European Journal of Business and Management www.iiste.

org
ISSN 2222-1905 (Paper) ISSN 2222-2839 (Online)
Vol.8, No.7, 2016

Portfolio Optimization Analysis with Markowitz Quadratic

Mean-Variance Model
Dr. Ihsan Kulali
Bahçeşehir University, Turkey

Abstract
In this study, Markowitz mean-variance approach is tested on Istanbul Stock Exchange (BIST). 252 days of data
belonging a year of 2015 are analyzed. First, a hypothetical portfolio is created. It involves ten securities with
equal weights. They are chosen from three different industries to minimize risk of portfolio. However, the
number of securities is not adequate for a well diversified portfolio alone. Markowitz model takes into account a
relation between return on financial assets investing in portfolio. In empiricial analysis, I followed mean-
variance model and created many portfolios. The model adjusted them as a minimum variance for a given
expected return. Investors choose any of them as their risk preferences. Because they are all efficient. My
optimal portfolio is constructed by eight assets with different weights. It provides more return comparing with a
portfolio with equal shares of ten stocks.
Keywords: Markowitz, mean-variance approach, modern portfolio theory, efficient frontier

1. Introduction
Risk and the expected return are the main parameters of any kind of investment. After emergence of capital
markets, individuals have a choice to earn money from different investment area. Investors are generally
accepted as risk averse. They are trying to minimize risk while maximizing their return as well. In that sense,
investment strategies of individuals heavily depend on how much risk they can take to achieve expected return.
Risk is measured by deviation from expected return and when the investor takes more risk, its profit or
loss is larger as well. To avoid risk, diversification of financial assets is the main investment strategy until
Modern Portfolio Theory of Harry Markowitz. Investors try to select securities with small variance and to add
number of securities in their portfolios. However, this kind of diversification ignores a relation between return on
financial assets. Modern Portfolio Theory asserts that an individual asset risk itself may be different when it is a
part of a portfolio. This model takes into account co-movement of stocks. Portfolio risk is measured by
covariance between return on financial assets. In that sense optimal portfolio is selected by minimizing variance
at given expected return or maximizing expected return at given variance. This minimum variance values of
given expected return constructs an efficient set and individuals can choose any of optimal portfolios as their risk
preferences.
The aim of this paper is briefly to test mean-variance model on Istanbul Stock Exchange (BIST) and to
select optimal portfolios. The rest of the paper is organized as follows. In section 2, Modern Portfolio Theory is
briefly introduced. In section 3, mean – variance approach of Markowitz is used in empirical analysis to choose
optimal portfolios from efficient frontier. Many statistical techniques are used in this section. Finally, section 4
concludes the paper.

2. Modern Portfolio Theory

Which portfolio is the best? This question has arisen since from the beginning of stock-market itself (Marling
and Emanuelsson, 2012: 2). All investors have a difficulty to choose best assets among many alternatives. This
problem is subject to portfolio theory (Elton and Gruber, 1997: 1743). However, Markowitz transformed
portfolio selection process as a mathematical optimization problem (Marling and Emanuelsson, 2012: 2).
Markowitz (1952, 1959) can be accepted as a father of Modern Portfolio Theory. He solved portfolio
optimization problem as taking into account the mean and variance of assets. Markowitz formulated theory as
holding constant variance, maximize expected return, and holding constant expected return minimize variance.
This two principles formulation give an efficient frontier which contains all possible optimal portfolio under
given expected return and variance combinations. Then the investors make a choice according to his preference
depending on risk behavior of himself. Markowitz dwelled on portfolio risks rather than individual assets’ risk.
In that sense, stocks could be selected on how much contribute the portfolio risk rather than individual risk value
itself. These interactions between the returns on financial assets are key parameters to create a well diversified
portfolio (Elton and Gruber, 1997: 1746).
Mean-variance model proposes not only diversification but also the "right kind" of diversification for
the "right reason''. Adding many more stocks are not enough for adequate diversification. For example, a
portfolio with same industries’ stocks would not be as well diversified as the same size portfolio with different
industries’ stocks. Similarly to find small variance stocks is not enough to construct diversified portfolio. It is
crucial to avoid investing in securities with high covariances among themselves. Generally, companies in

73
European Journal of Business and Management www.iiste.org
ISSN 2222-1905 (Paper) ISSN 2222-2839 (Online)
Vol.8, No.7, 2016

different industries have smaller covariances than firms within same industry (Markowitz, 1952, 89). In that
sense, Markowitz advanced portfolio diversification a step forward to take into account not only number of
assets but also their covariance relation (Mangram, 2012: 60). Because of the correlation between stocks,
diversification reduces a risk of portfolio but not to eliminate it entirely. Markowitz was the first who considered
the covariances between return on financial assets (Müller, 1988: 128). Modern Portfolio Theory contains many
assumptions about markets and investors. Some of them are as such(Mangram, 2012: 61):
1) Investors are rational which means that they seek to maximize returns while minimizing risk,
2) Investors are risk averse which means that they desire to accept higher amounts of risk only if they are
compensated by higher expected returns,
3) Investors reach all investment related information timely,
4) Investors are able to borrow or lend an unlimited amount of capital at a risk free rate of interest,
5) Investors have a one single period investment horizon,
6) There is no transaction costs or taxes in the market,
7) Markets are perfectly efficient
Consequently, Markowitz optimization methodology calculates mean-variance efficient portfolios. “It
is based on mean-variance analysis, where the variance of the overall rate of return is taken as a risk measure and
the expected value measures profitability” (Müller, 1988: 128).

3. Empirical Analysis
3.1. Data and Formulas
A hypothetical portfolio is created at first. Ten companies are selected arbitrarily from three different industries
to create a diversified portfolio. They have traded in Istanbul Stock Exchange (BIST). They are operating in a
petroleum, pharmaceutical and banking industry.
Table 1. Name and Operations of Companies
Stocks Company Name Industry Sector
1 LKMNH Lokman Hekim A.Ş. Pharmaceutical Hospital
2 RTALB RTA Laboratuvarları A.Ş. Pharmaceutical Laboratory
3 ATPET Atlantik Petrol Ürünleri A.Ş. Petroleum Industrial oil
4 TRCAS Turcas Petrol A.Ş. Petroleum Petroleum and electricity
5 MEPET Mepet Petrol A.Ş. Petroleum Service station
6 PETKM Petkim PetroKimya Holding A.Ş. Petroleum Petrochemistry
7 ISCTR İş Bankası A.Ş. Banking Deposit and credit
8 KLNMA Kalkınma Bankası A.Ş. Banking Deposit and credit
9 SKBNK Şekerbank A.Ş. Banking Deposit and credit
10 VAKBN Vakifbank A.Ş. Banking Deposit and credit
In this study, market return of companies are estimated by using daily returns (adjusted price for US dollar).
They are achieved from the Isyatirim database1. One year period (252 working days) data is used. It belongs to a
year of 2015.
To calculate for stocks daily return; the formula is applied as follows:
(1)
where “Ri” is a daily return of stock i, “Rit” is a closing price of stock i in t date and “Rit-1” is a closing price of
stock i in t - 1 date
To calculate average return of stocks; the formula is applied as follows:
(2)
Where “E (Ri)” is a average return for stock i, “Rit” is a market return in t date, “N” is a number of dates.
The general formulas of expected return for n assets is as belowed:
n
E (rP ) = å wi E ( ri ) (3)
i =1
Where “Σ wi = 1”, “n” is the number of stocks, “wi” is the proportion of the funds invested in stock i, “ri, rp” is
the return on ith stock and portfolio p, and “E (Rp)”the expectation of the variable in the parentheses
To calculate variance of stocks daily return and index return, following historical volatility formula is used:
n
1
s2 =
n -1
å ( Ri - Raverage)
i =1
2
(4)

1 http://www.isyatirim.com.tr/LT_isadata2.aspx. (22.02.2016)

74
European Journal of Business and Management www.iiste.org
ISSN 2222-1905 (Paper) ISSN 2222-2839 (Online)
Vol.8, No.7, 2016

Where “s2” is a variance of daily stock return, “Ri” is a daily return of stock i, “Raverage” is average daily return,
“n” is a sample size (252 days)
To measure how stocks vary together, standard formula for covariance can be used:
1 n
Cov (X,Y) = å[( X i - X ).(Yi - Y )]
n - 1 i =1
(5)

where the sum of the distance of each value X and Y from the mean is divided by the number of observations
minus one. The covariance calculation enables us to calculate the correlation coefficient, shown as:
Cov( X , Y )
Correlation Coefficient = (6)
s X .s Y
where s the standard deviation of each asset. However, if there are more than two financial assets in the portfolio,
then correlation and covariance matrices are needed to solve equations.
To calculate standard deviation of portfolio (position), the following formula is applied:
n n n
sp = å (wi2 .s i2 ) + 2(åå (wi .s i .w j .s J .rij )
i =1 i =1 j =1
(7)

Where “σp” is a standard deviation of portfolio, “σi” is a standard deviation of stocks, “wi” is a weight of stocks in
a portfolio and “ρij” is a correlation coefficient between stocks i and j.

3.2. Empirical Results

In this study, excel functions and data solver are used for all calculation. The calculation of mean –variance
optimal portfolio involves the following steps:
Step 1 – Determining return and standard deviation of stocks (table 2)
Step 2 – Creating correlations matrice (table 3)
Step 3 – Creating covariance matrice (table 4)
Step 4 – Creating variance - covariance matrices with equal weights (table 5)
Step 5 – Calculating the volatility and return of the portfolio (table 6)
Step 6 – Creating variance - covariance matrices with different weights (table 7)
Step 7 – Calculating the volatility and return of the new portfolio (table 8)
Step 8 – Creating efficient frontier (table 9 and figure 1)
Table 2. Risk and return of stocks (2015)
Average Average Standard Standard Variance Variance
Return Return Deviation Deviation (daily) (annualy)
Stocks (daily) (annualy) (daily) (annualy)
LKMNH 0,001295 0,32634 0,023909 0,379538 0,000572 0,009074
RTALB 0,000389 0,098028 0,029844 0,473757 0,000891 0,014139
ATPET 0,006905 1,74006 0,053654 0,85173 0,002879 0,045699
TRCAS 0,003184 0,802368 0,02569 0,407813 0,00066 0,010477
MEPET 0,006855 1,72746 0,060377 0,958456 0,003645 0,057869
PETKM 0,000481 0,121212 0,01948 0,30924 0,000379 0,006024
ISCTR 0,002517 0,634284 0,025382 0,402922 0,000644 0,010227
KLNMA -0,00029 -0,07308 0,044999 0,714342 0,002025 0,032145
SKBNK 0,001527 0,384804 0,023822 0,378169 0,000568 0,009009
VAKBN 0,002315 0,58338 0,027858 0,44223 0,000776 0,01232
We know from the modern portfolio theory that volatility of portfolio is less than sum of individual’s
because of correlation between assets. The relation between assets determines the degree of portfolio risk. When
there is negative (small) correlation between assets, risk of portfolio is smaller and when there is positive (big)
correlation between assets, risk of portfolio is greater. The degree of correlation coefficients value takes + 1 and
– 1.

75
European Journal of Business and Management www.iiste.org
ISSN 2222-1905 (Paper) ISSN 2222-2839 (Online)
Vol.8, No.7, 2016

Table 3. Correlation matrice

LKMNH RTALB ATPET TRCAS MEPET PETKM ISCTR KLNMA SKBNK VAKBN
LKMNH 1,000
RTALB 0,401 1,000
ATPET 0,196 0,130 1,000
TRCAS 0,525 0,425 0,190 1,000
MEPET 0,342 0,271 0,074 0,372 1,000
PETKM 0,520 0,385 0,137 0,535 0,368 1,000
ISCTR 0,615 0,488 0,215 0,649 0,412 0,717 1,000
KLNMA 0,192 0,077 0,041 0,162 0,102 0,173 0,142 1,000
SKBNK 0,466 0,371 0,096 0,549 0,388 0,443 0,634 -0,026 1,000
VAKBN 0,630 0,471 0,189 0,639 0,442 0,731 0,897 0,143 0,620 1,000
As seen from Table 3, all correlations except SKBNK-KLNMA between stocks are positive and not
low enough. Nearly all of them are greater than zero. In that sense, the portfolio is not well diversified.
Covariance measures similarity of assets. It shows us that how stocks move together. As seen from Table 4,
securities move together in certain degree except SKBNK-KLNMA.
Table 4. Covariation matrice
LKMNH RTALB ATPET TRCAS MEPET PETKM ISCTR KLNMA SKBNK VABN
LKMNH 0,00057 0,00028 0,00025 0,00032 0,00049 0,00024 0,00037 0,00021 0,00026 0,00042
RTALB 0,00028 0,00089 0,00021 0,00032 0,00049 0,00022 0,00037 0,00010 0,00026 0,00039
ATPET 0,00025 0,00021 0,00287 0,00026 0,00024 0,00014 0,00029 0,00010 0,00012 0,00028
TRCAS 0,00032 0,00032 0,00026 0,00066 0,00027 0,00027 0,00027 0,00027 0,00027 0,00027
MEPET 0,00049 0,00049 0,00024 0,00058 0,00363 0,00043 0,00063 0,00028 0,00056 0,00074
PETKM 0,00024 0,00022 0,00014 0,00027 0,00043 0,00038 0,00035 0,00015 0,00020 0,00040
ISCTR 0,00037 0,00037 0,00029 0,00042 0,00063 0,00035 0,00064 0,00016 0,00038 0,00063
KLNMA 0,00021 0,00010 0,00010 0,00019 0,00028 0,00015 0,00016 0,00202 -0,00003 0,00018
SKBNK 0,00026 0,00026 0,00012 0,00033 0,00056 0,00020 0,00038 -0,00003 0,00057 0,00041
VAKBN 0,00042 0,00039 0,00028 0,00046 0,00074 0,00040 0,00063 0,00018 0,00041 0,00077
It is very difficult to calculate volatility of portfolio which gets more than two assets for using formula
7. Variance-covariance matrice is a very simple way of measuring risk of portfolio. It is done by multiplied each
assets’ weights and covariance values for each cell.
For example, in Table 5, value of C3 cell is achieved by (0,1) * 0,00057 * (0,1). Value of C13 cell
shows risk contribution of LKMNH stock to the portfolio. Volatility (risk) of portfolio is achieved by sum of
each stocks’ risk. F15 cell shows volatility of portfolio (=sum (C13: L13)).
As seen from the Table 5, our portfolio is created by ten stocks with equal weights (10 %). Sum of
weights of stocks must be equal one.
Table 5. Variance-covariation matrice with equal weight
A B C D E F G H I J K L
1 LKMNH RTALB ATPET TRCAS MEPET PETKM ISCTR KLNMA SKBNK VABN
2 w 0,1 0,1 0,1 0,1 0,1 0,1 0,1 0,1 0,1 0,1
3 0,1 5,69E-06 1,62E-08 7,129E-09 8,03E-09 1,58E-08 1,18E-08 8,96E-09 7,65E-09 5,44E-09 1,1E-08
4 0,1 2,85E-06 2,53E-08 1,840E-08 6,74E-09 1,58E-08 1,08E-08 8,21E-09 3,78E-09 2,69E-09 1,02E-08
5 0,1 2,5E-06 5,19E-09 5,947E-08 7,47E-08 6,2E-09 3,41E-09 4,17E-09 2,88E-09 1,21E-09 3,44E-09
6 0,1 3,21E-06 1,04E-08 8,463E-09 1,71E-08 1,75E-08 7,12E-09 7,12E-09 7,12E-09 7,12E-09 7,12E-09
7 0,1 4,91E-06 2,39E-08 1,158E-08 1,37E-08 2,09E-07 1,56E-07 2,71E-08 1,73E-08 1,54E-08 4,12E-08
8 0,1 2,41E-06 5,37E-09 3,187E-09 3,82E-09 1,15E-08 1,63E-08 1,34E-08 5,32E-09 3,09E-09 8,09E-09
9 0,1 3,72E-06 1,37E-08 1,072E-08 1,23E-08 2,65E-08 2,22E-08 2,27E-08 1,03E-08 6,15E-09 2,41E-08
10 0,1 2,06E-06 2,11E-09 1,013E-09 1,85E-09 5,16E-09 4,16E-09 2,43E-09 3,25E-08 -5,7E-09 -5E-10
11 0,1 2,64E-06 6,94E-09 3,20E-09 4,09E-09 1,86E-08 1,14E-08 7,81E-09 -1,1E-09 -1,6E-09 2,32E-08
12 0,1 4,18E-06 1,63E-08 1,097E-08 1,28E-08 3,37E-08 2,93E-08 2,5E-08 1,12E-08 7,3E-09 3,17E-08
13 1 3,42E-05 1,25E-07 1,34172E- 1,55E-07 3,6E-07 2,73E-07 1,27E-07 9,71E-08 4,11E-08 1,6E-07
14 Daily
15 Variance of portfolio = 0,000410672
16 Standard deviation of portfolio = 0,0202650
17 Return of portfolio = 0,002518
The Mean-Variance model of Harry Markowitz (1952) asserts that investors are risk-averse and
efficient portfolios must be satisfied at least two conditions as “(1) minimize the variance of portfolio return,
given the expected return, and (2) maximize the expected return, given the variance.” (Miao, 2013: 6-7). In that
sense, minimizing the risk or variance of the portfolio is one of the main aims of optimal portfolio analysis. This
subject function of Markowitz mean – variance portfolio can be showed as belowed:
Min σ2 (8)
Where “Wi” and “Wj” are weights of stocks in the portfolio and “Covij” is covariance value between stocks i and
j.
There are mainly two constraints in Markowitz standard quadratic convex optimization problem. They are
written as
(9)

76
European Journal of Business and Management www.iiste.org
ISSN 2222-1905 (Paper) ISSN 2222-2839 (Online)
Vol.8, No.7, 2016

*
where E is the target expected return, E (Ri) is an expected return and
(10)
The portfolio manager determines a target return which must be equal to the expected return. The first constraint
above (formula 9) provides this condition. The second constraint asserts that the weights of the stocks invested in
the portfolio must sum to one (formula 10). A third constraint can be added if short sale restriction sets in
variance problem (formula 11).
Wi ≥ 0, i = 1, ……………, N (11)
If short sale is allowed, it is possible to sell these shares without owning it. In that sense, weight column of
shares can take negative sign if optimal portfolio captures short sale shares. However, in all circumstances, sum
of shares invested in portfolio must be 1.
Table 6. Parameters of Excel Solver
Target cell F 15 (variance of portfolio)
Equal to Minimum (variance of portfolio)
By changing cells $C$2 : $L$2
Constraints $C$2 : $L$2 ≥ 0 (short sale restriction)
$B$13 = 1
$F$17 = 0,027 (expected return)
In this study, all these three constraints are identified by an excel data solver. Optimal portfolio is
achieved by following Markowitz mean – variance model steps. Solutionss are seen from Table 7 and Table 8.
Table 7. Variance- covariation matrice without short sales
A B C D E F G H I J K L
1 LKMNH RTALB ATPET TRCAS MEPET PETKM ISCTR KLNMA SKBNK VAB
2 W 0,053 0 0,143 0,275 0,044 0,196 0,021 0,048 0,221 0
3 0,053 1,59E-06 0 1,89E-06 4,62E-06 1,12E-06 2,51E-06 4,03E-07 5,18E-07 3,09E-06 0
4 0 0 0 0 0 0 0 0 0 0 0
5 0,143 1,89E-06 0 5,84E-05 1,02E-05 1,48E-06 4,01E-06 8,54E-07 6,70E-07 3,84E-06 0
6 0,275 4,67E-06 0 1,024E-05 4,98E-05 3,20E-06 1,44E-05 1,51E-06 3,49E-06 1,62E-05 0
7 0,044 1,13E-06 0 1,483E-06 6,92E-06 6,92E-06 3,69E-06 5,64E-07 5,72E-07 5,36E-06 0
8 0,196 2,50E-06 0 4,010E-06 1,44E-05 3,69E-06 1,45E-05 1,42E-06 1,40E-06 8,881E-06 0
9 0,021 4,04E-07 0 8,542E-07 2,38E-06 5,64E-07 1,42E-06 2,71E-07 1,57E-07 1,73E-06 0
10 0,048 5,17E-07 0 6,70E-07 2,44E-06 5,72E-07 1,40E-06 1,57E-07 4,55E-06 -2,96E-07 0
11 0,221 3,08E-06 0 3,849E-06 2,03E-05 5,36E-06 8,88E-06 1,73E-06 -2,98E-07 2,75E-05 0
12 0 0 0 0 0 0 0 0 0 0 0
13 1
14 Daily
15 Variance of portfolio = 0,000366829
16 Standard deviation of portfolio = 0,0191528
17 Return of portfolio = 0,0027
For different choices of stocks, the investor will get different combinations of expected return and risk.
The set of all possible expected return and risk combinations is called the attainable set. Those risk and expected
return with minimum variance for a given expected return or more and maximum expected return for a given
variance or less are called the efficient set (or efficient frontier). Because an investor wants a high profit and a
small risk, he wants to maximize expected return and minimize variance and therefore he should choose a
portfolio which gives an expected return - risk combination in the efficient set (Marling and Emanuelsson, 2012:
2). While original portfolio contains ten stock with equal weights, under the three conditions shown in Table 7,
new optimal portfolio contains just eight stocks with different weights. This new portfolio involves % 5,3
LKMNH, % 14,3 ATPET, % 27,5 TRCAS, % 4,4 MEPET, %19,9 PETKM, %2 ISCTR, %4,8 KLNMA
and %22,1 SKBNK stocks. It is possible to create many portfolio under the different target expected return. All
these portfolios are efficient and are located over the efficient frontier. I created 11 portfolios a range between
0,00001 and 0,006 expected return.

77
European Journal of Business and Management www.iiste.org
ISSN 2222-1905 (Paper) ISSN 2222-2839 (Online)
Vol.8, No.7, 2016

Table 8. Weight and risk of different expected return portfolios

LKMNH RTALB ATPET TRCAS MEPET
standard deviation return Weights of stocks in a portfolio
Portfolio 1 0,028747169 0,00001 0,00000 0,00000 0,00000 0,00000 0,00000
Portfolio 2 0,017401218 0,0005 0,00000 0,14812 0,00000 0,00000 0,00000
Portfolio 3 0,016682763 0,001 0,09221 0,08072 0,04206 0,00000 0,00000
Portfolio 4 0,016910244 0,0015 0,08725 0,04695 0,07983 0,08823 0,00000
Portfolio 5 0,017625047 0,002 0,07747 0,01638 0,11004 0,17845 0,00936
Portfolio 6 0,018663879 0,0025 0,06394 0,00000 0,13456 0,24982 0,03502
Portfolio 7 0,0191528 0,002700 0,05291 0,00000 0,14274 0,27541 0,04366
Portfolio 8 0,019946118 0,003000 0,03648 0,00000 0,15458 0,31207 0,05604
Portfolio 9 0,023038145 0,004000 0,00000 0,00000 0,19951 0,43011 0,10377
Portfolio 10 0,027356428 0,005000 0,00000 0,00000 0,29370 0,48634 0,20052
Portfolio 11 0,033964434 0,006000 0,00000 0,00000 0,43739 0,23886 0,32374

Table 9. Weight and risk of different expected return portfolios (continue)

PETKM ISCTR KLNMA SKBNK VAKBN
standard deviation return Weights of stocks in a portfolio
Portfolio 1 0,028747169 0,00001 0,00000 0,00000 0,00000 0,00000 0,00000
Portfolio 2 0,017401218 0,0005 0,59839 0,00000 0,12790 0,12559 0,00000
Portfolio 3 0,016682763 0,001 0,44863 0,00000 0,09389 0,24249 0,00000
Portfolio 4 0,016910244 0,0015 0,37016 0,00000 0,07792 0,24967 0,00000
Portfolio 5 0,017625047 0,002 0,29815 0,00000 0,06372 0,24643 0,00000
Portfolio 6 0,018663879 0,0025 0,23156 0,00000 0,05180 0,23330 0,00000
Portfolio 7 0,0191528 0,002700 0,19632 0,02056 0,04751 0,22088 0,00000
Portfolio 8 0,019946118 0,003000 0,14106 0,05741 0,04116 0,20121 0,00000
Portfolio 9 0,023038145 0,004000 0,00000 0,14992 0,00768 0,10901 0,00000
Portfolio 10 0,027356428 0,005000 0,00000 0,01944 0,00000 0,00000 0,00000
Portfolio 11 0,033964434 0,006000 0,00000 0,00000 0,00000 0,00000 0,00000
Markowitz solution asserts that an investor should choose a portfolio from the efficient frontier,
depending on risk averse characteristics (Marling and Emanuelsson, 2012: 2).

Figure 1. Efficient frontier and portfolios

As stated above, efficient frontier shows us minimum – variance portfolios. Given expected return, all
efficient portfolios are located over efficient frontier curve. Investors or portfolio managers can choose any one
of them as their risk preferences. As seen from Figure 1, many portfolios are existed on efficient frontier curve
with different risk and return combinations.

78
European Journal of Business and Management www.iiste.org
ISSN 2222-1905 (Paper) ISSN 2222-2839 (Online)
Vol.8, No.7, 2016

Table 9. Comparison of portfolio return

Investment Daily Return Annual Return Daily standard Annual standard
Alternatives (%) (%) deviation (%) deviation (%)
Portfolio with equal
0,002518 0,634536 0,0202650
weights 0,321697
Optimal Portfolio 1 0,002700 0,680400 0,0191528 0,304041

4. Conclusion
Individuals are mainly two related diversification strategy as minimizing risk or maximize return. Investor aims
to select best diversification. In this paper, I tested workability of portfolio optimization and diversification on
Istanbul Stock Exchange (BIST). I created a portfolio which contains ten stocks from three different industry.
One year daily stock return data is used for the analysis. I followed mean – variance approach of Markowitz
involving best possible combination of expected return and risk. I believe that I constructed efficient portfolios in
my analysis. I made my choice from this efficient set. My investment portfolio which did not contain short sale
was achieved by excel data solver. I compared return and risk of optimal portfolio with an original portfolio with
equal weights of ten stocks. This portfolio involves eight assets with different weights. The return of my optimal
portfolio is greater than an original portfolio with equal weights of ten stocks as seen from Table 9. In that sense,
Markowitz mean-variance approach provide best solutions in many alternatives.

References
Elton, Edwin J. and Martin J. Gruber (1997). Modern portfolio theory, 1950 to date, Journal of Banking &
Finance, 21, 1743 – 1759.
Mangram, Myles E. (2013), A Simplified Perspective Of The Markowitz Portfolio Theory, Global Journal Of
Business Research, Volume 7, Number 1.
Markowitz, Harry (1952). Portfolio Selection, The Journal of Finance, Vol. 7, No. 1, 77-91.
Marling, Hannes and Sara Emanuelsson (2012). The Markowitz Portfolio Theory, November 25.
Miao, Dingquan (2013). Empirical Researches of the Capital Asset Pricing Model and the Fama-French Three-
factor Model on the U.S. Stock Market, Bachelor Thesis in Economics, Division of Economics, The
School of Business, Society and Engineering (EST) Mälardalens University Västerås.
http://www.diva-portal.org/smash/get/diva2:629540/FULLTEXT01.pdf
Müller, Heinz H. (1988). Modern Portfolio Theory: Some Main Results, Astin Bulletin, 18 (2), 127 – 145.

79