Why does a "normal" multiple look completely different by sector?

A P/E of 12 is cheap for a software company and expensive for a bank. That isn't a contradiction — it's the most important context there is. Multiples live inside sectors, and comparing across them without translation is one of the classic beginner errors.

Different economics, different multiples

Sectors carry structurally different multiples because their businesses are structurally different:

• Growth is priced. High-growth sectors (software, some healthcare) carry high multiples because the market pays for expected future earnings. Mature, slow-growth sectors (utilities, tobacco, banking) carry low ones because there's less future growth to pay for.
• Asset intensity shifts the tool. Asset-heavy sectors (banks, real estate, industrials) are often read on P/B, because their balance sheets are the business. Asset-light sectors are read on P/E or EV/EBITDA, because their book value barely reflects their worth.
• Stability commands a premium. Sectors with steady, predictable demand (consumer staples, utilities) often trade at a premium to their growth rate alone would justify, because reliability is itself valued.

So a utility at a P/E of 16 and a software firm at a P/E of 40 might both be trading exactly in line with their sectors. Neither is cheap or expensive until measured against its own kind.

The unbreakable rule

Only compare multiples within the same sector. Ranking a bank, a software company, and an oil producer by raw P/E is meaningless — you're comparing three different economic species on a scale that means something different for each. The comparable set from the last unit exists precisely to enforce this: it keeps the comparison inside one economic world.

Where the sector median comes from

A sector's "normal" multiple isn't a fixed law either. It reflects the sector's current growth prospects, its risk, and where interest rates sit. When rates rise, the multiples the whole market will pay tend to compress — but through two separate channels, and it pays to keep them apart:

• Through valuation. Future earnings are discounted harder, so the sectors whose value sits furthest in the future — high-growth software, say — feel that pull most. The longer the wait for the profits, the heavier the discount.
• Through the business itself. Bond-like sectors feel a different squeeze: their steady dividends look less special beside now-higher bond yields, and the borrowing costs of debt-heavy names climb.

Two mechanisms, both real, landing on different sectors — which is why "rates rose, so multiples fell" is never the whole sentence. That's the bridge to the next lesson: multiples move with the cycle, not just the company.

Try it now

Two sectors, the same seven multiples, and no comparison worth making between them:

1. Line the two trailing P/E figures up. Then ask the only question that matters — is either number high or low for its own kind? Neither table can tell you, and that is the lesson.
2. Notice which multiple each business is actually read on. The bank's price-to-book anchors to a balance sheet that is the business; for an asset-light company the same multiple is nearly decorative. Ranking a bank and a software firm on one ruler compares two economic species.
3. Add a third sector and check the pattern holds:

Its P/E is quoted the same way as the others and describes a completely different kind of demand. Say out loud why lining all three up on raw P/E would be an unfair comparison. Spotting that unfairness is the skill. 4. Build the honest version, a handful of names from one industry on the same multiples:

Take the median of each column. Then place each of the three companies above against the utility medians and say why that placement tells you nothing. A sector median means something; a cross-sector ranking does not.