Modern portfolio theory

Modern portfolio theory, or MPT, explains how an asset’s expected return, volatility and relationship with other assets affect portfolio-level risk. Its central insight is diversification: what matters is not only the risk of each holding, but how holdings move together. The framework shaped investment practice and broader risk thinking, yet its outputs depend heavily on uncertain estimates.

When to use it

• To structure discussion of asset allocation and diversification.
• To distinguish diversifiable asset-specific risk from market-wide risk.

This article is educational, not personalised investment advice. Suitability depends on circumstances, jurisdiction, horizon, liquidity, tax and capacity for loss.

Origins

Harry Markowitz introduced the approach in “Portfolio Selection,” published in the Journal of Finance in 1952. Earlier stock selection often ranked securities individually by expected risk and return. Markowitz showed that combining individually attractive assets from the same correlated group can leave a portfolio dangerously concentrated.

In the model, returns are random variables described by expected return, variance and covariance. Feasible portfolios can then be compared. Those offering the greatest expected return for a given variance—or the least variance for a given expected return—form the efficient frontier.

What it is

MPT uses diversification to reduce unsystematic, asset-specific risk. Systematic risk, such as broad market shocks, cannot be eliminated merely by holding more securities.

Standard deviation measures volatility in the model, while covariance or correlation captures how holdings move together. Beta is a measure of systematic sensitivity used in the capital asset pricing model; see Capital asset pricing model (Finance); thumbnail basis: capital-asset-pricing-model-figure-01.jpg.. It is not a complete measure of every risk relevant to inclusion.

Assets do not need to be negatively correlated for diversification to help; imperfect positive correlation can still reduce portfolio volatility. Correlations can rise in crises, however, so historical diversification may fail when protection is most needed.