Why the Type of Risk Matters
Two assets can have the same total volatility and still have different required returns under CAPM. The distinction is whether an asset tends to move with the market or mainly fluctuates for reasons specific to itself. A market-wide downturn can affect many holdings at once; an isolated operational setback may affect only one company. Holding a range of assets can reduce the impact of the second kind of fluctuation, but it cannot remove exposure shared with the market.
This is the reason CAPM links expected return to systematic risk, measured by beta, rather than to an asset’s total standard deviation. The explanation depends on covariance: the risk relevant to a diversified portfolio is how an asset’s returns move with the returns of the other holdings, especially the market portfolio.
The model is an equilibrium model: its pricing relation applies in the CAPM setting, in which investors can form diversified portfolios and the market portfolio represents market-wide risk. The equation is a required-return relation, not a promise about the return that will actually occur. It does not say that an asset with high total volatility is necessarily rewarded more; it says that, for a given market risk premium, the required return rises with beta.
Separating Market Exposure from Residual Risk
A useful way to interpret beta is to write an asset return as a market-related component plus a residual:
Here \(\varepsilon_i\) is the return component not explained by the market term. For the variance decomposition below, assume that the residual has zero covariance with the market return. That condition is consistent with beta being the market regression coefficient. Then the variance of the return is the sum of a systematic component and a residual component:
The first term is systematic variance: it comes from the asset’s market exposure. The second is residual variance: it is specific to the asset after accounting for market movement. Standard deviation is the square root of variance, so total standard deviation combines both sources; it does not reveal how much of the risk is market-related.
Why Diversification Reduces Residual Risk
Consider \(n\) assets held in equal proportions. Suppose each has the same beta and residual variance, and suppose the residuals of different assets are uncorrelated. These assumptions isolate the diversification effect: market exposure is shared, while residual fluctuations do not systematically occur together.
The portfolio’s beta is the average of the constituent betas. Its residual is the average of the constituent residuals. Because the residuals are uncorrelated, the variance of their average is the sum of their variances divided by \(n^2\), or the individual residual variance divided by \(n\). Thus, adding assets reduces residual variance, while the common market component remains.
If residuals are correlated, their covariances also enter the portfolio residual variance. Shared exposures that cause residuals to move together cannot be diversified away in the same way. The simple division-by-\(n\) result specifically relies on equal residual variances and zero pairwise residual covariances.
Worked Example: Same Beta, Different Total Volatility
Two hypothetical assets, A and B, have market covariance \(0.0200\). The market variance is \(0.0400\), the risk-free rate is \(0.0300\), and the expected market return is \(0.0900\). Asset A has return variance \(0.0900\), while asset B has return variance \(0.0225\). Find each beta, separate systematic and residual variance, and calculate each CAPM required return. Treat the supplied variances and covariances as the complete moment information; no parametric return distribution or mortality model is assumed.
Method and beta. Divide each asset’s covariance with the market by the market variance:
Both assets have the same market exposure, even though their total variances differ.
Systematic and residual variance. The systematic variance for either asset is beta squared times market variance:
Subtract this amount from each total variance. For A, the residual variance is \(0.0900-0.0100=0.0800\). For B, it is \(0.0225-0.0100=0.0125\). The decomposition checks: \(0.0100+0.0800=0.0900\) for A and \(0.0100+0.0125=0.0225\) for B.
Required returns. The market risk premium is \(0.0900-0.0300=0.0600\). Therefore, for each asset,
Each CAPM required return is \(0.0600\), or \(6.00\%\). The result is consistent with their equal betas. The common mistake is to assign A the higher required return because its total variance is larger. That would treat residual volatility as if it were additional market exposure.
Portfolio Risk: What Remains After Diversification
A diversified portfolio’s market component does not disappear when more securities are added. For example, if each holding has positive exposure to the same market movement, the portfolio continues to respond to that movement. Its beta is determined by its weighted market exposure, while residual risk depends on both the residual variances and residual covariances among the holdings.
Worked Example: Adding Holdings Reduces Residual Variance
A hypothetical portfolio initially holds four assets in equal weights. Each asset has beta \(1.00\) and residual variance \(0.0900\). Assume all residuals have zero covariance with the market and are mutually uncorrelated. The market variance is \(0.0400\), the risk-free rate is \(0.0200\), and the expected market return is \(0.0800\). Find the portfolio beta, systematic variance, residual variance, and total variance. Then find the total variance for an otherwise comparable equal-weight portfolio of nine such assets. Use only the stated first and second moments; no normality or mortality assumption is required.
Method and four-asset portfolio. The average beta is \((1.00+1.00+1.00+1.00)/4=1.00\). Its systematic variance is
Each weight is \(1/4\). With uncorrelated residuals, the portfolio residual variance is
Total variance is \(0.0400+0.0225=0.0625\), giving a portfolio standard deviation of \(\sqrt{0.0625}=0.2500\), or \(25.00\%\). Its CAPM required return is \(0.0200+1.00(0.0800-0.0200)=0.0800\), or \(8.00\%\).
Nine-asset portfolio. Its beta remains \(1.00\), so its systematic variance remains \(0.0400\). Its residual variance is
The nine-asset portfolio’s total variance is \(0.0400+0.0100=0.0500\), and its standard deviation is \(\sqrt{0.0500}\approx0.2236\), or \(22.36\%\). Its CAPM required return remains \(8.00\%\) because its beta has not changed. The trap is to assume that more holdings must reduce all risk: here they reduce residual variance, not the systematic variance.
Equal Total Volatility Does Not Mean Equal Required Return
Total volatility alone cannot indicate an asset’s beta. An asset may have substantial total variance because its residual component is large, while another asset with the same total variance may have a larger market-related component. Under CAPM, it is the latter exposure that determines the required return.
Worked Example: Same Standard Deviation, Different Betas
Two hypothetical assets, X and Y, each have return variance \(0.0400\). The market variance is \(0.0400\); their covariances with the market are \(0.0300\) and \(0.0100\), respectively. The risk-free rate is \(0.0200\), and the expected market return is \(0.0800\). Calculate their betas, systematic and residual variances, and CAPM required returns. Use the stated moments only; no particular return distribution or mortality assumption is needed.
Method and beta. The market variance is the same denominator for both beta calculations:
Variance components. For X, systematic variance is \((0.7500)^2(0.0400)=0.0225\), leaving residual variance \(0.0400-0.0225=0.0175\). For Y, systematic variance is \((0.2500)^2(0.0400)=0.0025\), leaving residual variance \(0.0400-0.0025=0.0375\). Both decompositions add to the given total variance of \(0.0400\).
Required returns. The market risk premium is \(0.0800-0.0200=0.0600\). Thus,
The required returns are \(6.50\%\) for X and \(3.50\%\) for Y. Their total standard deviations are both \(20.00\%\), but X has the larger market covariance and beta. A common mistake is to infer equal CAPM required returns from equal standard deviations; beta, not standard deviation by itself, is the relevant comparison.
Interpreting the Result Carefully
The CAPM result is not a claim that residual risk is harmless in every context. A person holding only one company’s shares still bears that company’s residual volatility. A firm evaluating its own cash-flow uncertainty may also care about risks that diversified investors can reduce. The narrower CAPM claim is that, in the model’s equilibrium pricing relation, the required expected return on an asset depends on its contribution to market risk.
Covariance is the key comparison. A high asset variance does not necessarily imply high covariance with the market; the asset may fluctuate for largely independent reasons. Conversely, an asset with modest total variance may have a meaningful beta if much of its variation aligns with market movements. For a portfolio, correlations among residual components matter too: if supposed asset-specific risks in fact move together, diversification will reduce them less than the uncorrelated-residual calculation suggests.
Check Your Understanding
For each problem, use the CAPM relation and the stated covariance or residual assumptions. Returns and variances are expressed in decimal form; no parametric distribution or mortality model is assumed.
- A hypothetical asset has covariance \(0.0100\) with a market of variance \(0.0250\). The risk-free rate is \(0.0250\), and the expected market return is \(0.0750\). Find beta and the CAPM required return. (Answer: \(\beta=0.4000\); required return \(=0.0450\), or \(4.50\%\).)
- A hypothetical asset has total variance \(0.1200\), covariance \(0.0450\) with a market of variance \(0.0360\), and the risk-free rate is \(0.0300\). The expected market return is \(0.0700\). Find beta, systematic variance, residual variance, and required return. (Answer: \(\beta=1.2500\); systematic variance \(=0.05625\); residual variance \(=0.06375\); required return \(=0.0800\), or \(8.00\%\).)
- Five hypothetical assets are held in equal weights. Each has beta \(0.8000\) and residual variance \(0.0500\); residuals are mutually uncorrelated and uncorrelated with the market. Market variance is \(0.0400\), the risk-free rate is \(0.0200\), and the expected market return is \(0.0800\). Find portfolio beta, systematic variance, residual variance, total variance, and required return. (Answer: \(\beta_p=0.8000\); systematic variance \(=0.0256\); residual variance \(=0.0100\); total variance \(=0.0356\); required return \(=0.0680\), or \(6.80\%\).)
- A hypothetical portfolio invests \(60\%\) in asset A and \(40\%\) in asset B. Their betas are \(1.2000\) and \(0.3000\), and their residual variances are \(0.0400\) and \(0.0900\); residuals are mutually uncorrelated and uncorrelated with the market. Market variance is \(0.0250\), the risk-free rate is \(0.0300\), and expected market return is \(0.0700\). Find portfolio beta, systematic variance, residual variance, total variance, and required return. (Answer: \(\beta_p=0.8400\); systematic variance \(=0.01764\); residual variance \(=0.0288\); total variance \(=0.04644\); required return \(=0.0636\), or \(6.36\%\).)