The Efficient Frontier & Minimum-Variance Portfolio

You have the pieces now: portfolios combine return as a weighted average and risk in a way that rewards low correlation. This lesson assembles them into the central picture of portfolio theory, the efficient frontier, and shows a surprising result along the way: sometimes adding a riskier asset makes your whole portfolio safer. This extends the modern portfolio theory you met in Unit 5 with the actual math of finding the best mix.

The feasible set: every portfolio you could build

Take a handful of assets. By varying the weights, you can build a huge number of different portfolios, and if you plot each one with risk on the horizontal axis and expected return on the vertical, they fill in a region shaped a bit like a bullet or an umbrella lying on its side. This region is the feasible set: every combination you could actually hold. Most points in it are poor choices, because for the same risk you could have earned more return by choosing different weights.

The efficient frontier: the upper-left edge

The portfolios worth considering all lie along the upper-left boundary of the feasible set. This edge is the efficient frontier. A portfolio on it delivers the highest expected return for its level of risk, or equivalently the lowest risk for its level of return. Any portfolio below the frontier is inefficient, because a portfolio directly above it offers more return for the same risk, and a portfolio to its left offers the same return for less risk. No rational investor knowingly holds a portfolio below the frontier, since a strictly better one always exists.

The minimum-variance portfolio

The leftmost point of the frontier is the minimum-variance portfolio, the combination of assets with the lowest possible risk. And here is the surprise. For two assets, the minimum-risk mix usually is not 100 percent in the safer asset. Mixing in some of the riskier asset can actually lower total risk, as long as the two are not perfectly correlated, because the riskier asset's offsetting swings cancel some of the safer asset's. For two uncorrelated assets, the weight in the first asset that minimizes risk follows a simple formula.

The safest portfolio is rarely the one holding only the safest asset. A dash of something wild, moving the other way, can steady the whole.

Mean-variance optimization

The formal procedure that finds these best portfolios is called mean-variance optimization. You feed it estimates of each asset's expected return, volatility, and the correlations between them, and it computes the weights that give the most return for any chosen level of risk. The output is the efficient frontier. It is the formal version of the search for the best risk-reward tradeoff, and it is used across the investment industry, though usually with many practical adjustments.

Where you sit, and the honest limitation

The frontier is a menu, not a single answer. A conservative investor picks a point low and to the left, near the minimum-variance portfolio, while an aggressive one picks a point higher and to the right. Where you sit depends on your risk tolerance. But there is a serious catch, familiar from Unit 5. The frontier is built entirely from estimates of returns, volatilities, and correlations, and those estimates are uncertain and drawn from a past that may not repeat. Small errors in the inputs can shift the frontier a lot, so an optimization that looks perfect on historical data can disappoint in the future. Use the efficient frontier as a way of thinking, not as a precise machine to trust blindly.

The frontier so far is built from risky assets only. The next lesson adds a risk-free asset, which straightens the curved frontier into a line and reveals the single best portfolio of risky assets to hold.