996 to
effect is studied over 5 indices. The returns
August 10th 2002. In study different
were collected for a period of 10 years. The
patterns are made for investment strategies
indices used were S&P CNX Nifty, CNX
using seasonality effect in order to make
Nifty junior, CNX Midcap, CNX IT and
greater gains. Study observed presence of
Bank Nifty. In study statistical tool like
seasonality in stock exchanges return.
descriptive statistics, dummy variable
Brooks & Persand (1999) studied
regression, Mann-Whitney U test, Kruskal-
seasonality in Southeast Asian stock
Wallis test is used and hypothesis is tested
markets. The study examine day of week
whether there exist effect of seasonality
effect over different stock exchanges. South
over stock exchanges. This study concluded
Korea, Malaysia, the Philippines, Taiwan
that there is presence of statistically
and Thailand stock exchanges are
significant effect over Nifty. January
considered. Market risk, proxied by the
Anomaly is specifically found over
return on the FTA World Price Index, is not
exchanges returns.
sufficient to explain this calendar anomaly.
Kumar & Jawa (2017) studies the effect of Although an extension of the risk-return
calendar events on stock market specifically equation to incorporate interactive seasonal
Indian Market. It studies the day of week, dummy variables can explain some
weekend and month of the year returns significant day-of-the week effects, market
effect over Nifty. The study uses the Arch risk alone appears insufficient to
Model and uses Hypothesis to check the characterise this phenomenon. Data is
test. The study concluded the presence of collected from 31 December, 1989 to 19
seasonal abnormalities over stock January, 1996 ( a total of 1581 observations).
exchanges returns. Harish & Sathyanaryana Tools like CAPM Model and market model
(2017) examined the calendar month is used as statistical tools to analyse the
anomaly returns over Sensex. For this, BSE data.

48
Archana et.al. (2014) conducted a study Sun Pharma, Tata Steel and Wipro. Data is
analyzing calendar effect and stock split collected for a period from January 2000 to
effect over stock exchanges return. The October 2015. Tools used to analyse the
study measures the existence of market data z- test for day of month effect and t-
anomalies with respect to BSE SENSEX test for day of week effect and week of
Index from 2008 – 2012. The closing price month effect over stock exchanges. Study
and turnover of BSE index for a period of found seasonal effects present in Indian
5years has been taken for the study. Daily stock Markets. It found day of the month
return, average return per month , average and week of the month effect to be present.
return per day for all 5 years were tabulated For securities tested, only WIPRO did not
for analysis. The closing price of the have any seasonality. Rest other stocks have
companies before and after stock split was seasonal effects in their returns.
taken to understand the impact of stock
Angelovska (2013) does an econometric
split on the market movements. For the
analysis of market anomaly - Day of the
study of stock split of 5 companies from
week effect on a small emerging market.
different industries are selected, say Bajaj
The selected stock exchange index is MBI10
Corporation, HDFC Bank,Jindal steel,
Index (Macedonian Stock Market). Data is
Oriental Hotels, Tata steel. Menahem (2004)
collected for a period January 4, 2005 to
conducted a study and in it he examines the
December 31, 2009.Tools like descriptive
effect of end of month returns over stock
statistics, regression and Annova and Garch
market. The data is collected daily returns
model are used to analyse this data. The
for the period from July 1962 through
study concludes returns are not evenly
December 1993, which includes a total of
distributed across days of the week. So
7,927 daily observations. They employ
there is presence of seasonality in stock
three indices in our analyses: the NYSE
exchanges. Investors can use these finding
value-weighted index, the NYSE equally-
to have abnormal gain.
weighted index, and the S&P 500 index.
Tools like GARCH Model are used to After studying the vast literature, it has
analyse the data. been observed that lots of work has been
done on seasonality in India and abroad.
Aggarwal & Tandon (1994) observe the
The results of the above-mentioned study
seasonal anomaly over different stock
are mixed. Some studies support the
exchanges. Study is conducted over 18
existence of seasonality and some studies
countries stock exchanges. Research paper
don‟t support it. Moreover, none of the
studies the day of week effect, turn of
studies were found that analysed the events
month effect, December end month effect ,
like Budget announcement and Diwali
month of year effect (January month). Data
festival. This study tries to understand and
is collected for a period from 1971 to 1987
analyse the impact of these events on the
on 18 exchanges. The countries taken into
returns of Indian stock market.
consideration are :- Australia , Brazil,
Belgium , Canada , Denmark, France , RESEARCH DESIGN AND METHODS
Germany, Hong Kong , Italy, Japan, The research study is conclusive in nature
Luxembourg , Mexico, Netherland, New as its motive to find the relationship
Zealand, Singapore, Sweden, Switzerland, between seasonality effects over stock
UK, USA. Desai & Josh (2015) see the exchanges. Nifty 50 has been taken as
seasonal anomalities in stock exchanges reference for stock exchanges returns. The
with respect to some blue chip stocks. population contains all the returns given by
Research is made to form trading strategies nifty 50 in pre-described duration..S&P
over following stocks in order to gain at the CNX Nifty 50 has been taken as sample.
period of seasonality. Study is made on Last 10 years returns have been taken. It
S&P Nifty Index, BSE Sensex, BSE 100 and comprises returns of nifty for the duration
stocks ACC, BHEL, Bank of India, Colgate, of 10 years from 2007-2017.Judgment
GAIL, Infosys, ONGC, RIL, SBI, Sesa Goa, sampling technique i.e. non random

49
sampling technique has been used. Table 2: Descriptive Results
Secondary data of stock exchanges i.e. The
Descriptive Bef. Diwali Aft Diwali
data has been collected from the national
Mean -0.003280093 -0.0460674
stock exchange website for nifty returns for
Standard
the period of 10 years i.e. 2007-2017.The
Error 0.001860014 0.0477296
method of data analysis used in this
Median -0.003248226 0.00040608
research work is the descriptive statistics
and paired sample t-test. In order to test the Standard
hypotheses concerning the relationship Deviation 0.004159117 0.10672662
between the dependent and independent Sample
variables was used. Variance 1.72983E-05 0.01139057
Kurtosis 1.721006092 4.96778972
Hypotheses formed Skewness -0.045656771 -2.2271945
H1: There is no significant difference in Range 0.011679449 0.24242721
nifty returns before and after Diwali event Minimum -0.009158677 -0.2368332
Maximum 0.002520773 0.00559403
H2: There is no significant difference in
Sum -0.016400467 -0.2303369
nifty returns before and after Budget event
Count 5 5
H3: There is no significant difference in Confidence
nifty returns before and after change in Level(95.0%) 0.005164226 0.13251861
financial year.
H4: There is no significant difference in Interpretation
Nifty returns before and after change in Avgbd = average before Diwali
calendar year. Avgad = Average after diwali
DATA ANALYSIS AND INTERPRETATION Paired sample statistics (table 1) indicates
The impact of 4 events i.e. budget week, that there are 5 observations before and
Diwali week, change in financial year, after the diwali event. It also shows their
change in calendar year has been assessed. mean and standard deviation. Pearson
The Nifty 50 is used as a representative of correlation (0.7790) indicates towards
stock exchanges. For analysis, logarithmic positive correlation between both variables.
returns of closing prices on 5 days before Table 2 gives the mean value of both the
and after the event have been assessed variables and standard error which is used
using MS- Excel. The nifty returns are being in computing both the test statistic and the
assessed using paired sample t-test. upper and lower bounds of the confidence
The first hypothesis states that there is no interval. The p-value is 0.433 which is
significant difference on Nifty returns greater than 0.05. This indicates that the
before and after the Diwali event. The null hypothesis is accepted and samples are
duration for analysis is taken from 2007 to not statistically significant. Therefore, there
2017. Paired sample t-test have been used to is no significant difference on Nifty 50
compute the results in excel and their returns due to this event.
results are as under : The second hypothesis states that there is
Table 1: Paired sample t-test no significant difference on Nifty returns
before and after the budget event. The
Hypothesized Mean Difference 0.000 duration for analysis is taken from 2007 to
Df 4 2017. Paired sample t-test has been used to
t Stat 0.869795053 compute the results in excel and their
P(T<=t) one-tail 0.216747136 results are as under:
t Critical one-tail 2.131846786
P(T<=t) two-tail 0.433494271
t Critical two-tail 2.776445105

50
 Table 3: Paired sample t-test before and after the change in financial year
event. The duration for analysis is taken
Hypothesized Mean Difference 0
from 2007 to 2017. Paired sample t-test has
Df 4 been used to compute the results in excel
t Stat 1.638453 and their results are as under:
P(T<=t) one-tail 0.088335
Table 5: Paired sample t-test
t Critical one-tail 2.131847
P(T<=t) two-tail 0.17667 Hypothesized Mean Difference 0
t Critical two-tail 2.776445 Df 4
t Stat 1.544942
Table 4: Descriptive Results P(T<=t) one-tail 0.098624
t Critical one-tail 2.131847
Descriptive Bef. Aft.
Budget Budget P(T<=t) two-tail 0.197248
Mean 0.056024 -0.07611 t Critical two-tail 2.776445
Standard Error 0.05594 0.074383
Median -0.01592 -0.00164 Table 6: Descriptive Results
Standard Deviation 0.125085 0.166326 Descriptive Bef. Fin Aft. Fin
Sample Variance 0.015646 0.027664 Year Year
Kurtosis 0.048212 -1.47833 Mean 0.09559 -0.00187
Skewness 1.188333 -0.49423 Standard Error 0.064413 0.002198
Range 0.288628 0.410615 Median 0.048222 -0.00093
Minimum -0.03916 -0.3003 Standard
0.144031 0.004914
Deviation
Maximum 0.249467 0.110315
Sample Variance 0.020745 2.42E-05
Sum 0.280119 -0.38055
Kurtosis 3.713214 -2.74221
Count 5 5
Skewness 1.891449 -0.25299
Confidence
0.155314 0.206521 Range 0.350023 0.010892
Level(95.0%)
Minimum -0.00528 -0.00763
Interpretation Maximum 0.344742 0.003261
Avgbb = average before budgetAvgab = Sum 0.477949 -0.00937
Average after budget Count 5 5
Confidence
Paired sample statistics (table 3) indicates 0.178838 0.006102
Level(95.0%)
that there are 5 observations before and
after the budget event. It also shows
Interpretation
variable‟s mean and standard deviation.
Pearson correlation (0.2593) indicates Avgbf = average before financial year
towards positive correlation between both Avgab = Average after financial year
variables. Table 4 gives the mean value of Paired sample statistics (table 5) indicates
both the variables and standard error which that there are 5 observations before and
is used in computing both the test statistic after the budget event. It also shows
and the upper and lower bounds of the variable‟s mean and standard deviation.
confidence interval. The p-value is 0.088 Pearson correlation (0.614) indicates
which is greater than 0.05. This indicates towards highly positive correlation
that the null hypothesis is accepted and between both variables. Table 6 gives the
samples are not statistically significant. descriptive value of both the variables and
Therefore, there is no significant difference standard error which is used in computing
on Nifty 50 returns due to this event. both the test statistic and the upper and
The third hypothesis states that there is no lower bounds of the confidence interval.
significant difference on Nifty returns The p-value is 0.197 which is greater than

51
0.05. This indicates that the null hypothesis Paired sample statistics (table 7) indicates
is accepted and samples are not statistically that there are 5 observations before and
significant. Therefore, there is no significant after the budget event. It also shows
difference on Nifty 50 returns due to this variable‟s mean and standard deviation.
event. Pearson correlation (-0.471) indicates
towards negative correlation between both
The fourth hypothesis states that there is no
variables. Table 8 gives the mean value of
significant difference on Nifty returns
both the variables and standard error which
before and after the change in calendar year
is used in computing both the test statistic
event. The duration for analysis is taken
and the upper and lower bounds of the
from 2007 to 2017. Paired sample t-test has
confidence interval. The p-value is 0.176
been used to compute the results in excel
which is greater than 0.05. This indicates
and their results are as under:
that the null hypothesis is accepted and
Table 7: Paired sample t-test samples are not statistically significant.
Hypothesized Mean Therefore, there is no significant difference
Difference 0 on Nifty 50 returns due to this event.
Df 4
1.64085561
CONCLUSION
t Stat 6
P(T<=t) one-tail 0.08808567 Seasonality study is one of the most tested
2.13184678 studies in the stock market. Investors are
t Critical one-tail 6 quite eager to know that whether there are
P(T<=t) two-tail 0.17617134 some possibilities on these events to earn
abnormal profits and beat the market. The
2.77644510
current study tried to analyse the impact of
t Critical two-tail 5
four major events on the returns of Indian
stock market. The events analysed are
Table 8: Descriptive Results
budget, Diwali, change in financial year,
Descriptive Bef Cal year Aft Cal Year change in calendar year. The results of the
Mean 0.001499885 -0.00243 research indicated that these have no
Standard significant effect on nifty 50 returns. The
0.000251818 0.002265 results thus do not support the existence of
Error
Median 0.001746091 -0.00073 seasonality in the Indian stock market as
Standard there was no significant difference in the
0.000563083 0.005064 returns of nifty before and after the four
Deviation
Sample events considered in the study. So it can be
3.17062E-07 2.56E-05 concluded that investors cannot take the
Variance
advantage of seasonality to book abnormal
Kurtosis 2.615658811 3.437266
returns. It was also found that Diwali and
Skewness -1.661917021 -1.80551
Change in calendar year events have an
Range 0.001369127 0.012653 inverse relationship with Nifty returns as
Minimum 0.000553578 -0.01111 they have negative correlation between
Maximum 0.001922704 0.00154 them. While Budget announcement and
Sum 0.007499424 -0.01214 change in financial year events have direct
Count 5 5 relationship with Nifty returns as there
Confidence exists a positive correlation between them.
0.00069916 0.006288 Overall, the study doesn‟t support seasonal
Level(95.0%)
effects.
Interpretation
Avgbc = average before calendar year
Avgac = Average after calendar year

52
REFERENCES
Agrawal, Anup, and Kishore Tandon(1994), Brooks, Chris, and Gita Persand (2001),
“Anomalies or Illusions Evidence from “Seasonality in Southeast Asian Stock
Stock Markets in Eighteen Countries,” Markets: Some New Evidence on Day-of-
Journal of International Money and Finance, the-Week Effects,” Applied Economic Letters
Vol. 13(1), pp.83-106. Vol.8 (3), pp.155-158 DOI:
10.1080/13504850150504504
Ariel, Robert. A.(1990), “High Stock Returns
before Holidays: Existence and Evidence on Choudhary, K and Sakshi Choudhary
Possible Causes,” Journal of Finance, Vol. (2008), “Day-of-the-Week Effect: Further
45(5),pp.1611-26. Empirical Evidence”, Asia-Pacific Business
Review, Vol. 4(3), pp.67-74
Sah, A.N. (2008). Stock Market
Seasonality: A Study of the Indian Stock Dash, Mihir and Sabharwal, Mohit and
Market. NSE Research Paper. Available at Dutta, Anirban, Seasonality and Market
www.nseindia.com/content/research/res_pa Crashes in Indian Stock Markets (March 14,
per_final228.pdf 2011). Available at SSRN:
https://ssrn.com/abstract=1785112 or
Bodla, B. S.; Jindal, Kiran. (2006) Seasonal
http://dx.doi.org/10.2139/ssrn.1785112.
Anomalies in Stock Returns: Evidence from
India and the US Decision (0304-0941).
Jan-Jun2006, Vol. 33 Issue 1, pp.163-178.

***

53