Modern Portfolio Theory — Why Combining Assets Reduces Risk

In 1952, a 25-year-old graduate student named Harry Markowitz published a 14-page paper titled “Portfolio Selection” that fundamentally rewired how the world thinks about investing. He won the Nobel Prize for it 38 years later. The core insight, that combining imperfectly correlated assets can produce a portfolio with lower risk than any of its components individually, is the mathematical foundation of every diversified portfolio in the world today. It also gets misunderstood and misapplied constantly. This lesson explains what it actually says.

The Pre-Markowitz World

Before Markowitz, professional investing was about picking individual securities. Find the cheapest stocks, buy them, hold them, repeat. Risk was barely quantified, investors talked about it qualitatively (this stock is “speculative,” that one is “safer”) but had no mathematical framework for thinking about how holdings combined.

Markowitz’s insight was that the risk of a portfolio is not just the average of the risks of its components. Two volatile stocks combined can produce a portfolio less volatile than either one, if their volatilities don’t move in the same direction at the same time. The key variable is correlation: the statistical relationship between how two assets move.

The Diversification Equation

For a two-asset portfolio: portfolio variance = (w1² × σ1²) + (w2² × σ2²) + (2 × w1 × w2 × ρ × σ1 × σ2). The third term is the magic: when correlation (ρ) is less than 1, the cross-term reduces total variance. When correlation is exactly 1, you get no benefit. When correlation is -1, you can build a perfectly riskless portfolio. Real-world correlations between asset classes typically range from -0.3 to +0.8, so meaningful diversification is achievable but never perfect.

The Efficient Frontier

For any given level of risk (volatility), there’s a maximum expected return achievable by combining the available assets in optimal proportions. Plot all those points and you get a curve called the efficient frontier. Any portfolio on this curve is efficient, you cannot improve return without taking more risk, and you cannot reduce risk without giving up return. Any portfolio below the curve is suboptimal, there’s some other combination that gives you more return at the same risk, or less risk at the same return.

The implication is striking: holding any single stock, even a great one, is almost always suboptimal compared to a portfolio. A single stock’s risk includes its idiosyncratic risk (company-specific issues like a fraud, a product recall, an unexpected lawsuit). A diversified portfolio diversifies away most of this idiosyncratic risk while keeping most of the expected return. You’re not paid to take idiosyncratic risk in efficient markets, only systematic (market-wide) risk.

How Many Stocks Diversify Away Idiosyncratic Risk?

Studies (Evans & Archer 1968, Statman 1987) suggest you capture most of the diversification benefit with roughly 15-30 stocks if they’re chosen across industries. Beyond that, the marginal benefit drops sharply, going from 30 to 100 stocks barely changes your risk-adjusted return. Going from 1 to 15 stocks reduces idiosyncratic risk by about 75%; going from 15 to 100 reduces it by another 5%. The practical takeaway: a portfolio of 20-30 carefully chosen stocks across sectors is “diversified enough” for most purposes.

The Texas Wildcatters Problem

Imagine 100 oil wildcatters, each spending $1M to drill a well. Each individual well has a 10% chance of striking oil worth $30M and a 90% chance of producing nothing. Each wildcatter individually has a high probability of going broke. But pool the 100 wells into one portfolio: expected value is 100 × ($3M − $1M) = $200M, with very high probability of meaningful profit. The pooling didn’t change any individual well’s economics, it just took the same expected return and made it much more reliable. That’s diversification in its purest form.

The Capital Market Line and the Single Best Risky Portfolio

An extension of MPT (called the Capital Asset Pricing Model, CAPM) showed that if all investors face the same opportunity set, there’s actually only one optimal risky portfolio, the “market portfolio” containing every available asset weighted by market capitalization. The only decision that matters is how much of your wealth to put in this market portfolio versus risk-free bonds.

This conclusion is the theoretical justification for index investing. If the optimal risky portfolio is the market itself, just buying it (via an S&P 500 or total-market index fund) gets you most of the way to optimal. Active management has to overcome a high bar, fees, taxes, transaction costs, to justify deviating from this passive approach.

What MPT Assumes (And Why Reality Disagrees)

Modern Portfolio Theory’s elegance comes from strong assumptions: returns are normally distributed, correlations are stable, investors have access to risk-free borrowing and lending at the same rate, transaction costs are zero, and information is symmetric. None of these are quite true. Returns have fat tails (extreme moves happen far more often than the normal distribution predicts). Correlations spike toward 1 in crises (exactly when diversification matters most). Real risk-free rates differ from real-world borrowing costs by hundreds of basis points. Information asymmetry is enormous. The framework is still useful, but applied without these caveats it leads to disasters like Long-Term Capital Management.

The Correlation-in-Crisis Problem

The single biggest practical limitation of MPT is that correlations between assets are not stable. In normal times, US stocks and emerging market stocks might have a correlation of 0.6, meaningful diversification. In crisis times (2008, 2020), correlations between essentially all risk assets spike toward 0.9-1.0 as everyone sells everything to raise cash. Exactly when you need diversification most, you tend to have it least.

This is why genuine diversification requires assets that maintain uncorrelated behavior in stress, typically long-duration government bonds (which often rally in crises as the Fed cuts rates), gold (which sometimes acts as a flight-to-safety asset), and potentially cash itself. A portfolio of “diversified” risk assets (US stocks + EM stocks + REITs + corporate bonds + commodities) often performs identically to all-stocks during a real crash because they all decline together.

Beyond Asset Classes. Diversifying Across Time, Strategy, and Manager

Modern thinking goes further than asset class diversification. A complete diversification framework includes:

Time diversification: Dollar-cost averaging into positions over months or years rather than lump-sum at one price reduces the risk of buying at a single bad moment.
Strategy diversification: Combining momentum, value, and quality strategies in equity allocations smooths out periods when any single factor underperforms.
Geographic diversification: Holding equities across multiple developed and emerging markets reduces single-country political and economic risk.
Currency diversification: Some exposure to non-home-currency assets hedges against the home currency’s potential debasement.
Tail-risk diversification: Allocating a small percentage to assets that explicitly perform in crises (long volatility positions, gold, long-duration treasuries) addresses the correlation-in-crisis problem.

“Diversification is the only free lunch in investing.”

. Harry Markowitz

The Concentration Counterargument

Despite the elegance of MPT, the greatest investors in history (Buffett, Munger, Lynch, Soros) all ran concentrated portfolios. Their argument: diversification protects against ignorance. If you actually know what you’re doing, if you’ve done deep work on a few businesses you understand, diversification beyond your highest-conviction ideas dilutes returns without improving risk-adjusted outcomes.

Both views are simultaneously correct. For the average investor without time or expertise to deeply analyze 30 businesses, diversification (especially via index funds) is overwhelmingly the right answer. For the rare investor who genuinely does the work to understand a small number of businesses, concentration produces dramatically better outcomes, at the cost of higher short-term volatility and the risk of being catastrophically wrong if their analysis fails. Choose honestly.

The Practical Takeaways

For most investors: own 20-30 stocks (or a low-cost index fund), spread across sectors, with some allocation to bonds and possibly gold to manage drawdowns. Don’t believe the elegance of the math more than the messiness of crisis correlations. Recognize that diversification doesn’t eliminate risk, it just changes which risks you’re taking. The market risk you can’t diversify away is the source of your long-term returns; the idiosyncratic risk you can diversify away is what would have ruined you if you’d been concentrated in the wrong stock at the wrong time.