THE CAPITAL ASSET PRICING MODEL (CAPM)

In the chronological development of modern financial management, portfolio theory came first with Markowitz in 1952. It was not until 19 ! that "illiam #harpe derived the capital Asset Pricing Model (CAPM)1 $ased on Markowitz%s portfolio theory. &or e'ample, a key assumption of the ()*M is that investors hold highly diversified portfolios and thus can eliminate a significant proportion of total risk. +he ()*M was a $reakthrough in modern finance $ecause for the first time a model $ecame availa$le which ena$le academic, financiers and investors to link the risk and return for an asset together, and which e'plained the underlying mechanism of asset pricing in capital markets. T PES O! IN"ESTMENT RIS# In the preceding chapter we have seen how the total risk ,as represented $y the standard deviation, - of a two.security portfolio can $e significantly reduced $y com$ining securities whose returns are negatively correlated, or at least have low positive correlation / the principle of diversification. )ccording to the ()M*, the total risk of a security or portfolio of securities can $e split into two specific types, s$ste%atic ris& and 'ns$ste%atic ris&. +his is sometimes referred to as risk partitioning, as follows 0 Total ris& ( S$ste%atic ris& ) *ns$ste%atic ris& Systematic (or market) risk cannot $e diversified away 0 it is the risk which arises from market factors and is also fre1uently referred to as undiversifiable risk. It is due to factors which systematically impact on most firms, such as general or macroeconomic conditions ,e.g. $alance of payments, inflation and interest rates-. It may help you remem$er which type is which if you think of systematic risk as arising from risk factors associated with the general economic and financial system.

Unsystematic (or specific) risk can $e diversified away $y creating a large enough portfolio of securities 0 it is also often called diversifiable risk or company-unique risk. It is the risk which relates, or is uni1ue, to a particular firm. &actors such as winning a new contract, an industrial dispute, or the discovery of a new technology or product would contri$ute to unsystematic risk.
+he relationship $etween total portfolio risk, p, and portfolio size can $e shown diagrammatically as in &igure $elow. 2otice that total risk diminishes as the num$er of assets or securities in the portfolio increases, $ut also o$serve that unsystematic risk does not disappear completely and that systematic risk remains unaffected $y portfolio size.

THE CAPM MODEL "e have previously descri$ed the ()*M as a method of e'pressing the risk.return relationship for a security or portfolio of securities0 it $rings together systematic ,undiversifia$le- risk and return. )fter all, for any rational, risk.averse investor it is only systematic risk which is relevant, $ecause if the investor creates a sufficiently large portfolio of securities, unsystematic or company.specific risk can $e virtually eliminated through diversification. It is therefore the measurement of systematic risk which is of primary importance for rational investors in identifying those securities which possess the most desired risk.return characteristics. It is the measurement of systematic risk which $ecomes critical in the ()*M $ecause the model relies on the assumption that investors will only hold well diversified portfolios, so only systematic risk matters. +he ()*M is 1uite a comple' concept so if you find it difficult to grasp at first do not $ecome disillusioned, stick with it. &or reasons of presentation and ease of understanding we will approach our study of the ()*M $y $reaking it down into five key components as follows0 +, T-e .eta coe//icient0 ( )1 2, T-e CAPM e3'ation1 4, t-e CAPM grap-5t-e sec'rit$ %ar&et line (SML)1 6, S-i/ts in t-e SML5in/lationar$ e7pectations and ris& a8ersion1 9, Co%%ents and criticis%s o/ t-e CAPM, 3et us e'amine each component in turn, $eginning with the key concept of $eta, , T-e .eta coe//icient ( ) 4ecall that the standard deviation, , is used to measure an asset or share%s total risk, while the $eta coefficient, , in contrast is used to measure only part of a share or portfolio%s risk, namely the part that cannot $e reduced $y diversification, that is the systematic or market risk of an individual share or portfolio of shares. #ystematic risk can $e further su$divided into $usiness risk and financial risk. 5usiness risk arises from the nature of the firm%s $usiness environment and the particular characteristics of the type of $usiness or industry in which it operates. &or e'ample the competitive structure of the industry, its sensitivity to changes in macroeconomic varia$les such as interest rates and inflation and the sta$ility of industrial relations all com$ine to determine a firm%s $usiness risk. +he level of $usiness risk in some industries, for e'ample catering and construction, is higher than in others and is a varia$le which lies largely outside management%s control.

Categories o/ .eta #hares or securities can $e $roadly classified as aggressive, average or defensive according to their $etas. #hares with a $eta61.7 are descri$ed as aggressive8 they are more risky than the market average, although they will tend to perform well in a rising or $ull market. (onse1uently investors would re1uire a rate of return from the share which is greater than the market average. #hares with a $eta 9 1.7 are descri$ed as average or neutral as their rate of return moves in e'act harmony with movements in the stock market average return8 they are of average risk and yield average returns. In contrast, shares with a $eta : 1.7 are classed as defensive. ) defensive share does not perform well in a $ulk market $ut conversely it does not fall as much as the average share in a falling or $ear market. 5ut if they are held in portfolio then they will set off the returns also. Alternati8e deri8ation o/ .eta ;sing past data on individual share and market returns over a sufficiently lengthy period, say, the most recent four to five years, $etas can also $e calculated statistically. &or e'ample the $eta ,-of a share ,#- is e1ual to the covariance $etween the share%s returns and the market%s returns ,(<=sm- divided $y the variance of the market%s returns ,=arm->which in turn is the standard deviation of the market%s returns s1uared, that is0 5eta, s (ovariancesm (<=sm (<=sm 9 >>>>>>> 9 >>>> 9 >>>> =ariancem =arm 2m

+he returns on a suita$le stock market inde' can $e used as a pro'y for the market returns. &or e'ample, su$stituting the &+#? 177 #hare Inde', the $eta , - of a share ,#- would $e calculated as0 (<=s,&+#?177 (<=s,&4#?177 5eta,s 9 >>>>>> 9 >>>>>> =ar&4#?177 2&+#?177 )s the covariance of each individual share is divided $y a common denominator, the variance of the market ,=arm- or a suita$le surrogate market inde', we end up with a standardised measure of risk, that is, the share%s $eta. 5eing a standardised measure we are a$le to directly compare the $eta of one share with the $eta of another. Port/olio .etas )s it is learned that a share%s $eta represents only part of a share%s risk, namely the element of systematic or market risk, which is the risk element that cannot $e diversified away. "hen it comes to including a share in a portfolio we are only concerned with the impact of that share%s market risk on the portfolio risk. In a portfolio conte't market risk is also the only relevant risk and $eta is the $est measure. +he portfolio $eta measures the portfolio%s responsiveness to macroeconomic varia$les such as inflation and interest rates.

+o determine the systematic risk for a portfolio, that is the portfolio $eta, we simply calculate a weighted average of the $etas of the individual securities making up the portfolio, as follows02 *ortfolio $eta, p 9 w11 @ w22 @ wAA B @ wnn where, p 9 the portfolio $eta ,i.e. risk of the portfolio relative to the marketwi 9 portfolio weightings of the individual securities ,where I 9 1, 2, B ni 9 $eta of the individual security ,where i 9 1,2,Bnn 9 num$er of securities in the portfolio O.taining and interpreting .etas 5etas can $e o$tained from pu$lished sources e.g. the 3ondon 5usiness #chool ,35#- and through $rokerage firms. +he 35# pu$lished values and other data for ;C and Irish companies listed on the 3ondon #tock ?'change every 1uarter in its 4isk Measurement #ervice pu$lication. Individual $etas are produced for all companies listed in the &inancial +imes ,&+- )ll. #hare Inde'. +he individual $etas are calculated from the monthly returns over the most recent five.year period related to the monthly returns from &+ )ll.#hare Inde' using a standard least s1uares regression computer programme. T-e CAPM e3'ation "e will now e'amine the actual e1uation for the capital asset pricing model. It is one of the most famous e1uations in financial management. +he ()*M e1uation links together risk and the re1uired return for a share. It shows, for e'ample, that the return a rational investor would re1uire on a particular share, 4,ri-, is a function of the share%s market or systematic risk ,$eta-, i, and a risk premium to compensate for investing in the risky market. +hus the higher the risk, the higher the return the investor will re1uire a vice.versa. #imply stated, the underlying precept of the ()*M is that the e'pected return on a security is composed of two elements as follows0 ?'pected return, ?,r- 9 a risk.free interest rate @ a risk premium ;sing the capital asset pricing model ,()*M- this relationship is e'pressed more formally as0 ?,ri- 9 rf @ i,?4m / 4fwhere, ?,rI- 9 re1uired return on assetDshare I 4f 9 risk.free rate of return i 9 $eta coefficient for assetDshare i

?4m 9 e'pected market return, that is the return e'pected on the market portfolio of share. )s it is seen a$ove, the ()*M e1uation can $e split into two segments0 1. 2. the risk.free rate of return, 4f8 and +he risk premium, i ,?4m / 4f-

"e will discuss the risk.free rate of return 4f, first. T-e CAPM grap- : t-e sec'rit$ %ar&et line (SML) Eaving now had some practice in using the ()*M to calculate the e'pected returns you will have noticed that the ()*M e1uation is in fact a straight line e1uation. (onventionally the e1uation for a straight line is usually given as0 y 9 a' @ c.

R'les o/ CAPM ;<<<<<<< E ( Rp ) ( R/ ) p =E ( R% ) < R/ > ?-ere0 E (Rp) ( E7pected ret'rn o/ t-e port/olio R/ ( Ris& /ree rate o/ Ret'rn p ( Port/olio @eta i,e, %ar&et sensiti8it$ inde7 ( c-ange in e7pected rate o/ ret'rn c-ange in %ar&et rate o/

ret'rn E(R%) ( E7pected Ret'rn on %ar&et port/olio =E (R%) : R/> ( Mar&et ris& pre%i'%, (Rs%) (S,D, s ) S,D,% S,D,s ( Standard de8iation o/ an asset , E(R%) ( ( DI" ) Cap, Gain ) Total

Port/olio @eta ( p) (

?-ere0 Rs% ( Correlation Coe//icient Ait- %ar&et , S,D,% ( Mar&et Standard de8iation , In8est%ent

(apital Fain 9 Market =alue . =alue of <riginal Investment .