Gyutai Kim, Suhee Jung Department of Industrial Engineering, Chosun University, Gwangju, Korea
description
Transcript of Gyutai Kim, Suhee Jung Department of Industrial Engineering, Chosun University, Gwangju, Korea
Powerpoint TemplatesPage 1
Powerpoint Templates
EXPLOITING THE DECISION-MAKING TECHNIQUE TO EXPLORE THE RELATIONSHIP BETWEEN THE
FANCIAL FACTORS AND THE STOCK PREFERECE
Gyutai Kim, Suhee JungDepartment of Industrial Engineering, Chosun University, Gwangju, Korea
Powerpoint TemplatesPage 2
Table of contents
Introduction1
The Decision-Making Framework for a Stock Investment2
Determine the Best Alternative Using the TOPSIS Technique3
Financial Analysis4
To Compare the TOPSIS Result with the Financial Analysis Result5
The Concluding Remarks6
Powerpoint TemplatesPage 3
Introduction
When investors make a decision which stocks to invest, they have to simultaneously take into consideration of a number of financial and nonfinancial factors affecting a stock price.
Suck an investment decision is to some extent extremely difficult to make.
In this paper, we employed the TOPSIS technique with which we considered only the financial factors due to the availability of obtaining relevant data.
Powerpoint TemplatesPage 4
A difference between TOPSIS and existing method◦ The existing method
· Consider only the financial ratios influencing the stock price.
◦ The TOPSIS method· Do grouping all the financial ratio using a factor analysis.
√ The financial ratios usually involve the subordinate relationship among them.
→ total rate of return = ratio of net income to net sales ⅹ total asset turnover ratio
We implemented a comparison analysis for the preference ordering determined by between the general four financial classifications and the TOPSIS.
Introduction
Powerpoint TemplatesPage 5
A Brief Liturature Survey T. C. Wang and J. C. Hsu, “Evaluating of the Business Operation Performance
of the Listing Company by Applying TOPSIS Method,” 2004 IEEE International Conference on System, Man and Cybernetics.
M. Guo and Y. B. Zhang, “A Stock Selection Model Based on Analytic Hierarchy Process, Factor Analysis and TOPSIS,” 2010 International Conference on Computer and Communication Technology in Agriculture Engineering.
T. C. Chu and C. T. Tsao, and Y. R. Shiue, “Application of Fuzzy Multiple Attribute Decision Making on Company Analysis for Stock Selection,” 1996 IEEE.
P. Xidonas and D. Askounis, “ Common Stock Portfolio Selection: A Multiple Criteria Decision Making Methodology and An Application to the Athens Stock Exchange,” Operations Research International Journal, Vol. 9, 2009, pp. 55-79.
I. Ertugrul and N. Karakasogu, “Performance Evaluation of Turkish Cement Firms with Fuzzy Analytic Hierarchy Process and TOPSIS Methods,” Expert Systems with Application, Vol. 36, 2009, pp. 702-715.
Powerpoint TemplatesPage 6
The Decision-Making Framework for a Stock Investment
Select the base factors and collect
data
- Financial data - Non-financial data
Normalize the values of the base
factors
- Minkowski metrics · Manhattan distance · Euclidean distance · Chebyshev distance
Factor analysis with the normalized
data
- Regroup the existing groups into the newly formed groups according to the result of the factor analysis
1,
/1
1
21
pxxd
pn
j
p
jjp
Powerpoint TemplatesPage 7
The Decision-Making Framework for a Stock Investment
Perform a regression analysis
- Group a benefit concept of the factors - Group a cost concept of the factors
Apply the TOPSIS technique
- Rank the preference ordering of the stocks based on the TOPSIS result.
Powerpoint TemplatesPage 8
The Procedure of Comparing result of the two techniques
Calculate the ranking of each ratio
- A 16 financial ratios
Determine the ranking of the financial
analysis
- Calculate the average of ranking of ratios
Compare the financial analysis result with
TOPSIS
- Perform the Spearman’s rank correlate analysis between each ratio and TOPSIS
Compare the ranking of each category with
TOPSIS
- Execute the Spearman’s rank correlate analysis between each category and TOPSIS
Calculate the ranking of each category
- Calculate the average of ranking of ratios by each category
Compare the ranking of each ratio with
TOPSIS
- In order to analyze in detail, execute the Spearman’s rank correlate analysis between each ratio and TOPSIS
Powerpoint TemplatesPage 9
The Selection of the Base Factors and Data Collection for the Alternative Analysis
Base factors
Data collection ◦ The financial statements of each company in eight years for the
communication and broadcasting equipment manufacturing companies (from 2001 to 2008)
Liquidity ratios Leverage ratios Activity ratios Profitability ratios Valuation ratios
- Current ratio - Acid-test ratio
- Debt ratio - Debt-to-equity ratio
- Total Asset Turnover ratio - Fixed Assets Turnover ratio - Inventory Turnover ratio
- Return on Total Asset - Return on Equity - Return on Net Income - Ratio of Ordinary profit - Ratio of Net Profit to Net Income
- Book-Value per Share - Price/Earnings Ratio - Earning per Share - Price on Book-Value ratio
Powerpoint TemplatesPage 10
Use the vectors normalization method with p=2 in the Minkowski’s lp metrics to transform the raw data into the normalized data to compare one with another alternative.
Transform the Raw Data into the Normalized Data
where, i : a company index for i=1,2,…,l j : a year index for j=2001,2002, …,m k : a base factor index for 1,2, …,n xijk : data of the kth factor for company i and period j
l
i
m
jijkijkijk xxr
1 2001
2/ (1)
Powerpoint TemplatesPage 11
Samsung Electronics
YearCurrent Ratio
Acid- Test Ratio
Debt Ratio
...Price/
Earnings Ratio
2008 1.517811 1.192102 0.24789 ... 13.88549
2007 1.538655 1.22966 0.265021 ... 12.73988
2006 1.527155 1.193011 0.279031 ... 13.17394
2005 1.711944 1.36331 0.27603 ... 14.73145
2004 1.605946 1.244249 0.272242 ... 7.496795
2003 1.466771 -1.19697 0.332792 ... 13.55107
2002 1.590061 -1.29062 0.36997 ... 7.918366
2001 1.069899 -0.75871 0.433694 ... 16.70282
Samsung Electronics
YearCurrent Ratio
Acid- Test Ratio
Debt Ratio
...Price/
Earnings Ratio
2008 0.048082 0.039903 0.005511 ... 0.01810
2007 0.048742 0.04116 0.005892 ... 0.01661
2006 0.048378 0.039933 0.006203 ... 0.01717
2005 0.054231 0.045634 0.006137 ... 0.01920
2004 0.050874 0.041648 0.006053 ... 0.00977
2003 0.046465 -0.04007 0.007399 ... 0.01766
2002 0.05037 -0.0432 0.008225 ... 0.01032
2001 0.033893 -0.0254 0.009642 ... 0.02177
The raw data of the factors The normalized data of the factors
Transform the Raw Data into the Normalized Data
Powerpoint TemplatesPage 12
The result of the factor analysis with eigenvector being more than “1”
- Those six factors were newly obtained from 16 independent variables based on the VARIMAX technique.
Factor Analysis
Total Explained Variance
Component
Initial Eigenvalue Extraction Sums of Squared Loadings Rotation Sums of Squared Loadings
Total % Variance % Cumulative Value % Variance % Cumulative Value % Variance % Cumulative
1 4.335 27.091 27.091 4.335 27.091 27.091 3.672 22.953 22.9532 3.096 19.349 46.440 3.096 19.349 46.440 3.062 19.135 42.0883 2.118 13.239 59.680 2.118 13.239 59.680 2.150 13.440 55.5284 1.576 9.851 69.531 1.576 9.851 69.531 2.055 12.841 68.3695 1.061 6.632 76.163 1.061 6.632 76.163 1.189 7.434 75.8036 1.014 6.340 82.504 1.014 6.340 82.504 1.072 6.700 82.5047 .893 5.582 88.0868 .612 3.827 91.9139 .482 3.014 94.926
10 .307 1.918 96.84411 .271 1.693 98.53812 .117 .732 99.27013 .070 .440 99.70914 .028 .175 99.88415 .010 .065 99.94916 .008 .051 100.000
Extraction Method: Principal Component Analysis
Powerpoint TemplatesPage 13
where, i : a company index for i=1,2,…,l, j : a year index for j=2001,2002, …,m k : a base factor index for 1,2, …,n, Aijk : a variable for combining k factors Zijk : kth common factor for the ith company in the jth period Uij : a factor related to only the variable of xij
Wijk : a coefficient of the kth factor for the ith company in the jth period xijk : a normalized value of the kth factor for the ith company in the jth period
ijijkijkijijijijijk UZAZAZAx 2211 (2)
l
i
m
jijkijkijk xWF
1 2001(3)
Calculate the factor value using a principal component analysis
The values of factors were calculated in a linear combination on the basis of the responses of the variables observed. The values of factors which were not observed could be derived in a linear combination using Equation(2) and the values of the factor for a specific year of each company could be estimated with Equation (3).
Powerpoint TemplatesPage 14
The converted factor value
No. Company Year Factor1 Factor2 Factor3 Factor4 Factor5 Factor6
1
Samsung Electronics
2008 -0.11362 0.24278 -0.17695 -0.38225 2.8002 0.16142
2 2007 -0.09888 0.34286 -0.16988 -0.3157 2.89493 0.22562
3 2006 -0.09584 0.39929 -0.19241 -0.26471 2.74001 0.27687
4 2005 -0.10015 0.43644 -0.15838 -0.22327 2.46718 0.3407
5 2004 -0.06387 0.60567 -0.16802 -0.20031 2.64874 0.24976
6 2003 -0.0853 0.37802 -0.52218 -0.16976 1.89792 0.09583
7 2002 -0.06383 0.4729 -0.49821 -0.16516 1.9433 -0.02273
8 2001 -0.09921 0.24594 -0.55912 -0.1616 1.039 0.01778
No. Company Factor1 Factor2 Factor3 Factor4 Factor5 Factor6
1Samsung
Electronics -0.09009 0.390488 -0.30564 -0.23535 2.30391 0.1681563
The arithmetic mean
Calculate the factor value using a principal component analysis
Powerpoint TemplatesPage 15
ijkkijijij FFFy 22110 (4)
where, i : a company index for i=1,2,…,l j : a year index for j=2001,2002, …,m k : a base factor index for 1,2, …,n k : a non-normalized value for the kth factor Fijk : a value of the kth factor for the ith company in the jth period
The Regression Analysis with the Values of the Factors
The main purpose of the work was to discriminate the factors into the group between a benefit concept and a cost concept.
Powerpoint TemplatesPage 16
A VARIANCE ANALYSIS FOR A REGRESSION TEST OF THE AUTOMOBILE PART MANUFACTURING INDUSTRY
Model Sum of SquaredDegrees of Freedom Mean Square F Sig.
1 Regression 0.733512 6 0.122252 61.40051 0.000 Residual 0.111499 56 0.001991Total 0.845011 62
a Predictors: (Constant), Factor6, Factor5, Factor4, Factor3, Factor2, Factor1
b Dependent Variable: Communication and broadcasting equipment manufacturing Stock Price
Coefficients(a)
ModelUnstandardized Coefficients
Standardized Coefficients
t Sig.B Std. Error Beta1 (constant) 0.049599048 0.005622 8.822702 0.000
Factor1 -0.001082429 0.005667 -0.009271801 -0.19101 0.849 Factor2 0.018949373 0.005667 0.162315285 3.343864 0.001 Factor3 -0.014160832 0.005667 -0.12129792 -2.49886 0.015 Factor4 -0.011119761 0.005667 -0.095248912 -1.96223 0.055 Factor5 0.105166183 0.005667 0.900825546 18.55795 0.000 Factor6 0.009299714 0.005667 0.079658874 1.641056 0.106
a Dependent Variable: Automobile Stock Price
The Regression Analysis with the Values of the Factors
Powerpoint TemplatesPage 17
Determine the Best Alternative Using the TOPSIS Technique
(Technique for Order Preference by Similarity to Ideal Solution)
Specified Factor X1
Specified Factor X2
A*
A2
A1
A-
A* : Positive ideal Solution A1 : Alternative plan 1A- : Negative ideal Solution A2 : Alternative plan 2
TOPSIS was developed under concept which the selected alternative is the nearest from the ideal solution and the farthest from the negative-ideal solution.
TOPSIS is the MADM method which select the alternative according to relative closeness to the ideal solution which considered simultaneously a distance about ideal solution and negative-ideal solution.
Powerpoint TemplatesPage 18
ijjij rwv (5)
where, i : a company index for i=1,2,…,l j : a year index for j=2001,2002, …,m vij : a normalized value of the jth factor for the i company wij : a value of the jth factor rij : a value of the jth factor of the ith company
Calculate a weighted-normalized value
It is necessary to convert the values of the factors into the product of the weight and the value.
Powerpoint TemplatesPage 19
Factor1 Factor2 Factor3 Factor4 Factor5 Factor6
Optimus -0.12646 -0.5561 0.154514 -0.24577 -0.399883 0.055055
DONGWON SYSTEMS -0.14434 0.176705 -0.43717 -0.08935 -0.453176 -0.013886
DAIDONG ELECTRONICS -0.05844 0.423976 1.882716 -0.25373 -0.183706 -0.302025
Kedcom -0.12749 -0.44611 -0.48579 -0.14472 -0.373801 -0.169644
LG Electronics 1.008873 0.258364 -0.44158 -0.11787 -0.040187 0.0205329
Huneed Technologies -0.21849 0.083301 -0.36473 0.045934 -0.416999 0.4592988
Samsung Electronics -0.09009 0.390488 -0.30564 -0.23535 2.30391 0.1681563
GS Instruments -0.11746 -0.29834 -0.05752 1.026124 -0.441183 -0.214925
Weight 3.027961 2.713943 2.585995 1.986546 1.8644621 1.0488159
Factor1 Factor2 Factor3 Factor4 Factor5 Factor6
Optimus -0.38291 -1.50922 0.399572 -0.48823 -0.745566 0.0577426
DONGWON SYSTEMS -0.43706 0.479567 -1.13053 -0.17751 -0.84493 -0.014564
DAIDONG ELECTRONICS -0.17695 1.150647 4.868694 -0.50404 -0.342513 -0.316769
Kedcom -0.38604 -1.2107 -1.25625 -0.2875 -0.696938 -0.177925
LG Electronics 3.054828 0.701186 -1.14192 -0.23416 -0.074927 0.0215352
Huneed Technologies -0.66158 0.226075 -0.94318 0.091249 -0.777478 0.4817198
Samsung Electronics -0.27278 1.059761 -0.79039 -0.46752 4.2955529 0.1763649
GS Instruments -0.35566 -0.80968 -0.14875 2.038442 -0.822568 -0.225417
<A foundation factor value>
<A weighted-normalized value>
Calculate a weighted-normalized value
Powerpoint TemplatesPage 20
miJkv
miJkvvvvvA
miJkv
miJkvvvvvA
iki
ikimk
iki
ikimk
,,2,1)(max
,,2,1)(min},,,,{
,,2,1)(min
,,2,1)(max},,,,{
2
121
2
1***2
*1
*
(6)
where, J1 : a benefit concept of the factors J2 : a cost concept of the factors A* : the ideal solution A- : the negative-ideal solution
Construct the Ideal and Negative-ideal Solution
Powerpoint TemplatesPage 21
Factor1 Factor2 Factor3 Factor4 Factor5 Factor6
A* -0.66158 1.150647 -1.25625 -0.50404 4.2955529 0.4817198
A- 3.054828 -1.50922 4.868694 2.038442 -0.84493 -0.316769
Construct the Ideal and Negative-ideal Solution
Powerpoint TemplatesPage 22
Calculate a separation measure
The separation of each company from the ideal and negative-ideal solutions.
l
ijiji
l
ijiji vvSvvS
1
2
1
2** )(,)( (7)
where, Si* : the separation measure from the ideal solution for the ith company
Si - the separation measure from the negative-ideal solution for the ith company
Powerpoint TemplatesPage 23
Calculate a separation measure
No. Company Si* No. Company Si
*
1 Optimus 0.423826 5 LG Electronics 0.641904
2 DONGWON SYSTEMS 0.567794 6 Huneed Technologies 0.545985
3DAIDONG
ELECTRONICS0.359001 7 Samsung Electronics 0.855031
4 Kedcom 0.530545 8 GS Instruments 0.441609
Powerpoint TemplatesPage 24
ii
ii
SS
SC
**
(8)
where, Ci* : the relative closeness of the ith company from the ideal solution
0 ≤ Ci* ≤ 1
if Ai = A-, Ci* = 0
if Ai = A*, Ci* = 1
Calculate the relative closeness the ideal solution
Powerpoint TemplatesPage 25
The analysis of the TOPSIS results
No. Company Ci* Ranking
1 Optimus 0.423826 7
2 DONGWON SYSTEMS 0.567794 3
3 DAIDONG ELECTRONICS 0.359001 8
4 Kedcom 0.530545 5
5 LG Electronics 0.641904 2
6 Huneed Technologies 0.545985 4
7 Samsung Electronics 0.855031 1
8 GS Instruments 0.441609 6
Powerpoint TemplatesPage 26
Financial Analysis
Basic classification
Definition Financial affair ratio
Stability analysis
-The measuring indices of the ability of repay the short-term debt
Current ratio, Quick ratio, Debt ratio, Equity ratio
Profitability analysis
-The evaluating indices of the ability of the produce profit.
Return on total assets, Return on equity,
Sales margin, Ordinary margin, Net profit margin
Activity analysis
-The measuring indices of the physical utilization of the total asset and inventory etc.
Total asset turnover, Inventory turnover, Fixed asset turnover
Market value analysis
-The ratios which are associated with the share price in stock market. -This ratios can measure the company performance because these reflect both the risk and rate of return.
Book value per share, Earnings per share, Price to equity ratio, Price to earnings ratio
Powerpoint TemplatesPage 27
Financial Analysis results
No. Company Ranking
1 Optimus 6
2 DONGWON SYSTEMS 4
3 DAIDONG ELECTRONICS 3
4 Kedcom 8
5 LG Electronics 2
6 Huneed Technologies 5
7 Samsung Electronics 1
8 GS Instruments 7
Powerpoint TemplatesPage 28
Compare the TOPSIS Result with the Financial Analysis Result
The investors evaluate the value of the companies from the financial statement to determine the best investment alternative.
The financial analysis is fundamental method which decides the best investment alternative, in the same way TOPSIS is one of the decision-making techniques selecting the best stock.
So, we will analyze that the financial analysis compare with the result of TOPSIS.
Powerpoint TemplatesPage 29
We execute the Spearman’s rank correlate analysis in order to evaluate measurably the relationship between financial analysis result and TOPSIS result.
The Spearman’s rank correlate analysis◦ The Spearman’s rank correlation coefficient is used to
analyze relationship between two continual variables, if they are the criterion of the rank.
◦ The Spearman’s rank correlate analysis coefficient can have the values from “-1” to “1”.
· If the value is “1”, it means that they have same order of ranking, on the other hand, the value is “-1”, it shows that they have completely reversed order.
Compare the TOPSIS Result with the Financial Analysis Result
Powerpoint TemplatesPage 30
A Comparisons of the Financial Result and TOPSIS Result
<The preference ordering of the TOPSIS and the financial analysis>
Powerpoint TemplatesPage 31
The Spearman’s rank correlate analysis between the result and TOPSIS result
◦ The correlate coefficient is “0.4” between financial analysis and TOPSIS.
◦ We can observe there are scarcely relation between financial analysis result and TOPSIS result.
Financial analysis
TOPSIS 0.4
A Comparisons of the Financial Result and TOPSIS Result
Powerpoint TemplatesPage 32
A Comparisons of the Financial Result and TOPSIS Result
< The preference ordering of the TOPSIS and stability analysis >
Powerpoint TemplatesPage 33
A Comparisons of the Financial Result and TOPSIS Result
< The preference ordering of the TOPSIS and profitability analysis >
Powerpoint TemplatesPage 34
A Comparisons of the Financial Result and TOPSIS Result
< The preference ordering of the TOPSIS and activity analysis >
Powerpoint TemplatesPage 35
A Comparisons of the Financial Result and TOPSIS Result
< The preference ordering of the TOPSIS and market value analysis >
Powerpoint TemplatesPage 36
The Spearman’s rank correlate analysis 4 categories and TOPSIS
◦ In all categories, the correlation coefficients are under “0.5”.
◦ Consequently, all categories have little relation with TOPSIS.
Stabilityanalysis
Profitabilityanalysis
Activityanalysis
Market valueanalysis
TOPSIS -0.4 0.3 0.3 0.5
A Comparisons of the Financial Result and TOPSIS Result
Powerpoint TemplatesPage 37
The concluding remarks
We present one unique method when choosing the best-investment-alternative, so called TOPSIS to make a determination of the order of priority between stocks.
Then, we compare the financial analysis result with TOPSIS result to figure out the relation between two.
As a result of correlation analysis, we know the financial analysis is low correlation with TOPSIS.
◦ It means the ranking of financial analysis is not equal to the ranking of TOPSIS, although we use the same base factors to determine the preference order in the stock market.
Powerpoint TemplatesPage 38
The concluding remarks
We can explain the differences between two methods through two.
◦ First, we can be explained depending on whether we conduct the factor analysis.· TOPSIS: we execute the factor analysis to reduce the
number of factors by grouping the factors which are same effect on stocks.
◦ Second, we can describe contingent upon whether to apply weight value in the stocks.· Financial analysis : the same weight in each factor. · TOPSIS : A different weight according to degree of effect
on stock price.
Powerpoint TemplatesPage 39
The concluding remarks
We regrettably failed to set up the benchmarking base to compare the TOPSIS result.
So, we need to find out the sound and acceptable benchmarking base which will be the following research.