‘ter you read this chapter, you should be abe to answer the following questions: ‘What do we mean by rik aversion, and what evidence indicates that investors are general risk averse? ‘What are the basic assumptions behind the Markowitz portfolio theory? ‘What do we mean by rs, and what are some measures of risk used in investments? How do we compute the expected ate of return for an individual risky asset ora portato of assets? How do we compute the standard deviation of rates of return for an individual risky asset? What do we mean by the covariance between rates of return, and how is it computed? ‘What is the relationship between covariance and correlation? ‘What is the formula for the standard deviation for a portfolio of risky assets, and how does it differ from the standard deviation of an individual risky asset? Given the formula for the standard deviation of a portfolio, how do we diversify a portfolio? What happens to the standard devition of a portolio when we change the conelation between the assets in the portfolio? Whats the sk-return efficient frontier of sky assets? Would you expect ifferent investors to selec diferent porto from the st of portfolios ‘on the ecient frontier? ‘What determines which portfolio on the ficient frontier ‘is selected by an individual investor? ‘One of the major advances in the investment field during the past few decades has been the recognition that you cannot create an optimum investment portfolio by simply combining numerous individual securities that have desirable risk-return characteristics. Specifically, it has been shown that an investor must consider the relationship among the investments to build ‘an optimum portfolio that witl meet investment objectives. The recognition of how to create an ‘optimum portfolio was demonstrated in the derivation of portfolio theory. In this chapter we explain portfolio theory step by step. We introduce the basic portfolio risk formula for combining different assets. Once you understand this forraula and its impli cations, you will understand why you should diversify your porfolio and also how you should iversify 7.1 Some Background ‘Assumptions Before presenting portfolio theory, we need to clarify some general assumptions ofthe theory. ‘This includes not only what we mean by an optimum portfolio but also what we mean by the terms risk aversion and risk, ‘One basic assumption of portfolio theory is that investors want to maximize the returns from your total set of investments for a given level of risk. To adequately deal with such an assumption requires certain ground rules. First. your portfolio should include all of your assets ‘and liabilities, not only your marketable securities bt also yout car, house, and less marketable142, PART -LIVILOPMF ITS Wt SIA TCE investments such as coins, stamps, art, antiques, and furniture. The full spectrum of invest- 3 iments must be considered because the returns (ror all these investments interact, and this 4 ‘relationship among the returns for assets inthe portfolio is important, Hence, a good portfolio is ‘ot simply a collection of individually good investments, 5 7.1.1 Risk Aversion Portfolio theory also assumes that investors are basically risk averse, meaning that, given a choice between two assets with equal rates of return, they will select the asset with the lower level of He isk, Evidence that most investors ate risk averse is that they purchase various types of insurance, Jon ineluding life insurance, car insurance, and health insurance. Buying insurance basically involves ‘an outlay of given known amount to guard against an uncertain, possibly larger, outlay in the future. Further evidence of risk aversion is the difference in promised yield (the required rate of On return) for different grades of bonds with different degrees of credit risk, Specifically, the prom: ene ised yield on corporate bonds increases from AAA (the lowest risk class) 0 AA to A, and so on, we indicating that investors require a higher rate of retura to accept higher risk. ‘This does not imply that everybody is visk averse, or that investors are completely risk avecae regarding all financial commitments The fact is, not everybody buys insurance for ta everything. Some people have no insurance against anything, cither by choice or because they rete ‘cannot aflord it In addition, some individuals buy insurance related to some risks such as auto et accidents or illness, but they also buy lotery tickets and gamble at race tracks or in easinos, ie where it i known thatthe expected returns are megative (hich means that participants are only willing o pay for the excitement ofthe risk involved). This combination of risk preference and ear risk aversion can be explained by an attitude toward rsk hat depends on the amount of money ret involved, Researchers such as Friedman and Savage (1948) speculate that this isthe case for ae people who like to gamble for small amounts (in lotteries or slot machines) but buy insurance » to protect themselves against large losses such as fire or accidents. issue ‘While recognizing such attitudes, we ascume that most investors with a large investment ata portfolio are risk averse. Therefore, we expect a positive relationship between expected return dard, and expected risk, which is consistent with historical results as shown in Chapter 3, ae 7.1.2 Definition of Risk mode Although there is a difference in the specific definitions of risk and uncertainty, for our pur- [poses and in most financial literature the two terms are used interchangeably. For most inves- Wea tos, rick means the uncertainty of future outcomes. An alternative definition might be the prab- Thee ability of an adverse outcome. tn our subsequent discussion of portiolio theory, we consider tong several measures of risk that are used when developing and applying the theory. tT 7.2 Markowitz Portfolio Theory propo In the early 1960s, the investment community talked about risk, but there was no specific mea- sure for the term. To build a portfolio model, however, investors had to quantify their risk vari- able. The basic portfolio model was developed by Harry Markowitz (1952, 1958), who derived the expected rate of return for a portfolio of assets and an expected risk measure. Markowitz showed that the variance ofthe rate of return was a meaningful measure of portfolio risk under a reasonable set of assumptions, More important, he derived the formula for compating the variance of a portfolio. This portfolio variance formula not only indicated the importance of diversifying investments to reduce the total risk ofa portilio but also shoved iow to effectively diversify, The Markowitz model is based on several assumptions regarding investor behavior: 1. Investors consider each investment alternative as being represented by a probal disteibution of expected returns over some holding period. 2. Investors maximize one-period expected utility, and their utiity curves demonstrate i diminishing marginal utility of wealth, ecm(CHAPTER -AN NTROOUCHION TOPORTTOVO MANAGEMENT 183 3. Investors estimate the risk of the portfolio on the bass ofthe variability of expected returns. Tvestors base decisions solely on expected return and risk, so their utility curvesarea function ‘of expected return and the expected variance (or standard deviation} of returns only. 5. Fora given risk level, investors prefer higher returns to lower returns, Simitarly, for s given level of expected return, investors prefer less risk (o more risk, Under these asumptions, a single asset or portfolio of assets is considered to he efficient if tno other asset or portfolio of assets offers higher expected return with the same (or tower} risk or lower risk with the seme for higher) expected return 7.2.1 Alternative Measures of Risk One of the best-known measures of risk is the variance, or standard deviation of expected returns! Iti a statistical measure of the dispersion of returns around the expected value whereby a larger variance or standard deviation indicates greater dispersion. The idea is that the more disperse the expected returns, the greater the uncertainty of future returns. Another measure of risks the range of returns. It isassuumed that a larger range of expected ‘returns, from the lowest tothe highest, means greater uncertainty regarding future expected returns, Instead of using measures that analyze all deviations from expectations, some observers believe that investors should be concerned only with returns below expectations, which means ‘only deviations below the mean value. A measure that only considers deviations below the mean isthe semivariance. An extension of the semivariance measure only computes expected returns below zero (that is, negative returns), or returns below the returns of some specific asset such as T-bill, the rate of inflation, or a benchmark. These measures of risk implicitly assume that investors want to minimize the damage (regret) from returns less than some target rate. ‘Assuming that investors would welcome returns above some target rate, the returns above such a target rate are not considered when measuring risk. ‘Although there are numerous potential measures of risk, we will use the variance or stan- dard deviation of rturns because (1) this measure is somewhat intuitive, (2) it isa correct and widely recognized risk measure, and (3) ithas been used in most ofthe theoretical asset pricing models. RSX4ARAP OMS a2 a 7.2.2 Expected Rates of Return We compute the expected rate of return for an individual investment as shown in Exhibit 7.1 ‘The expected return for an individual risky asset withthe set of potential returns and an assump- tion of the different probabilities used in the example would be 10.3 percent. ‘The expected rate of return for a portfolio of investments is simply the weighted average of the expected rates of return forthe individual investments in the portfolio. The weights are the proportion of total value for the individual investment. re der Expected Securty Return percent) ‘00280 ‘00300 0240 0210 HAp= 01030 "wie consider the variance and standstddevaion 3s one measure of rik because the standard deviation she suare aot ofthe areWA PART 2: DIVUOMMENTS MIRVESTMENT TIKORY Weighting (percent port 020 030 030 020 “the expected rate of return for a hypothetical portfolio with four risky assets is shown in Exhibit 72. The expected return for this portfolio of investments would be 11.5 percent. The cffect of adding or dropping any investment from the portfolio would be easy to determine; we ‘would use the new weights based on value andl the expected returns fr each ofthe investments. ‘We can generalize this computation ofthe expected return for the portfolio E(R,,,) 3s follows: a E(Ryn) = Sw, where: ‘w, = the weight of an individual asset in the portfolio, or the percent of the portfolio in Asset R, = the expected rate of return for Asset 7.2.3 Variance (Standard Deviation) of Returns for an Individual Investment ‘As noted, we willbe using the variance or the standard deviation of returns as the measure of risk, Therefor, at this point we demonstrate how to compute the standard deviation of returns for an individual investment. Subsequently, after discussing some other statistical concepts, we will consider the determination of the standard deviation for a portfolio of investments ‘The variance, or standard deviation, is a measure of the variation of possible rates of return R from the expected rate of return E(R) as follows: Bi ~ FFP, 72 Variance where: probability of the possible rate of return R, 73 Standard Deviation = & = yf DIR, — E(R)??, ‘The computation of the variance and standard deviation of returns for the individual risky asset in Exhibit 7.1 is set forth in Exhibit 7.3 for 74 when7.2.4 Variance (Standard Deviation) of Returns for a Portfolio ‘Two basic concepts In statistics, covariance and correlation, must be understood before we discuss the formula for the variance of the rate of return for a portfolio. Covariance of Returns In this subscetion we discuss what the covariance of returns is intended to measure, give the formula for computing it, an present an example of ts compu tion. Covariance isa measbre of the degree to which two variables move together relative 10 ther individual mean values overtime. In portfolio analysis, we usually are concerned with the ‘covariance of rates of return rather than prices or some other veriable.? A pasitive covariance sncans thatthe rates of return for two investments tend to move in the same direction rela tive to ther individual means during the same time period. In contrast, 2 negative covariance indicates thatthe rates of return for two investments tend to move in different directions ela- tive to their means ducing specified time 1. The magnitude ofthe covariance ‘depends on the variances ofthe individual retura series, as well as onthe relationship between the series Exhibit 74 contains the monthly rates of return values for U.S stocks (measued using the Wilshire 5000 index) and bonds (measured by the Lehman Brothers Treasury Bond Inde) oth indexes are total return indexes—that is, the stock index includes dividends paid and the bond index includes acerued interest, as discussed in Chapter 5. Using end-of-month values for each index, we compute the percentage change in the index each month, which equals its ‘monthly rates of return during 2007. Fxhibits 75 and 7.6 contain a time-series plot of these ‘monthly rates of return, Although the rates of return for the two assets moved together dur {ng some months, n other months they moved in opposite directions. The covariance statistic provides an absolute measure of how they moved together over time, For two assets, / and j, we define the covariance of rate of return as 7A Covy=E{[R, — E(R)I[R, ~ E(R)]} 7 16 feo 166 Moe a7 005 er 333 052 May W155 -a50 aun 347 ~o04 a uaz 16 hug Bas is sep 203 054 on 7473 079 Now 076 307 Dee 623 08 Mean 030 os Wiehe 5000 Stock lade {tehman Brothers Teasury Bonds index 2007 Monthly Rate of Return) ‘Monthly Rate of Return) “CHAPTER? -AN INTRODUCTION 10 #2180110 MANAGLIMLNT 185 ‘usa of counts canbe measured na vary of wap, depending on the rype of set. You vl cl shat we deine returns (8) in Chapter Vas +o BY sebere Bis nding value, BY is beaming value, and CF ithe cath low daring the period1086 PART2-DLVILOPMINIS IN RVESIMLNT THLORY Monthly Rates of Return Montly Rate ofReturn ‘When we apply ths formula o the monthly rates of return forthe Wilshire 000 stock and ‘he Treasury bond indexes during 2007, it becomes 12 ne [R, — RR, ~ Note that when we apply formula 7.4 to actual sample data, we use the sample mean (R) as an estimate of the expected return and divide the values by (n ~ 1) rather than by 1 0 avoid statistical bias. ‘As ean be seen ifthe rates of return for one asset are above (below) its mean rate of return (® during. given period and the returns forthe other asset at likewise above (below) its mean rate of return during this same period, then the product of these deviations from the mean is positive. If this happens consistently, the covariance of returns between these two assets will bbe some large positive value. Uf, however, the rate of retum for one ofthe securities is above its mean return, while the return on the other security is below its mean return, the product will oft ig itt and ing. 073 Det(CHAPTER 7 AVINTRODUCHON TOFURFIEMO MANAGEMENT 187 bbe negative. If this contrary movement happens consistently, the covariance between the rates of return forthe two asses wil be a large negative value. Exhibit 77 includes the monthly rates of return during 2007 contained in Exhibit 74, One ‘night expect returns fr the two asset indexes to have reasonably ow covariance because of the differences inthe nature ofthese assets. The arithmetic means ofthe monthly returns were =0.30 (stocks) and ‘We rounded al figures to the nearest hundredth of | percent, so there may be small round- ing ervors, The average monthly return was-0.30 percent for the Wilshire 5000 stock index and 0.73 percent for the Treasury bond index, The results in Exhibit 7.7 show that the covariance between the rates of return for these two assets was x 935 = 088 Interpretation of number such as 0.85 is dificult: ist high or low for covariance? We know the relationship between the two assets is generally postive, but it is not possible to be more spe cific, Exhibit 78 contains a scaterplot wit poired values of R, and R, plotted against each other, ~ ‘This plot demonstrates the linear nature and strength ofthe relationship. I isnot surprising that ® ‘the relationship during 2007 was not very strong, since during five months the two assets moved ao ‘counter to each other. Asa result, the overall cavariance was a sill postive value. am Covariance and Correlation Covariance is affected by the variability of the two indi- can vidual retuen indexes. Therefore, a sumber such as the 0.85 in our example might indicate a ais weak positive lationship ifthe two individual indexes were volatile, but would reflect a strong a positive relationship if the two indexes were stable. Obviously, we want to standardize this wellWea PART ae LOAM NTS IN STEIN THR Monthly Return for Lehman Brothers | | | "Treasury Bond Index 1 sol Money Returns for Wilshire 5000 Stock index ‘covariance measure. We do so by taking inte consideration the variability of the two individual return indexes, as follows the correlation coefficient of returns the standard deviation of R, ‘Standardizing the covariance by the product of the individual standard deviations yields the correlation coefficient r, which can vary only in the range ~1 to +1. A value of +1 indi- cates a perfect positive near relationship between R, and R,, meaning the returns forthe two assets move together in a completely linear manner, A value of ~1 indicates a perfect negative relationship between the two retura indexes, so that when one asset's rate of return is above its ‘mean, the other asset’ rate of return will be below its mean by a comparable amount. ‘To calculate this standardized measure of the relationship, we need to compute the standard deviation for the two individual return indexes. We already have the values for (By RD ad By ~ Rin Eni 77. We can equate each ofthese ves nd im them as shown in Exhibit 79 to calculate the variance of each return series; again, we divide by {n= 1) to avoid statistical bias Linas -199 and 13.05 = 1.19 ‘The standard deviation for each index is the square root of the variance for each, as follows: tions Trea Obvi Asno that had» they « insige perio. Portt covari of rete that tb return the po inthe the ine for the 16(CHAPTER 7 ASINTHOOUCTON TO PORTTONOMANALEMENT 189 ‘Wahire S000 Stock index Lehman Brothers Treasury Bond index 2007 mR) (Re a bo 12 146 08, 079 Fe 136 19 093 oar Mar «7 1632 O78 060 Ape 302 np “o2i 908 May 135 156 a det we aur 1008 “avr ase aa tan 202 093 87 ug 376 mM oat on Ser 232 S20 019 oot oct 40 wer 006 000 Now as 021 234 548 Dee 593 Par} 06s aaa wom 2091 ims 1205 vatincey©12081/11= 1099 viance= 1305/1 1s Standord Deviation = (109516332 ‘Starla Devotion, =(.19)"%= 1.09 ‘Thus, based onthe covariance between the two indexes ad the individual standard devi tions, we can calculate the correlation coefficient between returns for common stocks and ‘Treasury bonds during 2007: tpn 0808S we)” 352108) ~ 3.62 ‘Obviously, this formula also implies that = 0235 avy = nyrar ~ (0235)(3.32)(1.09) = 085, 7.2.5 Standard Deviation of a Portfolio ‘As noted, a correlation of +1. indicates perfect positive correlation, and a value of -1.0 means that the returns moved in completely opposite ditections. A value of zero means that the returns had no linear relationship, thats, they were uncorrelated statistically. That does not mean that they are independent. The value of r, ~ 0.235 is not significantly different from zero. This insignificant postive correlation is not unusual for stocks versus bonds during & short time period such as one year Portfolio Standard Deviation Formula Now that we have discussed the concepts of covariance and correlation, we can consider the formula for computing the standard deviation ‘of returns for a portfolio of assets, our measure of risk for a portfolio. In Exhibit 7.2, we showed that the expected rate of return of the portfolio was the weighted average of the expected returns for the individual assets in the portfolio; the weights were the percentage of value of the portfolio. One might assume itis possible to derive the standard deviation of the portfotio in the same manner, that i, by computing the weighted average of the standard deviations for the individual assets. This would be a mistake. Markowite (1959) derived the general formula for the standaed deviation of a portfolio as follows: 76 onan y Swiar + 3 Sw co190. PARE 2.11961 09M NI IN INVESTMENT TORY Oye, ™ the standard deviation of the portfolio e "w, = the weights of an individual asset in the portfolio, where weights are determined by the proportion of value in the portfolio he variance of rates of return for asset i co Cav, = the eawarkance between the rates of return for assets and j, where Cov, = 1479, ‘This formula indicates that the standard deviation for a portfolio of assets is function 2 of the weighted average wo the individual variances (where the weights are squared), plus the weighted covariances hetwven all the assets in the portfolio, The very Important point is that the standard deviation for a portfolio of assets encompasses not only the variances of the indi- ‘vidual assets but alzo includes the covariances between all the pairs of individual assets in the portfolio. Further, it can be shown that, in a portfolio with a large number of securities, this th formula reduces to the sum of the weighted covariances, Impact of a New Security in a Portfolio Although in most of the following discussion we will consider portfolios with only two assets (because itis possible to show the effect in two dimensions), we will also demonstrate the computations for a three-asset portfolio. Stil, i is important at this point to consider what happens in a large portfolio with many assets. Specifically, what happens to the portfolios standard deviation when we add @ new security to such a portfolio? As shown by the forrmula, we see two effects. The first is the assets own Int ‘variance of returns, and the second is the covariance between the returns of this new asset a and the returns of every other asset that is already i the portfolio, The relative weight ofthese x ‘numerous covariances is substantially greater than the assets unique variance; the more assets {nthe portfolio, the more this is true. This means that the important factor to consider when adding an investment to & portfolio that contains a number of other investments is not the new security’s own variance but the average covariance of this asset with all other investments in the portfolio, Portfolio Standard Deviation Calculation Because of the assumptions used in devel- oping the Markowitz portfolio model, any aset or portfolio of assets can be described by two characteristics: the expected rate of return and the expected standard deviation of returns. "Therefore, the following demonstrations can be applied to two individual assets. two portfolios of assets, or two asset classes with the indicated rate of return-standard deviation characteris. The tics and correlation coefficients. 0.00 0.10 Equat Risk and Return—Changing Correlations Consider first the casein which both assets have they the same expected return and expected standard deviation of return. As an example, let's assume 7 rt ER) = 020, Eo,) = 010 EQ} 0.20, E(o,) = 0.10 To show the effect of different covariances, we assume different levels of correlation between the two assets. We also assume that the two assets have equal weights in the port- folio (w, ~ 0.50: w, ~ 0.50), Therefore, the only value that changes in each example is the correlation between the returns for the two assets Now consider the following five correlation coefficients and the covariances they yield Since Cow, = r,o,9, the covariance will be equal tor, (0.10)(0.10) beceuse the standard devia- tion of both assets 160.10. 1.00)(0.10)(0.10) = 0.01 (0.50)(0.109(0.10) = 0.005 herearte the veld. Gov, 4 = (-0.80K0.10)(0.10) Cov) = (= 1.00){0.10)(0.10) = Now let's see what happens to the standard deviation of the portfolio under these five conditions. ‘When we apply the general partflo formula from Equation 7.6 10 a two-asset portfolio, iis 2 yan = Viner vr + Dowie es Som = Vai has n ise Veeasyniy + Toxsyoa1) = H525)C000) Voor =0.10 pe Ia this case, where the returns forthe two assets are perfectly positively correlated, the standard deviation forthe portfolio ts, infact, the weighted average of the individual standard deviations. "The important point is that we get no real benefit from combining two assets that are perfectly correlated; they are like one asset already because their returns move together. ‘Now consider Case b, where, , equals 0.50. =o 988 “The only term that changed from Case ais the last term, Cov,» which changed from 0.01 to 0.005. As result, the standard deviation of the portfolio declined by about 13 percent, from 0.10 100.0866, Note that the expected retura of the porfolio did not change because itis simply the weighted average ofthe individual expected returns; i is equal to 0.20 in both cases, You should be able to confirm through your own calculations that the standard deviations for Portfolios cand d are as follows: 0.0707 4.005 ‘The final case, where the correlation between the two assets is ~ 1.00, indicates the ultimate benefits of diversification S petes= W105)(0.10)" + (05)*(0.10)" + 2(05)(05)(—0.01) = V(e.0050) + (0.0080) vo Here, the negative covariance term exactly offsets the individual variance terms, leaving an overall standard deviation of the portfolio of zero. This would be a risk-free poryfalio, (CHAPTER 7 AN INVRODUCTION TO HORTIOLIO MANAGEMENT 191392. PART -DIViLOSMINTS IRMA STMT HILGIEY Exhibit7.10 illustrates a graph of such a pattern, Perfect negative correlation gives a mean ‘combined retuen for the two securities over time equal to the mean for each of them, so the returns for the portfolio show no variability. Any relurns above and below the mean for each of the assets are completely offct by the return for the other asset, so there iso variability in total retuens—that is, no risk-tor the portfolio, Thus, a pair of completely negatively corre- {ated assets provides the maximum beaelits of diversification by completely climinating risk ‘The graph in Exhibit 7.11 shows the difference inthe risk-return posture for our five eases As noted, the only effect ofthe change in correlation isthe change inthe standard deviation of this two-assel portfalio. Combining assets that are not perfectly correlated does nos affect the ‘expected return of the portfolio, but it does reduce the risk of the portfolio (as measured by its standard deviation). When we eventually ceach the ultimate combination of perfect negative correlation, risk is eliminated, Combining Stocks with Different Returns and Risk We have seen what happens when only the ‘correlation coefficient (covariance) differs between the assets. We now consider lwo assets (or portfolios) with different expected rates of return and individual standard deviations." We will ‘show what happens when we vary the correlations between them, We will assume two assets with the following characteristics Asset co) “ 7 ‘10 05000087 2 020 50 00100 ‘We will use the previous set of correlation coefficients, but we must recalculate the covariances because this time the standard deviations of the assets are different. The results are shown in this table. case [Correlation Coefficient.) Covariance > i90 ‘00070 b +080 9.0035, « 000 ‘9.0000 ¢ 050 00035 e 190 -20070 Returss rom Asset Aver Time Mean Return ram Portola of Ascets A and 8 Returns from Asset Bove Time asnted thee could be two set classes, For example, Astet | could below sisk-low rcum bonds and Asset? could behigherretars-bigher ek stocks Agai age ‘natCee 002 003 00% 005 006 007 008 an» a1 ‘Standard Deviation of Return (a) CHAPTER 7 ANINIRCOUCTION TO PORILOLO MANACEWENT 193 Because we are assuming the same weights in all cases (0.50 ~ 0.50), the expected return in every instance will be 50 (0.10) + 0.50 (0.20) FR o1s ‘The portfolio standard deviation for Case a will be peter V(05)*(0.07) + (0.5)(0.10)? + 2(0.5)(0.5)(0.0070) = 0.007225 =0.085 ‘Again, with perfect postive correlation, the portfolio standard deviation is the weighted aver- age ofthe standard deviations of the individual assets: (0.5) (0.07) + (0.8) (0.10) = 0.085 ‘As you might envision, changing the weights with perfect positive correlation causes the portfolio standard deviation to change in a linear fashion. This will be an important point to remember when we discuss the capital asset pricing model (CAPM) in the next chapter. For Cases b, , d, and ¢, the portfolio standard deviations ae as follows" Epon = VCO 007225) + (010028) + (038)(0.0085) Voo0s47s 07399 (0.001225) + (0.0025) + (0.5)(0.00) 0610 Sp vu “tml the allowing exampes, we wl sip some steps because you ae ov aware tht ony the ast er changes. Yoo are escomaged vo Work oo he ino seps to ensure hat Jou Understand the computation procedure.16 PARE2-DIVELOPMINTS IN RVESTIANT TRY a potas ® (0.001225) + (0.0025) + (05){ 0.0035) Laas panies W0.009725) + =oos Note that, in this example, with perfect negative correlation the portfolio standard deviation is nov 2cF0, This is because the different examples have equal weights, but the asset standard deviations are not equal Exhibit 7.12 shaws the results for the two individual assets and the portfolio of the tw assets assuming the correlation coefficients vary as set forth in Cases a through 2, As before, the expected return does not change because the proportions are always set at 0.50-00.50, so all the portfolios lie along the horizontal line at the return, R = 0.15. Constant Correlation with Changing Weights If we changed the weights of the two assets ‘while holding the correlation coefficient constant, we would derive a set of combinations that trace an ellipse starting at Asset 2, ging through the 0.50-0.50 point, and ending at Asset 1. We can demonstrate this with Case c, in which the correlation coefficient of zero eases the computations. We begin with 100 percent in Asset 2 (Case f) and change the weights follows, ‘ending with 100 percent in Asset 1 (Case D: cate “ “ etn T 300 0 a 020 a0 one h 040 080 016 ' 050 030 os i 060 ao 014 k om 020 on ' 00 000 o10 Lo (000 007 062 003 004 O05 O06 007 ONE Gos O10 OTT aN? Standard Deviation f Return to) “The ewo appendiees to this chapter show proos fo equal weights with equal valance ad sl forthe sppeapiae ‘welghtsto get ate sandard Gettin whe sadarédewtons are nt eq. ¥ ny by si folios occur bighe acon highe low. skit bys 58 pe the be exhiby a the tw the tw: itisto* “again.‘We already know the standard deviations (a) for portfolios fand| (only one asset) and portisio (i. In Cases gh, and ky the standard deviations arc* {020} (007 + (0.80)*(0.10)* + 2(0.20) (0.0) (0.00) = V(0.04) (0.0049) + (0.64) (0.07) + (0) = Vo.n06s96. own yavayy = W040)" (0.079 4 (0,60}°(0.10)" + 2(0,40) (0.60) (0.00) Vo.0043ae 0662 V(0.60)"(0.07)! + (0.40)*(0.i0}* + 216.60) (0.40){0.00) Vo.003366 += 0.0580 seca) = W(080)* (0.07 + (0.20) (0.10) + 20.80) (0:20) (0.00) = V0.003536 0595 ‘The various weights with a constant correlation yield the following risk-return combinations. cose cA ™ eR) Elo + 000 1.00 020 0.1000 9 020 80 ore oosi2 b 040 060 016 0.0662 i 050 oso ans 0610 i 060 40 one 00580 « 080 020 on 00595 1 1100 200 090 20700 {A graph of these combinations appears in Exhibit 7.13. We could derive e complete curve by simply varying the weighting by smaller increments, ‘A notable result is that with low, zero, or negative correlations, itis possible to derive port- folios that have lower risk than either single asset. In our set of examples where f, = 0.00, this ‘occurs in Cases hi sand k. This ability to reduce risk isthe essence of diversification. ‘As shown in Exhibit 7.13, assuming the normal rsk-return relationship where assets with higher risk (larger standard deviation of returns) provide high rates of return, itis possible for 1 conservative investor to experience both lower risk and higher return by diversifying nto a higher risk-higher retum asset, assuming that the correlation between the two asses is fairly low. Exhibit 7.13 shows that, inthe case where we used the correlation of zero (0,00), the low. ‘isk investor at Point 1—who would receive a return of 10 percent and risk of percent—could, by investing in portfolio j, imerease the retuen (0 14 percent and experience a decline i tsk to 538 percent by investing (diversifying) 40 percent of the portfolio in riskier Asset 2. As noted, the benefits of diversification are critically dependent on the correlation between assets. The exhibit shows that there is even some benefit when the correlation is 0.50 rather than zero. Exhibit 7.13 also shows thatthe curvature inthe graph depends on the correlation between the two assets or portfolios. With r, = + 1.00, the combinations li along a straight line between the two assets, When r, = 0.50, the curveis tothe right of ther, = 0.00 curve; when, © 0.50, itis to the left. Finally, when r, = —1.00, the graph would be two straight ines that would touch “again you ae encouraged of in the eps we sipped i dhe computations, (CHAPTER 7 AN UTADIRICIION I /OPKC9 GD MANAGEMENT 195Lig i 000 Gor OO 003 04 G05 Ons O67 ONE Om O10 ON O12 ‘Standard Deviation of Return) atthe vertical line (zero isk) with some combination. As shown in Appendix 7B, itis possible {solve for the specified set of weights that would give a portfolio with zero risk, In this case, 0.412 and w, = 0.588, which implies an E(R) of 0.1588. 7.2.6 AThree-Asset Portfolio {A demonstration of what occurs with a thre-asset portfolio is useful because it shows the dynamics ofthe portfolio process when assets are added. It also shows the rapid growin the computations required, which s wy we wil stop at three! Inthisexample, we will combine thee asset classes we have been discussing: stocks, bonds, and cash equivalents” We will assume the following characteristics: Asset Castes etn et ™, Stocks 6) 012 020 ‘080 Bonds (6) 008 aso 930 Cashequbalent(c) 004 0s ono ‘The correlations are tgp = 0.25; toc = 0.08: tye = O15 Given the weights specified, the E(R,9) E(Ryoct) = (0.60) (0.12) + (0.30}(6.08) + (0.10)(0.04) = (0.072 + 0.024 + 0.004) =0,100= 10.00% ‘When we apply the generalized formula from Equation 7.610 the expected standard devia ton of a three-asset portfolio, It is 7 hay= (wea + 9p} + sto) + gm rsruras + Imywcowrerye + weer eTerae) The asc allocation artis regoaly contained inthe Wal Set ournal general rele to thet thre ant aisFrom the characteristics specified, the standard deviation ofthis thre assetclass portfalio. (a) wort be org, ={(06) {0.20} + (0:3}°(0.10)8 4 (0.1)*(003)!) + {[2(0.63(03)(0.20) 0.10) (0.25) | + {2(0.6)(0.1) (0.20) (0.03)(-0.08)| + |2(0.3)(0.1}(0.10) (008) (0.15)]} = [o.019309 4 (0.0018) 4 (0.000076) + (0.000027) = 00170784 oi 70784)" 1306 = 13,0655 7.2.7 Estimation Issues It is important to keep in mind that the results of this portfolio asset allocation depend on the accuracy of the statistical inputs. In the current instance, this means that for every asset (or asset class) being considered for inclusion in the portfolio, we must estimate its expected returns and standard deviation. We must also estimate the correlation coefficient among the entire set of assets. The number of correlation estimates can be significant—for example, for 1 portfolio of 100 securities, the number is 4,950 (that is, 99 + 98 + 97 +...). The potential source of error that arises from these approximations is referred to as estimation risk We can reduce the number of correlation coefficients that must be estimated by assuming that stock returns can be described by the relationship of each stock to a market index—that i, a single index market model, as follows: 79 R DR, te the slope coefficient that relates the returns for Security {to the returns for the aggregete stock market R,, = the returns forthe aggregate stock market fall the securities are similarly related to the market and a slope coefficient b, is derived for ‘cach one, it can be shown that the correlation coefficient between two securities and jis om 7A0 ry = bbe iw, where: Fs = the variance of returns forthe aggregate stock market ‘This reduces the number of estimates from 4,950 to 100—that is, once we have derived 4 slope estimate b, foreach security, we can compute the correlation estimates, Notably, this assumes that the single index markel model provides a good estimate of security returns. 7.2.8 The Efficient Frontier we examined different two-asset combinations and derived the curves assuming all the pos- sible weights, we would have a geaph like that in Exhibit 7.14. The envelope curve that contains the best ofall these possible combinations is referce to asthe efficient frontier. Specifically, the elfcient frontier represents that set of porifotios that has the maximum rate of return for every given level of risk or the minimum risk for every level of return, An example of sich a {rontier is shown in Exhibit 7.15. Every portfolio that lies on the efficient frontier has either 9 higher rate of return for equal risk or lower risk for an equal rate of return than some portfolio beneath the frontier. Thus, we would say that Portfolio A in Exhibit 7.15 domizates Portfolio (CHAPTER 7 - AN INTROINICTON TO FOPIICH OMANACEWAENT 197198. PART2- LEV LOMINTS NIRASTMENE THK Standard Deviation of Return (o ‘Standard Deviation of Return) C because it has an equal rate of return but substantially less risk. Similarly, Portfolio B domi- nates Portfolio C because it has equal risk but a higher expected rate of return. Because of the benefits of diversification among imperfectly correlated assets, we would expect the efficient frontier to be made up of portfolio of investments rather than individual securities. Two pas- sible exceptions arise atthe end points, which represent the asset withthe highest return and the asset with the lowest risk, As an investor, you will target a point along the efficient frontier based on your utility function, which reflects your attitude toward risk. No portfotio on the efficient frontier ‘can dominate any other portfolio on the efficient frontier. All ofthese portfolios have dif- ferent return and risk measures, with expected rates of return that increase with higher risk calet my bern dete) are is addit thes, high, wher cow returhe nd ity it ner Caen 7.2.9 The Efficient Frontier and Investor Utility ‘The curve in Exhibit 7.15 shows thatthe slope ofthe efficient frontier curve decreases steadily as we move upvard. This implies that adding equal increments of risk as we move up the effi cient frontier gives diminishing increments of expected tetura, To evaluate this situation, we ‘aleulate the slope of the efficient frontier as fellows: AB(Ryon) BECO pn) ‘An individual investor's tility curves specify the trade-offs he or she is willing to make between expected return and risk. In conjunction with the efficient frontier, these utility curves determine which particular portfolio on the efficient frontier best suits an individual investor. ‘Two investors will choose the stme portfolio from the efficient set only if their utility curves are identical, Exhibit 7.16 shows two sets of utility curves along with an efficient frontier of investments, ‘The curves labeled U,, Uand U, are fora strongly risk-averse investor. These utili are quite steep, indicating that the investor will not tolerate much additional risk to obtain additional returns. The investor is equally disposed toward any E(R), E(o) combinations along the specific utility curve U, ‘The curves labeled (Uys U U,) characterize a less risk-averse investor. Such an investor is willing to tolerate abit more risk to get a higher expected return, ‘The optimal portfolio is the efficent portfolio that has the highest wilty for @ given investor. {lies atthe point of angency between the efficient frontier and the U, curve withthe highest possible utility. A conservative investor's highest uilty is at Point X ia Exhibit 7.16, ‘where the U, curve ust touches the efficent frontier. A ess risk-averse investor's highest wtlity ‘occurs at Point Y, which represents a portfolio on the efficient frontier with higher expected returns and higher risk than the portfolio at X. zay200 PART2-DLVALOPAANTS IN STRAHETTHE ORY 46 * SUMMARY eee +The hisie Markowitz porflio model detives the expected rate of return for a portfolio of asets and measure of ‘expected rik, which i the standard deviation of expected rales of retura. Markowits showed that the expected tate of return of a porifalio is the weighted average of the expected seturn far the individual investments inthe Portllio, The standard deviation ofa portfolin is a fun tion not only of the tandard deviatlons forthe sndividual westmens but als ofthe eavariance between the rates of ‘return forall the pairs of assets inthe portfolio, fn a large portfolio, these covariances are the important factors. + Dilferent weighs or amounts of portfolio held in various assets yield a curve of potential combinations. Correlation coefficients among asses ae the critical factor to consider when selecting investments. Investors can maintain their rate of eturn while reducing the risk lve of thei portfolio. by combining assets or portfolios that have low positive or negative correlation ‘+ Assuming numerous astets and a mubitude of combina- tion curves, the efficient frontier i the envelope curve that encompasses all ofthe best combinations I defines the set ‘of portfolios that has the highest expected retuen foreach _dven level of risk or the miisnum tsk foreach yiven level ‘of return, From this set of dominant portfolios, investors sect the ane that lies atthe point af tangency between the celficient Fret and their highest tity curve. Hecase tiskereturm utility Fanetioms dif, the point of tangeney and, therefore, the portflio choice wil difer among investors. AL this point, you understand that an optimum partie is 4 combination of investments, each having desirable id ‘dual isk-retuen characteristics that also fi together based ‘on their correlations. "This deeper understanding of porto lio theory should lead you w reflect back on our earlier dis- ‘cussion of global investing, Because many foreign stock and bond investments provide superior rates of return com- pared with U.S, securities and have low correlations with portfolios of US. stocks and bonds (es shown in Chapt 3), including these foreign securities in your portfolio will help you to reduce the overall risk of your portfolin while possibly increasing your rate of tetue, +e © SUGGESTED READINGS «+ -- Elton, Edwin J. Martin Gruber Stephen]. Brown,and Wiliam ohn L Masinn, Donald L, Tuttle, Jerold E. Pinto and Dennis 1N. Goetzmann, Modern Porfolio Theory and Invesiment W. Meleavy, Managing Investment Poifolio: A Dynamic Analysis 6th ed New York: Wey, 2003, Process, ed ed. Hoboken. NI: John Wiley & Son, 2007, Farrell, James L., Je. Portfolio Management: Theory and “Applicaton, 2nd ed, New York: McGraw Hil, 1997, ses QUESTIONS « «-- 1, Why do mostimesors hold diversified prttos? 2. Whats covariance, and why si important in polio theory? 3. Whydomostasst ofthese type show postve covariances of returns wih each other? Would you expect positive covariances of returns between diferent Spe fasts sec a ens on Treasury tis, General Electric eommon stock, nd commercial el estate? Why or why not? Wihatis the relationship between covariance andthe correlation coefiient? Espa the shape ofthe efficent rome. 6, Draw a propery labeled graph ofthe Markowits ecient frontier. Describe the efficient Foner Jn enact terms, Discuss the concept of dominant portly, and show an example of one on yout frsph 2. Assume you want o run a computer program to derive the efficient frontier for your fesbe set of stocks: What information must you input othe program? Why are imestor uty carves important in portfolio theory? 9. Explain how a given investor chooses an optinal porto, Wil this choice aay be 9 divested orto, or could it be single asset Explain your answer. 10, Assume that you anda busines asencnte develop en efficent frontier for a et of investments, Why tight the two af you set diferent podaloson the frontier? 11, Draw a hypothe graph ofan efficent frontier of US, common socks. On the same grep, daw ‘mcf frontier assuming he incisian of US, bonds wel aly onthe same graph dea anCHAPTER 7- AN NTICOUCTION TOPURTTOMO MANAGEMENT 201 2. Cficlent frontier that includes U.S. common stocks, U.S. bonds, and stocks and bonds from around the world. Dicuss the differences in these frontiers. Stocks; L,and M cach have the same expected ctuen and standaed deviation, The correlation cock ficients between cach pair ofthese stocks are: Kand Lcorelation coefficient = +08 K and M cortelation coelfciont = 402 1and M curelation coelficiemt = 0.4 ‘Given these core eviaton? Expl (CEA EXAMINATION LEVEL A three-asset portato has the following characteristics as, portfolio constructed of which pair of stacks will have the lawest standard Expected Standard Asset Return __Deviation x ons 022 sy y nie 08 40 z 108 oes Bio “The expected return on this three-asst portfolio is © 121% 4. 148% (2 minutes) ‘CFA EXAMINATION LEVEL 11 An investor is considering adding another investment to a portfolio, “To achieve the maximum diversification benefits, the investor should add, if possible, an invest ‘ment that has which of the following correlation coeficients with the other investments in the portfotio? a 10 b, -0s 00 2. +1001 minute) ++ 8 © PROBLEMS © e+ “The standard deviation of Shamrock Corp. stock is 19 percent. The standard devtation of Sophie Co. sock is 14 petcent The covariance between these two stocks i 100. What isthe coreation between Shamrock and Sophie stock? ‘You are considering two assets with the following characteristics, FR) =015 Eo) = 010 w,= 05 BR) = 020 Elo,) = 020 w,=05 : ‘Compute the mean and standard deviation of two portfolios if, = 040 and -0.60, respectively. Plot the two portfolios on arsk-return graph and beefy explain the results. ‘The following are monthly percentage price changes fo four market indexes Month DNA SKP500 Ruts! 2000 _ Mikkel 1 030g 902 ‘0 2 067008 210-002 3 <0 act om aor ‘ nor os 08 002 5 oo: Oot on 902 ‘ 008 868 -ace 008 ‘Compute the following, 3. Average monthly rate of return for each index be. Standard deviation for each index202 PART? «DVI CS/VFOIS NMA SIRART HCY © Covariance between the rates ofreuen fo the following Indes: DIIA-S&P S00 S&P 500-Russell 2000 SW? 500-Nikhei asl) 2000-Nikkct 4. The correation coefficients fe the sane four combinations Using the anscers foi pars fa), (band (a), calculate the eapevted retuee and standard devia tian of portoli comsising of equal parts of 1) the SAP al the Russell 2000 and (2) the S&P and the Nike. Discuss the twu partolins ‘he following are the musty rales af return for Mautison Cookies and for Sophie Hiei sic month peri. Month Medison Cookies __Sophle Elect 1 104 or 2 ‘te nay 3 or 10 * mn fas 5 -0o? tas 6 os ua Compt he lowing 2. Average monthiy rte of tur R for each stock &, Standard deviation of retin for ech sock Govarane tween the rate of ree 4. The corelation coefficient betwen the rtes of ern ‘What lee of coreton did you expect? How did your expectations compare withthe computed corttaton? Woul these two snc offer a good chance for diversification? Why or why not Give BRR) FR) = 016 Calculate the expected returns and expected standard deviations of 2 two-stock portfolio having a correlation coeficient of 0.70 under the fllowing conditions ew, = 005 Plot the results on a return—risk graph. Without calculations, draw in what the curve would Jook like firsif the correlation coefficient had been 0.00 and then ifit had been -0:70, Given the following market values of stocks in your portfolio and their expected rates of return, what 1s the expected rate of return for your common stock portfohoz Stode Macket Value ($ Mil ! Disey s1s00 ove Sorbus ‘7000 a0 . aey asian 32000 ons Inve Bow ote Walgreens 700 on Gives EAR ER) Fie, Ha, Calculate the expected returns and expected standard deviations of 2 two-stock portfolio in which Stock 1a a weight of 60 percent under the following conditions. any 100 ben = 075 © 437 025 d= 0.00CHAPTER AN NTH DUCTION TO PORTICO MANACENANT 203 ina eto 1 Gonsderinghe world econamie out othe coming yen sn estimates sk and ering fr the partic inte you eet ae fers fr Lauren a cnm eh age Ietmeer hi percen at pent a ening pba feobabily Pobre wa a ‘Compute the expected rate of return ECR) for Lauren Labs. THOMSON ONE iu: 1. Collect daily price data forthe past 30 trading days and compute the daily percentage price returns foreach: Avon Products, Inc. (AVP), Best Buy Inc. (BBY), and Cisco Systems (CSCO), 2. Compute the mean daly return for each stock and the standard deviation of daily returns. 3. Compute the covariance and corcelations between the three possible pairs. 4, Assuming equal weigh between each pair of stocks, compote the mean daity return and the standard eviation ofeach ofthe three porflios 5. Create a rck-returm scatterplot and enter the average daily return (vertical axis) and standard devia- tion (horizontal sx) forthe theee individual stocks and the three portfolios. Briefly discuss the ‘results hased on the material inthis chapter. ke A. Proof That Minimum Portfolio Variance Occurs with Equal Weights When Securities Have Equal Variance When @, = o,,we have: Coax lar? + (1 = mF CoP = 2m(1 = wn alo)? (asl +1 — 2m + we + Dwr 2 — WHE) = (oan +1 = 20 + 2wyres = 20rd For this toe a minima, > z ° x pon) a, a = (rae, 2-4 Ars Aner al Assuming (o,)*> 0, Aw, ~ 24 Drs — Awitig 0 4wi(1 = 2) 20 ~ ng)206 PART -DIviLC# MINTS IN Re SHARE THK from which 20 = nad 1 MAC ra) 2 regardless of r,,. Thus, if6,= oy a2, will always be minimized by choosing w; = = 1/2, regardless of the value ofr,» except when r,, = +1 {in which case u,,, = ef, ~ 7) This can bye vriied by checking the second-order condition * 5 z Alera) iwi Problem he follawing information applies to Questions La and 1b. The general equation forthe weight ofthe fist security 4 achieve minimum variance (in a twe-stock portfolio) is given by (on) ~ slower) (oy)? + (oy)? = Aral Mor) Ja, Show that w, = 05 when @,= a, 4b, What isthe weight of Security Uhat gives minimum portfolio variance when = 0.5.0, = 004, and a, = 0067 B, Derivation of Weights That Will Give Zero Variance When Correlation Equals -1.00 og= WO F (1 = won) + 2m(1 — w)ralerN(o) = wile)? + (ox? -2mor) — wilon}* + 2muns(oi)(ox) ~ 2winaloN(o2) Iir,g~ Is this can be rearanged and expressed as Shon = vAl(a,} + 260 (a2) + (0,9) ~2W{(o,)* + (o,)(o,)] + (oP = Wille.) + (ODF ~ 2wi(o2llen) ~ (or) + (os? ={willon) + (o)] ~ (orb ‘We want to find the weight, w,, which will reduce (@%,,) to zero; therefore, wl) + I~ (6) =0 which yields . (a) ey + (oD fo.) Mi Ga eames Problem 1 Given two assets with the following eharacterites: