📚 Diversification & Portfolio Risk
Brief Overview
This note covers Diversification and Portfolio Risk and was created from a PDF
containing 89 pages of lecture notes. It offers a concise walk‑through of key portfolio
concepts: expected returns, variance‑covariance calculations, correlation effects, the
efficient frontier, and the Capital Allocation Line.
Key Points
How diversification reduces unsystematic risk while leaving systematic risk
intact.
The math behind portfolio variance and covariance, including the role of
correlation.
How the efficient frontier is constructed and how the optimal risky portfolio is
identified via the Capital Allocation Line.
📈 Diversification and Portfolio Risk
Diversification – the practice of spreading investments across many assets to reduce
exposure to any single source of risk.
The classic proverb “Don’t put all your eggs in one basket” is illustrated both by
a cartoon chick and a Warren Buffett quote:
The chick symbolizes the need to allocate eggs (capital) across multiple baskets
(assets).
Buffett’s warning underscores the same principle for investors.
Empirical evidence: As the number of stocks in a portfolio grows, the average
portfolio standard deviation falls dramatically at first and then levels off.
The curve shows rapid risk reduction up to about 20 stocks; after that the
marginal benefit diminishes, stabilizing near a 20 %‑40 % standard deviation.
Takeaway: Diversification can lower risk, but systematic (market) risk remains even with
many assets.
📊 Portfolios of Two Risky Assets
5.2.1 Expected Return
Expected portfolio return – the weighted average of the component securities’
expected returns
$ E(R_p)=w_A,E(R_A)+w_B,E(R_B),\qquad w_A+w_B=1 $
Positive weights → long positions; negative weights → short‑selling.
Example 1 – Equal allocation (Alibaba 28.30 %, Google 15.10 %)
$ E(R_p)=0.5\times28.30%+0.5\times15.10%=21.70% $
Example 2 – Solving for weights given a target return of 21.70 %
$ 21.70% = w_A;28.30% + (1-w_A);15.10% ;\Longrightarrow; w_A = w_B = 0.5 $
Example 3 – Short‑selling Google
Invest 260 % of capital in Alibaba, ‑160 % (short) in Google.
$ w_A = 2.60,\qquad w_B = -1.60 $ $ E(R_p)=2.60\times28.30% +
(-1.60)\times15.10% = 49.42% $
5.2.2 Portfolio Risk (Variance)
Portfolio variance – the weighted sum of individual variances plus twice the weighted
covariance
$ \sigma_p^{2}=w_A^{2}\sigma_A^{2}+w_B^{2}\sigma_B^{2}+2,w_Aw_B,\sigma_{AB} $
where (\sigma_{AB}= \rho_{AB},\sigma_A\sigma_B).
Covariance definition
(\displaystyle \text{Cov}(R_A,R_B)=E!\big[(R_A-E(R_A))(R_B-E(R_B))\big])
Correlation coefficient
(\displaystyle \rho_{AB}= \frac{\text{Cov}(R_A,R_B)}
{\sigma_A\sigma_B},\qquad -1\le\rho_{AB}\le1)
Numerical illustration (σ_A = 10 %, σ_B = 20 %)
ρ_AB σ_p (portfolio σ)
–1 5%
–0.5 8.66 %
0 11.20 %
0.5 13.20 %
1 not shown (risk rises to the weighted
average of individual σ)
Insight: Lower (more negative) correlation yields lower portfolio risk.
The chart visualizes how portfolio risk changes as ρ varies, confirming the benefit of low
or negative correlation.
📚 Risk Measures & Formulas
Variance – the average squared deviation of returns from their mean, a primary
measure of investment risk.
Covariance – captures how two assets move together; positive indicates
same‑direction moves, negative indicates opposite.
Correlation – standardized covariance, ranging from –1 (perfect negative) to +1
(perfect positive).
General variance formula for an n‑asset portfolio:
$ \sigma_{P}^{2}= \sum_{i=1}^{n}\sum_{j=1}^{n} w_i w_j \sigma_{ij} $
The double‑summation captures every pairwise covariance weighted by portfolio shares.
Unique vs. Market Risk:
Diversification erodes unique risk (blue curve) but cannot eliminate market
risk (horizontal gray line).
🧭 Feasible Set of Portfolios
Feasible set – the collection of all attainable portfolios formed from a given set of
assets, depicted in the (expected return, standard deviation) plane.
Shape depends on correlation (ρ_AB):
ρ = 1 → straight line connecting the two assets.
ρ = –1 → two intersecting lines (one with negative slope) producing a
minimum‑variance point.
–1 < ρ < 1 → a quadratic curve (the efficient frontier).
The curved frontier shows optimal risk‑return trade‑offs for different ρ values; points A
and B are the single‑asset extremes, C is the efficient blend, D (ρ = –1) yields a
risk‑free‑like portfolio.
Example: Feasible Set Calculations (weights 0 %‑150 %)
Weight % A Expected Return % Std Dev %
–50 32.79 –
–25 22.50 26.34
0 20.00 20.00
25 17.50 13.92
50 15.00 8.66
75 12.50 6.61
100 10.00 10.00
125 7.50 15.61
150 5.50 21.79
Interpretation: As the allocation shifts from Asset A to Asset B, both expected return
and risk move along the feasible frontier.
Minimum‑Variance Portfolio (ρ = –1)
The optimal weight for Asset A is given by:
$ W_A^{\star}= \frac{\sigma_B^{2} - \rho_{AB}\sigma_A\sigma_B}{\sigma_A^{2} +
\sigma_B^{2} - 2\rho_{AB}\sigma_A\sigma_B} = \frac{\sigma_B^{2} - \sigma_{AB}}
{\sigma_A^{2} + \sigma_B^{2} - 2\sigma_{AB}} $
Derivation involves minimizing (\sigma_p^{2}) with respect to (w_A); the first
derivative and second derivative are provided in the transcript.
📈 The Power of Diversification (Chinese Proverb)
** proverb** – “Don’t put all the eggs in one basket.”
Risk declines sharply as stocks increase from 1 to ~20, then plateaus around 20 %.
Key conclusions:
Unique (unsystematic) risk diminishes with more assets,
asymptotically approaching zero.
Market (systematic) risk remains unchanged regardless of portfolio
size.
These principles underpin modern Portfolio Theory (Markowitz) and guide
asset‑allocation decisions.
📊 Feasible Set of Portfolios (Two Risky Assets) 🛠️
Feasible set – the collection of all portfolios that can be formed from a given set of
securities, plotted in the (expected return $E(R)$, risk $\sigma$) plane.
When assets are perfectly positively correlated ($\rho_{AB}=1$) the feasible
set collapses to a straight line joining the two single‑asset points.
With perfect negative correlation ($\rho_{AB}=-1$) the set consists of two
intersecting straight lines sharing the same intercept but with opposite
slopes.
For intermediate correlation ($-1<\rho_{AB}<1$) the set forms a quadratic
(curved) frontier.
The red dot (point C) marks a particular portfolio on the curved efficient portion; the
yellow line connects the minimum‑variance point D to the risk‑free asset (when introduced
later).
📈 Feasible Set of Portfolios (Three Risky Assets) 📐
Adding a third risky asset expands the feasible region from a curve to a
two‑dimensional solid area that “bulges” toward lower risk.
Key observations:
The region protrudes to the left, indicating the possibility of achieving lower
risk for a given level of return by mixing the three assets appropriately.
The efficient frontier will be the upper boundary of this shaded feasible
region.
🌐 General Shape of the Feasible Set (Multiple Risky
Assets)
With many risky securities the feasible set becomes a convex hull in the risk‑return
plane; its upper edge is the efficient frontier.
The minimum‑variance portfolio is the leftmost point on the frontier.
Any portfolio to the right of this point is less efficient because it carries higher
risk without a commensurate increase in expected return.
⚖️ Efficient Set of Portfolios (Mean‑Variance Criterion)
📈
Efficient portfolio – a portfolio that, for a given level of risk, provides the highest
expected return, or for a given expected return, yields the lowest risk.
Two Risky Assets
When $-1 < \rho_{AB} < 1$: the efficient set is the upper half of the feasible
curve, extending from the minimum‑variance point to the upper‑right extreme.
When $\rho_{AB}=1$: the entire feasible line is efficient because risk and
return move linearly together.
When $\rho_{AB}=-1$: the efficient set coincides with the upper straight‑line
segment of the two‑line feasible set.
Multiple Risky Assets
The efficient set remains the upper boundary of the convex feasible region, starting at
the minimum‑variance portfolio and stretching upward‑right.
The tangent CAL identifies the optimal risky portfolio (ORP) for a given
risk‑free rate.
🏦 Capital Allocation Line (CAL) with a Risk‑Free Asset
📉
Capital Allocation Line – the line joining the risk‑free rate point to any feasible risky
portfolio; its slope equals the Sharpe ratio of that risky portfolio.
Key formulas:
Expected return of a mixed portfolio:
$E(R_p)=w_f R_f + (1-w_f)E(R_A)$
Portfolio variance (since $ \sigma_f = 0 $):
$\sigma_p = (1-w_f)\sigma_A$
Slope of CAL = $\dfrac{E(R_A)-R_f}{\sigma_A}$ (the Sharpe ratio).
When the CAL is tangent to the efficient frontier, the point of tangency is the optimal
risky portfolio (ORP).
🎯 Optimal Risky Portfolio (ORP) vs. Optimal Portfolio
(OP) 🧭
ORP – the risky portfolio that maximizes the risk‑adjusted return (i.e., has the highest
Sharpe ratio).
OP – the overall portfolio (mix of risk‑free asset and ORP) that maximizes the
investor’s utility.
Distinctions
Aspect Optimal Risky Portfolio Optimal Portfolio (OP)
(ORP)
Criterion Maximize unit risk Maximize utility: $U = E(R)
premium: $\displaystyle - \frac{1}{2}\lambda
\max \frac{E(R)-R_f} \sigma^2$ (with
{\sigma}$
risk‑aversion parameter
$\lambda$)
Tangency CAL from the risk‑free rate Investor‑specific
is tangent to the efficient indifference curve is
frontier of risky assets. tangent to the overall
efficient frontier
(including the risk‑free
asset).
Investor dependence Independent of personal Depends on the investor’s
risk preferences; same for risk aversion; different
all investors. investors may have
different OPs.
Location Point M where the CAL Point where an
touches the risky‑asset indifference curve meets
frontier. the CAL (often also point M
for a particular risk
aversion level).
The orange box emphasizes that the OP is the portfolio offering the highest
utility given the investor’s risk appetite.
🧩 Separation Theorem (Two‑Fund Separation) 📐
Separation theorem – regardless of an investor’s risk preferences, the optimal overall
portfolio can be constructed by combining the risk‑free asset with a single optimal
risky portfolio (the ORP). The straight line (CAL) separates the efficient frontier from
the set of indifference curves.
The tangency point M on the CAL is the same for all investors; only the mixing
weight between the risk‑free asset and the ORP varies with risk tolerance.
This theorem implies that portfolio selection can be split into two independent
tasks:
1. Identify the ORP (purely a market problem, independent of investor
preference).
2. Choose the allocation between the ORP and the risk‑free asset
(an individual‑preference problem).
The CML is the CAL when the market portfolio is the ORP; it is the steepest
CAL achievable.
Any investor, regardless of risk aversion, will locate their optimal mix on the
straight line FM (risk‑free point F to market portfolio M).
📊 Example: Constructing the ORP with Real Assets 📈
Given:
• Index bond D Expected return $13%$, σ $12%$
• Stock E Expected return $20%$, σ $20%$
• Correlation $\rho_{DE}=0.3$
• Risk‑free rate $R_f=5%$
1. Compute the covariance: $\sigma_{DE}= \rho_{DE}\sigma_D\sigma_E = 0.3
\times 0.12 \times 0.20 = 0.0072$ (i.e., $0.72%$).
2. Form the variance‑covariance matrix and solve for the weights that maximize
the Sharpe ratio:
$ w_D = \frac{E(R_D)-R_f}{\sigma_D^2 - \rho_{DE}\sigma_D\sigma_E}
\qquad w_E = 1 - w_D $
Substituting the numbers yields approximately $w_D = 0.82$, $w_E = 0.18$
(rounded for illustration).
3. The resulting portfolio M lies on the CAL and is tangent to the efficient frontier
– the ORP.
The red line is the CAL through M; the black curve is the efficient frontier; the
blue curve shows the feasible set.