Mean-Variance Portfolio Optimization (Markowitz)
Also known as: Markowitz portfolio theory, modern portfolio theory, efficient frontier optimization, Ortalama-Varyans Portföy Optimizasyonu (Markowitz)
Mean-variance portfolio optimization is the foundational model of modern portfolio theory, introduced by Harry Markowitz in 1952. It describes portfolios in an expected-return versus risk (variance) plane and traces the efficient frontier of allocations that offer the highest expected return for each level of risk, covering the minimum-variance portfolio, the maximum-Sharpe-ratio portfolio, and constrained variants.
Key highlights
- Provides a rigorous, optimisation-based framework for diversification and the trade-off between expected return and risk.
- Generates the full efficient frontier, including the minimum-variance and maximum-Sharpe-ratio portfolios.
- Accommodates practical long-only and concentration constraints and accepts robust covariance estimators.
Intuition
This section is available to Pro members. Upgrade to Pro
How it works
This section is available to Pro members. Upgrade to Pro
When to use it
Use mean-variance optimization to construct or rebalance a portfolio across several assets when you have a reasonable history of returns (the model expects at least 60 observations, ideally far more) and a stationary time-series structure. Returns are assumed to be adequately described by their means and covariance matrix, with normality being the ideal case. It is best suited to prioritisation and prediction goals over continuous return series. It is unreliable when the return history is short, when returns are non-stationary, or when the covariance estimate is too noisy to trust.
Strengths & limitations
- Provides a rigorous, optimisation-based framework for diversification and the trade-off between expected return and risk.
- Generates the full efficient frontier, including the minimum-variance and maximum-Sharpe-ratio portfolios.
- Accommodates practical long-only and concentration constraints and accepts robust covariance estimators.
- Output weights are extremely sensitive to estimation error in the inputs — the model acts as an 'error maximiser'.
- The covariance matrix estimate is noisy; shrinkage (Ledoit-Wolf) is needed to stabilise it.
- With fewer than 250 observations the covariance matrix is unreliable and a risk-parity model is preferable.
- Non-stationary returns make the optimisation misleading and require differencing or an ARIMA-style treatment first.
Common pitfalls
This section is available to Pro members. Upgrade to Pro
Applications
This section is available to Pro members. Upgrade to Pro
Frequently asked
What is the efficient frontier?
It is the curve of portfolios that deliver the highest expected return for each level of risk (variance). Portfolios below the frontier are dominated; you can always find another mix with the same risk but higher expected return.
Why are the optimised weights so sensitive to the inputs?
Mean-variance optimization behaves like an 'error maximiser': small errors in the estimated expected returns or covariances are amplified into large swings in the weights. Robust covariance estimation and constraints reduce this instability.
Why use Ledoit-Wolf shrinkage for the covariance matrix?
The sample covariance matrix is noisy, especially when the number of assets is large relative to the history. Ledoit and Wolf's (2004) shrinkage estimator produces a well-conditioned matrix that stabilises the optimisation and yields more reliable weights.
When should I use a risk-parity model instead?
When the return history is short — fewer than about 250 observations — the covariance matrix becomes unreliable and the mean-variance weights are untrustworthy. A risk-parity model, which allocates by risk contribution rather than estimated returns, is then a more robust alternative.
Sources
- 1.Markowitz, H. (1952). Portfolio Selection. The Journal of Finance, 7(1), 77-91.
- 2.Ledoit, O. & Wolf, M. (2004). A Well-Conditioned Estimator for Large-Dimensional Covariance Matrices. Journal of Multivariate Analysis, 88(2), 365-411.
You have read it. What now?
Cite this page
ScholarGate. (2026, June 1). Mean-Variance Portfolio Optimization. ScholarGate. https://scholargate.app/finance/portfolio-optimization-mean-variance