In 1952, a 25-year-old graduate student at the University of Chicago named Harry Markowitz published a 14-page paper called "Portfolio Selection." It was, by his own admission, mostly mathematics. But it changed how the world thought about investing — and earned him a Nobel Prize four decades later.
The central insight was strange and powerful: under certain conditions, you can combine risky investments and end up with a portfolio that is less risky than any of its individual parts. Risk is not just about each investment; it is about how investments behave together. Markowitz called the principle that exploits this the only free lunch in finance.
Most people learn the word "diversification" early — don't put all your eggs in one basket — and assume they understand it. They usually don't.
The Easy Half: Reducing Idiosyncratic Risk
Every individual stock carries two kinds of risk. The first is idiosyncratic risk — the risk specific to that company. A pharmaceutical company's lead drug fails its trial. A retailer suffers a data breach. A CEO is arrested. These are real risks, and they can wipe out a single company at any time.
Idiosyncratic risk can be eliminated by holding many stocks. If you own 30 stocks across different industries, the failure of any single company costs you less than 4% of your portfolio. If you own 500 stocks (think of an S&P 500 index fund), it costs you 0.2%. The math is straightforward.
Meir Statman's 1987 paper "How Many Stocks Make a Diversified Portfolio?" found that the marginal benefit of additional diversification dropped sharply after about 30 stocks, and was essentially gone after 100. This is one reason index funds work so well: they hold hundreds or thousands of stocks at minimal cost, eliminating idiosyncratic risk almost entirely.
The Hard Half: Reducing Correlated Risk
But there is a second kind of risk that diversification across stocks alone cannot fix: systematic risk, also called market risk. This is the risk that affects all stocks together — recessions, interest rate changes, financial crises, pandemics. Owning 500 stocks in the same market won't protect you when the entire market drops 30 percent.
This is where Markowitz's deeper insight comes in. To reduce systematic risk, you need to combine investments that don't move together — assets with low or negative correlation. When one falls, the other holds steady or rises. The portfolio's overall risk is lower than what you would get from any single asset class.
The classic example is stocks and bonds. Stocks tend to deliver higher returns but with much more volatility. Bonds tend to be steadier and often rise when stocks fall (during a flight to safety). A portfolio that holds both will be less volatile than a pure-stock portfolio without sacrificing as much return as you might expect — because the bonds cushion the worst stock years.
This is the principle behind asset allocation, and it is the foundation of modern portfolio theory.
The Counterintuitive Math
Here is what makes diversification feel like magic. Suppose you have two assets, each with 10 percent expected return and 20 percent volatility. If they are perfectly correlated (move in lockstep), combining them gives you 10 percent return and 20 percent volatility — no benefit. But if they have zero correlation (move independently), combining them in equal proportion gives you 10 percent return and only 14 percent volatility. The expected return stays the same; the risk drops by 30 percent.
This is the free lunch. There is nothing else in finance that offers this trade-off. Nothing else lets you reduce risk without giving up return. Markowitz's mathematics showed that for any given level of expected return, there is a portfolio that minimizes risk — and for any given level of risk, there is a portfolio that maximizes return. He called the set of these optimal portfolios the efficient frontier.
What Diversification Cannot Do
Three honest caveats are worth knowing.
Correlation rises in crises. During major market crashes, correlations between asset classes tend to spike. Things that normally move independently start moving together — usually all downward. The 2008 financial crisis is the textbook example: stocks, real estate, commodities, and corporate bonds all fell sharply together. Diversification still helps, but it helps less precisely when you need it most.
Diversification has limits at the global level. You can't diversify away the risk of the entire global economy contracting. If world GDP falls, almost everything falls with it. The only assets that consistently rise in those conditions are very specific hedges (some forms of cash, gold sometimes, certain government bonds), and they typically deliver low long-term returns.
Over-diversification has costs too. Holding too many overlapping funds, paying multiple layers of fees, or owning hundreds of individual stocks introduces complexity and friction without adding much benefit. The goal is enough diversification, not maximum diversification.
What This Means for Most Investors
For someone building a long-term portfolio, the lessons of Markowitz translate into a few practical principles:
- Hold a broad mix of stocks rather than betting on a few picks
- Combine asset classes that don't move together (typically stocks plus bonds)
- Adjust your stock-bond mix based on your timeline and risk tolerance
- Use low-cost index funds to capture broad diversification efficiently
- Rebalance periodically so the mix stays where you want it
None of this is exotic. Most of it can be done with two or three index funds. The free lunch is available to anyone willing to take it.
The thing about a free lunch is that it does not stop being free just because most people do not show up to eat. Markowitz's insight is still on the table. Take it.
