The columned front of the New York Stock Exchange building on Wall Street, hung with a large American flag
Photo: Arild Vågen / Wikimedia Commons, CC BY-SA 4.0

Valuation, from first principles

Do better businesses cost more?

We checked 83 US industries. The price tag did not care how good the business was.

vizmaya desk · 16 Oct 2026 · Research, not investment advice

Step 1 · The price tag

A price tag has a simple meaning.

You buy a shop for $20. It earns $1 of profit a year.

You wait 20 years to get your $20 back. That is a 20x multiple.

Experts call this EV/EBIT: the price of the whole firm divided by its yearly operating profit. We use it all deck.

Step 2 · The quality test

A good shop earns more from the same money.

Put $100 in. If it pays you $15 a year, the return is 15%.

Experts call this return on capital. The weakest quarter of US industries earns about $6.5 per $100. The best quarter earns $29.7.

DERIVEDAverages of 83 industries, Jan 2026. Data: Damodaran, NYU Stern.

Step 3 · The obvious guess

So a better shop should cost more.

If one shop turns $100 into $30 and another into $6, buyers should pay extra for the first.

Plot return on the bottom and price up the side. The dots should climb.

Step 4 · The test

I bet on a rising line. The dots had other plans.

Each dot is one of 83 US industries in January 2026. They land like spilled rice.

The link between return and price is -0.05. A perfect link would be 1.0.

REPORTEDDamodaran, Jan 2026. DERIVEDr = -0.05; 95% range -0.24 to +0.17.

Same test, simpler

The best quarter costs a bit less than the worst.

Industries in the top quarter for return: 23.5x yearly profit. Bottom quarter: 24.8x.

If quality were paid for, the top bar would be the tall one.

DERIVED20 industries per group, Jan 2026. Damodaran multiples.

Case 1 · Steel

Steel earned half as much. Its price more than doubled.

Return fell from 12.5% to 6.7%. The multiple rose from 11.6x to 28.5x.

My reading: buyers priced the recovery, not the dip.

REPORTEDJan 2025 vs Jan 2026. ESTIMATEDthe reading.

Two tall blast furnaces and their pipework at the Pittsburgh Steel Company works in Monessen, Pennsylvania
Photo: Historic American Engineering Record, Library of Congress / Wikimedia Commons, Public domain

Case 2 · Chips

Chips earned more. Their price fell.

Return rose from 21.1% to 27.2%. The multiple slipped from 49.3x to 44.5x.

My reading: the good news was in the price long before the profit showed up.

REPORTEDJan 2025 vs Jan 2026. ESTIMATEDthe reading.

Close-up of a bare silicon chip die showing its fine metal circuit pattern
Photo: Mister rf / Wikimedia Commons, CC BY-SA 4.0

Step 5 · Speed, not level

Compare changes, and a pattern shows up.

Industries whose returns rose more saw their multiple rise less.

The link is -0.31 across 83 industries. The 95% range, -0.54 to -0.14, sits fully below zero.

DERIVEDChanges from Jan 2025 to Jan 2026, 2025 returns and 2025 multiples.

A trap in the maths

Part of that is a see-saw built into the numbers.

Profit sits inside both measures. When profit rises, return goes up, and price divided by profit goes down.

So some of that -0.31 is plain arithmetic, not insight.

ESTIMATEDToy firm: $200 price, $100 capital.

Step 6 · Is it priced in?

The market paid for gains before they arrived.

Take industries priced high in January 2025. A year later, their returns were up 1.8 points on average.

The cheap half saw returns fall 1.3 points. Price pointed at the future.

DERIVED40 vs 40 industries. Link r = +0.35, 95% range +0.18 to +0.60.

Step 7 · The whole market

Same returns. A higher price.

US firms outside finance earned 14.7% in 2024 and 15.1% in 2026. The multiple went from 22.9x to 26.5x.

Returns rose 3%. The price rose 16%. The gap is a bet on what comes next.

REPORTEDDamodaran, Jan 2024, 2025, 2026. DERIVEDindex.

Step 8 · Other places

Each place charges a different price per point of return.

Divide the multiple by the return. US 1.76, Europe 1.19, Japan 2.46.

Emerging markets are missing. I did not get that data in time.

DERIVEDEx-financials, Jan 2026. Europe, Japan.

Mostly priced in. Today's quality does not set today's price. Expected change does. Level link: -0.05. Change link: -0.31. Expensive industries then saw returns rise: +0.35. Industry data, one date, no cause proven. Research, not investment advice.

What would change my mind

Three checks I will rerun in January 2027.

  • The level link turns positive.If return and price reach r above +0.25, quality is being paid for.Now: -0.05
  • The change link flips.If rising returns start to bring rising multiples, the see-saw is not the story.Now: -0.31
  • Expensive stops meaning early.If the 1.8 vs -1.3 point gap shrinks below 1 point, price has lost its edge on the future.Now: a 3.0 point gap

Sources & method

Data (all REPORTED)

Method (DERIVED)

  • 83 industries after dropping 3 with extreme values. Correlations run twice, in plain Python and numpy, with 95% bootstrap ranges.
  • Return = after-tax operating income over invested capital.

Blind spots

  • Industries, not companies. One firm can break the pattern.
  • Profit sits in both numbers, so the change link is partly arithmetic.
  • Three January snapshots. Emerging markets missing. No share prices pulled.

Photo credits

vizmaya desk · 16 Oct 2026 · Research, not investment advice · More at vizmaya stories