Chapter 4: Introduction to Modern Portfolio Theory

1. Summary

Chapter 4 establishes the mathematical and statistical foundations of Modern Portfolio Theory
(MPT), originally introduced by Harry Markowitz in 1952. MPT provides a quantifiable framework for
constructing and selecting portfolios based on expected performance and investor risk appetite.
Prior to MPT, the benefits of diversification were known intuitively, but Markowitz quantified this
concept using statistical notions of covariance and correlation. The chapter details the assumptions
underlying MPT, such as investors maximizing a one-period expected utility and making decisions
based solely on expected return and risk. It distinguishes between risk-averse, risk-neutral, and risk-
seeking behaviors, emphasizing how individual risk aversion indexes help compute a "Utility score"
for competing portfolios.

+2

The core of the chapter delves into the mathematical computation of Ex-Ante (expected) returns and
Ex-Ante risk (variance and standard deviation) for both individual securities and multi-asset
portfolios. Crucially, it demonstrates that a portfolio's risk is not merely the weighted average of its
constituents' risks, but is heavily influenced by the co-movement (correlation) between the assets. A
correlation coefficient of less than +1.0 indicates diversification benefits, and a correlation of -1.0
theoretically allows for the creation of a zero-risk portfolio. The chapter concludes by mapping these
potential portfolios on a graph to identify the "Efficient Frontier"—an umbrella-shaped curve
representing optimal portfolios that offer the maximum return for a given level of risk —and
highlights the mathematical estimation issues inherent in the optimization process.

+4

2. Key Points by Section

4.1 Framework for constructing portfolios

 Harry Markowitz published "Portfolio Selection" in the Journal of Finance in 1952 and later
won the Nobel Prize (1990) for this framework.

 MPT quantifies the concept of diversification using covariance and correlation, establishing
that the variance of the rate of return is a meaningful measure of portfolio risk.

4.2 Assumptions of the theory

 Investors aim to maximize return for a given level of risk.

 Investments are viewed as a probability distribution of expected returns.

 Investors maximize one-period expected utility and base decisions strictly on expected return
and risk.

4.3 Definition of risk averse, risk seeking and risk neutral investor

 Risk-averse investors demand a risk premium for bearing risk.

 Utility Score formula: $U = E(r) - \frac{1}{2} A \sigma^2$ (where A is the index of risk
aversion).
  Unlike the Sharpe Ratio, which assumes homogenous risk perceptions, Utility explains why
investors make heterogeneous choices based on individual risk aversion.

4.4 & 4.5 Expected rate of return and Variance for individual security

 Ex-Ante return is calculated by summing the expected returns multiplied by their

corresponding probabilities.

 Risk is measured by variance ($\sigma^2$) and standard deviation ($\sigma$), representing

the dispersion around the expected return.

4.6 & 4.7 Expected rate of return and Variance for a portfolio

 Portfolio expected return is the simple weighted average of individual expected returns.

 Portfolio variance depends on the weights, individual risks, and the correlation/covariance
between the assets.

+2

 Correlation ranges from -1 (perfect negative) to +1 (perfect positive).

4.8 & 4.9 Graphical presentation and Efficient Frontier

 If correlation is +1, combinations plot as a straight line (no diversification benefit).

 If correlation is < +1, the curve bulges to the left, indicating risk reduction.

 The Efficient Frontier is an umbrella-shaped curve representing portfolios that yield the
highest return for a given risk or lowest risk for a given return.

+1

4.10 & 4.11 Portfolio Optimization process and Estimation issues

 Optimization requires estimates of expected returns, standard deviations, and correlation

coefficients for all pairs in the universe.

 The number of correlation estimates grows exponentially (e.g., $n(n-1)/2$, meaning 1225
estimates for 50 securities), leading to "estimation risk".

3. Numerical Problems

Problem 1: Calculating Utility Score

Scenario: An investor has a risk aversion index ($A$) of 4. They are evaluating a portfolio with an
expected return of 18% and a standard deviation (risk) of 12%. Calculate the investor's Utility Score.
If the risk-free rate is 6%, should the investor choose this portfolio?

Formula: $U = E(r) - \frac{1}{2} A \sigma^2$ Step-by-step Solution:

1. Identify variables: $E(r) = 0.18$, $A = 4$, $\sigma = 0.12$.

2. Calculate variance ($\sigma^2$): $0.12 \times 0.12 = 0.0144$.

3. Plug into formula: $U = 0.18 - (0.5 \times 4 \times 0.0144)$.

 4. Calculate penalty: $2 \times 0.0144 = 0.0288$.

5. Utility Score = $0.18 - 0.0288 = 0.1512$ or 15.12%.

Conclusion: Since the Utility Score (15.12%) is significantly higher than the risk-free rate (6%), the
investor should choose this investment.

Problem 2: Ex-Ante Return and Variance of a Security

Scenario: Stock X has three possible states for the upcoming year:

 Boom (Probability 0.25): Return 20%

 Normal (Probability 0.50): Return 12%

 Recession (Probability 0.25): Return 4%

Calculate the Ex-Ante Expected Return and Standard Deviation.

Step-by-step Solution:

1. Expected Return $[E(R)]$: $(0.25 \times 0.20) + (0.50 \times 0.12) + (0.25 \times 0.04)$.

2. $E(R) = 0.05 + 0.06 + 0.01 = 0.12$ or 12%.

3. Variance ($\sigma^2$): $\sum [R_i - E(R)]^2 \times P_i$.

o Boom: $(0.20 - 0.12)^2 \times 0.25 = 0.0064 \times 0.25 = 0.0016$

o Normal: $(0.12 - 0.12)^2 \times 0.50 = 0$

o Recession: $(0.04 - 0.12)^2 \times 0.25 = 0.0064 \times 0.25 = 0.0016$

4. Variance = $0.0016 + 0 + 0.0016 = 0.0032$.

5. Standard Deviation: $\sqrt{0.0032} = 0.05656$ or 5.66%.

Problem 3: Two-Asset Portfolio Risk

Scenario: A fund manager allocates 60% of funds to Security A and 40% to Security B. Security A has
a standard deviation of 8%. Security B has a standard deviation of 14%. The correlation coefficient
between their returns is 0.30. Calculate the portfolio's standard deviation.

Formula: $\sigma_{port} = \sqrt{w_A^2 \sigma_A^2 + w_B^2 \sigma_B^2 + 2 w_A w_B r_{A,B} \

sigma_A \sigma_B}$ Step-by-step Solution:

+1

1. $w_A^2 \sigma_A^2 = 0.60^2 \times 0.08^2 = 0.36 \times 0.0064 = 0.002304$.

2. $w_B^2 \sigma_B^2 = 0.40^2 \times 0.14^2 = 0.16 \times 0.0196 = 0.003136$.

3. $2 w_A w_B r_{A,B} \sigma_A \sigma_B = 2 \times 0.60 \times 0.40 \times 0.30 \times 0.08 \
times 0.14 = 0.0016128$.

4. Portfolio Variance ($\sigma_{port}^2$) = $0.002304 + 0.003136 + 0.0016128 = 0.0070528$.

5. Portfolio Standard Deviation ($\sigma_{port}$) = $\sqrt{0.0070528} = 0.08398$ or 8.40%.

Problem 4: Estimation Risk Calculations

Scenario: An AIF manager wants to construct a globally diversified portfolio containing exactly 65
distinct risky securities. Based on Markowitz's Modern Portfolio Theory, exactly how many separate
correlation coefficients must the manager estimate to build the variance-covariance matrix?

Formula: $n(n-1) / 2$ Step-by-step Solution:

1. Identify $n$: $n = 65$.

2. Calculate $n - 1 = 64$.

3. Multiply $n \times (n-1)$: $65 \times 64 = 4160$.

4. Divide by 2: $4160 / 2 = 2080$.

5. The manager must estimate 2080 correlation coefficients.

4. Hard-Level Multiple Choice Questions (MCQs)

1. Harry Markowitz's seminal 1952 article that laid the foundations for Modern Portfolio Theory
(MPT) was published in which academic journal?

A) The Journal of Portfolio Management

B) The Journal of Finance

C) The Financial Analysts Journal

D) The American Economic Review

2. Which of the following is an explicit assumption of Modern Portfolio Theory?

A) Investors maximize multi-period expected utility.

B) Investors evaluate portfolios exclusively on the basis of dividend yields.

C) Investors maximize a one-period expected utility.

D) All investors can borrow at the risk-free rate.

3. In the MPT framework, if an investor is given a choice between two assets offering an identical
expected rate of return, the investor will systematically choose:

A) The asset with the highest variance.

B) The asset with the lowest risk.

C) The asset with the highest positive correlation to the market.

D) The asset that pays regular dividends.

4. The Utility Score formula $U = E(r) - \frac{1}{2} A \sigma^2$ penalizes a portfolio's expected
return based on:

A) The systemic market risk premium.

B) The individual investor's unique index of risk aversion and the portfolio's variance.

C) The risk-free rate prevailing in the economy.

D) The lack of liquidity in the underlying securities.

5. How does the concept of 'Utility' in MPT fundamentally differ from the Sharpe Ratio?

A) Utility uses standard deviation, while the Sharpe ratio uses variance.

B) Utility assumes all investors have homogenous risk perceptions, unlike the Sharpe ratio.

C) Utility incorporates the unique individual risk aversion of the investor, whereas the Sharpe ratio
assumes homogenous risk perceptions.

D) Utility does not penalize for risk, while the Sharpe ratio does.

6. If a portfolio's calculated Utility ($U$) is strictly less than the prevailing risk-free rate ($Rf$), a
rational risk-averse investor should:

A) Invest entirely in the risky portfolio.

B) Short the risk-free asset to buy the portfolio.

C) Reject the risky investment in favor of the risk-free asset.

D) Increase their risk aversion index.

7. "Ex-Ante Return" of an individual security is best defined mathematically as:

A) The historical average of past returns.

B) The sum of expected returns across different states multiplied by their corresponding
probabilities.

C) The median return over a defined holding period.

D) The return adjusted for inflation and taxes.

8. Under MPT, what does the variance of an expected rate of return measure?

A) The statistical dispersion of possible returns around the expected value.

B) The exact downside potential of the investment.

C) The covariance of the stock relative to the index.

D) The liquidity premium demanded by the market.

9. The standard deviation of a security's return is calculated as:

A) The square of its variance.

B) The geometric mean of its historical returns.

C) The square root of its variance.

D) The probability-weighted sum of its deviations.

10. For a portfolio containing multiple assets, the expected Ex-Ante return of the portfolio is:

A) The sum of the variances multiplied by the weights.

B) Always lower than the lowest expected return of its constituent assets.
C) The simple unweighted average of the expected rates of return.

D) The weighted average of the expected rates of return for the individual investments.

11. Which statistical concept did Markowitz use to mathematically quantify the exact benefits of
diversification?

A) Standard Error

B) Covariance or Correlation

C) R-squared

D) Tracking Error

12. The correlation coefficient ($r_{ij}$) between two assets is calculated by:

A) Multiplying the covariance by the product of individual variances.

B) Dividing the covariance by the product of the individual standard deviations.

C) Subtracting the standard deviations from the covariance.

D) Adding the respective utility scores.

13. A correlation coefficient of exactly +1.00 between two securities indicates that:

A) Their returns move in completely opposite directions.

B) Their returns move together in a completely linear, positive manner.

C) There is zero relationship between their returns.

D) Their risk is entirely diversifiable.

14. When computing the variance of a two-asset portfolio, which of the following variables does
NOT directly affect the result?

A) The weights of the individual investments.

B) The standard deviation of the individual investments.

C) The expected rate of return of the overall market.

D) The co-movement (correlation) between the investments.

15. If two securities in a portfolio have a correlation of +1.0, what does their risk-return plot look
like when the weights are varied?

A) A curve bulging sharply to the left.

B) A curve bulging sharply to the right.

C) A straight line connecting the returns of the two securities.

D) A perfect circle.

16. What happens to the shape of the portfolio risk-return curve as the correlation coefficient
between two assets drops below +1.0?
A) The curve becomes perfectly horizontal.

B) The curve bulges to the right, indicating increased risk.

C) The curve bulges to the left, indicating that portfolio risk is less than the weighted average risk of
the assets.

D) The curve breaks into disjointed segments.

17. According to Markowitz's theory, the "essence of diversification" is achieved as long as:

A) The correlation between assets is exactly +1.0.

B) The correlation between assets is less than perfect 1.0.

C) Both assets have identical standard deviations.

D) The expected returns of both assets are equal.

18. What specific condition allows for the theoretical creation of a "zero risk" portfolio from two
risky assets?

A) A correlation coefficient of 0.00.

B) A correlation coefficient of +0.50.

C) A correlation coefficient of -1.00.

D) Identical expected returns and zero variance.

19. When plotting all theoretically infinite combinations of risky securities on a risk-return graph,
the resulting upper boundary is called the:

A) Security Market Line (SML)

B) Capital Market Line (CML)

C) Efficient Frontier

D) Utility Curve

20. Which of the following accurately describes a portfolio that lies ON the Efficient Frontier?

A) It offers the maximum rate of return for a given level of risk.

B) It has a correlation coefficient of +1.0 with all other portfolios.

C) It represents a sub-optimal allocation compared to portfolios inside the curve.

D) It relies entirely on risk-free assets.

21. If Portfolio A and Portfolio C have the same expected return, but Portfolio A lies on the Efficient
Frontier while Portfolio C lies below it, which statement is true?

A) Portfolio A has higher risk than Portfolio C.

B) Portfolio C has higher risk than Portfolio A.

C) Both portfolios have identical risk.

D) Portfolio C has a higher Utility Score than Portfolio A.

22. During the "Portfolio Optimization Process", which of the following is strictly required as an
input?

A) The expected return of every asset class in the investment universe.

B) The precise date of the next central bank rate cut.

C) The historical dividend yield of the overall market over 50 years.

D) The exact liquidation value of the investment manager's firm.

23. For a portfolio containing $n$ securities, the formula to calculate the number of required
correlation estimates is:

A) $n^2$

B) $(n^2 - n) / 2$

C) $n / (n-1)$

D) $2n - 1$

24. The term "estimation risk" in the context of Modern Portfolio Theory refers to:

A) The risk that an investor will default on a capital call.

B) The potential source of error arising from inaccurate statistical inputs (returns, standard
deviations, correlations) required for portfolio allocation.

C) The risk of calculating utility incorrectly due to fluctuating risk-free rates.

D) The risk of using an inaccurate benchmark index.

25. An investor has a risk aversion index of 2. Portfolio X has an expected return of 10% and a
variance of 5% (0.05). What is the Utility score?

A) 9.50%

B) 5.00%

C) 15.00%

D) 0.00%

26. If the covariance between Asset A and Asset B is zero, it definitively means that:

A) The two assets are perfectly negatively correlated.

B) The correlation coefficient is zero, implying no linear co-movement.

C) The portfolio variance will be zero.

D) Neither asset possesses any standard deviation.

27. What is the fundamental shape of the Efficient Frontier in a standard expected return (y-axis)
vs. standard deviation (x-axis) graph?

A) A downward sloping straight line.

B) A U-shaped curve facing downwards.

C) An umbrella-shaped curve arching upwards and to the right.

D) A perfectly vertical line.

28. If an investor considers adding a new security to an existing portfolio to achieve the highest
level of risk diversification, they should seek a security with a correlation coefficient closest to:

A) +1.00

B) +0.50

C) 0.00

D) -1.00

29. Which of the following best defines "Covariance" as used in MPT?

A) A measure of the degree to which two variables move together relative to their individual mean
values.

B) The squared deviation of a single asset's return from its mean.

C) The difference between an asset's return and the risk-free rate.

D) The standardized value of correlation.

30. Why is calculating the variance-covariance matrix critical for portfolios with more than two
securities?

A) Because risk-free assets must be priced accurately.

B) Because MPT dictates that portfolio risk must account for the correlated risks between every
single pair of constituent assets.

C) Because it removes the need to estimate individual standard deviations.

D) Because it guarantees a correlation of -1 across the board.

31. In the utility function $U = E(r) - \frac{1}{2} A \sigma^2$, what does the $\frac{1}{2} A \
sigma^2$ portion mathematically represent?

A) The risk-free rate.

B) The alpha generated by the manager.

C) The penalty deducted from expected return based on the risk involved and the investor's unique
risk aversion.

D) The capital asset pricing premium.

32. According to the MPT assumptions, investors define the "risk" of a portfolio specifically on the
basis of:

A) The absolute maximum drawdown over a 10-year period.

B) The variability of expected returns of the constituent assets.

C) The likelihood of a total default by the issuer.

D) The liquidity constraints in the secondary market.

33. If two distinct portfolios offer the identical level of risk (standard deviation) but different
expected returns, a rational investor governed by MPT will:

A) Choose the portfolio with the higher expected return.

B) Choose the portfolio with the lower expected return to be conservative.

C) Indifferently allocate 50% to each portfolio.

D) Borrow at the risk-free rate to short the higher-return portfolio.

34. To calculate Ex-Ante risk (variance), the squared deviation of each state's return from the
expected return is multiplied by:

A) The risk-free rate.

B) The correlation coefficient.

C) The probability of that state occurring.

D) The beta of the market.

35. If the weights of all assets in a portfolio sum to exactly 1.0 (100%), and the assets have a
correlation coefficient of +1.0, the portfolio standard deviation will equal:

A) Zero.

B) The simple arithmetic average of the individual standard deviations.

C) The weighted average of the individual standard deviations.

D) The geometric mean of the individual variances.

36. A highly risk-averse investor will mathematically possess:

A) A lower risk aversion index (A).

B) A higher risk aversion index (A).

C) A guaranteed negative utility score.

D) A preference for investments with high variance.

37. Which concept explains why a portfolio plotting directly on the Efficient Frontier dominates a
portfolio plotting directly below it?

A) The sub-optimal portfolio provides higher risk for the exact same return.

B) The sub-optimal portfolio provides higher return for the exact same risk.

C) The optimal portfolio provides lower risk for the exact same return.

D) Both A and C are manifestations of dominance on the Efficient Frontier.

38. If the standard deviation of Security A is 10%, the standard deviation of Security B is 20%, and
their covariance is 0.01, what is their correlation coefficient?
A) 0.50

B) 1.00

C) 0.05

D) -0.50

(Hint: $r = Cov / (\sigma_A \times \sigma_B) = 0.01 / (0.10 \times 0.20)$)

39. If an asset's actual return turns out exactly as expected 100% of the time, its variance under
MPT would be:

A) +1.0

B) -1.0

C) Equal to the risk-free rate.

D) Zero.

40. Under MPT, an optimal portfolio is defined as a combination of investments having:

A) Desirable individual risk-return characteristics for a given set of constraints.

B) Only high-yield debt securities.

C) Zero correlation with the risk-free asset.

D) Maximum possible expected return regardless of risk.

Answer Key for Chapter 4 MCQs

1. B (The Journal of Finance)

2. C (Investors maximize a one-period expected utility)

3. B (The asset with the lowest risk)

4. B (Individual's risk aversion and portfolio variance)

5. C (Utility incorporates unique individual risk aversion)

6. C (Reject the risky investment in favor of the risk-free asset)

7. B (Sum of expected returns multiplied by probabilities)

8. A (Statistical dispersion around expected value)

9. C (Square root of its variance)

10. D (Weighted average of expected rates of return)

11. B (Covariance or Correlation)

12. B (Dividing covariance by product of standard deviations)

13. B (Move together in completely linear, positive manner)

 14. C (Expected rate of return of overall market)

15. C (A straight line connecting the returns)

16. C (Curve bulges to the left)

17. B (Correlation is less than perfect 1.0)

18. C (Correlation coefficient of -1.00)

19. C (Efficient Frontier)

20. A (Maximum return for a given level of risk)

21. B (Portfolio C has higher risk than Portfolio A)

22. A (Expected return of every asset class)

23. B ($(n^2 - n) / 2$)

24. B (Error arising from inaccurate statistical inputs)

25. B ($0.10 - (0.5 \times 2 \times 0.05) = 0.05$ or 5.00%)

26. B (Correlation coefficient is zero)

27. C (Umbrella-shaped curve arching upwards/right)

28. D (-1.00 provides maximum diversification)

29. A (Degree two variables move together relative to means)

30. B (Must account for correlated risks between every pair)

31. C (Penalty deducted from expected return)

32. B (Variability of expected returns)

33. A (Choose portfolio with higher expected return)

34. C (Probability of that state occurring)

35. C (Weighted average of individual standard deviations)

36. B (Higher risk aversion index)

37. D (Both A and C are manifestations of dominance)

38. A ($0.01 / 0.02 = 0.50$)

39. D (Zero variance)

40. A (Desirable characteristics for given constraints)

Would you like me to proceed with Chapter 5: Introduction to Capital Market Theory?