Vous êtes sur la page 1sur 29

Contents

“List of Abbreviations” ........................................................................................................................... 1


“Table of Figures” ................................................................................................................................... 2
1. Introduction ........................................................................................................................................ 3
2. Literature review of the Capital Asset Pricing Model (CAPM) .......................................................... 4
2.1 Markowitz’s portfolio theory and Tobin’s separation theorem ................................................. 4
2.2 The Sharpe-Lintner-Mossin CAPM ............................................................................................... 6
2.3 The first critiques and further extensions of the CAPM ............................................................. 7
2.4 The three-factor model of Fama and French............................................................................. 11
“3. The CAPM model” .......................................................................................................................... 13
“3.1 Data description”...................................................................................................................... 13
3.1.1 Introduction ......................................................................................................................... 13
3.1.2 The four moments of distribution ...................................................................................... 13
“3.2 Methodology of OLS” ............................................................................................................... 15
“4. Explanation of empirical evidence” ............................................................................................... 17
4.1 The test procedure ..................................................................................................................... 17
4.2 The empirical results .................................................................................................................. 18
4.2.1 The 2002-2017 period ......................................................................................................... 18
4.2.2 The two subsamples............................................................................................................ 18
4.2.3 An equally-weighted portfolio............................................................................................ 18
5. Conclusion ........................................................................................................................................ 20
Bibliography.......................................................................................................................................... 21
“APPENDICES” ...................................................................................................................................... 25
“Appendix A” .................................................................................................................................... 25
“List of Abbreviations”

“Capital Asset Pricing Model” CAPM


“Modern Portfolio Theory” MPT
“Mean Variance” MV
“Security Market Line” SML
“New York Stock Exchange” NYSE
“Intertemporal CAPM” ICAPM
“Arbitrage Pricing Theory” APT
“Price to earnings” P/E
“Consumption-based CAPM” CCAPM
“Debt to equity” D/E
“Book to market” B/M
“Small minus Big” SMB
“High minus Low” HML
“Robust minus Weak” RMW
“Conservative minus Aggressive” CMA
“Financial Times Stock Exchange” FTSE
“Ordinary Least Squares” OLS

1
“Table of Figures”

"Figure 1: Minimum-Variance frontier of risky assets" .......................................................................... 5


"Figure 2: The CAPM" ............................................................................................................................. 6
"Figure 3: The Security Market Line" ...................................................................................................... 7

2
1. Introduction

In this paper, I attempt to examine the theory behind the Capital Asset Pricing Model (CAPM)
of Sharpe, Lintner and Mossin from a critical point of view. At the same time, in order to
discover whether the CAPM applies into the real world, I will empirically test its correctness.
Firstly, in the next section, I will explain how these economists built upon Markowitz’s earlier
findings and created their own model. The CAPM afterwards became a milestone in modern
finance and it is still being used as a tool to quantify the relationship between risk and
expected return. However, there were many economists who challenged the CAPM and its
implications and assumptions. One of the most important papers over the years was Fama
and MacBeth’s in 1973, in which they supported a two-factor model as a better explaination
of the trade-off between risk and return. Then, exactly two decades later, Fama and French
(1993) developed their multifactor model which aims to better describe this relationship.
Then, in section three and four, I will describe the data that I am going to use, explain the
methodology and run the tests of the Sharpe-Lintner CAPM. I will be regressing a 16-year
mothly sample of 30 stocks in the UK and the market.

3
2. Literature review of the Capital Asset Pricing Model (CAPM)

2.1 Markowitz’s portfolio theory and Tobin’s separation theorem

From my point of view, a critical review of the CAPM has to start with the model of portfolio
evaluation established by Harry Markowitz in 1952. As Rossi (2016) highlighted, the CAPM
which was founded later on, builds upon the model of Markowitz. Elton and Gruber (1998),
among other economists, strongly believe that the first one who introduced the Modern
Portfolio Theory was Markowitz. This is because he has influenced the practice of portfolio
management. More explicitly, the American economist published his paper with the title
‘Portfolio Selection’ in 1952. Furthermore, Keller (2014) characteristically mentions that the
Mean Variance (MV) analysis, which played a significant role in modern finance, is part of
Markowitz’s theory. Basically, the MV analysis attempts to quantify the variance (risk) of an
asset in regard to the expected return. This approach helps the investors to be aware of the
risks of a potential investment and how much return they will earn at certain levels of risk.
The MV model contains some assumptions such as that investors are risk averse (West, 2006),
which means that investors will choose an asset that has lower variance when the expected
return is the same. Moreover, Rossi (2016) points out that their decisions are based on the
mean and the variance of the portfolio of their one-period investment returns. Therefore,
investors will prefer portfolios with the least mean variance that maximise the expected
return at the same time.

At this point, it is essential to mention the effect of diversification because it helps investors
to select the optimal portfolio. Mangram (2013) defines diversification as a risk reduction tool
which involves the investors’ capital allocation among various financial assets. In other words,
diversification can be explained by the saying ‘’don’t put all your eggs in one basket’’. This
means that an investor should not use all of their capital in a single financial instrument.
Hence, investors can achieve diversification by investing in different asset sectors, products
and commodities and dividing their investing capital. In my opinion, diversification is the key
component of Markowitz’ (1952) portfolio theory.

Another key concept of the MPT of Markowitz, according to Mangram (2013), is the efficient
frontier.

4
Figure 1: MV frontier
Source: Bodie et al. (2013)

Figure 1 explains quite convincingly the MV frontier of risky assets. The vertical axis represents
the expected rate of return and the horizontal axis shows the investor’s risk tolerance. The
efficient frontier is the line curve which illustrates the potential yield of an investment given
a certain level of risk. All the optimal portfolios obtained through diversification should lie on
this curve. From the graph it is easily observed that the portfolios that are below the line
curve, do not represent the best investment choice because for the same amount of risk an
investor could achieve higher amount of return. On the other hand, any portfolio that is above
the curve is impossible.

Another great economist that contributed to the MV analysis and asset pricing theory is Jim
Tobin (Buiter, 2003). As Mangram (2013) highlighted, Sharpe and Lintner carried upon
Markowitz’ and Tobin’s previous works to the develop the CAPM. Tobin (1958) developed the
Separation Theorem, which underlines that if there is a safe asset and lots of risky assets,
portfolio selection by any portfolio holder can be explained as a choice between the risk-free
asset and the exact same portfolio of risky assets (Buiter, 2003). Sharpe (1964) also argues
that Tobin analysed the Markowitz model into two phases; firstly, the choice of an optimum
combination of risky commodities and secondly, a separate choice regarding the capital
allocation between this combination and a single safe risk asset. It is now easily observed that
all efficient portfolios are combinations of risky portfolios and a riskless asset (Fama and
French, 2004).

5
Figure 2: The CAPM
Source: Fama and French (2004)

Figure 2 illustrates the portfolio investment opportunities and describes the CAPM story
according to Rossi (2016). Consequently, I believe that the line of the CAPM is now
straightforward.

2.2 The Sharpe-Lintner-Mossin CAPM

According to Rossi (2016), the CAPM was first developed by Sharpe (1964), Lintner (1965) and
Mossin (1966) which marks the start of the asset pricing theory. The CAPM is widely
considered as a pioneering tool in modern academic finance because it estimates the cost of
capital for firms and the investors’ returns. The CAPM tries to quantify the relationship
between the expected return of an asset and its corresponding beta (Womack and Zhang,
2003). The authors further argue that the CAPM model has several assumptions, but the most
noteworthy one, in my opinion, is about the investor behaviour and the existence of a risk-
free asset. Furthermore, it must be mentioned here that there are two types of risk an
investor will face in the future (Bodie et al., 2013). Firstly, it is the unsystematic risk, or
diversifiable risk which means that this type of risk can be reduced through diversification, as
it was discussed above. Then, there is also the systematic risk. This is the type of risk that
cannot be diversified away no matter what the investor does with their portfolio of choice; it

6
is also called market risk. This level of uncertainty will always have to be considered by an
investor. In my opinion, a better understanding of the model and the risks requires a graphical
illustration.

Figure 3: The “Security Market Line”


Source: Bodie et al. (2013)

Figure 3 presents the “Security Market Line” (SML) of the CAPM. The SML shows how the
expected return changes along with the beta. Beta is calculated as the ratio of the excess
return of an asset compared to the overall excess market return (Womack and Zhang, 2003).
By excess return, it is the return of a given asset less the risk-free asset return. In other words,
a security that has a beta of one indicates that the security’s return moves along with the
market. If the beta is less than one, then the security is less volatile than the market. On the
other hand, a security with zero beta means that the asset is not dependent on the market
reactions (Rossi, 2016).

2.3 The first critiques and further extensions of the CAPM

At this point, I believe it is essential to discuss how other economists evaluate the Sharpe-
Lintner-Mossin CAPM over the years. The broad consensus argues that the CAPM model is a
significant tool in modern finance because the calculations to estimate the risk-return trade-
off are quite straightforward. On the other hand, the main drawbacks of the CAPM model are

7
its own assumptions. As Fama and French (2014) have stated, the simplicity of the model is
what raised many questions on whether the CAPM explains the trade-off between risk and
expected return. Rossi (2016) clearly underlines that it is difficult to find a risk-free asset,
which is one of the factors of the CAPM model. More specifically, a riskless asset would most
likely be a government bond or a treasury bill which makes the possibility of default extremely
small, but on the contrary, inflation cannot be predicted and hence the real rate of return is
uncertain. He further argues that the assumption of Lintner that the borrowing rates have to
be equal with the lending rates cannot be realistic. Finally, another drawback of the model
stems from the assumption that betas are stable over time. Fernandez (2017) states that
betas are unstable because they change from one day to the next. Therefore, this shows that
calculating historical betas is not the best indicator of the future risk assets.

To begin with, the finance literature is divided by those who built upon the CAPM and those
who exceeded the single-factor model of the CAPM and created their own multifactor
models. As Rossi (2016) highlights, the first problems of the CAPM started right away. The
empirical results of Lintner (1965) and Douglas (1969) were not that encouraging. To be exact,
both of them found that the coefficient of intercept has larger values than the return of the
riskless security, when at the same time the coefficient of beta has smaller value. A few years
later, an empirical study by Miller and Scholes (1972) rejected the validity of the CAPM. They
conducted multiple regression tests with individual company stock returns. Miller and Scholes
(1972) recognised many drawbacks in the CAPM implications, and although they tried to
overcome them, there was still a negative relation between performance and risk. These
problems of the CAPM were related to the nature of the structure of returns’ securities (Rossi,
2016). He further argues that this problem was overcome by other studies in the future. The
very first of these studies was conducted by Black et al. (1972); in their tests, they formed
portfolios of stocks in contrast with Miller and Scholes (1972). Their results were not in favour
of the CAPM (Black, 1972). As Black et al. (1972) conclude, this evidence is strong enough to
reject the traditional model of Sharpe and Lintner. A year later in 1973, Fama and MacBeth
extended the research of Black et al. (1972) and highlighted some interesting facts. More
specifically, they formed 20 portfolios of all the stocks in the NYSE and conducted a regression
for the 1935-1968 period on monthly data. Their concluding remarks state that it exists an
intercept that is larger than the rate of the risk-free asset. Also, the linear relationship of the

8
beta and average returns holds when the data are for longer time periods (Fama and
MacBeth, 1973). On the other hand, Blume and Friend (1973) examined if the CAPM holds
both empirically and theoretically and found that it has to be rejected. They analysed the
relationship between the risk and rate of return and concluded that the returns in the stock
market cannot be explained by the market line theory. They argued that the failure of the
theory may stem from the unrealistic assumption of a perfect short-selling mechanism.

After the first theoretical and empirical studies on the CAPM, Black (1972) introduces his two-
factor model which develops another version of CAPM, but without the risk-free return on
assets. He further shows that the market portfolio return that achieves the least risk can be
obtained by allowing short sales on the risky assets. Womack and Zhang (2003) state that
since the Black CAPM adds another factor to the traditional form of the model, the
explanatory power is rather larger than the single factor model of the CAPM. Furthermore,
by adding new risk factors, it allows more specific explanation of the risks to which a firm is
exposed. They argue that in terms of the power of regression, it is more significant if the
model has more than one factors to regress. Generally, less evidence has been found against
the Black CAPM than the Sharpe-Lintner CAPM, since the first one has an extra factor to
consider and makes it more powerful.

A year later, Merton (1973) came along and presented his development of the Sharpe-Lintner-
Mossin CAPM. In particular, Merton developed an equilibrium model of the CAPM under
some assumptions which is called the Intertemporal CAPM (ICAPM). Merton stated that
investors plan ahead for more than a single period which makes his model more dynamic,
compared to the static model of CAPM. Merton emphasises that the ICAPM is based on
consumer-investor behaviour which expands the traditional form of CAPM. Finally, in his
concluding marks, it is underlined that more factors (i.e. wage income and many consumption
goods) could be included to achieve better results with the CAPM, which explains the later
studies on multifactor models. Another important theory that was presented as an alternative
to the existing capital asset pricing theory by Sharpe and Lintner is the Arbitrage Pricing
Theory (APT) by Ross in 1976, which became very popular over the years. The APT is a
multifactor technical model that is based on the relationship between an asset’s expected
return and risk (Ross, 1976). His main purpose was to illustrate the sensitivity of an asset’s
return to changes in various macroeconomic variables. The assumptions of his model were

9
more lenient than the ones of the CAPM, so it can better reflect the real world. However, the
APT is quite difficult to apply in paper, because it involves complex data and statistical analysis
(Ross, 1976). Ross presents a no-arbitrage framework as an extension to the Sharpe-Lintner-
Mossin and Black model. As it was mentioned above, Sharpe and Lintner assume that the
investor does not face borrowing constraints and Black assumes that the investor is allowed
to short-sale, while Ross assumes that at least one investor faces the no-borrowing or short-
sale constraint.

However, after Ross’s APT in 1976, Roll raised serious doubts and questions in testing the
CAPM. Until now, the proxies used for the return on market portfolio in the Sharpe-Lintner-
Mossin model have not been tested (Roll, 1977). Continuing his work, he states that the
regression tests have limited explanatory power and he points out some objections. In my
opinion, the biggest one is that the Sharpe-Lintner-Mossin model uses proxies such as the
S&P500 and Roll (1977) considers this as a ‘’benchmark error’’. Almost two decades later,
both Roll and Ross (1995) and Kandel and Stambaugh (1995) took this critique one step
further. More explicitly, they highlighted that tests which reject the positive relationship
between returns and beta stem from the inefficiencies in proxies and not from the theoretical
part of the model. Furthermore, one of the first ones that argued that a single-factor model
like the CAPM does not fully explain the relationship between returns and beta is Basu (Rossi,
2016). Basu (1977) attempts to determine empirically if the investment performance of stocks
is related to their price to earnings (P/E) ratios. Basu finds that stocks with high P/E ratio have
higher expected return than the CAPM’s predictions. Therefore, the financial ratios may be
better indicators of investment performance in the future.

A few years after Merton’s version of CAPM, a paper from Breeden comes to further develop
this intertemporal extension in a continuous-time model. This model uses the same
continuous-time framework as Merton does, but it is presented that the multi-beta pricing
equation by Merton can be distorted into a single-beta pricing equation (Breeden, 1979).
Breeden concludes that his model also permits both stochastic investment opportunities and
stochastic consumption goods and that is simpler to test it empirically than Merton’s. This is
the reason why his theory is widely known as the Consumption-based CAPM (CCAPM).
Roughly two decades later, Campbell (1996) develops an alternative model to the CCAPM
derived by Breeden. He explains that his model is based on log-linear approximations and has

10
a multifactor form. These factors include a proxy’s return as in the traditional CAPM, new
variables that help to predict future labour income and new variables that help to predict
future returns as it was discussed earlier in Merton’s version. Lettau and Ludvigson (2001)
conclude that the CCAPM models generally perform far better that other models with
unconditional specifications and about as the multifactor model of Fama and French that I
will analyse later in the review.

The 80’s literature proved in some cases that a single-factor model like CAPM is violated and
it does not explain the linear relationship between the excess market return and beta. Basu
(1977) first stated that non-market factors (i.e. P/E ratio) contribute to the risk-return
relationship. Four years later, Banz (1981) examined the relationship between the market
value of the stock of a company and the return. Over the 1936-1975 period, Banz observed
higher returns on the stocks of smaller firms compared to the larger ones. Finally, he argues
that over this forty-year period, the size effect is not related linearly to the market proportion.
On the same page, Bhandari (1988) believes that beta cannot explain the expected stock
return. Therefore, he proposes to use the Debt to Equity (D/E) ratios. Bhandari also underlines
that this additional variable will be able to explain the stock returns. For example, an increase
in the D/E of a company will increase the risk of its equity. The stocks’ expected returns are
positively related to the D/E (Bhandari, 1988). All these previous researches show that a
single-factor model does not hold, and other factors need to be taken into account. Chan et
al. (1991) attempt to explore the cross-sectional forecasting of equity returns in Japanese
firms using four variables; earnings yield, cash flow yield, size of equity and book to market
(B/M) ratio. Their results make clear that their variables have a positive impact on expected
returns and especially the B/M ratio and cash flow yield.

2.4 The three-factor model of Fama and French

Following the studies that underlined that market factors explain adequately the returns on
common stocks, two famous researchers, Fama and French, did an extensive research in other
significant factors (Womack and Zhang, 2003). These factors are the Small Minus Big (SMB)
and the High Minus Low (HML), to describe the size risk and value risk respectively. These
findings were first published in 1992 and the two authors have continued to extend their work
since. Fama and French (1992) argue that the beta is not the best indicator and, hence, other

11
financial indicators have to be considered. They later formed 25 portfolios on the basis of the
B/M ratios and market capitalisation of stocks (Fama and French, 1993). The data period was
from 1963 to 1991 and they regressed the returns of these portfolios on the three variables.
The R-squared values that they got from the regressions were from 0.83 to 0.94. This means
that these independent variables almost fully explain the changes in the dependent variable.
The next three years, these two variables were being tested to extract better estimations of
their “three-factor” model (Fama and French 1995; Fama and French 1996). A year later after
gathering all their results, they concluded that the evidence does not make the three-factor
model as a perfect model to explain the stock returns (Fama and French, 1997). However,
over the years, Fama and French’s work has been criticised. More specifically, their study
keeps ignoring positive evidence in historical betas and tends to exaggerate the importance
of the Price to Book (P/B) ratio (Kothari and Shanken, 1995). They further argue that the “P/B”
ratio is a weak factor of the “cross-sectional” variation in expected returns. Davis, Fama and
French (2000) claim that the multifactor model, along with the Sharpe-Lintner-Mossin model,
underestimates the expected returns to assets with low beta and overestimates the returns
to assets with high beta. Sattar (2017) later highlights that the multifactor model of Fama and
French is not being violated in markets with high volatility and if not, the model is about as
effective as the Sharpe-Lintner CAPM. In their recent work, Fama and French (2014) moved
from their three-factor model and extended it. They presented” a five-factor model and the
newly added variables are a profitability factor and an investment factor; the Robust Minus
Weak (RMW) and the Conservative Minus Aggressive (CMA) respectively.

12
“3. The CAPM model”

“3.1 Data description”

In this section I will be presenting the data that will be used later on in the empirical analysis.
More specifically, I use 30 company stocks from the United Kingdom (UK) and the Financial
Times Stocks Exchange All-Share (FTSE ALL-SHARE) as well.

3.1.1 Introduction

These data describe the period from the 31st of January in 2002 until the 29th of December in
2017 in a monthly basis. Therefore, my sample has 192 observations in total. Furthermore, it
is important to be mentioned that I assume that the index FTSE All-Share represents well the
market. The FTSE All-Share consists of the largest companies in the UK, as of their market
value. Hence, since it includes the biggest proportion of the market capitalisation in the UK, I
believe it is a strong approximation of the market. It is important to point out that I use the
excess company stock returns and the excess market return to derive all my data. As it was
previously discussed, the difference of the risk-free asset and the stock return or the market
gives the excess return on the stock and market respectively.

3.1.2 The four moments of distribution

At this point, I find it essential to describe briefly the mean of my dataset, the variance, the
skewness and the kurtosis. These are also known as the four moments of distribution.
However, I also extract the standard deviation, the minimum, the maximum, the range and
the median of my dataset1.

3.1.2.1 The mean

Starting with the mean, I believe that the mean is one of the most crucial indicators of a
dataset. It is extremely easy to calculate and quick. The mean refers to the expected value of

1
“See Appendix A.1”

13
a sample and it shows the central tendency of the random variable. However, the mean is
sensitive to the most extreme values, which are the least representative of the dataset. In my
sample, I find that the highest mean figure is 2.1461% by Hill & Smith which is significantly
higher than the 0.4540% of the market. On the other hand, the lowest mean return is the one
of Carpetright with -0.1046%, indicating mostly negative performances over the 16-year
period. Carpetright is the only company out of the 30 in my sample that has a negative mean
figure.

3.1.2.2 The variance

Except from the mean of a sample, another essential moment of distribution is the variance;
or the standard deviation in our case because it measures the level of volatility of a stock
which is an indicator of risk. Therefore, it is significantly important for a potential investor to
be aware of the challenges and risk and that makes the calculation of the standard deviation
crucial. It is obvious that a company stock with high figures of variance, and subsequently of
standard deviation, will be riskier than one with low standard deviation. Based on my
calculations, the most volatile company stock appears to be Oxford Biomedica with 19.9702%,
which is quite higher than the rest ones and substantially higher than the market’s risk
(3.9183%). The lower figures of the FTSE All-Share could be explained by the diversification
that it was discussed earlier, since it consists of over 600 companies and, hence, the sample
is larger. In contrast, Barr (AG) achieves the lowest standard deviation of my sample of 30
companies with 5.8808%, which is still relatively higher than the one of FTSE All-Share.

3.1.2.3 The skewness

Thirdly, skewness is another necessary tool that needs to be considered and analysed.
Skewness measures the lop-sidedness of the sample distribution and it can be either positive
or negative. A positive figure of skewness means that the data lean upon the left side of the
bell curve, while a negative skewness means that the data lean upon the right-hand side.
When an investor knows which way the data lean, they can estimate if the future stock
returns will be more or less than the mean in the future. A positive skewness means that the
sample mean, and median are larger than the mode, while a negative figure of skewness

14
shows the exact opposite. My dataset has 13 companies with negative skewness and the rest
of them (17) with positive, while the market skewness is negative with -0.7479.

3.1.2.4 The kurtosis

Last but not least, kurtosis is the fourth moment of distribution and it measures how many
extreme values are in both tails. Therefore, a large kurtosis distribution means that the data
are 5 or more standard deviations from the mean, which makes the distribution to be called
as fat-tailed. On the other hand, distributions with low kurtosis means that the data are less
extreme than the ones of the normal distribution. My calculations of the sample show that
the highest kurtosis is 14.0678 of the company Renewi, implying that the kurtosis is rather
high, and except Renewi’s there are 2 other companies that have also high kurtosis; Travis
Perkins and Ferguson with 7.9245 and 6.9853 respectively. However, the index kurtosis is
1.4103, which means that most of the observations are approximately 1.4 standard deviations
away from the mean. In business and finance, high kurtosis means that the investor will most
likely receive extreme returns, either positive ones or negative.

“3.2 Methodology of OLS”

In this part, I believe that it is essential to describe the methodology that will be used below
to test my sample. To begin with, the methodology of “ordinary least squares (OLS)” has been
developed as a technique to estimate the unknown variables “in a linear regression model”
like the CAPM; it was first developed by a German mathematician, Carl Friedrich Gauss. This
method has been widely used over the years because it has some very appealing statistical
properties and it has been proved that it is one of the most popular and strongest methods
of regression analysis. The OLS method is a mathematical regression analysis that illustrates
the line of best fit for a set of data and shows the relationship between the variables. Firstly,
I believe I need to set up the least squares equation:
"𝑌𝑖 = 𝛽1 + 𝛽2𝑋𝑖 + 𝑢𝑖", (1)
where 𝛽1 is the coefficient of the intercept, 𝛽2 is the coefficient of the X variable and 𝑢𝑖 is
the residuals.

15
However, in our case, the regression analysis will be between the excess market returns and
the excess stock returns that were discussed earlier. Therefore, the equation that will be
regressed is the following:
"𝐸(𝑅𝑖) = 𝑎 + 𝑅𝑓 + 𝛽(𝑅𝑚 − 𝑅𝑓) + 𝑢𝑖", (2)
The equation (5) can be re-arranged to best illustrate the regression process into the
following:
"𝐸(𝑅𝑖) − 𝑅𝑓 = 𝑎 + 𝛽(𝑅𝑚 − 𝑅𝑓) + 𝑢𝑖", (3)
The OLS method, in our case, chooses the parameters of a linear function by minimising the
sum of the squares of the differences between the excess stock returns (dependent variable)
in the given dataset and those predicted by the linear equation. It is obvious that as the
differences get smaller, the better the model describes the data.
However, in order for this method to be applicable, some assumptions are needed to be set.
More specifically, Gujarati (2003, p.66-75) points out 10 assumptions in his book ‘’Basic
Econometrics’’, but I will only report some of them that I believe are important. These are
that is a linear regression model, the covariance between the residuals and the X variable is
zero and there is no perfect multicollinearity.

16
“4. Explanation of empirical evidence”

4.1 The test procedure

In this section, I will analyse the empirical results extracted from the regression analysis I
conducted. First, it is important to state what I am testing, the parameters, the hypotheses,
the confidence level etc. The main purpose of my analysis is to show whether the traditional
model CAPM states or not.
To start with, I ran a regression analysis multiple times and for 3 different periods over the
years. In the first one, I took the monthly excess returns of 30 stocks and I regressed them
with the excess market returns from 2002 to 2017. Therefore, my dataset was consisted of
192 observations. The summary output of the regression extracts lots of data, such as
“multiple R, R-squared, adjusted R-squared, standard error” and the intercept’s coefficient,
standard error, t-stat and p-value and many more. However, I will only report some of those.
At this point, I believe it is important to recall the equation (2), which was re-arranged into
equation (3). Then, I state the two hypotheses of my analysis. The “null hypothesis argues
that the there is no α, meaning that the CAPM holds, while the alternative hypothesis argues
that α” is not zero and, hence, the CAPM is violated. More specifically,
i. "𝐻𝑜: 𝛼 = 0", “where 𝐻𝑜 is the null hypothesis and 𝛼 is the intercept” and
ii. "𝐻𝑎: 𝛼 ≠ 0", “where 𝐻𝑎 is the alternative hypothesis.”
Now, when it comes to rejecting or do not rejecting the null hypothesis, the t-stat or the p-
value are very important indicators. If the test statistic is between the area of acceptance
(−1.96 ≤ 𝑡 − 𝑠𝑡𝑎𝑡 ≤ +1.96), then the 5% significance level test procedure do not rejects the
null hypothesis. As far as the p-value is concerned, if it is lower or equal to 0.05 (≤ 0.05). I
would like to mention here that if the significance level changes, either increases or fall, the
test statistics and the p-values do not change; what changes is the rejection area.

17
4.2 The empirical results

4.2.1 The 2002-2017 period

In the period 2002-2017 I found that 252 companies have their calculated test statistic within
the area of acceptance and only 5 represent the ones that are not within the area of
acceptance. This means that the null hypothesis is not being rejected and so the CAPM does
hold for almost 85% of the time. This can be interpreted as well by observing the p-values of
the regression; those 5 companies have p-values under 5%. However, it is obvious here that
these particular companies do not have high R-squared values (a maximum of roughly 27%),
which makes the explanatory power of the regression weak. The highest R-squared across the
companies is almost 44% (Johnson Matthey), which is still relatively low.

4.2.2 The two subsamples

The next two periods are two equally portioned subsamples of the 2002-2017 period. In the
first subsample (2002-2009), I found that the CAPM holds 233 times out of the 30.
Furthermore, the highest R-squared is almost at 50% (Weir Group) and at the same time the
test statistic of this company rejects the CAPM. Continuing with the second subsample (2010-
2017), the CAPM is being rejected only 34 times with highest R-squared of 21.2754% of DCC.
The highest R-squared in this subsample does not also exceed the 50% mark (45.5343%).

4.2.3 An equally-weighted portfolio

Finally, I combined all company stock returns and formed a single equally weighted portfolio.
I ran the same tests for the 2002-2017, 2002-2009 and 2010-2017 period. The results were
quite different to the ones previously. I rejected the CAPM 2 times out of the 3. More
specifically, in the entire 16-year period and the 2002-2009 period I rejected the null
hypothesis. Moreover, I have to mention here that the R-squared figures are significantly
higher than the ones before when I regressed the individual stock returns. These figures are

2
“See Appendix A.2”
3
“See Appendix A.3”
4
“See Appendix A.4”

18
at 70.7829% and 74.4714% respectively. In contrast, the second subsample indicates that the
CAPM is not rejected since the calculated test statistic is between -1.96 and 1.96. The R-
squared here is at 65.7987%.

19
5. Conclusion

The Sharpe-Lintner CAPM has never been an empirical success according to Fama and French
(2004). I strongly believe that this has happened because, like Markowitz’s portfolio selection
model, is nevertheless a theoretical tool. The assumptions of the model were criticised
because of their unrealistic nature. Therefore, I conducted multiple tests for different time
periods throughout the last 16 years to see whether my results will be different.

To summarise, my results indicate that the Sharpe-Lintner CAPM holds most of the times
when company stock returns are being tested. However, the R-squared figures observed are
rather small which makes the explanatory power of the sample weak. On the other hand,
when I formed an equally-weighted portfolio of the exact same stocks, I found that the CAPM
is violated 2 out of 3 times, proving earlier studies that when regressing a portfolio there are
strong evidence against the CAPM (Black et al., 1972; Fama and MacBeth, 1973; Blume and
Friend, 1973 and Roll, 1977).

20
Bibliography

Banz, R. (1981) ‘The relationship between return and market value of common stocks’.
Journal of Financial Economics, 9(1), pp. 3-18.

Basu, S. (1977) ‘Investment performance of common stocks in relation to their price earnings
ratios: a test of efficient market hypothesis’. The Journal of Finance, 32(3), pp. 663-682.

Bhandari, L. (1988) ‘Debt/equity ratio and expected common stock returns: empirical
evidence’. The Journal of Finance, 43(2), pp. 507-528.

Black, F. (1972) ‘Capital market equilibrium with restricted borrowing’. The Journal of
business, 45(3), pp. 444-455.

Black, F., Jensen, M. and Scholes, M. (1972) ‘The capital asset pricing model: Some empirical
tests’. In Studies in the theory of capital markets, Michael Jensen, ed. New York: Praeger, pp.
79-121.

Blume, M. and Friend, I. (1973) ‘A new look at the capital asset pricing model’. The journal of
finance, 28(1), pp. 19-34.

Bodie, Z., Kane, A. and Marcus, A. (2013) Investments and Portfolio Management. 9th edn.
New York: McGraw – Hill Irwin.

Breeden, D. (1979) ‘An intertemporal asset pricing model with stochastic consumption and
investment opportunities’. Journal of Financial Economics, 7(3), pp. 265-296.

Buiter, W. (2003) James Tobin: an appreciation of his contribution to economics, Working


paper. NBER. Available at: http://www.nber.org/papers/w9753

Campbell, J. (1996) ‘Understanding risk and return’. The Journal of Political Economy, 104(2),
pp. 298-345.

Chan, L., Hamao, Y. and Lakonishok, J. (1991) ‘Fundamentals and stock returns in Japan’.
Journal of Finance, 46(5), pp. 1739-1789.

Davis, J., Fama, E. and French, R. (2000) ‘Characteristics, covariances, and average returns:
1929 to 1997’. The Journal of Finance, 55(1), pp. 389-406.

21
Douglas, G. (1969) ‘Risk in the equity markets: an empirical appraisal of market efficiency’.
Yale Economic Essays, 9(1), pp. 3-45.

Elton, E. J. and Gruber, M. J. (1998) ‘Modern portfolio theory, 1950 to date’. Journal Banking
& Finance, 21(11-12), pp. 1743-1759.

Fama, E. and French, K. (1992) ‘The cross-section of expected stock returns’. Journal of
Finance, 47(2), pp. 427-465.

Fama, E. and French, K. (1993) ‘Common risk factors in the returns on stocks and bonds’.
Journal of Financial Economics, 33(1), pp. 3–56.

Fama, E. and French, K. (1995) ‘Size and book-to-market factors in earnings and returns’.
Journal of Finance, 50(1), pp.131–156.

Fama, E. and French, K. (1996) ‘Multifactor Explanations of Asset Pricing Anomalies’. Journal
of Finance, 51(1), pp. 55–84.

Fama, E. and French, K. (1997) ‘Industry costs of equity’. Journal of Financial Economics, 43(2),
pp. 153-193.

Fama, E. and French, K. (2004) ‘The capital asset pricing model: theory and evidence’. Journal
of Economic Perspectives, 18(3), pp. 25-46.

Fama, E. and French, K. (2014) A Five-Factor Asset Pricing Model, Working paper. Fama-Miller.
Available at: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2287202

Fama, E. and MacBeth, J. (1973) ‘Risk, return, and equilibrium: Empirical tests’. Journal of
political economy, 81(3), pp. 607-636.

Fernandez, P. (2017) Are calculated betas good for anything?, Working paper. IESE Business
School. Available at: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=504565

Gujarati, D. (2004) Basic Econometrics, 4th edn, New York: McGraw Hill.

Kandel, S. and Stambaugh, R. (1995) ‘Portfolio inefficiency and the cross-section of expected
returns’. Journal of Finance, 50(1), pp. 185-224

22
Keller, W. (2014) Momentum, Markowitz, and smart beta: A tactical, analytical and practical
look at modern portfolio theory, Working paper. Available at:
https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2450017

Kothari, S., Shanken, J. and Sloan, R. (1995) ‘Another look at the cross-section of expected
stock returns’, Journal of Finance, 50(1), pp. 185-224.

Lettau, M. and Ludvigson, S. (2001) ‘Resurecting the (C) CAPM: A cross-sectional test when
risk premia are time-varying’. Journal of Political Economy, 109(6), pp. 1238-1287.

Lintner, J. (1965) ‘The valuation of risk assets and the selection of risky investments in stock
portfolios and capital budgets’. Review of Economics and Statistics, 47(1), pp. 13-37.

Mangram, M. (2013) ‘A simplified perspective of the Markowitz Portfolio Theory’. Global


Journal of Business Research, 7(1), pp. 59-70.

Markowitz, H. (1952) ‘Portfolio selection’. Journal of Finance, 7(1), pp. 77-99.

Merton, R. (1973) ‘An intertemporal capital asset pricing model’. Econometrica: Journal of the
Econometric Society, 41(5), pp. 867-887.

Miller, M. and Scholes, M. (1972) ‘Rates of return in relation to risk: a re-examination of some
recent findings’. Studies in the theory of capital markets, pp. 47-78.

Mossin, J. (1966) ‘Equilibrium in a capital asset market’. Econometrica: Journal of the


Econometric Society, 34(4), pp. 768-783.

Roll, R. (1977) ‘A critique of the asset pricing theory’s tests – part 1: on past and potential
testability of the theory’. Journal of Financial Economics, 4(1), pp. 129-176.

Roll, R. and Ross, S. (1995) ‘On the cross-sectional relation between expected return and
betas’. Journal of Finance, 50(1), pp. 185-224.

Ross, S. (1976) ‘The arbitrage theory of capital asset pricing’. Journal of Economic Theory,
13(3), pp. 341-360.

Rossi, M. (2016) ‘The capital asset pricing model: a critical literature review’. Global Business
and Economics Review, 18(5), pp. 604-617.

23
Sattar, M. 2017 ‘CAPM vs Fama-French three-factor model: An evaluation if effectjve ess in
explaining excess return in Dhaka Stock Exchange’. International Journal of Business and
Management, 12(5), pp. 119-129.

Sharpe, W. (1964) ‘Capital asset prices: a theory of market equilibrium under conditions’.
Journal of Finance, 19(3), pp. 425-442.

Tobin, J. (1958) ‘Liquidity preference as behaviour toward risk’. Review of Economic Studies,
25(2), pp. 65-86.

West, G. (2006) An introduction to modern portfolio theory: Markowitz, CAPM, APT and Black-
Litterman, Working paper. University of the Witwatersrand. Available at:
http://janroman.dhis.org/finance/Portfolio/Modern%20Portfolio%20Theory.pdf

Womack, K. L. and Zhang, Y. (2003) Understanding Risk and Return, the CAPM, and the Fama-
French Three-Factor Model, Working paper. Tuck School of Business at Dartmouth. Available
at: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=481881

24
“APPENDICES”

“Appendix A”

“A.1 Table of the calculated four moments of distribution and four additional indicators for
every company in the dataset and the market portfolio.”

"Mean" "Median""Standard Deviation""Sample Variance" "Kurtosis" "Skewness" "Range" "Minimum" "Maximum"


SIG 0.005572811 1.6957% 0.114557026 0.013123312 2.4822751 -0.067512932 0.880196984 -0.352142901 52.8054%
HAYS 0.005139148 0.6287% 0.087291638 0.00761983 0.488851874 -0.08058118 0.536348682 -0.282869342 25.3479%
ROYAL DUTCH SHELL B 0.00645784 1.0326% 0.060466895 0.003656245 1.11217905 0.034122166 0.411568028 -0.170128718 24.1439%
HUNTSWORTH 0.005394762 -0.0232% 0.110827723 0.012282784 1.922002153 0.118227801 0.800211924 -0.394067909 40.6144%
BHP BILLITON 0.013030158 1.0313% 0.087432115 0.007644375 -0.103195744 -0.020717428 0.511976834 -0.260267769 25.1709%
TRAVIS PERKINS 0.010179487 0.6373% 0.110711611 0.012257061 7.924503111 1.138961249 1.029839365 -0.436453223 59.3386%
TESCO 0.001613066 0.0103% 0.065375093 0.004273903 0.915769064 0.013001097 0.425725733 -0.208032945 21.7693%
BOOT (HENRY) 0.013789494 0.7817% 0.078907871 0.006226452 1.309336156 -0.279502392 0.511983041 -0.282972343 22.9011%
WEIR GROUP 0.016872886 1.8405% 0.095578756 0.009135299 2.832522354 -0.759377111 0.67458019 -0.42989359 24.4687%
DAIRY CREST GROUP 0.007496172 1.1565% 0.086069584 0.007407973 1.588508416 -0.255639311 0.613559151 -0.361843354 25.1716%
BODYCOTE 0.0147983 0.9851% 0.108263456 0.011720976 1.198899065 -0.094106482 0.703768686 -0.327474683 37.6294%
RENEWI 0.006482497 0.2570% 0.113226741 0.012820295 14.06789301 1.801752416 1.16462266 -0.407418838 75.7204%
CHARLES TAYLOR 0.004170787 0.0515% 0.079689598 0.006350432 4.859077893 0.53354454 0.719651897 -0.278036852 44.1615%
FERGUSON 0.009392534 1.1326% 0.095604675 0.009140254 6.985313694 -0.892295868 0.849083223 -0.551245933 29.7837%
BARR (AG) 0.013703107 1.3042% 0.058808318 0.003458418 0.686277665 0.241616098 0.3398338 -0.127571187 21.2263%
TULLOW OIL 0.011392879 1.0827% 0.115254566 0.013283615 1.414942795 0.368573236 0.73768335 -0.276541636 46.1142%
MOSS BROTHERS GROUP 0.011385234 -0.1841% 0.113525574 0.012888056 1.977971236 0.558103618 0.793135374 -0.353804281 43.9331%
CARNIVAL 0.009747946 1.4340% 0.074608186 0.005566381 -0.003847384 -0.317579502 0.38236602 -0.199925922 18.2440%
AVEVA GROUP 0.019766891 1.5299% 0.100878259 0.010176423 2.627858573 -0.074898124 0.745829456 -0.336737547 40.9092%
AGGREKO 0.009179858 1.0162% 0.100773208 0.01015524 1.740776053 -0.206065904 0.76540292 -0.391381211 37.4022%
DCC 0.014964206 1.7375% 0.070467297 0.00496564 0.620698435 0.254525119 0.44470315 -0.181555304 26.3148%
LAIRD 0.013015429 1.5490% 0.139891249 0.019569561 4.438967559 0.218313608 1.124283307 -0.543172427 58.1111%
OXFORD BIOMEDICA 0.010807158 -0.8189% 0.19970215 0.039880949 2.210073873 0.857427922 1.434250939 -0.626593588 80.7657%
MJ GLEESON 0.011466921 0.2711% 0.102619002 0.01053066 9.425152319 1.292121766 0.966992648 -0.286503797 68.0489%
HILL & SMITH 0.021460819 1.8663% 0.095270031 0.009076379 2.246546088 0.404530364 0.701144978 -0.252945406 44.8200%
BABCOCK INTERNATIONAL 0.014011364 1.4859% 0.073414153 0.005389638 1.666037663 0.448300046 0.510635724 -0.228766013 28.1870%
JOHNSON MATTHEY 0.009032804 0.4998% 0.07281749 0.005302387 2.805649969 0.100601098 0.593360911 -0.313433308 27.9928%
GREENE KING 0.007663295 1.2217% 0.072212216 0.005214604 2.544466178 -0.290192545 0.577618041 -0.303365764 27.4252%
FIRST GROUP 0.001419822 0.6502% 0.08979604 0.008063329 2.999970301 -0.859284077 0.652922702 -0.408602019 24.4321%
CARPETRIGHT -0.001045839 -0.8067% 0.103537714 0.010720058 1.016136412 0.435625481 0.67568256 -0.272705303 40.2977%
"FTSE (Market)" 0.004539651 0.009938 0.039182998 0.001535307 1.410330044 -0.74794751 0.235470662 -0.136548554 0.0989221

25
"A.2 Regression output for the 2002-2017 period."
"Intercept" "X variable 1"
"Multiple R" "R-square" "Adj R-Square" "Standard Error" "Observations" "Coefficient" "Standard Error" "T-stat" "P-value" "Coefficient" "Standard Error" "T-stat" "P-value"
SIG 0.56880 32.3533% 0.31997 0.09447 192 -0.00198 0.00686 -0.28797 77.3684% 1.66297 0.17445 9.53263 7.51362E-18
HAYS 0.51091 26.1028% 0.25714 0.07524 192 -0.00003 0.00547 -0.00510 99.5935% 1.13820 0.13894 8.19232 3.69067E-14
ROYAL DUTCH SHELL B 0.63484 40.3019% 0.39988 0.04684 192 0.00201 0.00340 0.59074 55.5397% 0.97968 0.08650 11.32555 4.70216E-23
HUNTSWORTH 0.29261 8.5623% 0.08081 0.10626 192 0.00164 0.00772 0.21212 83.2243% 0.82765 0.19622 4.21804 3.8114E-05
BHP BILLITON 0.60118 36.1413% 0.35805 0.07005 192 0.00694 0.00509 1.36366 17.4288% 1.34145 0.12936 10.36975 2.98541E-20
TRAVIS PERKINS 0.60409 36.4921% 0.36158 0.08846 192 0.00243 0.00643 0.37825 70.5669% 1.70685 0.16336 10.44871 1.76057E-20
TESCO 0.43455 18.8836% 0.18457 0.05903 192 -0.00168 0.00429 -0.39130 69.6014% 0.72503 0.10902 6.65067 3.02015E-10
BOOT (HENRY) 0.38546 14.8579% 0.14410 0.07300 192 0.01027 0.00530 1.93549 5.4414% 0.77625 0.13481 5.75816 3.36966E-08
WEIR GROUP 0.60059 36.0711% 0.35735 0.07662 192 0.01022 0.00557 1.83626 6.7881% 1.46502 0.14149 10.35398 3.31706E-20
DAIRY CREST GROUP 0.38603 14.9021% 0.14454 0.07961 192 0.00365 0.00578 0.63051 52.9117% 0.84796 0.14701 5.76822 3.20301E-08
BODYCOTE 0.59662 35.5960% 0.35257 0.08711 192 0.00731 0.00633 1.15574 24.9237% 1.64848 0.16087 10.24757 6.74573E-20
RENEWI 0.43840 19.2195% 0.18794 0.10203 192 0.00073 0.00741 0.09867 92.1502% 1.26684 0.18842 6.72349 2.01922E-10
CHARLES TAYLOR 0.36212 13.1132% 0.12656 0.07448 192 0.00083 0.00541 0.15292 87.8625% 0.73648 0.13753 5.35494 2.45389E-07
FERGUSON 0.58609 34.3506% 0.34005 0.07767 192 0.00290 0.00564 0.51404 60.7818% 1.43004 0.14342 9.97077 4.23381E-19
BARR (AG) 0.29751 8.8514% 0.08372 0.05629 192 0.01168 0.00409 2.85484 0.4784% 0.44653 0.10395 4.29544 2.77934E-05
TULLOW OIL 0.32050 10.2718% 0.09800 0.10946 192 0.00711 0.00795 0.89443 37.2225% 0.94272 0.20214 4.66376 5.841E-06

26
MOSS BROTHERS GROUP 0.34028 11.5788% 0.11113 0.10703 192 0.00691 0.00778 0.88855 37.5366% 0.98589 0.19765 4.98805 1.37143E-06
CARNIVAL 0.47014 22.1034% 0.21693 0.06602 192 0.00568 0.00480 1.18499 23.7503% 0.89520 0.12192 7.34256 5.97547E-12
AVEVA GROUP 0.46886 21.9830% 0.21572 0.08934 192 0.01429 0.00649 2.20116 2.8929% 1.20710 0.16497 7.31687 6.93816E-12
AGGREKO 0.48370 23.3963% 0.22993 0.08843 192 0.00353 0.00642 0.54981 58.3093% 1.24400 0.16330 7.61772 1.1859E-12
DCC 0.51803 26.8356% 0.26450 0.06043 192 0.01073 0.00439 2.44490 1.5401% 0.93163 0.11160 8.34799 1.41312E-14
LAIRD 0.50211 25.2111% 0.24818 0.12130 192 0.00488 0.00881 0.55347 58.0594% 1.79262 0.22399 8.00303 1.17282E-13
OXFORD BIOMEDICA 0.22158 4.9097% 0.04409 0.19525 192 0.00568 0.01419 0.40044 68.9283% 1.12931 0.36056 3.13210 0.002010251
MJ GLEESON 0.21305 4.5392% 0.04037 0.10053 192 0.00893 0.00730 1.22321 22.2767% 0.55798 0.18564 3.00577 0.003006814
HILL & SMITH 0.45225 20.4527% 0.20034 0.08519 192 0.01647 0.00619 2.66072 0.8464% 1.09960 0.15732 6.98940 4.54583E-11
BABCOCK INTERNATIONAL 0.41807 17.4783% 0.17044 0.06687 192 0.01046 0.00486 2.15218 3.2645% 0.78331 0.12348 6.34368 1.60194E-09
JOHNSON MATTHEY 0.65947 43.4907% 0.43193 0.05488 192 0.00347 0.00399 0.87002 38.5386% 1.22556 0.10135 12.09247 2.46208E-25
GREENE KING 0.43855 19.2322% 0.18807 0.06507 192 0.00399 0.00473 0.84491 39.9226% 0.80822 0.12016 6.72624 1.98871E-10
FIRST GROUP 0.39667 15.7348% 0.15291 0.08265 192 -0.00271 0.00600 -0.45082 65.2633% 0.90906 0.15262 5.95640 1.22674E-08
CARPETRIGHT 0.31231 9.7540% 0.09279 0.09862 192 -0.00479 0.00716 -0.66885 50.4405% 0.82526 0.18211 4.53162 1.03321E-05
Equally Weighted Portfolio 84.1326% 0.7078 70.6291% 0.0272 192 0.0051 0.1977% 2.5559 1.1373% 1.0779 5.0241% 21.4547 0.0000%
"A3. Regression output for the 2002-2009 period."
"Intercept" "X variable 1"
"Multiple R" "R-square" "Adj R-Square" "Standard Error" "Observations" "Coefficient" "Standard Error" "T-stat" "P-value" "Coefficient" "Standard Error" "T-stat" "P-value"
SIG 0.6086 37.0375% 0.3637 0.1042 96 -0.0021 0.0106 -0.2002 84.1747% 1.7708 0.2381 7.4361 4.77949E-11
HAYS 0.5126 26.2793% 0.2550 0.0743 96 -0.0036 0.0076 -0.4702 63.9310% 0.9824 0.1697 5.7886 9.3157E-08
ROYAL DUTCH SHELL B 0.6387 40.7913% 0.4016 0.0500 96 0.0021 0.0051 0.4152 67.8940% 0.9190 0.1142 8.0474 2.53933E-12
HUNTSWORTH 0.3772 14.2307% 0.1332 0.1165 96 -0.0018 0.0119 -0.1484 88.2343% 1.0512 0.2662 3.9492 0.000151432
BHP BILLITON 0.6290 39.5584% 0.3892 0.0715 96 0.0199 0.0073 2.7304 0.7554% 1.2817 0.1634 7.8436 6.7877E-12
TRAVIS PERKINS 0.5743 32.9829% 0.3227 0.1087 96 0.0067 0.0111 0.6010 54.9289% 1.6897 0.2484 6.8017 9.48479E-10
TESCO 0.5171 26.7416% 0.2596 0.0551 96 0.0056 0.0056 0.9863 32.6504% 0.7382 0.1260 5.8577 6.87542E-08
BOOT (HENRY) 0.4531 20.5293% 0.1968 0.0834 96 0.0092 0.0085 1.0845 28.0934% 0.9395 0.1907 4.9277 3.56209E-06
WEIR GROUP 0.7069 49.9652% 0.4943 0.0707 96 0.0143 0.0072 1.9840 5.0176% 1.5645 0.1615 9.6886 8.4386E-16
DAIRY CREST GROUP 0.3785 14.3278% 0.1342 0.0958 96 0.0032 0.0098 0.3281 74.3600% 0.8680 0.2189 3.9649 0.00014315
BODYCOTE 0.6077 36.9349% 0.3626 0.0951 96 0.0020 0.0097 0.2038 83.8939% 1.6122 0.2173 7.4197 5.16619E-11
RENEWI 0.4700 22.0913% 0.2126 0.1268 96 0.0067 0.0129 0.5209 60.3658% 1.4958 0.2897 5.1628 1.35496E-06
CHARLES TAYLOR 0.4468 19.9637% 0.1911 0.0731 96 -0.0044 0.0075 -0.5836 56.0906% 0.8093 0.1671 4.8422 5.03511E-06
FERGUSON 0.5922 35.0683% 0.3438 0.0955 96 -0.0023 0.0097 -0.2367 81.3383% 1.5548 0.2182 7.1251 2.08559E-10

27
BARR (AG) 0.2806 7.8764% 0.0690 0.0536 96 0.0151 0.0055 2.7489 0.7170% 0.3476 0.1226 2.8349 0.005611434
TULLOW OIL 0.4053 16.4292% 0.1554 0.0914 96 0.0310 0.0093 3.3166 0.1296% 0.8982 0.2089 4.2988 4.19173E-05
MOSS BROTHERS GROUP 0.4506 20.3023% 0.1945 0.1179 96 0.0008 0.0120 0.0667 94.7001% 1.3191 0.2696 4.8934 4.09382E-06
CARNIVAL 0.5149 26.5117% 0.2573 0.0696 96 0.0052 0.0071 0.7383 46.2153% 0.9258 0.1590 5.8234 7.9982E-08
AVEVA GROUP 0.4897 23.9834% 0.2317 0.0999 96 0.0224 0.0102 2.2000 3.0254% 1.2428 0.2282 5.4458 4.10791E-07
AGGREKO 0.5734 32.8789% 0.3216 0.0949 96 0.0131 0.0097 1.3558 17.8411% 1.4717 0.2169 6.7857 1.02171E-09
DCC 0.5485 30.0858% 0.2934 0.0655 96 0.0094 0.0067 1.4118 16.1319% 0.9519 0.1497 6.3601 7.23867E-09
LAIRD 0.6315 39.8825% 0.3924 0.1287 96 0.0093 0.0131 0.7073 48.1131% 2.3222 0.2941 7.8969 5.25171E-12
OXFORD BIOMEDICA 0.3633 13.1989% 0.1228 0.2117 96 0.0074 0.0216 0.3409 73.3915% 1.8292 0.4838 3.7807 0.000274612
MJ GLEESON 0.2624 6.8869% 0.0590 0.1150 96 -0.0030 0.0117 -0.2527 80.1065% 0.6928 0.2627 2.6368 0.009793602
HILL & SMITH 0.5585 31.1948% 0.3046 0.0883 96 0.0209 0.0090 2.3162 2.2720% 1.3182 0.2019 6.5282 3.35666E-09
BABCOCK INTERNATIONAL 0.4301 18.5018% 0.1763 0.0778 96 0.0201 0.0079 2.5340 1.2930% 0.8210 0.1777 4.6195 1.2204E-05
JOHNSON MATTHEY 0.6978 48.6878% 0.4814 0.0544 96 0.0049 0.0056 0.8848 37.8530% 1.1752 0.1244 9.4442 2.79395E-15
GREENE KING 0.4542 20.6272% 0.1978 0.0760 96 0.0059 0.0078 0.7576 45.0593% 0.8589 0.1738 4.9425 3.35427E-06
FIRST GROUP 0.5576 31.0935% 0.3036 0.0760 96 0.0064 0.0078 0.8265 41.0615% 1.1311 0.1737 6.5128 3.6026E-09
CARPETRIGHT 0.4389 19.2603% 0.1840 0.0896 96 0.0082 0.0091 0.8941 37.3531% 0.9697 0.2048 4.7353 7.72181E-06
Equally Weighted Portfolio 0.8630 74.4714% 0.7420 0.0313 96 0.0078 0.0032 2.4271 1.7124% 1.1851 0.0716 16.5594 0.0000%
"A4. Regression output for the 2010-2017 period."
"Intercept" "X variable 1"
"Multiple R" "R-square" "Adj R-Square" "Standard Error" "Observations" "Coefficient" "Standard Error" "T-stat" "P-value" "Coefficient" "Standard Error" "T-stat" "P-value"
SIG 0.4909 24.0954% 0.2329 0.0843 96 -0.0004 0.0088 -0.0475 96.2202% 1.4567 0.2667 5.4626 3.82479E-07
HAYS 0.5192 26.9566% 0.2618 0.0759 96 0.0016 0.0080 0.2072 83.6298% 1.4143 0.2401 5.8899 5.96611E-08
ROYAL DUTCH SHELL B 0.6325 40.0050% 0.3937 0.0438 96 0.0011 0.0046 0.2424 80.9015% 1.0958 0.1384 7.9170 4.76471E-12
HUNTSWORTH 0.1308 1.7100% 0.0066 0.0945 96 0.0081 0.0099 0.8162 41.6464% 0.3820 0.2987 1.2788 0.204115254
BHP BILLITON 0.6006 36.0712% 0.3539 0.0665 96 -0.0074 0.0070 -1.0618 29.1038% 1.5311 0.2102 7.2828 9.90106E-11
TRAVIS PERKINS 0.6748 45.5343% 0.4495 0.0629 96 -0.0022 0.0066 -0.3362 73.7504% 1.7642 0.1990 8.8649 4.75978E-14
TESCO 0.3616 13.0789% 0.1215 0.0624 96 -0.0091 0.0065 -1.3845 16.9488% 0.7422 0.1974 3.7609 0.000294219
BOOT (HENRY) 0.2407 5.7928% 0.0479 0.0604 96 0.0135 0.0063 2.1234 3.6349% 0.4593 0.1910 2.4042 0.01816927
WEIR GROUP 0.4575 20.9281% 0.2009 0.0824 96 0.0072 0.0086 0.8378 40.4272% 1.2994 0.2605 4.9879 2.7875E-06
DAIRY CREST GROUP 0.4004 16.0336% 0.1514 0.0603 96 0.0044 0.0063 0.6907 49.1435% 0.8072 0.1905 4.2367 5.29074E-05
BODYCOTE 0.5716 32.6735% 0.3196 0.0790 96 0.0124 0.0083 1.4997 13.7033% 1.6865 0.2497 6.7541 1.18301E-09
RENEWI 0.3840 14.7426% 0.1384 0.0679 96 -0.0026 0.0071 -0.3627 71.7640% 0.8658 0.2148 4.0317 0.000112517
CHARLES TAYLOR 0.2368 5.6075% 0.0460 0.0760 96 0.0072 0.0080 0.9020 36.9340% 0.5676 0.2402 2.3631 0.020182873
FERGUSON 0.5744 32.9992% 0.3229 0.0540 96 0.0100 0.0057 1.7579 8.2016% 1.1620 0.1708 6.8042 9.37532E-10
BARR (AG) 0.3422 11.7093% 0.1077 0.0587 96 0.0069 0.0061 1.1161 26.7231% 0.6548 0.1854 3.5308 0.000643807
TULLOW OIL 0.3000 9.0000% 0.0803 0.1210 96 -0.0184 0.0127 -1.4476 15.1050% 1.1668 0.3827 3.0490 0.002981882
MOSS BROTHERS GROUP 0.1111 1.2340% 0.0018 0.0921 96 0.0176 0.0097 1.8236 7.1394% 0.3154 0.2911 1.0837 0.281269401

28
CARNIVAL 0.3967 15.7397% 0.1484 0.0630 96 0.0065 0.0066 0.9904 32.4506% 0.8343 0.1991 4.1903 6.287E-05
AVEVA GROUP 0.4464 19.9262% 0.1907 0.0776 96 0.0062 0.0081 0.7674 44.4783% 1.1866 0.2453 4.8365 5.15179E-06
AGGREKO 0.3344 11.1840% 0.1024 0.0796 96 -0.0036 0.0083 -0.4253 67.1592% 0.8663 0.2518 3.4405 0.000867854
DCC 0.4613 21.2754% 0.2044 0.0556 96 0.0123 0.0058 2.1197 3.6664% 0.8855 0.1757 5.0402 2.24953E-06
LAIRD 0.2389 5.7052% 0.0470 0.1074 96 0.0071 0.0113 0.6326 52.8552% 0.8096 0.3395 2.3848 0.019094814
OXFORD BIOMEDICA 0.0369 0.1361% -0.0093 0.1714 96 0.0130 0.0180 0.7221 47.1997% -0.1939 0.5418 -0.3579 0.721194859
MJ GLEESON 0.0914 0.8358% -0.0022 0.0824 96 0.0231 0.0086 2.6762 0.8786% 0.2318 0.2604 0.8901 0.375682744
HILL & SMITH 0.2746 7.5416% 0.0656 0.0810 96 0.0147 0.0085 1.7283 8.7220% 0.7090 0.2561 2.7690 0.006774485
BABCOCK INTERNATIONAL 0.4282 18.3392% 0.1747 0.0529 96 0.0008 0.0055 0.1532 87.8539% 0.7680 0.1672 4.5946 1.34559E-05
JOHNSON MATTHEY 0.6144 37.7531% 0.3709 0.0557 96 0.0013 0.0058 0.2232 82.3887% 1.3299 0.1761 7.5506 2.76766E-11
GREENE KING 0.4097 16.7829% 0.1590 0.0525 96 0.0027 0.0055 0.4868 62.7515% 0.7226 0.1660 4.3540 3.40127E-05
FIRST GROUP 0.1978 3.9130% 0.0289 0.0872 96 -0.0094 0.0091 -1.0235 30.8684% 0.5394 0.2757 1.9565 0.053371666
CARPETRIGHT 0.1895 3.5918% 0.0257 0.1058 96 -0.0165 0.0111 -1.4851 14.0864% 0.6259 0.3345 1.8714 0.064401402
Equally Weighted Portfolio 0.8112 65.7987% 0.6543 0.0209 96 0.0036 0.0022 1.6462 10.3052% 0.8895 0.0661 13.4478 0.0000%

Vous aimerez peut-être aussi