Multiples and Comparables

Why Multiples?¶

Relative valuation anchors value to market prices of comparable firms. Instead of forecasting cash flows explicitly, we ask: “What are investors paying for similar businesses right now?”

Multiples are:

Fast — no multi-year forecast required
Market-anchored — reflects current sentiment and pricing
Intuitive — easy to communicate to non-specialists

But they are also:

Circular — if markets are wrong, multiples inherit that error
Comparability-sensitive — garbage in, garbage out if peers are chosen poorly
Accounting-sensitive — different depreciation policies distort earnings-based multiples

Enterprise Value vs Equity Value Multiples¶

Multiple	Numerator	Denominator	Who bears it?
EV/EBITDA	Enterprise Value	EBITDA	All capital providers
EV/EBIT	Enterprise Value	EBIT	All capital providers
EV/Sales	Enterprise Value	Revenue	All capital providers
P/E	Market Cap	Net Income	Equity holders only
P/B	Market Cap	Book Equity	Equity holders only

EV/EBITDA: The Practitioner’s Workhorse¶

EV/EBITDA is preferred because:

It is capital structure-neutral (before interest)
It is less affected by depreciation policy differences than EV/EBIT
It is positive for most profitable firms (P/E can be negative for loss-making firms)

Typical EV/EBITDA ranges by sector (2020s, approximate):

Sector	Typical Range
Technology / Software	15–30×
Healthcare	12–20×
Consumer Staples	10–15×
Industrials	8–12×
Utilities	7–10×
Energy	4–8×

These ranges shift significantly with interest rates and market cycles.

Adjustments for Comparability¶

Raw multiples are misleading when peers differ in growth or risk. The PEG ratio adjusts P/E for growth:

PEG = \frac{P/E}{g}

(1)

where $g$ is the expected EPS growth rate (%). A PEG of 1 is often used as a heuristic for fair value.

For EV multiples, a regression-based approach regresses the multiple on fundamentals across the peer group:

\frac{EV_i}{EBITDA_i} = \alpha + \beta_1 g_i + \beta_2 \text{ROIC}_i + \beta_3 \text{Leverage}_i + \varepsilon_i

(2)

The fitted value for the target firm becomes the “warranted” multiple.

Python: Comparable Company Analysis¶

import pandas as pd
import numpy as np

# Hypothetical peer group data
peers = pd.DataFrame({
    "Company":       ["Alpha Co", "Beta Corp", "Gamma Inc", "Delta Ltd", "Epsilon SA"],
    "EV_m":          [4200, 2800, 6100, 1500, 3300],     # Enterprise value €m
    "EBITDA_m":      [380,  210,  490,  140,  280],       # EBITDA €m
    "Revenue_m":     [1200, 850,  1600, 500,  950],       # Revenue €m
    "EBIT_margin":   [0.18, 0.15, 0.19, 0.17, 0.18],     # EBIT / Revenue
    "Growth_3y":     [0.08, 0.12, 0.07, 0.15, 0.09],     # Revenue CAGR
})

peers["EV_EBITDA"] = peers["EV_m"] / peers["EBITDA_m"]
peers["EV_Revenue"] = peers["EV_m"] / peers["Revenue_m"]

print("=== Peer Group Multiples ===")
cols = ["Company", "EV_EBITDA", "EV_Revenue", "EBIT_margin", "Growth_3y"]
print(peers[cols].to_string(index=False, float_format="{:.2f}".format))

# Summary statistics
print("\n=== Multiple Summary ===")
print(peers[["EV_EBITDA", "EV_Revenue"]].describe().round(2))

# Apply to target company
target_ebitda   = 320  # €m
target_revenue  = 1050 # €m

median_ev_ebitda  = peers["EV_EBITDA"].median()
median_ev_revenue = peers["EV_Revenue"].median()

ev_from_ebitda  = median_ev_ebitda  * target_ebitda
ev_from_revenue = median_ev_revenue * target_revenue

print(f"\n=== Target Valuation ===")
print(f"  EV/EBITDA method: €{ev_from_ebitda:.0f}m  (median {median_ev_ebitda:.1f}x × €{target_ebitda}m EBITDA)")
print(f"  EV/Revenue method: €{ev_from_revenue:.0f}m (median {median_ev_revenue:.2f}x × €{target_revenue}m revenue)")
print(f"  Implied range: €{min(ev_from_ebitda, ev_from_revenue):.0f}m – €{max(ev_from_ebitda, ev_from_revenue):.0f}m")

Pitfalls and Best Practices¶

Peer selection is the most important and most subjective step. Peers should be similar in business model, geography, size, and growth profile. Including a high-growth tech firm in a peer group of mature industrials will distort every multiple.

Trailing vs forward multiples: Forward multiples (based on consensus earnings forecasts) are more appropriate for valuation because they reflect the future. Trailing multiples mix past performance with current prices.

LTM vs FY multiples: Use the last twelve months (LTM) figure to get the most current snapshot, especially when there are seasonal effects or recent acquisitions.