Why Beta Selection Matters
The CAPM cost of equity depends directly on beta: holding the risk-free rate and market risk premium fixed, a higher beta produces a higher estimated required return. Yet beta is not a fixed physical characteristic that can be read without judgment. It is an estimate based on a particular market index, return frequency, and historical period, and the company’s business or financing risk may have changed since that period.
A useful beta selection therefore asks more than “What number is published?” It asks whether the estimate measures the relevant systematic risk, whether it is based on a suitable market proxy and period, and whether it represents the company as it exists now. The goal is a reasoned input, not false precision.
Here \(R_E\) is the company’s equity return and \(R_M\) is the return on the chosen market proxy, measured over matching periods. A beta of 1 indicates market sensitivity equal to that proxy’s; a beta above 1 indicates greater sensitivity, and a beta below 1 indicates less. Beta measures systematic risk in this model, not total return variability or the chance of a loss.
The CAPM relation used to translate beta into a required equity return is:
This relation assumes the CAPM’s framework, including investors pricing market-wide risk and being able to diversify away company-specific risk. In practice, the market return is represented by an index, and the inputs are estimates. The beta choice is meaningful only when its market proxy and interpretation are consistent with the risk-free rate and market risk premium used with it.
Three Sources of Beta
A beta estimate may come from a regression of the company’s historical returns against a market index, from a financial data provider, or from comparable companies. These are not necessarily independent methods: a published beta may itself be a regression estimate that has been adjusted or calculated with a different index and period. The source label alone does not establish that an estimate is appropriate.
| Source | Potential value | Questions to check |
|---|---|---|
| Historical regression estimate | Uses the company’s observed equity returns | What period, return frequency, and market index were used? Does the period reflect current operations? |
| Published beta | Provides a convenient estimate and may include a stated adjustment | Is it raw or adjusted? What index and estimation choices underlie it? |
| Peer-company estimate | Offers evidence from companies with relevant business exposures | Are operations, market exposure, and financing risk sufficiently comparable? |
A regression estimate is sensitive to the observation window. A longer period generally supplies more observations, but older observations may describe a business that has since changed. A shorter period can reflect recent conditions more closely, but may be more affected by temporary events or noisy returns. Neither “longer” nor “shorter” is automatically better.
The market index matters because beta is defined relative to that market proxy. Estimates based on different indexes need not match: the indexes contain different assets and can respond differently to economic news. A beta is not fully described by its numerical value alone; the underlying index and estimation convention are part of its meaning.
Adjustments and Current Risk
Some published estimates adjust a raw historical beta toward 1. A common empirical adjustment is:
This adjustment reflects a tendency for estimated betas to move toward the market average over time. It is a convention, not a CAPM identity or a guarantee about a particular company. Before using an adjusted figure, identify whether the purpose calls for the raw estimate or the provider’s adjusted estimate, and avoid applying an adjustment twice.
Current risk also requires judgment. A historical estimate may combine periods before and after a major acquisition, divestiture, regulatory change, or shift in product mix. Peer companies can help indicate whether the historical estimate remains plausible, but a peer estimate is not automatically interchangeable with the target’s beta. In particular, companies with materially different financing risk should not be treated as equivalent on the basis of their raw equity betas alone.
Why the CAPM Effect Is Linear
The beta term scales the market risk premium. This follows directly by subtracting the CAPM return at one beta from the return at another. If the risk-free rate is unchanged, then:
Thus a beta difference of 0.10 changes the estimated cost of equity by 0.10 times the market risk premium—not by 0.10 percentage points in every case. This is why small differences in beta can matter when the premium is substantial.
Worked Example: Reconciling Historical Estimates After a Business Change
A hypothetical manufacturer reports a 60-month regression beta of 0.82 and a 36-month beta of 1.08, both measured against the same broad-market index using monthly returns. The company completed an acquisition 24 months ago that increased its exposure to cyclical demand. A review of three operationally similar companies, with broadly similar financing risk, suggests a current beta near 1.02. For a CAPM illustration, assume a risk-free rate of 4.00% and a market risk premium of 5.50%. Choose a beta and estimate the associated cost of equity.
Method. Compare the estimates in light of the company’s changed risk, rather than simply choosing the longest or most recent period. The 60-month estimate includes substantial pre-acquisition history. The 36-month estimate is more recent but still includes a year before the acquisition and may reflect temporary return movements. The comparable-company estimate provides an additional current-risk check.
The three estimates do not prove that the current beta is exactly 1.02. However, the peer estimate is close to the recent company estimate and is based on companies described as operationally and financially similar. A defensible provisional choice is therefore \(\beta_E=1.02\), with the period limitations disclosed.
CAPM calculation. The selected beta times the market risk premium is \(1.02\times5.50\%=5.61\%\). Add the risk-free rate:
For comparison, using 0.82 would give \(4.00\%+0.82(5.50\%)=4.00\%+4.51\%=8.51\%\). Using 1.08 would give \(4.00\%+1.08(5.50\%)=4.00\%+5.94\%=9.94\%\). The selected result lies between those estimates. Check: the 0.26 difference between 1.08 and 0.82 times the 5.50% premium is 1.43 percentage points, and \(9.94\%-8.51\%=1.43\) percentage points.
Common mistake avoided: choosing 0.82 solely because it uses more observations would give older, pre-acquisition risk substantial weight without addressing whether it remains relevant. The selection instead weighs period coverage and current peer evidence; it does not pretend that the estimate is exact.
Worked Example: Interpreting a Published Adjustment
A hypothetical data provider reports a raw beta of 1.20 and states that its adjusted beta uses two-thirds of the raw estimate and one-third of 1. A second source reports 1.05, but uses a different market index and does not state whether its figure is adjusted. Assume the first provider’s index matches the market proxy chosen for the analysis. Use the stated adjusted beta with a risk-free rate of 3.80% and a market risk premium of 6.00%.
Method. First apply the provider’s stated adjustment once, then use that adjusted estimate in CAPM. Do not assume the second value is comparable simply because it is also called a beta: its index and adjustment status are unclear.
The adjusted estimate is:
Using the unrounded adjusted beta, its market-premium component is \(1.1333\ldots\times6.00\%=6.80\%\). Therefore:
Check: the adjusted beta is between the raw beta of 1.20 and 1.00, as a weighted average with positive weights summing to 1 must be. Common mistake avoided: using the raw 1.20 after deciding to use the published adjustment would produce \(3.80\%+1.20(6.00\%)=11.00\%\), an overstatement of 0.40 percentage points relative to the calculation using 1.1333.
Worked Example: Using Peer Betas as a Current-Risk Check
A hypothetical company has a historical beta of 0.78 from a period before it expanded into a more cyclical line of business. Three peer companies now have betas of 0.94, 1.00, and 1.12. They operate in similar markets and have broadly similar financing risk. A risk-free rate of 4.10% and market risk premium of 5.20% are assumed. Use the peer estimates to select a beta and calculate the illustrative CAPM return.
Method. Sort the peer estimates and use their median as a robust summary: \(0.94,\ 1.00,\ 1.12\). The middle observation is 1.00. The peer group is described as comparable, while the company’s 0.78 estimate predates the change in business mix. Select \(\beta_E=1.00\) as a current-risk proxy, while recognizing that the target’s risk may still differ from its peers.
The market-premium component is \(1.00\times5.20\%=5.20\%\). Thus:
Check: because the selected beta is 1.00, the CAPM return is exactly the 4.10% risk-free rate plus the 5.20% market risk premium. Common mistake avoided: treating the stale 0.78 estimate as definitive would ignore the stated change in business exposure. Conversely, the peer median is only a proxy; it should not be presented as a directly observed beta for the target.
A Practical Selection Process
A concise, defensible selection can be organized around the following checks:
Record the index underlying each candidate beta. Give greater weight to estimates whose market proxy matches the analysis.
Ask whether the observations are current and whether the period spans significant changes in the company. Note that a short window can be noisy and a long one can be stale.
Determine whether the estimate is raw or adjusted and whether the adjustment is documented. Do not mix conventions without explanation.
Consider changes in business exposure and whether peer operations and financing risk are comparable. Treat peer evidence as a reasonableness check, not automatic substitution.
State the chosen beta, why it best represents current systematic risk, and the important limitations. Use it consistently with the CAPM inputs.
Common Traps
- Comparing unlike estimates as though they were identical. Different indexes, periods, return frequencies, or adjustment conventions can produce different beta values.
- Assuming more data always means a better estimate. More historical observations can include business conditions that no longer describe the company.
- Treating beta as total risk. Beta captures sensitivity to the market proxy; company-specific risk is not measured by beta in the CAPM relation.
- Applying an adjustment twice. Check whether a provider has already adjusted its reported estimate before modifying it.
- Using a peer beta without checking comparability. Similar industry labels do not guarantee similar business exposures or financing risk.
- Confusing a beta change with a percentage-point change in cost. The cost difference equals the beta difference multiplied by the market risk premium.
Check Your Understanding
For each problem, make the selection or calculation requested and compare your result with the answer shown.
- A hypothetical retailer has candidate betas of 0.86 over 60 months and 1.04 over 24 months, both against the same index. It shifted toward online sales 18 months ago, and comparable firms suggest a beta of 1.00. With a risk-free rate of 4.20% and market risk premium of 5.00%, select 1.00 as a current-risk proxy and calculate the CAPM return. (Answer: 9.20%)
- A provider’s raw beta is 1.35, and its stated adjustment is two-thirds raw beta plus one-third of 1. With a 3.60% risk-free rate and a 5.40% market risk premium, calculate the adjusted beta and CAPM return. (Answer: adjusted beta \(1.2333\); cost of equity \(10.26\%\))
- Three suitably comparable peer betas are 0.88, 1.04, and 1.16. Use their median as the selected proxy. If the risk-free rate is 4.00% and the market risk premium is 5.50%, calculate the CAPM return. (Answer: beta \(1.04\); cost of equity \(9.72\%\))
- Using a 3.50% risk-free rate and 5.50% market risk premium, compare the CAPM returns for betas 0.95 and 1.00. State the difference in percentage points. (Answer: \(8.725\%\) and \(9.00\%\), respectively; the 0.95-beta estimate is 0.275 percentage points below the 1.00-beta estimate.)
- A company’s published beta is 0.92 based on a narrow-sector index and a three-year daily-return window. A second estimate is 1.03 based on the market proxy used in the valuation and a five-year monthly-return window. The company has recently broadened its operations. Name two checks needed before selecting either estimate. (Answer: verify whether each period reflects the current business and whether the index and adjustment conventions are appropriate and comparable; the facts alone do not determine a uniquely correct beta.)