Multiples and Comparables

The shortcut

Building a full DCF takes hours. A comparables analysis takes minutes. That’s why it dominates day-to-day equity research, M&A pitches, and IPO pricing.

The premise: if comparable firms trade at 18× earnings on average, then a firm with $5 of earnings per share is “worth” 18 × $5 = $90.

The premise is right when the comparables are truly comparable and the multiple reflects fundamentals you can defend. The premise is wrong otherwise. Knowing the difference is the skill.

The four most-used multiples

P/E (Price / Earnings) — most familiar; what most people mean by “the multiple.”

• Numerator: market price per share.
• Denominator: trailing or forward earnings per share.

A P/E of 18× means investors are paying $18 for every $1 of annual earnings. Average S&P 500 P/E sits between 16-22× over the long run. Higher for growth, lower for cyclicals or distressed firms.

EV/EBITDA (Enterprise Value / EBITDA) — preferred by analysts because it’s capital-structure-neutral.

• Numerator: Enterprise value = Market cap + Debt − Cash.
• Denominator: EBITDA = Earnings before interest, taxes, depreciation, amortization.

Why EV: makes leveraged firms comparable to unleveraged ones. P/E is distorted by capital structure (debt → interest → smaller E); EV is not.

Why EBITDA: strips out non-cash D&A and tax differences, getting at operating performance. Useful for cross-border comparisons (different tax regimes) and capital-intensive industries (different depreciation policies).

EV/Sales — useful when earnings are negative or volatile.

Used heavily for early-stage tech, biotech, and other firms with negative profitability. Less informative than EBITDA-based multiples because it doesn’t reflect operating margin, but it’s the only multiple that always exists.

Price-to-Book (P/B) — historically used for banks and asset- heavy firms.

• Numerator: market price per share.
• Denominator: book value of equity per share.

P/B near 1.0 says “the market values the firm at what the books say.” Significantly above means investors expect ROE > cost of equity. Below 1.0 historically signaled distress. Less useful for asset- light businesses where book value misses most of the firm’s economic value (brand, IP, customer base).

How to do comparables right

1. Pick a peer set you can defend. Same industry, similar size, similar growth profile, similar geography. 5-10 firms is the sweet spot. Too few and the median is noisy; too many and the group stops being comparable.

2. Calculate each peer’s multiple. Pull from latest 10-K / quarterly reports. Decide on trailing vs forward (analysts’ consensus EPS for the next 12 months).

3. Compute the median. Use median, not mean — the mean is dragged around by one outlier. Some practitioners also report the inter-quartile range to show dispersion.

4. Apply the multiple to your target. Median P/E × target EPS = implied price.

5. Sanity-check against absolute valuation. If your DCF says $40 and your comparables say $80, somebody’s wrong. Reconcile before you publish.

How multiples mislead

Negative earnings. P/E is meaningless if the denominator is negative. Drop earnings-based multiples entirely; use EV/Sales or forward P/E based on consensus that excludes one-time losses.

Growth differences. A 5% grower at 15× and a 25% grower at 30× might be priced equally fairly — the higher multiple just reflects higher growth. The PEG ratio (P/E ÷ growth rate) partly corrects for this, but the relationship isn’t linear.

Cyclicality. A cyclical firm at the peak of its cycle has artificially low P/E (earnings are unusually high). Buying at “cheap” P/E at cyclical peaks is a classic value trap. Use mid-cycle earnings instead.

Accounting differences. Aggressive vs conservative revenue recognition can move EPS by 10-20% with no underlying business change. Stock-based compensation treatment varies between GAAP and “adjusted” earnings. Read the footnotes.

Apples-to-oranges peers. “Tech” is not a peer set. Cloud infrastructure is not the same as social media is not the same as fintech. Tighten the cohort.

Mid-cap vs large-cap discount. Smaller firms trade at lower multiples for reasons (illiquidity, less analyst coverage). Don’t apply S&P 500 multiples to a $500M firm without an explicit size discount.

Transaction multiples vs trading multiples

Trading multiples come from secondary-market prices of public firms.

Transaction multiples come from completed M&A deals. These typically embed a control premium of 20-40% above trading prices — acquirers pay extra for the right to set strategy.

Use trading multiples to value a stake you’d buy in the public market. Use transaction multiples to estimate what a whole company might fetch in an acquisition. They are not interchangeable.

A worked example

You’re valuing a mid-sized consumer software company. Five peers with median EV/EBITDA of 12× and median forward P/E of 25×. Your target has expected next-year EBITDA of $80M and EPS of $3.20. Net debt = $200M, 50M shares outstanding.

• EV implied by EBITDA multiple: 12 × $80M = $960M.

• Equity value: 960 − 200 = $760M.

• Per share: 760 / 50 = $15.20.

• Equity value implied by P/E: 25 × $3.20 = $80 per share.

Wait — those disagree by a factor of 5×. What happened? Capital structure. The target has $200M of debt on a $760M EV; that’s 26% leverage. Other peers might be unleveraged. The P/E (which depends on EPS, which is hurt by interest expense) understates equity value; the EV-based approach correctly normalizes. Use the $15.20.

The lesson: when multiples disagree, the disagreement usually points to a real economic difference between your target and the peer set. Investigate before averaging.

Where multiples sit in valuation work

In an actual valuation memo or pitchbook you’d present:

• DCF (intrinsic): your best estimate, but sensitive to assumptions.
• Trading multiples: market view, but reflects current sentiment.
• Transaction multiples: what an acquirer would pay.
• 52-week high/low: range of recent market views.

A “football field” chart shows these as horizontal bars; the overlap is the defensible valuation range. No single method is right; the multi-method approach acknowledges that all valuation involves judgment.