methodological problems associated with the investment risk
TRANSCRIPT
Methodological Problems Associated
with the Calculation of Investment
Risk – The ASE Case
Supervisor: Gklezakos Michael
Student: Bektsi Maria
Piraeus, 2015
PORTGRADUATE PROGRAM IN
APPLIED STATISTICS
UNIVERSITYOF PIRAEUS
Introduction
Dissertation
– Aim
testing the problems of investment risk
estimation
– Methodology
implementation of Market & Dimson
models
Problems of Investment Risk
Estimation
Problems of Investment Risk
Estimation
Methodological
Problems of Investment Risk
Estimation
Return Calculation
1
1
tt t t
t
DR P P
P
1
1
t tt
t
P PR
P
,
Methodological
Problems of Investment Risk
Estimation
Methodological
Return Interval
Problems of Investment Risk
Estimation
Methodological
Market Index
Problems of Investment Risk
Estimation
Horizon
Methodological
Problems of Investment Risk
Estimation
Methodological
Thin Trading
Problems of Investment Risk
Estimation
Statistical
Problems of Investment Risk
Estimation
Multicollinearity
Statistical
Problems of Investment Risk
Estimation
Statistical
Multicollinearity
Problems of Investment Risk
Estimation
Heteroscedasticity
Statistical
Problems of Investment Risk
Estimation
Statistical
Heteroscedasticity
Problems of Investment Risk
Estimation
Autocorellation
Statistical
Problems of Investment Risk
Estimation
Bias
Statistical
Problems of Investment Risk
Estimation
Lack of Normality
Statistical
Methodology
• Sample
– Selection Criterion
Marketability
– Stocks Number
20
– Examination Periods
01/00-12/02, 07/03-06/06, 01/09-12/11
• Return Calculation: 1
1
t tt
t
P PR
P
Methodology
Market & Dimson Models
• Market Model
Researchers: Sharpe(1964), Bey & Pinches(1980),
Huang & Jo(1988)
• Dimson Model
Researchers: Fowler et al(1979), Cohen et al(1983)
it i i Mt itR a b R e
,
n
it i k M t k it
k n
R a b R e
Market & Dimson Models
ASE Research
– Superiority of Market Model
Haritou(2000), Vazakides(2006)
[monthly data]
– Superiority of Dimson Model
Karathanassis & Patsos(1997),
Karathanassis et al(1999)
[daily data]
Marketability⟹ Dispersion & Skewness
⟹ Growth: Platykurtic Distribution
Distribution Characteristics
Portfolio Period Mean Max Min RangeStandard
DeviationKurtosis Skewness
Low
Marketability
01/00-12/02 -.255 11.793 -13.107 24.900 3.583 1.738 .237
07/03-06/06 .133 15.562 -12.853 28.415 3.139 4.582 .615
01/09-12/11 -.054 20.215 -18.579 38.794 4.531 3.889 .262
High
Marketability
01/00-12/02 -.107 10.685 -9.764 20.449 2.273 3.448 .380
07/03-06/06 .133 8.417 -8.097 16.514 1.823 2.226 .195
01/09-12/11 -.093 20.032 -13.859 33.891 3.375 3.357 .390
Testing Regression Assumptions
Tests
– Normality
Jarque-Bera
– Homoscedasticity
Levene
– Indepedence
Runs
22 1
36 4
nJB S K
*S: Skewness, Κ: Kurtosis
Stocks: Testing Regression
Assumptions
• Assumptions
– Normality
– Homoscedasticity Economy & Merketability⟹ Homoscedasticity
– Indepedence Economy & Marketability⟹ Depedence
• Economy⟹ Volatility
Low HighMarketability Marketability
Volatility < Volatility
Market & Dimson Models
Results Matching
(High Marketability)
Assumptions: Normality
Homoscedasticity
Independence
Portfolios: Testing Regression
Assumptions - Market Model
Portfolio PeriodNormality
(Jarque Bera test)
Homoscedasticity
(Levene’s test)
Independence
(Run’s test)
Volatility
Coefficient
(beta)
Low
Marketability
01/00-12/02 95.6330 .0958 .2077 1.237
07/03-06/06 726.8508 .0034 .3225 1.131
01/09-12/11 104.7290 .3307 .2312 .493
High
Marketability
01/00-12/02 140.3856 .1646 .3223 .998
07/03-06/06 80.9274 .0862 .3591 1.091
01/09-12/11 102.3306 .4604 .3739 1.068
*Level of Significance: 1%
Range
Standard
Deviation
38.794 4.531
33.891 3.375
Assumptions: Normality
Homoscedasticity
Independence
Multicollinearity
Portfolios: Testing Regression
Assumptions – Dimson Model
Portfolio PeriodNormality
(Jarque Bera test)
Homoscedasticity
(Levene’s test)
Independence
(Run’s test)
Volatility
Coefficient
(beta)
Low
Marketability
01/00-12/02 95.6330 .0279 .2191 1.220
07/03-06/06 726.8508 .0034 .3882 1.128
01/09-12/11 104.7290 .3249 .2186 .494
High
Marketability
01/00-12/02 140.3856 .1665 .2904 .991
07/03-06/06 80.9274 .0946 .3466 1.078
01/09-12/11 102.3306 .5054 .4572 1.068
<1.237
<1.131
<1.493
<1.998
<1.091
=1.068
*Level of Significance: 1%
Conclusions
• Stocks:– Normality & Homoscedasticity
– Recession & Marketability⟹ Homoscedasticity
– Recession & Marketability⟹ Independence
• Portfolio:– Normality
– Homoscedasticity & Independence
Conclusions
• Marketability Level: ATTENTION
• Marketability & Betas Relation:
• Inefficient Implementation of Dimson
Model
• Crisis: Imponderable Factor
Thanks for your
attention