Fama-MacBeth Risk Premium Calculator
Estimate factor risk premia from cross-sectional portfolio data.
Calculated Risk Premia
What is the Fama-MacBeth Regression?
The Fama-MacBeth regression is a two-step procedure used in financial economics to test asset pricing models and estimate the risk premium associated with various risk factors. Developed by Eugene Fama and James MacBeth, this method is a cornerstone of empirical finance. Its primary goal is to determine whether exposure to a systematic risk factor (like the market, company size, or value) is rewarded with higher average returns. A key objective is to calculate a different risk premium for each portfolio using Fama-MacBeth, by understanding how different factor exposures contribute to returns.
The process involves two main stages:
- Time-Series Regression: For each individual asset or portfolio, a regression is run over a period of time. The portfolio’s excess returns (returns above the risk-free rate) are regressed against the proposed risk factors. This step estimates the ‘beta’ or ‘factor loading’ for each portfolio on each risk factor, which measures the portfolio’s sensitivity to that factor.
- Cross-Sectional Regression: For each time period (e.g., each month), a single regression is run across all the portfolios. In this step, the excess returns of all portfolios for that month are regressed against the betas estimated in the first step. The resulting coefficient for each beta represents the risk premium for that factor in that specific month.
Finally, the method involves averaging the monthly risk premia calculated in the second step to get an overall estimate of the factor’s risk premium. The statistical significance of this average premium is then tested to see if the factor is genuinely priced by the market.
The Fama-MacBeth Formula and Explanation
The methodology can be broken down into two core regression equations. The process aims to ultimately calculate a different risk premium for each portfolio using Fama-MacBeth principles by identifying the price of each risk factor.
Step 1: Time-Series Regression
For each portfolio i, run the following regression over time t=1 to T:
Ri,t - Rf,t = αi + βi,F1(F1,t) + βi,F2(F2,t) + ... + εi,t
This provides the beta estimates (βi,F1, βi,F2, etc.) for each portfolio.
Step 2: Cross-Sectional Regression
For each time period t, run the following regression across all portfolios i=1 to N:
Ri,t - Rf,t = γ0,t + γ1,t(βi,F1) + γ2,t(βi,F2) + ... + ui,t
The coefficients (γ1,t, γ2,t) are the risk premia for each factor in period t.
Final Step: Averaging
The final risk premium (λ) for a factor is the time-series average of the gamma coefficients: λF1 = (1/T) * Σγ1,t
| Variable | Meaning | Unit / Type | Typical Range |
|---|---|---|---|
Ri,t |
Return of portfolio i at time t. | Percentage (%) | -20% to +20% (monthly) |
Rf,t |
Risk-free rate at time t. | Percentage (%) | 0% to 5% (annualized) |
βi,F1 |
Beta (sensitivity) of portfolio i to Factor 1. | Unitless ratio | -1.0 to 3.0 |
γ1,t |
Cross-sectional coefficient, or risk premium for Factor 1 at time t. | Percentage (%) | -5% to +5% (monthly) |
λF1 |
The average risk premium for Factor 1. | Percentage (%) | -1% to +2% (monthly) |
Practical Examples
To properly calculate a different risk premium for each portfolio using Fama-MacBeth, you need to understand how inputs translate to outputs. Let’s walk through a simplified, one-period example, which is what our calculator demonstrates.
Example 1: Positive Market Premium
- Inputs:
- Portfolio A: Excess Return = 1.0%, Market Beta = 1.2
- Portfolio B: Excess Return = 0.6%, Market Beta = 0.9
- Portfolio C: Excess Return = 0.3%, Market Beta = 0.7
- Logic: We regress the returns [1.0, 0.6, 0.3] on the betas [1.2, 0.9, 0.7]. The slope of this regression line is the market risk premium.
- Results: The calculator would find a positive market risk premium (lambda), indicating that higher beta was associated with higher returns in this period. The intercept (alpha) would be close to zero if the model explains returns well.
Example 2: Multi-Factor Model
- Inputs:
- Portfolio X: Excess Return = 1.2%, Mkt Beta = 1.1, Size Beta = -0.5 (Large Cap)
- Portfolio Y: Excess Return = 1.5%, Mkt Beta = 1.0, Size Beta = 0.8 (Small Cap)
- Logic: A multiple regression is performed. The calculator estimates the premium for market risk and size risk simultaneously.
- Results: The calculator might find a market premium of 0.5% and a size premium of 0.4%. This implies that for every unit of market beta, return increased by 0.5%, and for every unit of “small-cap” beta, it increased by another 0.4%.
How to Use This Fama-MacBeth Calculator
This calculator performs the cross-sectional (second stage) part of a Fama-MacBeth analysis for a single period. It helps you understand how risk premia are derived from portfolio betas and returns.
- Enter Portfolio Data: For each portfolio, input its average excess return (return minus the risk-free rate) and its factor betas. The betas are the sensitivities of the portfolio to the risk factors, which you would typically get from a time-series regression (Step 1 of the Fama-MacBeth process).
- Add Portfolios: You must have at least as many portfolios as you have factors plus one for the intercept. Use the “Add Portfolio” button to add more data points for a more robust regression.
- Review the Results: The calculator automatically performs a multiple linear regression. The primary results are the “Risk Premia” (gamma coefficients) for each factor. This tells you the estimated return premium per unit of beta for that factor.
- Interpret the Chart: The chart plots the actual excess returns you entered against the returns predicted by the model using your betas and the calculated risk premia. Points close to the 45-degree line indicate the model is a good fit for the data provided.
Key Factors That Affect Fama-MacBeth Results
The results of a Fama-MacBeth regression are sensitive to several methodological choices.
- Choice of Factors: The most important choice. Common factors include the market (Mkt-Rf), size (SMB), value (HML), momentum (UMD), and profitability (RMW). Including irrelevant factors can add noise, while omitting relevant ones can bias the results.
- Portfolio Construction: The way portfolios are formed can impact the results. Typically, assets are sorted into portfolios based on characteristics like size or book-to-market ratio to create a wide dispersion in betas, which gives the regression more statistical power.
- Time Period: The length of the time series and the specific historical period chosen can lead to different premium estimates. Some periods are more volatile or have different economic regimes.
- Data Frequency: Using daily, weekly, or monthly returns can affect beta estimates and the final risk premia. Monthly data is most common in academic studies to reduce noise.
- Errors-in-Variables: The betas used in the second stage are *estimates* from the first stage, not the true values. This introduces an “errors-in-variables” problem, which can bias the results. Using portfolios instead of individual stocks helps mitigate this, as portfolio betas are more stable.
- Number of Portfolios: Having a larger cross-section of portfolios (e.g., 25 or 100) generally leads to more precise estimates of the risk premia in the second-stage regression.
Frequently Asked Questions
1. What is a “risk premium”?
A risk premium is the excess return that an investment is expected to yield above the risk-free rate to compensate an investor for taking on a specific type of risk.
2. Why use a two-step approach instead of one big regression?
The two-step approach was designed to handle the fact that asset returns are correlated with each other at any given point in time. By running a cross-sectional regression for each time period, Fama and MacBeth could use the standard deviation of the resulting premium time-series to get a more robust standard error for the average premium.
3. What does it mean if a risk premium is not statistically significant?
It suggests that the market does not consistently reward investors for taking on that specific factor risk. The observed premium could be due to random chance rather than a systematic relationship.
4. Why does this calculator only do the second step?
A full Fama-MacBeth regression requires extensive time-series data (e.g., monthly returns for many portfolios over many years) and is computationally intensive. This calculator demonstrates the core cross-sectional logic of how betas and returns are used to derive the premium for a single period, which is the conceptual heart of the second stage.
5. What are SMB and HML?
They are two factors from the Fama-French three-factor model. SMB stands for “Small Minus Big” and represents the size risk factor. HML stands for “High Minus Low” and represents the value risk factor (based on book-to-market ratios).
6. What is the intercept (Alpha) in the results?
The intercept, often called alpha, is the portion of the return that is not explained by the risk factors in the model. In a well-specified asset pricing model, the alpha should be statistically indistinguishable from zero.
7. Can I use this calculator for my investment decisions?
This tool is for educational purposes to illustrate a financial-economic model. It is not investment advice. Real-world financial analysis requires more comprehensive data and sophisticated tools.
8. How many portfolios do I need to use the calculator?
You need at least one more portfolio than the number of factors you are testing. Since this calculator uses two factors plus an intercept, you need a minimum of three portfolios to get a result.