Markowitz's 1952 paper turned portfolio risk into a solvable equation

When Harry Markowitz chose the stock market as his dissertation subject at the University of Chicago, finance had no operational theory of risk. 'Don't put all your eggs in one basket' was the sum of diversification doctrine, and securities were pitched one at a time. His 1952 essay 'Portfolio Selection' replaced the adjectives with algebra.

The move was to treat the portfolio, not the security, as the unit of analysis. A portfolio's risk is its variance, which depends on every pairwise covariance — so an asset's risk is what it contributes to the whole, and a jumpy stock that moves against the rest can make the portfolio safer. Choice reduces to a trade-off between expected return and variance, which a 1956 essay showed how to solve by formulating it as quadratic programming; the 1959 book 'Portfolio Selection: Efficient Diversification' carried the framework to practitioners.

The Nobel committee in 1990 credited the work with establishing financial micro analysis as a respectable research area, sharing the prize with William Sharpe, who built CAPM on this foundation, and Merton Miller. Mean-variance logic became the spine of institutional asset allocation — the efficient frontier entered every textbook, and Britannica records that Markowitz's techniques became standard investment procedures.

An asset's risk is its covariance, not itself

Before 1952 risk was an adjective; converting it to variance made portfolios optimizable rather than merely felt.

Covariance reframed the question — not 'is this asset risky?' but 'what does it do to the whole?'

The 1956 quadratic-programming formulation made the theory computable with the machines of its day.

Sitting at RAND put the problem next to the optimization techniques and computing that could solve it.

Quantifying a fuzzy variable — turning 'risk' into variance — converts taste into engineering. Choose the unit of analysis first: Markowitz's breakthrough was measuring the portfolio, not the stock.

Mean-variance analysis became standard practice for investment managers, and CAPM rose directly on top of it. Markowitz spent his career on implementation as much as theory: he co-created the SIMSCRIPT simulation language, founded Consolidated Analysis Centers, ran Arbitrage Management Company in the early 1970s, worked at IBM's Watson lab, and taught into his nineties at UC San Diego until his death in June 2023.