Relative Valuation Explained: Picking the Right Multiple
The short answer
Relative valuation estimates what a stock is worth by comparing its price multiples, such as P/E or EV/EBITDA, with similar companies and with its own history. It is fast and grounded in real market prices, but it only tells you whether a stock is cheap compared with something else, so the quality of the comparison decides the quality of the answer.
Key takeaways
- Relative valuation compares a stock's multiples with true peers and with its own long-term range.
- Different business models need different multiples; P/E for steady earners, EV/EBITDA where debt matters, P/B for banks.
- A low multiple is a fact that usually has a reason. Find the reason before calling it cheap.
- Growth closes multiple gaps; a fast grower at 30 times earnings can be cheaper than a stagnant business at 10.
- Every multiple is shorthand for a forecast the market has already made.
What is relative valuation?
Relative valuation prices a business the way an appraiser prices a house: by looking at what comparable things sell for. Instead of price per square foot, the units are multiples, price divided by some fundamental the business produces.
Multiple = price / a fundamental (earnings, cash flow, book value)
A P/E of 15 means paying $15 for each $1 of annual earnings
The price-to-earnings ratio is the most familiar, but the family is large: enterprise value to EBITDA, price to book, price to free cash flow, price to sales. All of them compress a valuation into one number that can be compared across companies and across time.
Understand what the method does and does not answer. A multiple tells you whether a stock is cheap compared with something else, a peer group or its own past. It does not tell you what the business is worth in absolute terms; that is the job of a discounted cash flow. The two approaches check each other, which is why practitioners run both.
The strengths are real, which is why the method never goes out of fashion. It is fast. It anchors to prices other investors actually paid rather than to your imagination. And it functions where a cash flow model cannot go, in young industries, messy transition years, and businesses whose forecasts would be fiction. Used with discipline, multiples are a compact way of borrowing the market's collective work; used lazily, they are a way of borrowing its errors.
Which multiple fits which business model?
The right multiple is the one that captures how the business actually creates value, and using the wrong one produces confident nonsense. A quick map:
| Business model | Multiple that fits | Why |
|---|---|---|
| Steady, profitable operator | P/E or free cash flow yield | Earnings are real, stable and comparable |
| Heavy debt or heavy assets | EV/EBITDA | Counts the debt in the price and strips financing noise |
| Bank or insurer | Price-to-book | The balance sheet is the business |
| Young, unprofitable grower | EV/Sales, cautiously | Nothing else exists yet, and margins are a guess |
Two entries deserve a note. EV/EBITDA uses enterprise value, which includes debt, so it lets you compare a leveraged business with a debt-free one on even terms; a plain P/E is blind to the difference. And price-to-book survives in banking long after it stopped meaning much for software companies, because a bank's assets are financial instruments marked near reality while a software firm's best assets never reach the balance sheet.
Match the numerator to the denominator as well. Enterprise value belongs over EBITDA or revenue, flows that accrue to lenders and shareholders together; plain share price belongs over earnings or book value, which belong to shareholders alone. Cross the streams, price over EBITDA, say, and the output is a number with no economic meaning that still looks perfectly authoritative in a spreadsheet.
The classic wrong-tool error is applying a P/E to a deeply cyclical business at the top of its cycle. Peak earnings make the multiple look tiny exactly when the earnings are least sustainable. Commodity producers routinely look cheapest right before the cycle turns.
Who counts as a real peer?
A peer is a business with similar economics, not just the same sector label, and this is where most comparisons quietly fail. Growth rate, returns on capital, capital intensity and debt load matter more than sharing an industry code. A luxury brand and a discount chain are both retailers; they are not comparables.
A short checklist finds the economic twins: similar gross margins, similar capital intensity, similar growth, similar exposure to the cycle. Expect to end up with two or three honest comparables rather than fifteen loose ones. Averages taken across the wrong companies converge on nothing useful, however wide the sample gets.
Growth is the difference that distorts comparisons most, because a multiple is a snapshot while growth compounds. Watch three hypothetical companies, each earning $1 per share today:
| Company | Price today | Earnings growth | EPS in year 5 | Today's price / year-5 EPS |
|---|---|---|---|---|
| A | $10 | 0% | $1.00 | 10.0x |
| B | $15 | 6% | $1.34 | 11.2x |
| C | $30 | 15% | $2.01 | 14.9x |
Hypothetical figures; earnings compounded for five years, rounded.
On today's earnings, C looks three times as expensive as A. Measured against year-five earnings, if the growth arrives, the gap has collapsed from 30-versus-10 to about 15-versus-10. That "if" is the entire game: a multiple is shorthand for a forecast, and comparing multiples without comparing forecasts is comparing nothing. The PEG ratio is a crude first correction for this.
What does a company's own history add?
A second, often better comparison set is the company itself over time. A business that traded between 12 and 20 times earnings across a decade, and now sits at 26, is making an implicit claim: something fundamental has improved. Sometimes that claim is true. Your job is to name what changed, not to assume the old range must reassert itself.
Reversion to the historical band is a reasonable base case only when the business is the same business. A company whose revenue has shifted toward recurring, high-margin software deserves a higher band than its hardware past. A retailer losing share to online rivals deserves a lower one. The history frames the question; the business answers it. When a genuinely great company sustains a premium for years, the interesting problem becomes whether a great business can be too expensive, which is its own article.
A quick audit separates deserved re-ratings from mood. List what would have to be true for the new multiple to be correct, then check the filings for it: a margin structure that visibly improved, revenue mix shifting toward recurring contracts, returns on capital that climbed and stayed up. If the evidence is there, the old band is obsolete. If the business looks the same and only the enthusiasm changed, the old band is usually the better forecast, though it can take years to prove it.
Where do multiple comparisons go wrong?
The method fails quietly, so it pays to know the standard traps before trusting any screen.
- The peak-earnings trap. Cyclicals look cheapest at the top, because the E in P/E is temporarily fat. Average the earnings across a cycle before comparing.
- The expensive-neighborhood trap. Cheapest among peers means little if the whole group is dear. Relative cheapness is not absolute value; anchor the group against its own history too.
- The leverage blind spot. Two identical P/Es can hide wildly different debt loads. Check an enterprise-value multiple alongside any equity multiple.
- The single-year trap. One year of earnings, good or bad, is a weak base. Use several years, or normalized figures, for anything volatile.
- The forecast-free comparison. As the table above showed, multiples embed growth expectations. Comparing a 10x business with a 30x business tells you nothing until you compare what each must deliver.
- The precision trap. A multiple of 14.2 against a peer at 15.1 is noise, not signal. Multiples earn their keep flagging large gaps, a business at half or double its comparables, and turn into astrology when read to one decimal place.
Where to go from here
Relative valuation is half of a complete toolkit; it tells you how the market prices similar forecasts, while a DCF tests the forecast itself, and how to value a company shows how the pieces fit together. Practice reading multiples in context rather than in isolation: put two similar businesses side by side in the compare tool and work out which differences the market is paying for. When a multiple and a DCF disagree sharply about the same business, treat that disagreement as the most useful finding of the day and hunt down the assumption responsible.
Frequently asked questions
It is pricing a business by comparison, the way an appraiser prices a house from what similar houses sold for per square foot. You divide the company's price by a fundamental such as earnings, then judge that multiple against close peers and the company's own history.
None wins everywhere. P/E suits steady, profitable businesses. EV/EBITDA handles companies with different debt loads. Price-to-book fits banks and insurers, whose balance sheets are the business. Serious work uses two or three multiples at once, because each corrects a blind spot in the others.
There is no universal number. A P/E is only readable against the company's growth, its peers and its own past range. Fifteen can be expensive for a shrinking business and thirty can be reasonable for a durable compounder. Treat any multiple as the start of a question, never as the verdict.
Relative valuation prices a business against other assets; intrinsic valuation, usually a discounted cash flow, prices it against its own future cash. Relative is faster and anchored to the market, intrinsic is more fundamental and more assumption-heavy. When the two disagree sharply, something is worth investigating.
Educational content, not investment advice. Tenet explains concepts; it does not recommend securities. Do your own research before you invest.