13.1 Markowitz Portfolio Theory, Mean-Variance Optimization & the Efficient Frontier

13.1 Markowitz Portfolio Theory, Mean-Variance Optimization & the Efficient Frontier

Modern Portfolio Theory (MPT), pioneered by Harry Markowitz in his seminal 1952 paper, revolutionized investment management by shifting the analytical focus from individual security selection to the risk and return characteristics of the entire portfolio. By quantifying diversification through variance, covariance, and correlation, MPT provides the mathematical foundation for institutional asset allocation, the Capital Allocation Line, and quantitative portfolio optimization.

                    Modern Portfolio Theory & Capital Allocation Architecture
                                                │
         ┌──────────────────────────────────────┼──────────────────────────────────────┐
         │                                      │                                      │
Mean-Variance Optimization            Two-Fund Separation & CAL               Institutional Enhancements
  • Multi-Asset Covariance Matrix       • Tangency Portfolio (Max Sharpe)       • Error Maximizer Critique
  • Correlation Effects (ρ = -1, 0, 1)  • Risk-Free Asset Lending / Borrowing   • Black-Litterman Model
  • Global Minimum Variance Portfolio   • Quadratic Utility U = E(r)-0.5·A·σ²   • Resampled Efficient Frontier
  • Markowitz Efficient Frontier        • Optimal Complete Allocation (y*)      • Corner Solution Constraints

---

1. Markowitz Portfolio Theory & Mean-Variance Optimization

Foundational Assumptions

Classical Markowitz Mean-Variance Optimization (MVO) rests on several core economic and behavioral assumptions:

1. Risk Aversion: Investors are rational and risk-averse; given two portfolios with identical expected returns, they strictly prefer the one with lower variance (risk).
2. Single-Period Horizon: Investors optimize their allocations over a uniform, single holding period.
3. Distributional Characterization: Asset return distributions are fully described by their first two statistical moments: expected return ($E(R)$) and variance ($\sigma^2$) or standard deviation ($\sigma$).
4. Frictionless Capital Markets: Assets are infinitely divisible, and there are no transaction costs, taxes, or restrictions on short sales (unless explicitly constrained).
5. Non-Satiation: Investors exhibit positive marginal utility of wealth; more expected return is always preferred to less for a given level of risk.

Portfolio Expected Return and Covariance Matrix

For a portfolio composed of $n$ risky assets with weights $w_i$ (where $\sum_{i=1}^n w_i = 1$):

$E(R_p) = \sum_{i=1}^{n} w_i E(R_i) = w_1 E(R_1) + w_2 E(R_2) + \dots + w_n E(R_n)$

Portfolio variance is governed not merely by individual asset variances, but by the pairwise covariances between all asset pairs. For an $n$-asset portfolio, the variance is expressed as:

$\sigma_p^2 = \sum_{i=1}^{n} w_i^2 \sigma_i^2 + \sum_{i=1}^{n} \sum_{j \ne i}^{n} w_i w_j \text{Cov}(R_i, R_j) = \sum_{i=1}^{n} \sum_{j=1}^{n} w_i w_j \sigma_{ij}$

Where $\sigma_{ij} = \text{Cov}(R_i, R_j) = \rho_{ij} \sigma_i \sigma_j$. In matrix notation, $\sigma_p^2 = \mathbf{w}^T \mathbf{\Sigma} \mathbf{w}$, where $\mathbf{w}$ is the vector of asset weights and $\mathbf{\Sigma}$ is the $n \times n$ variance-covariance matrix.

                        The n x n Variance-Covariance Matrix

                  Asset 1       Asset 2       Asset 3  ...    Asset n
               ┌─────────────┬─────────────┬─────────────┬─────────────┐
       Asset 1 │  w₁² σ₁²    │  w₁w₂ σ₁₂   │  w₁w₃ σ₁₃   │  w₁wn σ₁n   │ ──► n Variance Terms
               ├─────────────┼─────────────┼─────────────┼─────────────┤     (Individual Asset Risk)
       Asset 2 │  w₂w₁ σ₂₁   │  w₂² σ₂²    │  w₂w₃ σ₂₃   │  w₂wn σ₂n   │
               ├─────────────┼─────────────┼─────────────┼─────────────┤ ──► n(n - 1) Covariance Terms
       Asset 3 │  w₃w₁ σ₃₁   │  w₃w₂ σ₃₂   │  w₃² σ₃²    │  w₃wn σ₃n   │     (Pairwise Co-Movements)
               ├─────────────┼─────────────┼─────────────┼─────────────┤
       Asset n │  wnw₁ σn₁   │  wnw₂ σn₂   │  wnw₃ σn₃   │  wn² σn²    │
               └─────────────┴─────────────┴─────────────┴─────────────┘

As the number of assets $n$ grows large, the contribution of individual asset variances to portfolio risk approaches zero, while the contribution of pairwise covariances dominates:

• Number of variance terms: $n$
• Number of unique covariance terms: $\frac{n(n - 1)}{2}$
• In an equally weighted portfolio ($w_i = 1/n$), portfolio variance converges to average covariance: $\lim_{n \to \infty} \sigma_p^2 = \overline{\text{Cov}}$.

The Mathematical Power of Correlation ($\rho$)

For a two-asset portfolio ($w_1 + w_2 = 1$):

$\sigma_p = \sqrt{w_1^2 \sigma_1^2 + w_2^2 \sigma_2^2 + 2 w_1 w_2 \sigma_1 \sigma_2 \rho_{12}}$

The correlation coefficient $\rho_{12} \in [-1.0, +1.0]$ dictates the degree of risk reduction:

1. Perfect Positive Correlation ($\rho_{12} = +1.0$): $\sigma_p = \sqrt{(w_1 \sigma_1 + w_2 \sigma_2)^2} = w_1 \sigma_1 + w_2 \sigma_2$ Portfolio standard deviation is the exact weighted arithmetic average of individual asset volatilities. Zero diversification benefit is achieved.

2. Imperfect Correlation ($\rho_{12} < +1.0$): $\sigma_p < w_1 \sigma_1 + w_2 \sigma_2$ Portfolio standard deviation is strictly less than the weighted average volatility. Diversification occurs without sacrificing expected return.

3. Uncorrelated Assets ($\rho_{12} = 0.0$): $\sigma_p = \sqrt{w_1^2 \sigma_1^2 + w_2^2 \sigma_2^2}$ Significant volatility reduction is realized across uncorrelated asset classes.

4. Perfect Negative Correlation ($\rho_{12} = -1.0$): $\sigma_p = |w_1 \sigma_1 - w_2 \sigma_2|$ A completely riskless portfolio ($\sigma_p = 0$) can be constructed by setting: $w_1 = \frac{\sigma_2}{\sigma_1 + \sigma_2}, \quad w_2 = \frac{\sigma_1}{\sigma_1 + \sigma_2}$

Worked Example: Correlation Impact on Portfolio Risk

An investment consultant constructs a two-asset portfolio allocating 60% to Global Equities ($E(R_1) = 10%$, $\sigma_1 = 18%$) and 40% to Core Fixed Income ($E(R_2) = 4%$, $\sigma_2 = 6%$).

• Portfolio Expected Return: $E(R_p) = (0.60 \times 10%) + (0.40 \times 4%) = 6.0% + 1.6% = 7.60%$.
• Weighted Average Volatility: $(0.60 \times 18%) + (0.40 \times 6%) = 10.80% + 2.40% = 13.20%$.

---

2. Minimum Variance Frontier vs. Efficient Frontier

The Markowitz Opportunity Set

Plotting all possible asset weight combinations in expected return-standard deviation space ($E(R)$ vs. $\sigma$) generates the investment opportunity set (the "Markowitz bullet").

   Expected Return E(R)
     ▲
     │                                              / Capital Allocation Line (CAL)
     │                                            / 
     │                             .-''''''-.   / Tangency Portfolio (P*)
     │                         .-'          `●-----------------
     │                       .'             /  \ Efficient Frontier (Upper Limb)
     │                     .'              /    \
     │                   .'               /      \
     │     GMVP ───────●                 /        \
     │                  `.              /          \ Opportunity Set
     │                    `.           /            \
     │                      `.        /              \ Inefficient / Dominated
     │                        `-.    /               / (Lower Limb)
     │                           `--/--------------'
     │                             /
     │  Risk-Free (Rf) ───────────●
     └────────────────────────────┴──────────────────────────────► Standard Deviation (σ)
                                 σ_GMVP   σ_P*

Minimum Variance Frontier (MVF)

The Minimum Variance Frontier represents the set of all portfolios that minimize portfolio variance for every attainable level of expected return. Mathematically, it is obtained by solving the quadratic optimization problem:

$\min_{\mathbf{w}} \frac{1}{2} \mathbf{w}^T \mathbf{\Sigma} \mathbf{w} \quad \text{subject to} \quad \mathbf{w}^T \mathbf{E(R)} = E_0 \quad \text{and} \quad \mathbf{w}^T \mathbf{1} = 1$

Global Minimum Variance Portfolio (GMVP)

The Global Minimum Variance Portfolio (GMVP) is the unique point on the Minimum Variance Frontier possessing the lowest possible variance across all feasible risky asset combinations. It serves as the mathematical inflection point dividing the frontier into two distinct limbs.

The Efficient Frontier

The Markowitz Efficient Frontier is strictly defined as the upper half of the minimum variance frontier that lies at or above the Global Minimum Variance Portfolio (GMVP).

• Efficiency Principle: Portfolios on the efficient frontier offer either the maximum expected return for a given level of standard deviation, or the minimum standard deviation for a targeted level of expected return.
• Dominated (Inefficient) Portfolios: Any portfolio lying below the efficient frontier or on the lower limb of the minimum variance frontier (below the GMVP) is economically inefficient. An investor can achieve a strictly higher return for the exact same risk, or lower risk for the same return, by moving vertically or horizontally to the efficient frontier.

---