Selecting Peer Groups for Multiples-Based Equity Valuation

The credibility of any relative valuation analysis hinges on one critical question: “Compared to whom?” The choice of comparator companies-the peer group-is as consequential as the selection of the valuation multiple itself. While the mathematics of calculating a P/E or EV/EBITDA ratio is straightforward, the judgment required to identify which firms belong in the comparator set is where the true art and science of relative valuation converge.

As established in previous sections, valuation multiples are driven by fundamental company characteristics: expected growth rates, risk profiles (reflected in required returns), profitability metrics such as ROE and profit margins, and reinvestment requirements. The ideal peer group consists of companies that mirror the target firm across these fundamental dimensions. However, because no two companies are identical, analysts must make reasoned judgments about which differences are material enough to warrant exclusion or adjustment.

The process of peer group construction typically follows one of two philosophical approaches:

1. Industry-Based Peer Groups: Selecting companies that share the same principal business activities, thereby facing similar competitive dynamics, customer bases, and supply chains
2. Factor-Based Peer Groups: Selecting companies that exhibit similar statistical characteristics-such as beta, size, or value/growth orientation-regardless of their industry classification

This section provides a comprehensive framework for constructing appropriate peer groups, addressing the practical challenges that arise, and offering guidance on when to adjust, exclude, or weight comparator firms.

The Industry-Based Approach to Peer Selection

Companies that produce similar goods or services, serve comparable customer segments, and face analogous competitive pressures are said to operate within the same industry. The logic underpinning industry-based peer groups is straightforward: firms confronting similar economic forces should exhibit similar cash flow dynamics, risk characteristics, and growth prospects. Consequently, their valuation multiples should converge if markets are pricing them efficiently.

Practical Implementation

When constructing an industry-based peer group, analysts typically follow a multi-stage screening process:

1. Stage 1: Classification System IdentificationAnalysts begin with established industry classification frameworks such as:
   
   
   
   • Global Industry Classification Standard (GICS)
   
   
   • Industry Classification Benchmark (ICB)
   
   
   • Refinitiv Business Classification (TRBC)
   
   
   
   

   These systems provide hierarchical structures-from broad sectors down to highly specific sub-industries-that serve as initial filters.

2. Stage 2: Geographic and Scale CalibrationEven within the same industry classification, companies may differ significantly in size and geographic reach. A global conglomerate competing across 30 countries is not directly comparable to a regional operator serving a single domestic market. Analysts typically apply filters based on:
   • Market Capitalization: Often excluding firms below a certain threshold
   • Revenue Scale: Ensuring similar operational size
   • Geographic Revenue Distribution: Comparing exposure to similar regulatory, economic, and competitive environments
3. Stage 3: Business Mix RefinementCompanies classified in the same industry often have different segment compositions. For example, within the utility sector, some firms focus primarily on power generation, others on transmission and distribution, and others on trading activities. Furthermore, within power generation, the mix of traditional (coal, gas) versus renewable (wind, solar) energy sources significantly affects risk, regulatory exposure, and growth prospects.

Illustrative Example: Constructing a Peer Group for a European Energy Company

Nordic Energy Group operates electric and natural gas utility businesses across 30 countries, with its revenue generated primarily in three European nations. An analyst seeking to establish a peer group for relative valuation might begin with an initial screen of European utilities, yielding the following data:

$$
\begin{array}{l|c|c|c|c}
\textbf{Company} & \textbf{Market Cap (EUR M)} & \textbf{Enterprise Value (EUR M)} & \textbf{EV/EBITDA} & \textbf{P/E} \\
\hline
\text{Nordic Energy Group} & 63,033 & 144,005 & 6.60 & 16.41 \\\hline
\text{Central European Power} & 18,311 & 33,237 & 5.13 & 11.65 \\\hline
\text{Baltic Energy Corp} & 11,090 & 12,487 & 6.38 & 7.23 \\\hline
\text{Mediterranean Utilities} & 5,869 & 11,120 & 6.08 & 8.95 \\\hline
\text{Iberian Power Group} & 74,864 & 139,222 & 8.93 & 15.00 \\\hline
\text{Franco-Belgian Energy} & 39,573 & 73,779 & 5.58 & 17.65 \\\hline
\text{Alpine Energy AG} & 24,632 & 27,475 & 6.14 & 10.87 \\\hline
\text{Hellenic Energy SA} & 5,438 & 7,402 & 7.24 & 8.45 \\\hline
\text{Northern Power plc} & 21,471 & 33,753 & 184.27 & 31.40 \\\hline
\text{Swiss Energy Holdings} & 7,406 & 8,896 & 9.01 & 15.11 \\\hline
\text{Danube Utility Group} & 5,163 & 6,591 & 8.56 & 9.76 \\
\end{array}
$$

Initial Observations

Nordic Energy’s EV/EBITDA multiple (6.60) aligns precisely with the peer median of 6.60, suggesting a fair valuation on an enterprise basis. However, its trailing P/E of 16.41 substantially exceeds the peer median of 11.65, indicating potential overvaluation on an equity basis.

Refining the Peer Set

Observing that Nordic Energy is considerably larger and more geographically diversified than many firms in the initial screen, the analyst excludes firms with market capitalization below EUR 10 billion:

$$
\begin{array}{l|c|c|c|c}
\textbf{Company} & \textbf{Market Cap (EUR M)} & \textbf{Enterprise Value (EUR M)} & \textbf{EV/EBITDA} & \textbf{P/E} \\
\hline
\text{Nordic Energy Group} & 63,033 & 144,005 & 6.60 & 16.41 \\ \hline
\text{Central European Power} & 18,311 & 33,237 & 5.13 & 11.65 \\ \hline
\text{Baltic Energy Corp} & 11,090 & 12,487 & 6.38 & 7.23 \\ \hline
\text{Iberian Power Group} & 74,864 & 139,222 & 8.93 & 15.00 \\ \hline
\text{Franco-Belgian Energy} & 39,573 & 73,779 & 5.58 & 17.65 \\ \hline
\text{Alpine Energy AG} & 24,632 & 27,475 & 6.14 & 10.87 \\ \hline
\text{Northern Power plc} & 21,471 & 33,753 & 184.27 & 31.40 \\ \hline
\textbf{Revised Median:} & \textbf{24,632} & \textbf{33,753} & \textbf{6.38} & \textbf{15.00} \\
\end{array}
$$

Revised Median: | 24,632 | 33,753 | 6.38 | 15.00

Interpretation

The perceived P/E overvaluation has nearly disappeared (16.41 versus the revised median of 15.00). However, Northern Power plc’s extraordinary multiples (EV/EBITDA of 184.27 and P/E of 31.40) represent a significant outlier. Further investigation reveals Northern Power suffered an extraordinary loss in the prior year, depressing EBITDA and earnings to unusually low levels. The analyst should consider:

1. Excluding Northern Power from the peer set
2. Using the median (which is less sensitive to outliers) rather than the mean
3. Adjusting the income measures to exclude the non-recurring loss

Challenges in Industry-Based Peer Group Construction

Challenge 1: Unique or Concentrated Industries
Some industries contain very few public companies, making the construction of a meaningful peer group difficult. In such cases, analysts may place less emphasis on relative valuation and rely more heavily on absolute valuation methods such as discounted cash flow analysis.

Challenge 2: Life-Cycle Stage Heterogeneity
Industries such as biotechnology, pharmaceuticals, and technology contain firms at vastly different stages of development-from early-stage, revenue-less companies to mature, cash-generating giants. Comparing across life-cycle stages introduces significant distortions. Analysts may choose to exclude firms at materially different stages, though this must be balanced against creating an overly small peer group.

Challenge 3: Cross-Border Comparability
Companies in the same industry but operating in different jurisdictions face varying regulatory frameworks, tax regimes, labor laws, and competitive intensities. These differences affect profitability, growth, and risk-and thus the appropriate multiples. Adjusting for these differences requires careful consideration of each jurisdiction’s specific characteristics.

Challenge 4: Multi-Industry Operations
Many modern corporations operate across multiple industries. Comparing such companies to single-industry peers overlooks the distinct risk, growth, and profitability characteristics of each business segment. The following section addresses this challenge in detail.

Handling Multi-Industry Companies

When a target company operates in multiple industries, the analyst must develop a methodology that accounts for the distinct characteristics of each business segment.

The Weighted Average Approach

The most common approach is to calculate a composite peer group multiple as the weighted average of segment-specific multiples:

$$
\begin{array}{c}
Composite\ Multiple = \sum_{j=1}^{n} w_j \times M_j
\end{array}
$$

Where:

• \(w_j\) = The proportion of the company’s operations (typically measured by revenue, operating income, or assets) in industry $j$
• \(M_j\) = The representative peer group multiple for industry $j$
• \(n\) = The number of industries in which the company operates

Segment-by-Segment Valuation

An alternative approach involves valuing each business segment separately using its own peer group multiple, then aggregating the segment values to arrive at total firm value:

$$
\begin{array}{c}
Firm\ Value = \sum_{j=1}^{n} (EBITDA_j \times M_j) – Net\ Debt
\end{array}
$$

This approach is particularly useful when the company’s segments have significantly different growth prospects, risk profiles, or capital intensity.

Illustrative Example: Valuing a Diversified Healthcare Company

OmniHealth Solutions is a diversified healthcare company with two distinct operating segments:

$$
\begin{array}{l|c|c}
\textbf{Segment} & \textbf{Proportion of Sales} & \textbf{Peer Median EV/EBITDA} \\
\hline
\text{Generic Pharmaceuticals} & 65\% & 10.5x \\ \hline
\text{Patented Specialty Pharmaceuticals} & 35\% & 18.3x \\
\end{array}
$$

Step 1: Calculate the Composite Multiple

$$
\begin{array}{c}
\text{Composite EV/EBITDA} = (0.65 \times 10.5) + (0.35 \times 18.3) = 6.825 + 6.405 = 13.23x
\end{array}
$$

Step 2: Apply the Multiple

If OmniHealth’s consolidated EBITDA is EUR 2.4 billion:

$$
\begin{array}{c}
Enterprise\ Value = 2.4 \times 13.23 = EUR 31.75\ billion
\end{array}
$$

Step 3: Segment-by-Segment Validation (Alternative Approach)

$$
\begin{array}{l|c|c|c}
\textbf{Segment} & \textbf{EBITDA (EUR M)} & \textbf{Multiple} & \textbf{Segment Value (EUR M)} \\
\hline
\text{Generic Pharmaceuticals} & 1,560 & 10.5x & 16,380 \\ \hline
\text{Patented Pharmaceuticals} & 840 & 18.3x & 15,372 \\ \hline
\textbf{Total Enterprise Value} & \textbf{2,400} & & \textbf{31,752} \\
\end{array}
$$

Both approaches yield consistent results, providing confidence in the valuation conclusion.

The OmniRetail Dilemma: A Practical Case

OmniRetail Corporation illustrates the challenge of classifying multi-industry companies. While the company established its reputation in e-commerce retail, a substantial portion of its value derives from its high-margin cloud computing division:

$$
\begin{array}{l|c|l}
\textbf{Segment} & \textbf{Revenue Proportion} & \textbf{Operating Profit Margin} \\
\hline
\text{North America Retail} & \sim 60\% & 0-5\% \text{ (sometimes negative)} \\ \hline
\text{International Retail} & \sim 25\% & \text{Often negative} \\ \hline
\text{Cloud Computing Services} & \sim 16\% & 25-30\% \\
\end{array}
$$

Despite representing only 16% of revenue, the cloud computing division typically accounts for well over half of OmniRetail’s total operating profit. An analyst comparing OmniRetail to traditional retailers would materially undervalue the company, while comparing it solely to technology companies would overstate its retail operations.

Recommended Approaches:

1. Weighted Average: Apply a weighted average of retail and technology peer group multiples based on segment contribution to operating profit (not just revenue)
2. Sum-of-the-Parts: Value each segment separately using distinct peer groups, then sum the segment values

Factor-Based Peer Groups

Conceptual Foundation

Factor-based peer grouping relies on the premise that securities with similar statistical characteristics-particularly those explaining risk and expected return-should be considered comparables regardless of their industry classification. The underlying logic is that if markets are efficient, factor exposures reflect compensation for bearing systematic risk, making factor-based groups appropriate benchmarks for relative value comparisons.

The Capital Asset Pricing Model Framework

The most widely recognized single-factor model is the Capital Asset Pricing Model (CAPM), which establishes expected returns as a function of market beta:

$$
\begin{array}{c}
E(R_i) = R_f + \beta_i (E(R_m) – R_f)
\end{array}
$$

Where:

• \(E(R_i)\) = Expected return on the target company
• \(R_f\) = Risk-free rate
• \(E(R_m)\) = Expected market return
• \(\beta_i\) = Beta coefficient

The beta coefficient measures the sensitivity of a security’s returns to market movements:

$$
\begin{array}{c}
\beta_i = \frac{Cov(R_i, R_m)}{Var(R_m)}
\end{array}
$$

Securities with \(\beta > 1\) exhibit higher expected returns than the market (and higher risk), while securities with \(\beta < 1\) exhibit lower expected returns.

Advantages of Factor-Based Peer Groups

1. Direct Risk Alignment: Establishes a direct relationship between grouping criteria and expected returns
2. Overcomes Industry Limitations: Enables comparison across industries when valid peers within the same industry are limited
3. Consistent Cost of Equity: Firms with similar betas should have similar costs of equity
4. Dynamic Classification: Can be updated as company characteristics evolve

Disadvantages of Factor-Based Peer Groups

1. Model Misspecification Risk: Factors that explain returns in one period may not be predictive in another
2. Factor Instability: Betas and other factor loadings change over time
3. Cyclical Factor Performance: Many factors underperform for extended periods
4. Multiple Factor Complexity: Companies may load differently on multiple factors simultaneously, complicating classification

Illustrative Example: High-Beta Screening

A factor-based investment strategy might screen for companies with the highest beta coefficients, regardless of industry:

$$
\begin{array}{l|l|c|c}
\textbf{Company} & \textbf{Sector} & \textbf{Beta} & \textbf{Forward P/E} \\
\hline
\text{TechNova Systems} & \text{Information Technology} & 1.85 & 36.78 \\ \hline
\text{Atlantic Financial} & \text{Financials} & 1.76 & 10.99 \\ \hline
\text{Energy Solutions Corp} & \text{Industrials} & 1.71 & 23.73 \\ \hline
\text{MedTech Innovations} & \text{Healthcare} & 1.68 & 16.50 \\ \hline
\text{DataStream Analytics} & \text{Information Technology} & 1.62 & 55.46 \\ \hline
\text{TravelVentures Inc} & \text{Consumer Discretionary} & 1.59 & 11.53 \\ \hline
\text{MetroBank Financial} & \text{Financials} & 1.58 & 14.79 \\ \hline
\text{GreenEnergy Partners} & \text{Utilities} & 1.52 & 12.42 \\
\end{array}
$$

Despite representing diverse industries-from technology to financials to healthcare-these companies share a common characteristic of high systematic risk relative to the broad market. An investor employing a factor-based approach would consider them a peer group for relative valuation comparisons, using beta as the primary grouping criterion.

Comparative Framework: Industry-Based vs. Factor-Based Approaches

$$
\begin{array}{l|l|l}
\textbf{Dimension} & \textbf{Industry-Based} & \textbf{Factor-Based} \\
\hline
\text{Primary Grouping} & \text{Same business activity} & \text{Similar statistical characteristics} \\ \hline
\text{Criterion} & & \\
\hline
\text{Underlying Logic} & \text{Similar competitive forces produce} & \text{Similar risk exposures should earn} \\
& \text{similar fundamentals} & \text{similar returns} \\
\hline
\text{Ideal Application} & \text{Companies with clear,} & \text{Companies in concentrated industries} \\
& \text{single-industry focus} & \text{or with blurry industry boundaries} \\
\hline
\text{Key Challenge} & \text{Multi-industry operations,} & \text{Factor stability, model misspecification} \\
& \text{cross-jurisdictional differences} & \\
\hline
\text{Examples} & \text{Pharmaceuticals, banks, retail} & \text{High-beta portfolios, value portfolios,} \\
& & \text{small-cap portfolios} \\
\end{array} $$

Practical Peer Group Construction Methodology

1. Step 1: Establish Initial Industry Universe
   
   Using a recognized classification system (GICS, ICB, or TRBC), identify all companies in the same primary industry as the target firm.
2. Step 2: Apply Fundamental Characteristic Filters
   
   Refine the peer set using filters based on key fundamental drivers:
   • Market Capitalization: Exclude firms that are materially smaller or larger
   • Revenue Scale: Ensure similar operational size
   • Growth Rates: Match historical and expected growth patterns
   • Profitability Levels: Consider ROE, ROIC, and profit margins
   • Financial Leverage: Compare debt-to-equity and debt-to-EBITDA ratios
   • Geographic Exposure: Match regional revenue distributions
   • Business Segment Mix: Align with the target’s segment composition
3. Step 3: Identify and Treat Outliers
   
   Review the remaining peer set for potential outliers:
   • Companies with negative earnings (P/E not meaningful)
   • Firms with non-recurring items distorting financial metrics
   • Companies at significantly different life-cycle stages
   • Firms with extraordinary capital structures or events
4. Step 4: Conduct Sensitivity Analysis
   
   Test the robustness of relative value conclusions by:
   • Using multiple multiples (P/E, EV/EBITDA, P/S, P/B)
   • Comparing mean versus median values
   • Running regressions to control for fundamental differences
   • Assessing the impact of including or excluding marginal peers
5. Step 5: Apply Adjustments Where Appropriate
   
   When peers differ in fundamental characteristics, either:
   • Adjust the Multiple: Using tools such as the PEG ratio to control for growth differences
   • Apply Regression: Using multivariate regression to control for multiple factors simultaneously
   • Use Weighted Averages: For multi-industry companies

Comprehensive Case Study: Peer Group Construction for a Global Logistics Company

An analyst is evaluating Global Logistics Corp, a transportation and logistics firm with operations across North America, Europe, and Asia.

$$
\begin{array}{l|l}
\text{Metric} & \text{Global Logistics Corp} \\
\hline
\text{Market Capitalization} & \$42.5\text{ billion} \\
\hline
\text{Revenue} & \$28.3\text{ billion} \\
\hline
\text{EBITDA Margin} & 14.2\% \\
\hline
\text{Revenue Growth (3-year average)} & 6.8\% \\
\hline
\text{Geographic Distribution} & \text{North America: 45%, Europe: 30%, Asia: 25%} \\
\hline
\text{Business Segments} & \text{Freight Transport (55%), Warehousing (25%), Supply Chain} \\
& \text{Services (20%)} \\
\hline
\text{Debt-to-Equity} & 0.65 \\
\hline
\text{ROE} & 16.2\% \\
\hline
\text{Beta} & 1.15 \\
\end{array}
$$

1. Step 1: Initial Industry Screen
   
   Using the GICS Transportation & Logistics classification, the analyst identifies 18 potential peers.
2. Step 2: Fundamental Characteristic Filters
   
   Filter 1: Market Capitalization \(\geq\) $10 billion (Excludes 7 smaller firms. Remaining: 11 firms.)
   
   Filter 2: At Least 30% Revenue Outside Home Region (Excludes 3 firms with primarily domestic operations. Remaining: 8 firms.)
   
   Filter 3: EBITDA Margin Between 10% and 18% (Excludes 2 firms with unusually high margins and 1 firm with very low margins. Remaining: 5 firms.)
3. Step 3: Segment Analysis
   
   Given Global Logistics Corp’s three business segments, the analyst calculates segment-specific peer multiples:
   
   $$
   
   \begin{array}{l|c|c}
   
   \textbf{Segment} & \textbf{% of Revenue} & \textbf{Peer Median EV/EBITDA} \\
   
   \hline
   
   \text{Freight Transport} & 55\% & 8.5x \\ \hline
   
   \text{Warehousing} & 25\% & 11.2x \\ \hline
   
   \text{Supply Chain Services} & 20\% & 14.0x \\
   
   \end{array}
   
   $$
   
   $$
   
   \begin{array}{c}
   
   \text{Composite EV/EBITDA} = (0.55 \times 8.5) + (0.25 \times 11.2) + (0.20 \times 14.0) = 10.275x
   
   \end{array}
   
   $$
4. Step 4: Regression-Based Control
   
   To refine the comparison, the analyst runs a regression of EV/EBITDA on growth and ROE across the 5 selected peers:
   
   $$
   
   \begin{array}{c}
   
   Predicted\ EV/EBITDA = 2.3 + 0.6 \times Growth + 0.4 \times ROE
   
   \end{array}
   
   $$
   
   For Global Logistics Corp, with Growth = 6.8% and ROE = 16.2%:
   
   $$
   
   \begin{array}{c}
   
   Predicted\ EV/EBITDA = 2.3 + (0.6 \times 6.8) + (0.4 \times 16.2) = 12.86x
   
   \end{array}
   
   $$
5. Step 5: Triangulation
   
   The analyst now has three benchmarks:
   
   $$
   
   \begin{array}{l|c}
   
   \textbf{Method} & \textbf{EV/EBITDA} \\
   
   \hline
   
   \text{Simple Peer Median} & 10.6x \\ \hline
   
   \text{Segment-Weighted Composite} & 10.275x \\ \hline
   
   \text{Regression-Adjusted} & 12.86x \\
   
   \end{array}
   
   $$

Global Logistics Corp’s superior profitability and growth justify a premium to the simple peer median. If the company’s actual EV/EBITDA is below 12.86x, it would appear undervalued on a factor-adjusted basis.

Question 1

Which of the following best describes the preferred approach for estimating a multiple for a company operating in multiple industries?

1. Estimate the multiple as a weighted average of industry peer group multiples based on the proportion of firm revenues, income, or assets in each industry.
2. Estimate the multiple by compiling a list of the firm’s peers across all industries and taking a simple average of the multiples across all peers.
3. Estimate the multiple using the median of the peers across all industries with similar risk, growth, and cash flow characteristics.

Solution

The correct answer is A.

A weighted average of industry peer group multiples best mirrors the risk, growth, and cash flow characteristics of a multi-industry firm.

B is incorrect. A simple average fails to account for the proportional contribution of each segment.

C is incorrect. It also fails to consider the proportional contribution of each segment to the company’s overall operations.

Question 2

Which of the following statements appropriately describes the composition of a factor-based peer group?

1. The final peer group consists of all firms in the same industry classification.
2. The final peer group consists of firms with similar beta coefficients, regardless of industry classification.
3. The final peer group consists of firms in the same industry with similar fundamental characteristics.

Solution

The correct answer is B.

This reflects a factor-based approach where firms with similar beta coefficients are grouped regardless of industry.

A is incorrect. The inclusion of “all firms” in an industry is rarely appropriate; analysts typically apply additional filters.

C is incorrect. Differences in fundamental characteristics can be incorporated through adjustment or regression methodology rather than requiring identical characteristics for inclusion.