Did the Market Rise Because Companies Earned More, or Because You Pay More?
A company's market value is its revenue, times the share of that revenue it keeps, times what the market pays for each dollar kept. Those three multiply back to the whole exactly, so splitting a rise between them is a measurement and not a model. Read it beside the Market P/E Ratio, which says what the market costs; this says how it got there.
Latest read
Over the five years to October 9, 2026 the 312 companies in the base grew their market value 1.65×, though 1.48× of that is the 279 companies present at both ends — 33 later entrants supply 12% of ending value and no holder earned it. Revenue accounts for 83% of that rise and wider margins for 48%, while the multiple contributed −31%: investors are paying less per dollar of earnings than five years ago. The market did not re-rate. It earned. Net margin is 14.3%, against 8.9% in 2011, and that widening is doing more of the work each year — 85% of the last twelve months against 19% across the whole record. The gain is broad, on the measure that can say so, and no handful of winners carries it: of the 279 companies present at both ends, 73% grew earnings and 59% widened their margin. (A profitability count cannot support that claim — a company profitable at both ends can have shrunk.) Whether 15%-plus margins hold is now the question that matters more than the multiple.
Sources, methodology & freshnessSEC EDGAR XBRL filings for the covered companies — revenue and net income from the statements, share counts from filing cover pages — joined to our daily price data · Rebuilt every day in the pipeline; the fundamental legs change as each 10-Q and 10-K landsData as of 2026-10-09 · Open ↓Close ↑
Net margin 14.3%, from 8.9% in 2011. 308 of 326 companies profitable. The multiple is the smallest of the three drivers, and over five years it worked against the market.
The exact split, by window
What the first column is, and what it is not. It is the change in the BASKET’S market value, and the basket gains members: a company whose filing history reaches our base after the start date arrives carrying its whole market value, which no holder earned by holding. “Same companies” repeats the identical decomposition over only the companies present at both ends, and the gap between the two columns is the composition effect. It is small over a year and large over fifteen. Neither column is a shareholder return: both exclude dividends, and the basket is drawn from today’s survivors, so companies that failed along the way are missing from the early years entirely.
| Window | Market cap | Same companies | Revenue | Margin | Multiple | Per year |
|---|---|---|---|---|---|---|
| 1Y0.8y | ×1.17 | ×1.184 joined | ×1.0955% | ×1.1485% | ×0.94−40% | 22.5% |
| 3Y2.8y | ×1.78 | ×1.7521 joined | ×1.2438% | ×1.3450% | ×1.0712% | 23.2% |
| 5Y4.8y | ×1.65 | ×1.4833 joined | ×1.5283% | ×1.2748% | ×0.85−31% | 11.0% |
| 10Y9.8y | ×6.26 | ×4.23111 joined | ×2.8357% | ×1.7430% | ×1.2813% | 20.7% |
| Since 201115.5y | ×11.72 | ×7.11179 joined | ×3.8254% | ×1.6019% | ×1.9226% | 17.2% |
Every row multiplies back exactly: revenue × margin × multiple equals the market-cap column. The multiple is the only driver that has gone backwards, and it has done so over both the last year and the last five. Over the full record it still added 26%, so the re-rating that happened belongs to the 2010s and not to the recent market.
Reading the split
- Shares can exceed 100%. The drivers compound rather than add, so they are split in logs. When one leg goes backwards it takes a negative share and the others sum past 100 — which is exactly what a falling multiple looks like.
- Margins are the swing factor. Revenue growth is steady across every window. What changes is how much of it companies keep.
- This is not a forecast. A margin that widened for a decade can narrow, and it would pull earnings down without any change in the multiple.
Net margin
How many companies actually make money
Revenue and the multiple, indexed
Method
Market cap = revenue × (earnings ÷ revenue) × (market cap ÷ earnings). The middle term is net margin and the last is the multiple, so the three legs are exact and leave no residual. Shares are taken in logs because the drivers compound.
Every leg is computed over the 312 companies whose revenue AND earnings are both known at the date. Summing earnings across the whole base while summing revenue across only the companies that report it would put a different denominator under each leg and overstate the margin — the one number this page exists to get right.
The basket is today's largest US companies, and a decomposition is more exposed to that than a ratio: these firms grew revenue faster than the market did, which flatters the revenue leg specifically. The margin and multiple legs are ratios and travel better between baskets. Read the split as a description of these companies rather than of the whole market.
Each quarter enters on the date it was first filed, never on the date it ended, so the timing carries no look-ahead. The values are latest-vintage: a figure later restated appears here as restated, sitting at its original filing date. That makes this a restated history dated by first disclosure rather than a point-in-time series — someone reconstructing what was knowable on a past date would have seen the original print, not this one.
Net margin 14.3%, from 8.9% in 2011. 308 of 326 companies profitable. The multiple is the smallest of the three drivers, and over five years it worked against the market.
Reading the split
- Shares can exceed 100%. The drivers compound rather than add, so they are split in logs. When one leg goes backwards it takes a negative share and the others sum past 100 — which is exactly what a falling multiple looks like.
- Margins are the swing factor. Revenue growth is steady across every window. What changes is how much of it companies keep.
- This is not a forecast. A margin that widened for a decade can narrow, and it would pull earnings down without any change in the multiple.
How Market Return Drivers Works
- 1Start from an identity rather than a modelA company's market value equals its revenue, times the share of revenue it keeps as profit, times what the market pays per dollar of profit. That is arithmetic rather than theory: the three terms multiply back to market value exactly, with no residual to explain away.
- 2Aggregate the basket before dividingRevenue, earnings and market value are summed across the companies first, then divided. Every leg is computed over the companies whose revenue and earnings are both known on the date, so one basket sits under all three and the margin cannot be inflated by a mismatched denominator.
- 3Split the change in logsThe drivers compound rather than add, so their shares are taken as logs of the ratios. That lets a driver which moved backwards take a negative share while the others sum past one hundred, which is exactly what a falling multiple looks like and is hidden by any method forced to sum to one hundred.
- 4Check whether the gain is broadA widening margin can mean every company improved or a few enormous ones did. The share of companies earning anything at all over the trailing year is published beside it, which distinguishes the two.