“A P/E of 30 is expensive” is one of those statements that sounds like analysis and functions as a reflex. Expensive compared to what? We took every US-listed company above ten billion dollars — roughly seven hundred and forty of them — and calculated the median earnings multiple and the median return on capital for each sector. The answer is not what most people assume.
What the market actually charges
Median trailing P/E, by sector, highest to lowest:
- Technology 33.2 · Healthcare 32.0 · Industrials 31.3
- Real Estate 26.5 · Consumer Defensive 23.8 · Communication Services 22.2
- Consumer Cyclical 21.2 · Utilities 20.5 · Basic Materials 20.1
- Financial Services 15.3 · Energy 15.3
A 30x multiple is roughly the median in technology, healthcare, and industrials. The same multiple in financial services would put a company at twice its sector norm. The number on its own carries almost no information.
The part that should surprise you
Now put returns next to price. Median return on capital employed, same universe:
- Consumer Cyclical 18.7% — median P/E 21.2
- Consumer Defensive 14.9% — median P/E 23.8
- Industrials 12.7% — median P/E 31.3
- Basic Materials 12.6% — median P/E 20.1
- Technology 11.9% — median P/E 33.2
- Real Estate 5.0% — median P/E 26.5
The sector earning the highest median return on its capital is consumer cyclical, and it trades at a third less than technology. The sector the market charges the most for — technology, at 33.2 times earnings — earns a median 11.9% on capital, below four other sectors that cost less.
Why this is not a buy signal
The obvious conclusion is that technology is overpriced and consumer cyclicals are a bargain. We do not think that follows, for two reasons worth understanding.
First, a median hides the distribution. Technology's median is dragged down by a long tail of unprofitable software companies, while its best names earn returns nothing in consumer cyclical approaches. The sector is bimodal; the median describes neither half.
Second, the market is paying for the direction of travel, not the current level. A technology company at 11.9% returns and 20% growth will out-earn a cyclical at 18.7% returns and 4% growth within a few years. Multiples price the future; returns measure the past.
Where the scores land
Investily's own median score per sector reorders things again. Basic Materials posts the highest median at 71, ahead of Industrials (69), Technology (70) and Healthcare (68). At the bottom sit Real Estate (47) and Utilities (48) — both capital-heavy, both earning around 5% on that capital, and real estate still charging 26.5 times earnings for the privilege.
That real estate line is the clearest mispricing-shaped thing in the data: the second-lowest returns in the market at the fourth-highest multiple.
The bottom line
Use these as reference points, not rules. A 25x multiple is cheap for technology, ordinary for consumer defensive, and rich for a bank — and knowing which of those you are looking at is most of the work.
The broader lesson is that a P/E only means something next to a peer group and a return on capital. On its own it is a number without a denominator you can reason about, which is why a low P/E so often isn't a discount.