How Diversification Reduces Risk Without Lowering Returns
This article provides a clear breakdown of the portfolio mathematics behind diversification. It explains how combining multiple assets eliminates idiosyncratic (asset-specific) risk without reducing a portfolio's expected return. By examining the mathematical formulas for expected value and variance, you will understand why returns simply average out while independent risks actively cancel each other out.
The Linearity of Expected Returns
To understand why diversification does not sacrifice returns, we first look at how a portfolio's expected return is calculated. According to portfolio mathematics, the expected return of a portfolio is simply the weighted average of the expected returns of its individual assets.
The formula is: E(Rp) = ∑ wi * E(Ri)
Where:
- E(Rp) is the expected return of the portfolio.
- wi is the weight (percentage of total capital) of asset i.
- E(Ri) is the expected return of asset i.
Because this equation is linear, adding more assets does not dilute the return, provided the new assets have similar expected returns. For example, if you invest 100% of your money in a single stock with an expected return of 10%, your expected portfolio return is 10%. If you instead divide your money equally among 50 different stocks, each with an expected return of 10%, your expected portfolio return remains 10%. You lose no expected upside by spreading your capital.
Total Risk: Systematic vs. Idiosyncratic
In finance, risk is measured as the variance (or standard deviation) of returns. An individual asset's total risk is composed of two parts:
- Systematic Risk: Market-wide risk (e.g., inflation, interest rate changes, recessions) that affects all assets. This cannot be diversified away.
- Idiosyncratic Risk: Firm-specific risk (e.g., a failed product launch, a CEO scandal, a factory fire) that affects only a single asset or a small group of assets.
Idiosyncratic risks are statistically independent, meaning the occurrence of a specific bad event at one company has no bearing on whether a specific bad event happens at another unrelated company. In mathematical terms, the correlation between the idiosyncratic risks of two different companies is zero.
The Non-Linear Math of Portfolio Risk
Unlike expected return, portfolio risk is not a simple weighted average of individual asset risks. It must account for how the assets move in relation to one another, a metric known as covariance.
If we construct an equally weighted portfolio of N assets, the mathematical formula for the portfolio's total variance (\(\sigma_p^2\)) can be simplified to:
\(\sigma_p^2\) = (1 / N) * Average Variance + ((N - 1) / N) * Average Covariance
This formula splits the portfolio's risk into two distinct buckets:
- The Idiosyncratic Component: (1 / N) * Average Variance. This represents the individual, isolated risks of the assets.
- The Systematic Component: ((N - 1) / N) * Average Covariance. This represents how the assets move together in response to broader market forces.
How Idiosyncratic Risk Vanishes
The magic of diversification happens when we increase N (the number of assets in the portfolio) and observe what happens to the two parts of the risk equation.
As N grows larger—say, from 1 stock to 100 stocks:
- The fraction 1 / N becomes 1 / 100 (or 0.01). The "Average Variance" term is divided by 100, effectively shrinking it toward zero. Because idiosyncratic risk has zero covariance with other assets, it lives entirely inside this first term. Therefore, as N approaches infinity, idiosyncratic risk is mathematically divided into nothingness.
- Meanwhile, the fraction (N - 1) / N becomes 99 / 100 (or 0.99), which approaches 1. This means the "Average Covariance" term remains intact. You are left only with the baseline systematic risk of the market.
Summary of the Mathematical Asymmetry
Diversification works—and is often called the only "free lunch" in finance—because of the mathematical asymmetry between expected value and variance:
- Returns are additive: The weights in the expected return formula always sum to 1, preserving the average return regardless of how many assets you hold.
- Independent risks are divisive: Uncorrelated fluctuations offset one another. For every asset that experiences a random negative shock, another asset is likely experiencing a random positive shock.
By holding a large number of assets, the random noise of idiosyncratic risk cancels out through the Law of Large Numbers, leaving the expected return untouched and the investor exposed only to the broader movements of the market.