The Single-Index Model
Understanding Systematic Risk and Stock Return Decomposition
1 Introduction
In our earlier discussions of portfolio theory, we discovered a powerful insight: holding a diversified portfolio allows investors to eliminate some—but not all—of the risk associated with individual stocks. This naturally raises a question that lies at the heart of modern asset pricing: if some risk can be diversified away, should investors be compensated for bearing it? And if not, what exactly determines the expected return on a stock?
The single-factor model provides an elegant framework for answering these questions. It decomposes a stock’s return into two distinct components: one driven by broad market movements that affect all firms (systematic risk) and another driven by firm-specific events that can be diversified away (idiosyncratic risk). This decomposition is not merely an academic exercise—it forms the foundation for understanding how securities are priced, how portfolios should be constructed, and how investment performance should be evaluated.
The key insight we will develop in this lecture is that the market rewards investors only for bearing systematic risk. Since idiosyncratic risk can be eliminated through diversification, rational investors will not pay a premium to avoid it, and therefore it earns no additional expected return. This principle has profound implications for both investors seeking to understand expected returns and portfolio managers attempting to beat the market.
2 The Single-Factor Model
2.1 The Core Idea
The single-factor model rests on a simple but powerful assumption: all sources of systematic risk—the macroeconomic forces that affect every firm to some degree—can be captured by the returns on the market portfolio. If this assumption holds, then we can express any stock’s excess return as a linear function of the market’s excess return.
The model is expressed mathematically as follows:
\[R_{i,t} - R_{f,t} = \alpha_i + \beta_i (R_{m,t} - R_{f,t}) + \epsilon_{i,t}\]
where \(R_{i,t}\) denotes the return on stock \(i\) at time \(t\), \(R_{f,t}\) is the risk-free rate at time \(t\), and \(R_{m,t}\) is the return on the market portfolio at time \(t\). The difference between any return and the risk-free rate is called the excess return—it represents the additional return earned above what could have been earned with zero risk.
This equation tells us that a stock’s excess return has three components. The term \(\beta_i (R_{m,t} - R_{f,t})\) captures the portion of the stock’s return that moves with the market. The coefficient \(\beta_i\) measures the sensitivity of the stock to market movements. The term \(\alpha_i\) represents the average excess return the stock earns beyond what is explained by its market exposure. Finally, \(\epsilon_{i,t}\) captures the random, firm-specific shock that affects the stock’s return in period \(t\).
2.2 Understanding Beta
The beta coefficient is perhaps the most important parameter in the single-factor model. It tells us how much systematic risk a stock carries relative to the average stock in the market. Since the market portfolio, by definition, has average market sensitivity, its beta equals exactly 1.
A stock with \(\beta > 1\) is more sensitive to market movements than the average stock. When the market rises by 1%, we expect this stock to rise by more than 1%. Conversely, when the market falls, high-beta stocks tend to fall more sharply. These stocks are sometimes called “aggressive” because they amplify market movements.
A stock with \(\beta < 1\) is less sensitive to market movements. These “defensive” stocks provide some insulation against market downturns but also participate less in market rallies. A stock with \(\beta = 0\) would theoretically have no sensitivity to market movements at all, behaving like a risk-free asset in terms of its systematic risk exposure.
The risk-free asset has a beta of exactly zero—its return does not vary with market conditions. The market portfolio has a beta of exactly one and, by construction, an alpha of zero.
2.3 Understanding Alpha
The alpha coefficient tells us whether a stock has historically delivered returns above or below what its beta would predict. A positive alpha indicates that the stock has outperformed its beta-adjusted benchmark, while a negative alpha indicates underperformance.
In an efficient market where the single-factor model correctly captures all systematic risk, we would expect alpha to be zero for all stocks. Any non-zero alpha would represent either a mispricing that should be quickly arbitraged away or compensation for some additional risk factor not captured by the model. This is why alpha is often interpreted as a measure of “abnormal” return—it represents performance that cannot be explained by exposure to market risk.
For portfolio managers, generating positive alpha is the holy grail. It represents genuine skill in stock selection that adds value beyond what could be achieved by simply adjusting market exposure.
2.4 Estimating the Model
The parameters of the single-factor model—alpha and beta—are estimated using ordinary least squares (OLS) regression. By regressing a stock’s historical excess returns on the market’s historical excess returns, we obtain estimates of both the intercept (alpha) and the slope (beta).
Example: Estimating Beta from Historical Data
Suppose you have collected monthly excess return data for Stock A and the market portfolio over the past 60 months. Running a regression of Stock A’s excess returns on market excess returns yields the following output:
- Intercept (\(\hat{\alpha}\)) = 0.002 (0.2% per month)
- Slope (\(\hat{\beta}\)) = 1.35
- \(R^2\) = 0.42
How do we interpret these results?
The estimated beta of 1.35 tells us that Stock A is an aggressive stock. On average, when the market rises by 1%, Stock A rises by 1.35%. When the market falls by 1%, Stock A falls by 1.35%.
The estimated alpha of 0.2% per month suggests that Stock A has historically outperformed its beta-adjusted benchmark by 0.2% per month, or roughly 2.4% per year. If markets are efficient, this could be due to statistical noise, or it might suggest the stock was underpriced during the estimation period.
The \(R^2\) of 0.42 indicates that 42% of Stock A’s return variance is explained by market movements. The remaining 58% is due to firm-specific factors.
3 Systematic versus Idiosyncratic Risk
3.1 Decomposing Total Risk
One of the most valuable applications of the single-factor model is its ability to decompose a stock’s total risk into systematic and idiosyncratic components. Taking the variance of both sides of the single-factor equation yields:
\[\text{Var}(R_{i,t} - R_{f,t}) = \beta_i^2 \cdot \text{Var}(R_{m,t} - R_{f,t}) + \text{Var}(\epsilon_{i,t})\]
This equation states that total variance equals systematic variance plus idiosyncratic variance. In terms of risk (measured by variance), we have:
\[\sigma_i^2 = \beta_i^2 \sigma_m^2 + \sigma_{\epsilon,i}^2\]
The term \(\beta_i^2 \sigma_m^2\) represents systematic risk—the portion of the stock’s variance that comes from market-wide movements. This risk cannot be diversified away because it affects all stocks simultaneously (though to different degrees).
The term \(\sigma_{\epsilon,i}^2\) represents idiosyncratic risk—the portion of the stock’s variance that comes from firm-specific events. This risk can be diversified away because the firm-specific shocks affecting different companies are typically uncorrelated.
3.2 The Role of R-Squared
The regression \(R^2\) from the single-factor model has a direct and intuitive interpretation: it tells us the fraction of a stock’s total variance that is systematic. Mathematically:
\[R^2 = \frac{\beta_i^2 \sigma_m^2}{\sigma_i^2}\]
If a stock has an \(R^2\) of 0.30, then 30% of its total variance is systematic (related to market movements) and 70% is idiosyncratic (related to firm-specific events). This means that in a well-diversified portfolio, roughly 70% of this stock’s standalone risk would be eliminated.
Stocks in different industries tend to have very different \(R^2\) values. Utility stocks, which are relatively insulated from economic cycles, often have lower \(R^2\) values. Technology stocks, which are highly sensitive to economic conditions and growth expectations, tend to have higher \(R^2\) values.
Example: Calculating Risk Decomposition
Stock B has a beta of 1.2 and a total return standard deviation of 40% per year. The market portfolio has a standard deviation of 18% per year. What fraction of Stock B’s variance is systematic?
First, we calculate the systematic variance: \[\beta^2 \sigma_m^2 = (1.2)^2 \times (0.18)^2 = 1.44 \times 0.0324 = 0.0467\]
Next, we calculate the total variance: \[\sigma_i^2 = (0.40)^2 = 0.16\]
The fraction of variance that is systematic (which equals \(R^2\)) is: \[R^2 = \frac{0.0467}{0.16} = 0.292 \approx 29\%\]
This tells us that about 29% of Stock B’s risk is systematic and cannot be diversified away. The remaining 71% is idiosyncratic and can be eliminated through diversification.
We can also back out the idiosyncratic standard deviation: \[\sigma_{\epsilon}^2 = \sigma_i^2 - \beta^2\sigma_m^2 = 0.16 - 0.0467 = 0.1133\] \[\sigma_{\epsilon} = \sqrt{0.1133} = 0.337 = 33.7\%\]
4 Adjusted Betas
4.1 Why Betas Change Over Time
Empirical research has consistently documented that estimated betas tend to revert toward the market average of 1.0 over time. Stocks with high estimated betas today tend to have somewhat lower betas in the future, while stocks with low estimated betas tend to have somewhat higher betas in the future.
Several factors explain this phenomenon. First, there is regression to the mean in the statistical sense—extreme estimates often contain measurement error that tends to correct over time. Second, companies themselves change. High-growth, high-beta companies mature and become more stable. Struggling, low-beta companies either improve their operations or exit the market.
4.2 Calculating Adjusted Beta
To account for this mean reversion when forecasting future betas, practitioners commonly use adjusted betas. The most widely used formula is:
\[\beta_i^{adj} = \frac{2}{3} \times \beta_i + \frac{1}{3} \times 1.0\]
This formula shrinks the estimated beta toward 1.0 by giving two-thirds weight to the estimated value and one-third weight to the market mean. The 2/3 and 1/3 weights are based on empirical research by Marshall Blume, though other weighting schemes are sometimes used.
Example: Computing Adjusted Beta
Stock C has an estimated beta of 1.6 based on the past 60 months of returns. Calculate its adjusted beta.
Using the standard adjustment formula: \[\beta^{adj} = \frac{2}{3} \times 1.6 + \frac{1}{3} \times 1.0\] \[\beta^{adj} = 1.067 + 0.333 = 1.40\]
The adjusted beta of 1.40 is lower than the raw beta of 1.6, reflecting the expectation that this high-beta stock will likely become less aggressive over time. An analyst forecasting the stock’s future sensitivity to market movements would use 1.40 rather than 1.6.
4.3 When to Use Adjusted Betas
Adjusted betas are most appropriate when the goal is to forecast future sensitivity to market movements. This is relevant for portfolio construction, risk management, and cost of equity estimation. Raw (unadjusted) betas are more appropriate when the goal is to describe historical performance or decompose past returns.
5 Portfolio Betas
5.1 Betas Combine Linearly
One of the convenient properties of beta is that it combines linearly across a portfolio. The beta of a portfolio is simply the weighted average of the betas of the individual securities:
\[\beta_p = \sum_{i=1}^{N} w_i \beta_i = w_1 \beta_1 + w_2 \beta_2 + \cdots + w_N \beta_N\]
where \(w_i\) is the portfolio weight of security \(i\) (the fraction of portfolio value invested in security \(i\)) and \(\beta_i\) is the beta of security \(i\).
This property makes portfolio risk management straightforward. If you want to increase your portfolio’s market sensitivity, overweight high-beta stocks. If you want to reduce sensitivity, overweight low-beta stocks or add risk-free assets.
5.2 Adjusting Portfolio Beta with the Risk-Free Asset
Since the risk-free asset has a beta of zero, allocating funds between a risky portfolio and the risk-free asset provides a simple way to dial portfolio beta up or down. If you invest fraction \(w\) in a risky portfolio with beta \(\beta_p\) and fraction \((1-w)\) in the risk-free asset, your overall beta is simply \(w \times \beta_p\).
You can even use leverage (borrowing at the risk-free rate to invest more than 100% in the risky portfolio) to achieve a beta greater than the risky portfolio’s beta alone.
Example: Calculating Portfolio Beta
You manage a portfolio with the following positions:
| Stock | Portfolio Weight | Beta |
|---|---|---|
| D | 25% | 0.8 |
| E | 35% | 1.2 |
| F | 40% | 1.5 |
What is the portfolio beta? If you wanted to reduce the portfolio beta to 1.0 by adding a position in Treasury bills (which have zero beta), what fraction of the portfolio would need to be reallocated to T-bills?
Part 1: Calculate the current portfolio beta
\[\beta_p = (0.25 \times 0.8) + (0.35 \times 1.2) + (0.40 \times 1.5)\] \[\beta_p = 0.20 + 0.42 + 0.60 = 1.22\]
Part 2: Determine the T-bill allocation needed
Let \(w\) be the fraction invested in the original risky portfolio and \((1-w)\) be the fraction in T-bills. We want the combined beta to equal 1.0:
\[w \times 1.22 + (1-w) \times 0 = 1.0\] \[1.22w = 1.0\] \[w = 0.82\]
Therefore, you would keep 82% of the portfolio in the original stock positions (maintaining their relative weights) and move 18% into Treasury bills. This would reduce the portfolio beta from 1.22 to 1.0.
6 Key Takeaways
The single-factor model provides a powerful lens for understanding stock returns and risk. At its core, the model recognizes that stock returns are driven by two fundamentally different types of forces: market-wide movements that affect all stocks and firm-specific events that affect only individual companies.
The beta coefficient measures a stock’s sensitivity to market movements and determines its systematic risk contribution. High-beta stocks amplify market movements and carry more undiversifiable risk, while low-beta stocks provide relative stability but participate less in market rallies. Because systematic risk cannot be diversified away, investors require compensation for bearing it, which is why beta plays a central role in asset pricing models.
The alpha coefficient captures performance above or beyond what beta would predict. In efficient markets, alpha should be zero on average, making it a useful benchmark for evaluating whether a manager or strategy truly adds value. The regression \(R^2\) tells us what fraction of a stock’s total risk is systematic—the remainder being idiosyncratic risk that diversification can eliminate.
When using betas for forecasting, practitioners typically adjust them toward the mean of 1.0 to account for documented mean reversion in beta estimates. Finally, the linearity of portfolio betas—the portfolio beta equals the weighted average of individual betas—makes risk management intuitive and tractable. By adjusting portfolio weights, investors can target any desired level of market sensitivity.
7 Key Formulas Summary
| Concept | Formula | When to Use |
|---|---|---|
| Single-Factor Model | \(R_{i,t} - R_{f,t} = \alpha_i + \beta_i(R_{m,t} - R_{f,t}) + \epsilon_{i,t}\) | Decomposing stock returns into systematic and firm-specific components |
| Variance Decomposition | \(\sigma_i^2 = \beta_i^2 \sigma_m^2 + \sigma_{\epsilon,i}^2\) | Separating total risk into systematic and idiosyncratic portions |
| Systematic Risk Fraction | \(R^2 = \frac{\beta_i^2 \sigma_m^2}{\sigma_i^2}\) | Determining what percentage of risk can vs. cannot be diversified away |
| Adjusted Beta | \(\beta^{adj} = \frac{2}{3}\beta + \frac{1}{3}(1)\) | Forecasting future beta for risk management or cost of equity estimation |
| Portfolio Beta | \(\beta_p = \sum_{i=1}^{N} w_i \beta_i\) | Calculating the systematic risk of a portfolio |
8 Practice Problems
Practice Problem 1: Interpreting Regression Output
You run a single-factor regression for Stock G using 48 months of data and obtain the following results:
- Intercept (\(\hat{\alpha}\)) = -0.003 (-0.3% per month)
- Slope (\(\hat{\beta}\)) = 0.75
- \(R^2\) = 0.28
Interpret each of these statistics. Is Stock G aggressive or defensive? What fraction of its risk is diversifiable?
Beta interpretation: The beta of 0.75 tells us Stock G is a defensive stock. When the market rises by 1%, Stock G tends to rise by only 0.75%. When the market falls by 1%, Stock G tends to fall by only 0.75%. It has less systematic risk than the average stock.
Alpha interpretation: The negative alpha of -0.3% per month (roughly -3.6% annually) suggests Stock G has underperformed what its beta would predict. Either the stock was overpriced during this period, the negative alpha is due to statistical noise, or there are other risk factors not captured by the model.
\(R^2\) interpretation: Only 28% of Stock G’s variance is explained by market movements. This means 28% of its risk is systematic (undiversifiable), while 72% of its risk is idiosyncratic (diversifiable). In a well-diversified portfolio, most of Stock G’s standalone risk would be eliminated.
Practice Problem 2: Risk Decomposition
Stock H has the following characteristics:
- Beta = 0.9
- Total standard deviation = 35% per year
- Market standard deviation = 20% per year
Calculate the systematic and idiosyncratic variance of Stock H. What percentage of Stock H’s total variance is systematic?
Step 1: Calculate systematic variance \[\beta^2 \sigma_m^2 = (0.9)^2 \times (0.20)^2 = 0.81 \times 0.04 = 0.0324\]
Step 2: Calculate total variance \[\sigma_i^2 = (0.35)^2 = 0.1225\]
Step 3: Calculate idiosyncratic variance \[\sigma_{\epsilon}^2 = \sigma_i^2 - \beta^2\sigma_m^2 = 0.1225 - 0.0324 = 0.0901\]
Step 4: Calculate the systematic fraction (\(R^2\)) \[R^2 = \frac{0.0324}{0.1225} = 0.265 \approx 26.5\%\]
Approximately 26.5% of Stock H’s variance is systematic, and 73.5% is idiosyncratic. The idiosyncratic standard deviation is \(\sqrt{0.0901} = 30.0\%\).
Practice Problem 3: Adjusted Beta Calculation
Stock I has an estimated (raw) beta of 0.55 based on historical data. Calculate its adjusted beta using the standard Blume adjustment. Explain why the adjusted beta is higher than the raw beta.
Calculate adjusted beta: \[\beta^{adj} = \frac{2}{3} \times 0.55 + \frac{1}{3} \times 1.0\] \[\beta^{adj} = 0.367 + 0.333 = 0.70\]
Explanation: The adjusted beta of 0.70 is higher than the raw beta of 0.55 because the adjustment formula shrinks all betas toward the market mean of 1.0. For below-average betas, this shrinkage pulls the estimate upward. This reflects the empirical finding that low-beta stocks tend to see their betas increase over time (mean reversion), so the adjusted beta provides a better forecast of future market sensitivity.
Practice Problem 4: Portfolio Beta and Risk Adjustment
You are constructing a portfolio with the following three stocks:
| Stock | Investment Amount | Beta |
|---|---|---|
| J | $30,000 | 1.4 |
| K | $50,000 | 1.0 |
| L | $20,000 | 0.6 |
Calculate the portfolio beta. If you then add $25,000 in Treasury bills to the portfolio, what is the new portfolio beta?
Step 1: Calculate portfolio weights for the original portfolio
Total investment = $30,000 + $50,000 + $20,000 = $100,000
- \(w_J = 30,000 / 100,000 = 0.30\)
- \(w_K = 50,000 / 100,000 = 0.50\)
- \(w_L = 20,000 / 100,000 = 0.20\)
Step 2: Calculate original portfolio beta \[\beta_p = (0.30 \times 1.4) + (0.50 \times 1.0) + (0.20 \times 0.6)\] \[\beta_p = 0.42 + 0.50 + 0.12 = 1.04\]
Step 3: Calculate new weights after adding T-bills
New total = $100,000 + $25,000 = $125,000
- Weight in original stocks = $100,000 / $125,000 = 0.80
- Weight in T-bills = $25,000 / $125,000 = 0.20
Step 4: Calculate new portfolio beta
Since T-bills have \(\beta = 0\): \[\beta_{new} = (0.80 \times 1.04) + (0.20 \times 0) = 0.832\]
Adding T-bills reduced the portfolio beta from 1.04 to 0.832, making it less sensitive to market movements.
9 Ask an LLM
Here are three questions you might ask an AI assistant to deepen your understanding of these concepts:
- What happens to a stock’s alpha and beta estimates if I use different time periods (e.g., 36 months vs. 60 months) or different return frequencies (daily vs. monthly)? Why might the estimates differ?
- Can a stock have a negative beta, and if so, what would that mean for its role in a portfolio? Are there real-world examples of negative-beta assets?
- The single-factor model assumes market returns capture all systematic risk. What are the limitations of this assumption, and how do multi-factor models like the Fama-French model address them?
- If the CAPM holds and markets are efficient, alpha should be zero for all stocks. Why do we still observe many stocks with statistically significant positive or negative alphas in practice?
- How do practitioners use beta in corporate finance applications like estimating a company’s cost of equity or making capital budgeting decisions?