Diversification and the Mean-Variance Frontier

The mean-variance framework, introduced by Harry Markowitz (1952), characterises each portfolio by two numbers: its expected return and its standard deviation. The mean-variance frontier traces out, for every level of expected return, the portfolio with the lowest achievable standard deviation. Investors prefer portfolios that lie on the upper portion of this frontier — the efficient frontier.

Holding multiple assets rather than a single asset can push the frontier to the left, meaning the same expected return is achievable with less risk. To see why, consider a two-asset portfolio with weights \(w_A\) and \(w_B = 1-w_A\). Its variance is

$$\sigma_p^2 \;=\; w_A^2\,\sigma_A^2 \;+\; w_B^2\,\sigma_B^2 \;+\; 2\,w_A\,w_B\,\rho_{AB}\,\sigma_A\,\sigma_B.$$

When \(\rho_{AB} < 1\), the cross-term \(2\,w_A\,w_B\,\rho_{AB}\,\sigma_A\,\sigma_B\) is smaller than under perfect co-movement, so portfolio variance falls below the weighted average of the individual variances — the essence of diversification. The lower the correlation, the greater the reduction; when \(\rho_{AB} < 0\), the cross-term turns negative and actively offsets risk. But how much you gain from adding a third asset depends critically on how correlated it is with what you already hold. The interactive chart below makes this concrete.

The visualization

The portfolio universe contains three assets:

  • Asset A — Domestic Stocks: expected return 9%, standard deviation 18%
  • Asset B — Government Bonds: expected return 3%, standard deviation 6%; correlation with stocks: −0.20
  • Asset C — A New Asset: expected return 7%, standard deviation 15%; its correlation with stocks (\(\rho_{AC}\)) is controlled by the slider

The light-grey dashed curve is the complete mean-variance frontier for assets A and B — including its lower, inefficient half — and you can see how it joins A and B along the full parabola. The solid grey curve highlights only the efficient portion of that frontier (from the minimum-variance portfolio upward, where higher return is achievable for the same risk). The solid orange curve is the efficient frontier achievable with all three assets, and the lightly-shaded region is the full feasible set of long-only three-asset portfolios. Drag the slider and observe how much — or how little — the orange efficient frontier expands relative to the grey baseline.

Correlation of Asset C with Stocks (\(\rho_{AC}\)) 0.05
Asset A (Stocks)
Asset B (Bonds)
Asset C (New Asset)
Full M-V Frontier: A+B
Efficient Frontier: A+B
Efficient Frontier: A+B+C
Feasible set: A+B+C

Key takeaways

  1. Diversification reduces risk. Even with two assets, combining them (when \(\rho_{AB} < 1\)) shifts the frontier to the left of both individual assets.
  2. Correlation determines the gain. The lower the correlation of the new asset with the existing portfolio, the farther the frontier moves leftward.
  3. Near-perfect correlation ≈ near-zero benefit. When \(\rho_{AC} \to 1\), asset C is essentially redundant — it duplicates what stocks already provide and does not expand the opportunity set.
  4. Negative correlation is especially valuable. When \(\rho_{AC} < 0\), the new asset hedges the existing portfolio, producing the largest frontier expansion and the lowest attainable risk.

Reference: Markowitz, H. (1952). Portfolio selection. The Journal of Finance, 7(1), 77–91.

← Return to Investment Management Materials