Diversification and Portfolio Risk Explained

Why diversification matters: portfolio risk basics

Diversification is the practice of holding multiple assets so that the overall portfolio risk is lower than the risk of any single holding. The key reason diversification works is that assets do not move perfectly together — their returns have different magnitudes and directions over time. That relationship is captured by correlation between assets.

Two central concepts to keep in mind:

• Risk (volatility) of an asset: usually measured as the standard deviation of returns.
• Correlation between assets: a number between -1 and +1 that measures how returns move in relation to each other.

When assets are less than perfectly correlated, combining them can reduce portfolio variance (and therefore volatility). This is the mechanical reason diversification can improve a risk-return profile.

Core formulas (simple and general)

For two assets A and B with weights w_A and w_B (w_A + w_B = 1), standard deviations σ_A and σ_B, and correlation ρ_{A,B}, the portfolio variance is:

Variance: Var_p = w_A^2 σ_A^2 + w_B^2 σ_B^2 + 2 w_A w_B σ_A σ_B * ρ_{A,B}

Portfolio volatility (standard deviation): σ_p = sqrt(Var_p)

In matrix form for n assets, if w is the weight vector and Σ is the covariance matrix (Σ_{i,j} = Cov(R_i,R_j)), then:

Variance: Var_p = w^T Σ w

This compact form generalizes to any number of assets and is what portfolio-optimization routines use.

Covariance can be computed from standard deviations and correlation:

Cov(R_i,R_j) = σ_i σ_j ρ_{i,j}

How correlation drives diversification benefit

• If ρ = 1 (perfect positive correlation): no diversification benefit — portfolio variance is the weighted average of variances.
• If ρ = 0: cross-term disappears, and variance declines relative to a single asset depending on weights.
• If ρ < 0: the cross-term reduces variance further; negative correlation is the most powerful diversifier.

In practice, correlations change over time and can increase in market stress, reducing diversification benefit exactly when it's most needed. That’s why diversification across sources (asset classes, geographies, styles, macro drivers) matters — not just adding many stocks from the same sector.

Worked example: two-asset portfolio step-by-step

Hypothetical inputs (illustrative numbers):

• Asset A: expected annual volatility σ_A = 20% (0.20)
• Asset B: expected annual volatility σ_B = 12% (0.12)
• Correlation between A and B: ρ_{A,B} = 0.3
• Weight in Asset A: w_A = 60% (0.6)
• Weight in Asset B: w_B = 40% (0.4)

Step 1 — compute variances:

• σ_A^2 = 0.20^2 = 0.04
• σ_B^2 = 0.12^2 = 0.0144

Step 2 — compute covariance term:

• σ_A σ_B ρ = 0.20 0.12 0.3 = 0.0072

Step 3 — plug into the two-asset variance formula:

Var_p = (0.6)^2 0.04 + (0.4)^2 0.0144 + 2 0.6 0.4 * 0.0072

Compute each term:

• (0.6)^2 0.04 = 0.36 0.04 = 0.0144
• (0.4)^2 0.0144 = 0.16 0.0144 = 0.002304
• 2 0.6 0.4 0.0072 = 0.48 0.0072 = 0.003456

Sum: Var_p = 0.0144 + 0.002304 + 0.003456 = 0.02016

Portfolio volatility: σ_p = sqrt(0.02016) ≈ 0.1420 or 14.20%

Interpretation: the weighted average of volatilities is 0.620% + 0.412% = 16.8%. Because assets are not perfectly correlated (ρ = 0.3), the combined portfolio volatility is 14.2% — lower than the weighted average and lower than Asset A alone. That is diversification at work.

Try the same exercise with different correlations or weights to see how the σ_p moves. If correlation rose to 0.8, the variance would increase; if correlation dropped to -0.2, σ_p would fall more.

Practical steps to apply this as a retail trader

1. Measure or estimate: collect historical returns for candidate holdings and compute σ and pairwise correlations (or use covariance matrix).
2. Decide objective: target volatility, target sector exposure, or target return. Be explicit about constraints (max weight per position, liquidity limits).
3. Build candidate portfolios: start with simple two- or three-asset mixes to understand interactions, then expand.
4. Compute Var_p = w^T Σ w and σ_p to confirm the portfolio meets risk goals.
5. Stress-test: increase correlation values (e.g., add 0.2 or move all correlations toward 1) and recompute to see concentration risk under market stress.
6. Rebalance rule: set rebalancing thresholds or calendar rules; rebalancing itself can be a source of risk and return.

Limitations and real-world caveats

• Historical correlations are only estimates; correlations can spike during crises (correlation breakdown). Always stress-test.
• Diversification reduces idiosyncratic (stock-specific) risk but cannot eliminate systematic risk (market-wide moves).
• Transaction costs, taxes, and liquidity constraints matter when implementing many small positions.

No guarantees are made by the calculations above — they are mechanical examples to build intuition.

Where to practice safely

If you want a risk-free way to test weightings, correlations, and rebalancing rules, try AIYUG's free paper-trading race: https://aiyug.us/race

No financial advice is provided here — treat these mechanics as educational content to help you form and test your own hypotheses.