Why Comparable-Firm Selection Matters
A bottom-up beta estimate starts with the question, which firms have operating businesses that resemble the business being valued? A firm’s historical beta reflects how its returns have moved with a market benchmark over a particular period. A firm with similar products but very different customers, revenue drivers, or markets may therefore be a poor guide. Conversely, two firms with different financing can still provide useful evidence about the target’s operating exposures.
The aim is not to find a company with the same name, size, or share-price pattern. It is to identify firms whose business activities are exposed to similar economic forces, then use only evidence that is relevant and sufficiently reliable. Comparable selection involves judgment; no single numerical screen establishes that two businesses have the same beta.
What Makes Operating Businesses Comparable?
Begin with what each firm sells and how it earns revenue. A useful comparison goes beyond broad industry labels. Two firms classified as “technology” could have very different businesses: one might sell recurring cybersecurity subscriptions to large corporations, while another earns most of its revenue from advertising. Their customers, revenue stability, and sensitivity to economic activity may differ substantially.
For each candidate, investigate the following features in a consistent way:
- Products and services: Identify the activities that generate revenue, including material secondary businesses. Similar labels do not guarantee similar products or economics.
- Customers and end markets: Compare customer types, such as households, small businesses, or governments, and the industries or needs that drive their purchases.
- Revenue drivers: Consider whether sales depend on recurring contracts, transaction volumes, commodity prices, project awards, or discretionary spending.
- Geography and regulation: Compare markets served, currency and economic exposures, and relevant regulatory environments. A major difference may matter even when product descriptions match.
- Scale and maturity: Consider whether the firm is an established operator or a rapidly expanding entrant, and whether its operating scope resembles the target’s. Scale is a context for judgment, not a substitute for activity-level comparison.
- Business changes: Check whether acquisitions, divestitures, or a recent strategic shift make historical returns unrepresentative of the current business.
A useful starting point is to map both the target and candidates into the same set of operating categories. Revenue shares can reveal a material difference that an industry label conceals. But a similar revenue mix is only one piece of evidence: firms in the same category may still face different competitive, geographic, or regulatory conditions.
A Transparent Revenue-Mix Screen
When the relevant categories can be defined consistently, a simple comparison can help organize the evidence. Let \(w_{T,j}\) be the target’s share of revenue in category \(j\), and let \(w_{C,j}\) be a candidate’s share. Each set of shares totals 1. Define the half-sum of absolute mix differences as:
This is a descriptive screening measure, not a standard beta estimator or a universal acceptance cutoff. The division by 2 has a straightforward justification. Because both firms’ shares sum to 1, a category whose share is lower at the candidate must be offset by one or more categories whose shares are higher. Summing absolute differences counts each share reallocation twice: once where revenue share is lower and once where it is higher. Dividing by 2 expresses the total share reallocated. A smaller value indicates a closer match by the chosen categories, but does not prove that the firms’ economic exposures are alike.
The comparison is meaningful only if the categories are consistently defined and measured over comparable periods. A candidate with sparse disclosures may not support a reliable mix calculation. Likewise, a low mix difference should prompt further checks, not automatic inclusion.
Worked Example: Comparing Subscription Software Businesses
A hypothetical target sells payroll and related software. Its latest-year revenue mix is 70% small-business payroll subscriptions, 20% human-resources software, and 10% payment services. Three candidates report these same categories: Firm A has a mix of 65%, 25%, and 10%; Firm B has 70%, 10%, and 20%; Firm C has 35%, 45%, and 20%. Treat the reported mixes as comparable for this preliminary screen. Which candidate is closest by revenue mix, and what else should be checked?
Method. Calculate \(D(T,C)\) for each candidate using the half-sum of absolute differences. All shares are in percentage points, and each firm’s three shares sum to 100%.
Firm A: The absolute differences are \(|70-65|=5\), \(|20-25|=5\), and \(|10-10|=0\) percentage points. Their sum is \(5+5+0=10\), so \(D(T,A)=10/2=5\) percentage points.
Firm B: The absolute differences are \(|70-70|=0\), \(|20-10|=10\), and \(|10-20|=10\). Their sum is \(0+10+10=20\), so \(D(T,B)=20/2=10\) percentage points.
Firm C: The absolute differences are \(|70-35|=35\), \(|20-45|=25\), and \(|10-20|=10\). Their sum is \(35+25+10=70\), so \(D(T,C)=70/2=35\) percentage points.
Firm A is closest by this screen: its difference is 5 percentage points, compared with 10 for Firm B and 35 for Firm C. Before including Firm A, check whether it serves similar customer sizes, sells through similar channels, and operates in comparable markets. The mix calculation does not test those features.
Check: Each candidate’s mix sums to 100%, and the half-sum results are 5, 10, and 35 percentage points. The common mistake is to interpret the smallest mix difference as proof of equal beta; it is evidence about one dimension of business similarity only.
Financing Is Not a Substitute for Business Similarity
A company’s observed equity beta is an equity-return measure. Its historical estimate can reflect both the business it operates and the financing through which equity holders bear risk. Consequently, financing information matters in a bottom-up analysis, but it answers a different question from whether the firm’s operating activities resemble the target’s.
If the target and a candidate have similar debt-to-equity ratios but sell different products to different customers, the financing match does not make their operating exposures comparable. If their businesses are similar but their financing differs, the candidate may still provide useful information about the operating business; financing can be considered separately in the appropriate stage of the analysis. Do not select a firm solely because its financing looks familiar, and do not discard a strong business match solely because its financing differs.
Worked Example: Rejecting a Financing-Only Match
A hypothetical regional discount grocer has three main revenue categories: food and household staples, discretionary merchandise, and fuel. Its revenue mix is 75%, 15%, and 10%, and its debt-to-equity ratio is 0.40. Candidate P, a department-store retailer, has a mix of 40%, 50%, and 10%, and a debt-to-equity ratio of 0.42. Candidate Q, another grocery operator, has a mix of 72%, 18%, and 10%, and a debt-to-equity ratio of 1.10. Assume the categories are consistently defined. Which candidate is the better operating comparable?
Method. Compare operating mixes using \(D(T,C)\), then use the financing figures only to identify the tempting but insufficient match.
Candidate P: The absolute mix differences are \(|75-40|=35\), \(|15-50|=35\), and \(|10-10|=0\) percentage points. Thus the sum is \(35+35+0=70\), and \(D(T,P)=70/2=35\) percentage points. Its debt-to-equity ratio is close to the target’s, but its business mix is substantially different.
Candidate Q: The absolute mix differences are \(|75-72|=3\), \(|15-18|=3\), and \(|10-10|=0\). Thus the sum is \(3+3+0=6\), and \(D(T,Q)=6/2=3\) percentage points. Its financing differs, but its revenue mix is much closer to the target’s.
Candidate Q is the stronger preliminary operating comparable. This conclusion does not assert that Q’s equity beta can be used unchanged: the financing difference must be addressed separately when the analysis calls for it. It means only that financing similarity does not outweigh the evident difference in business activity.
Check: Candidate P’s mix difference is 35 percentage points, and Candidate Q’s is 3. The common mistake is to choose P because 0.42 is nearer to 0.40 than 1.10 is. That comparison answers a financing question, not the operating-comparability question.
How to Screen and Document Candidates
A disciplined screen makes the selection reviewable. Use the same definitions for the target and each candidate, record the evidence supporting the comparison, and distinguish a clear mismatch from a lack of information. A missing disclosure is not evidence that the candidate matches.
List its material products or services, customer groups, revenue drivers, and geographic or regulatory exposures. Identify which activities are central rather than incidental.
Search for firms with overlapping products, customers, and economic drivers. Treat industry classifications as a starting point, not a final screen.
Review segment descriptions, revenue shares, markets served, and material business changes. Use the same categories and periods wherever possible.
Check whether the candidate’s business changed materially during the period relevant to its return history and whether its disclosures support the comparison. A company with the right current business but a different historical business may provide misleading historical evidence.
State why each candidate is included, excluded, or retained with a qualification. Keep operating similarity distinct from financing considerations.
A useful comparison table can make the reasoning concise:
| Screening question | Evidence to compare | Reason it matters |
|---|---|---|
| What does the firm sell? | Products, services, and material segments | Broad industry labels may hide different business activities |
| Who buys it? | Customer types and end markets | Different buyers can expose revenue to different economic conditions |
| What drives revenue? | Contracts, transactions, projects, or other drivers | Revenue sources can respond differently to changes in economic activity |
| Where and under what rules? | Geography and relevant regulation | Markets and constraints can differ even when products match |
| Has the business changed? | Acquisitions, divestitures, and strategic shifts | Historical returns may not represent the current operations |
Why the Beta Estimate Still Needs Judgment
For returns measured over a common period against the same market benchmark \(R_m\), beta is defined when \(\operatorname{Var}(R_m)>0\) by:
This definition explains why operating similarity is relevant: the goal is to find firms whose businesses face comparable forces that may affect their returns’ co-movement with the market. It does not say that similar products guarantee equal covariance, or that revenue-mix similarity mechanically determines beta. The observed estimate also depends on the return period and market benchmark. The earlier material on beta estimation and structural breaks is relevant when reviewing those data limitations.
Common traps include:
- Matching only the industry label: Inspect the activities and customers beneath the label.
- Matching only financing: Financing similarity does not repair a poor operating match.
- Treating a screen as a rule: A mix-distance measure can organize evidence but cannot prove equal market exposure.
- Ignoring business history: A major change in operations can make past returns a poor guide to the current firm.
- Using inconsistent categories: A comparison is unreliable if one firm’s “services” category means something materially different from another’s.
Check Your Understanding
For each problem, use revenue-mix distance only as a preliminary screen. Treat all figures as hypothetical and assume categories are consistently defined unless stated otherwise.
- A target apparel retailer has revenue shares of 50% essential clothing, 30% footwear, and 20% accessories. Candidate A has shares of 45%, 35%, and 20%; Candidate B has 70%, 15%, and 15%. Calculate each half-sum distance and identify the closer mix. (Answer: A has a 5-percentage-point distance; B has a 20-percentage-point distance; A is closer.)
- A target software firm has shares of 60% small-business subscriptions, 25% enterprise subscriptions, and 15% transaction fees. Candidate A has 55%, 30%, and 15%; Candidate B has 60%, 15%, and 25%. Calculate both distances and state which is closer by mix. (Answer: A is 5 percentage points; B is 10 percentage points; A is closer.)
- A target equipment maker earns 50% of revenue from utility customers, 30% from industrial customers, and 20% from residential customers. Candidate R has shares of 45%, 35%, and 20%, but sells in a similar geography and regulatory environment. Candidate S has 50%, 30%, and 20%, but serves markets with materially different regulation. Which is the exact mix match, and why is it not automatically the better overall comparable? (Answer: S is the exact mix match, but its regulatory mismatch requires further assessment; the mix alone does not establish overall comparability.)
- A farm-supply retailer has a 70% agricultural-input, 20% household-goods, and 10% services mix, with debt-to-equity of 0.50. Candidate P has a 40%, 50%, and 10% mix and debt-to-equity of 0.52. Candidate Q has a 68%, 22%, and 10% mix and debt-to-equity of 1.20. Compute the mix distances and identify the stronger preliminary operating match. (Answer: P is 30 percentage points; Q is 2 percentage points; Q is the stronger preliminary operating match despite its different financing.)
- A target earns 80% from recurring maintenance contracts and 20% from equipment sales. Candidate A reports the same shares but recently shifted into a different customer market; Candidate B reports 75% and 25% and serves similar customers and geography. What evidence should be reviewed before selecting a comparable? (Answer: Review the nature and timing of A’s market shift, customer and geographic exposures, and whether its historical business represents current operations; compare those facts with B’s modest mix difference. Shares alone do not settle the choice.)