FIN460: Investment Management

Group Assignment
Section 3
Faculty: Dr. Md. Taslim Uddin

Report on
Presence and Sources of Contrarian Profits in the
Bangladesh Stock Market

Group-4:
Abdullah Al Saif 1930419
Mohammed Shoaib 1930258
Kazi Muhammad Husain 1921681
Afsana Mimi 1 1921281
Jannatul Fardous Bushra 1920216

Date of Submission: 07.12.22

1.Statement of the Problem

The paper focusses on presence of Contrarian and momentum profits,

their sources, robustness of results to various market factors and
changes in behavior of sources and such profits.
Extensive Research on Contrarian and momentum profits for
developed markets have been done, in comparison these profits for
frontier and emerging markets had been largely ignored. Although
in recent years a few studies have been done on Contrarian and
momentum returns Chowdhury, Hassan, and Kabir (2007) and Hossan
and Park (2013), they do not attempt to find the sources of
contrarian/momentum profits.
Bangladesh’s market falling under the category of frontier market
and having hardly any studies on momentum and contrarian profits,
while these profits have shown return regularities and different
factors may affect these profits, gives us an importance to
investigate these two profits for institutional investors. Finding
out the sources would help regulators and policy makers to improve
overall efficiency of the market.
This study could also help foreign investors device investment
strategies moreover it can help investors having short term
investment horizon.

2.Research Questions

Unlike other studies our article takes into account a short-term

horizon of 1 week to medium-term horizon of 8 weeks.
Even though a few investigations have been done on the presence
and source of momentum/contrarian profits in the Bangladesh, but
their finding suffered limitations like:

• Did not consider the size of firm

• Did not use SMB (small minus big)
• Did not use sentiment analysis to check robustness of findings
and asset pricing.
The study incorporated all these factors for the findings and it
aimed to answer these 5 questions:
1. Are there any contrarian and momentum profits in the
Bangladesh stock market?
2. Are the contrarian and momentum profits still present after
market risk factors are introduced?
3. What are the sources of such profits?
4. How does firm size affect contrarian and momentum profits?
5. How does the market learn over time?

3.Data and Methodology

The strategy used in this research to create portfolios with a
weighted relative strength scheme is taken from Lo and Mackinlay
(1990).
The contrarian/momentum profits are broken down into three
components using Jagadeesh and Titman's (1995) methodology:
compensation for cross-sectional risk, lead-lag impact in time
series with regard to the common factor, and time series pattern
of stock returns.
By taking market risk variables into account, the robustness of
outcomes is examined.

Data

Historical Data of period January 2002 to August 2013 from Thomson

DataStream was used to collect the weekly stock price index and
market capitalization data, Historical data being messy and having
missing values needed cleaning because of which some firms had to
be dropped this due to frontier market being recent phenomenon in
Bangladesh.
Only stocks having regular activity and held up for the whole study
period was used, a total of 178 firms were in the final dataset
after Data cleaning. 137 Firms were used for size-related analysis
as it needed firms that traded frequently.
Methodology

Construction of Portfolios

Using WRSS of Lo and Mackinlay (1990) a riskless portfolio was

made first. Performance of a fixed number of past months/weeks was
used to construct the portfolio after which the performance was
seen for the same number of months/weeks (formation period =
holding period).
Formation and Holding period of 1,2,3,4, and 8 weeks used to find
the stocks that did better and worse than the market return. In
the portfolio stocks with positive return over the formation period
are bought and stocks with negative return are sold.
Stocks are termed as winners or loser based on their performance,
Weight is assigned to a stock based on their performance, stocks
having higher(lower) return gets a higher(lower) weight.
Throughout the tracking period all the stocks were maintained at
a constant weight. In the study period each stock weight was
assigned a weight using
1
𝑊𝑖,𝑡 = 𝑁 (𝑟𝑖,𝑡−1 − 𝑟̅𝑡−1 ) ,
(1)

Where 𝑟𝑖,𝑡−1 is the return of stock i at time t-1, N is the number

of stocks at period t-1, and is the market return at time t-1.
Thus, the total weight of the portfolio becomes zero if individual
stock weights are added. The momentum or contrarian profit, 𝜋𝑡 , is
given by:
1
𝜋𝑡 = 𝑁 ∑𝑁
𝑖=1 𝑟𝑖,𝑡 (𝑟𝑖,𝑡−1 − 𝑟̅𝑡−1 )
(2)
Short to medium run trading times were used by constructing
portfolio by considering the performance j =1, 2, 4, 6 and 8 weeks
(formation or ranking period), after which the performance of the
portfolio is evaluated during the next 1, 2, 4, 6, and 8 weeks.
This duration is called the evaluation, holding or tracking period.
After the portfolio is made, the cumulative return in the holding
period is calculated. Respective portfolio’s momentum/contrarian
profits in the evaluation period k = 1, 2, 4, 6, and 8 weeks is
given by:
𝑁𝑗
𝜋𝐽,𝑡 (𝑘) = ∑𝑖=1 𝑊𝑖.𝑡 𝑟𝑖,𝑡 + 𝑘 ,
(3)
where J = L (loser portfolio), W (winner portfolio), C (contrarian
portfolio), 𝑊𝑖,𝑡 is the weight of respective stocks in the
portfolio, and Nj is the number of stocks in the portfolio during
the ranking (formation) period. The weight of individual stocks
does not change during the evaluation (holding) period.
Systematic risk and Contrarian/Momentum Profits

Winner and loser portfolio return may change because of evolution

in risk. So, to test the results' robustness, Small minus Big (SMB)
and market return are employed for a robustness assessment of
momentum and contrarian profits since market capitalization and
return data are trustworthy and readily available.
Since Sentiment element splays a significance in asset pricing.
TRIN ratio is used as a stand-in for market sentiment. Robustness
of the contrarian/momentum profit was calculated using the model:

𝜋𝐽,𝑡 = 𝛼𝐽 + 𝛽𝐽 𝑟𝑚,𝑡 + 𝑌𝐽 𝑟𝑆𝑀𝐵,𝑡 + 𝛿𝐽 𝑇𝑅𝐼𝑁𝑡 + 𝑒𝑡

(4)

where 𝜋𝐽,𝑡 is weekly WRSS, winner and loser portfolio return.

Decomposition of Contrarian/Momentum Profits

The decomposition of momentum/contrarian profits given by

Jagadeesh and Titman(1995) is:

𝜋 𝑚 = 𝜎𝜇2 + 𝛿𝜎𝑓2 + Ω
(5a)

𝜋 𝑐 = −𝜎𝜇2 − 𝛿𝜎𝑓2 − Ω,
(5b)

where 𝜋 𝑚 and 𝜋 𝑐 are momentum and contrarian profit, respectively,

and 𝜎𝑓2 is the variance of the factor (market portfolio return).
Jagadeesh and Titman develop the following framework to find the
sources of momentum and contrarian profits. They estimate
𝑟𝑖,𝑡 = 𝜇 + 𝑏0,𝑖𝑓𝑡 + 𝑏1,𝑖𝑓𝑡−𝑘 + 𝜀𝑖.𝑡
(6)

Where 𝑟𝑖,𝑡 is the return of individual stock i at time t; 𝑓𝑡 is the

market return (equally weighted) at time t, which happens to be
the common factor for all the stocks; 𝑓𝑡−𝑘 is the market return
(equally weighted) during t-k period; k is the observation period;
and 𝑏0,𝑖 and 𝑏1,𝑖 are the estimated parameters. From this factor model
we can estimate the following components of contrarian/momentum
returns:
Cross-sectional risk component:
1
𝜎𝜇2 = 𝑁 ∑𝑁
𝑖=1(𝜇𝑖 − 𝜇̅ ),

Lead-lag effect component:

1
𝛿 = 𝑁 ∑𝑁 ̅ ̅
𝑖=1(𝑏0,1 − 𝑏0 ) (𝑏1,𝑖 − 𝑏1 ),

Time-series pattern component:

1
Ω = 𝑁 ∑𝑁
𝑖=1 𝐶𝑜𝑣(𝜖𝑖,𝑡 , 𝜖𝑖,𝑡−1 )

where 𝜇𝑖 is the intercept of the regression for an individual stock;

𝑏0,1 and 𝑏̅0 are the regression coefficient of the first factor and
cross-sectional mean of regression coefficients, respectively; 𝑏1,𝑖
and 𝑏 ̅1 are the regression coefficient of the lagged factor and
cross-sectional mean of that, respectively; and is the error-term
of the regression equation.
Equations of Cross-sectional risk, lead-lag effect, time-series
pattern along with the decomposition equation gives us three
components of Contrarian/momentum phenomenon:

• Profits related to cross sectional Variance of expected

returns
• Profits related to time difference in reacting to a common
factor
• Profits related to stock price adjustment to idiosyncratic
information.
4.Findings

Presence of Contrarian/Momentum Profits

Loser portfolio continued to produce significant negative
returns for all strategies indicating strong contrarian profits.
Winner portfolio showed no significant returns in short term
investment (4 weeks) but for longer investments of 6-8 weeks,
which is evidence pointing to momentum profits.
The WRSS portfolio having zero weight by construction shouldn’t
have produced profits, but all strategies used on it showed
notable negative profits which proofs the existence of
contrarian profits in the market. Resulting in an easy
investment decision of short selling the losing portfolio and
buying the winning portfolio.

Effect of Firm Size on Contrarian/Momentum Profits

Loser portfolio for all 5 strategies and firm sizes produced
negative profits, while winner portfolios in medium and large
sized firms had momentum profits.
For the WRSS portfolio, small firms had higher return than
medium and large firms, with large firms having the least return
showing definite behavior against medium and small firms.

Robustness Check for the presence of Contrarian and Momentum

Profits
WRSS and loser portfolios produced significant profits even
after risk was taken into account, also the constants were still
negative, implying existence of Contrarian profits in the
market.
It was found that significant return remained even after market
risk was taken into account.
Sources of Contrarian/Momentum Profits
Time series pattern was seen as the most important source of
momentum/contrarian profits for every sub period, but this was
due to unreliable information and noisy trading habit of
investors resulting in mispricing due to firm-specific
information.
Overall, the study found that firm-specific information
dominates as the source of contrarian profits which matched
finding from other studies.
Lead-lag effect on small firms for contrarian profits also
showed strong impact, this happens due to slow information
consumption by the smaller firms compared to large firms.
Small firms were found to be the main contributor of
contrarian/momentum profits in Bangladesh’s stock market.
Effect of Time-Series Momentum on Cross-sectional Momentum

From the table we can see that in Panel A strong and positive
influence of time-series momentum on cross-sectional momentum
profits, especially in the period 2009-2013. Good intercept
indicates good presence of contrarian profits even after time-
series momentum has been taken into account.
Panel B introduced other risk factors like SMB , Market return
and TRIN ,the results are still similar , intercept is fairly
significant in all the periods, proving that additional risk
factors and time-series momentum cannot determine eh reason for
cross-sectional contrarian profits

5.Conclusion
In this article, the Bangladesh stock market from January 2002
to August 2013 is evaluated for the presence and source of
momentum and contrarian gains. For a short length of time of one
to eight weeks, contrarian gains are visible on the Bangladesh
stock market. In comparison to large firms, the size of
contrarian profits is significantly greater for small and
medium-sized businesses. This period of opportunity for making
money in the opposite direction vanished after the stock market
crash in December 2010. The most recent phenomenon of a lack of
momentum or contrarian gains does not indicate that the market
has evolved into an efficient one.
The development of a trading strategy to capitalize on the
presence of contrarian profits in the Bangladesh stock market
has already been covered. Academics may employ additional
methods to identify momentum and contrarian profits. A serious
problem is any form of inefficiency, such as the existence of
contrarian profits.
Data on transaction costs that become available in the future
could reveal the true picture of contrarian profits. Future
studies may provide additional insight into such profits as more
data becomes available in the years to come. To ensure that such
profits are not the result of market manipulation, policymakers
and regulators must collaborate. To ensure that such profits are
not the result of market manipulation, policymakers and
regulators like the SEC must cooperate. Otherwise, regulators
may face a significant challenge in detecting manipulation. Due
to the lack of necessary data, it is impossible to consider all
the issues that are important and relevant to contrarian
profits.