Chapter 2: Where stock returns actually come from
From 1926 through 2016, the US stock market created about $16 trillion of wealth beyond what plain Treasury bills would have paid. Half of that entire mountain came from roughly 86 companies. Hold those two numbers gently for a few pages, because by the end of this chapter they will change what you expect from any single stock you ever buy.
First, the machinery. A stock pays you through exactly three pipes, and once you can name them, most market commentary becomes much easier to filter.
The three pipes
A share's return over any stretch of time comes from three sources and no others.
The first is earnings growth. Earnings are the company's profit, and earnings per share (EPS) is that profit divided by the share count: the profit your one slice commands. When a company earns more per share, your claim is worth more, in the most literal sense.
The second is dividends, cash the board pays out to shareholders, usually quarterly. Dividends look small in any single year and turn out to be enormous over decades, especially when reinvested into more shares that pay their own dividends. Chapter 6 gives them, and their twin the buyback, a full chapter.
The third is the multiple: what the market pays for each dollar of annual earnings, most commonly written as the price-to-earnings ratio, or P/E. A stock earning $5 per share and trading at $100 carries a multiple of 20. The multiple is the market's mood about the future, and it moves, sometimes violently. Chapter 5 teaches you to read it.
Price equals earnings per share times the multiple. That one sentence does a lot of work, because it means every price change splits cleanly into an earnings part and a mood part, with dividends collected on the side. Trace those three on a real holding and you know precisely where your money came from.
A decade of Harbor Foods
So trace them we will. Harbor Foods is a fictional packaged-food company with conveniently round numbers. You buy one share for $100: it earns $5.00 per share and trades at a multiple of 20.
A decade passes. The business performs well: earnings double to $10.00 per share, which works out to growth of about 7.2% a year. The board pays out 40% of each year's earnings as dividends along the way, and those checks total $29.87 per share over the ten years. The market's mood, though, cools: investors now pay 16 times earnings instead of 20. The share trades at $10.00 times 16, which is $160.
| Piece | What happened | Dollars per share |
|---|---|---|
| You pay | EPS of $5.00 at a multiple of 20 | $100.00 |
| Earnings growth | EPS doubles to $10.00; at the old multiple of 20, the share would sit at $200 | +$100.00 |
| Multiple change | the market pays 16 per dollar of earnings instead of 20, so $160 instead of $200 | -$40.00 |
| Dividends collected | ten years of payouts at 40% of each year's earnings | +$29.87 |
| What you end with | $160.00 share price plus $29.87 in cash | $189.87 |
Check the sum: $100.00 plus $100.00, minus $40.00, plus $29.87 equals $189.87. Your $100 became $189.87 in ten years, a total return of about 6.6% a year.
Sit with the shape of that result. The company did its job superbly, compounding earnings at 7.2% and mailing you cash the whole time, yet you earned 6.6%, because the multiple quietly took $40 off the table. Had the mood improved from 20 to 24 instead, the same business would have handed you $240 plus dividends, and you might have congratulated yourself on your stock-picking genius. Earnings and dividends are the engine, and they compound. The multiple is borrowed money from the future's opinion: when it expands, you are spending return that some later holder gives up, and when it contracts, you are the later holder.
The same decomposition runs on any stock and any period, and the dividend piece matters far more than most price charts admit, because price charts simply leave it out. Try it yourself on $10,000, with the dividend reinvestment toggled on and off. The calculator states its assumptions inside the widget.
The century record, with its haircut
Zoom all the way out and the engine's long-run output becomes visible. From 1926 through 2025, large US stocks returned roughly 10% a year before inflation. Inflation ate about three percentage points of that, leaving something near 7% a year in real purchasing power. Both numbers describe one specific century in one specific country, and neither is a promise about your next decade. They are, however, the most honest baseline available, and the Personal Finance Guide already built your plan on the conservative side of them.
One detail of the record deserves a flag. For most of that century, dividends supplied a large share of the compounding, with yields often in the 3% to 5% range. The S&P 500's dividend yield in 2026 sits near 1.2% to 1.5%, low by historical standards, which means future returns lean harder on earnings growth than your grandparents' market did. Nobody knows whether that makes the next decades better or worse; it does make the third pipe, the multiple, a larger share of short-run noise.
Elaine inherited 800 shares of a utility her father bought in the 1980s and held through every scare since. Family legend says the stock "always came through." Curious after reading an early draft of this chapter, she pulled the records and ran the decomposition. The share price itself had grown a little over 2% a year for two decades, barely ahead of inflation. The position had compounded at closer to 6%, because every quarterly dividend had been reinvested automatically, growing the holding from 800 shares to just over 1,560 without anyone adding a dollar. (Numbers simplified for the illustration.) Her father's genius, it turned out, was less about the utility he picked and more about the checks he never spent.
The part nobody tells beginners
Now return to the two numbers from the top of the chapter. They come from a study by finance professor Hendrik Bessembinder with a dry title and an explosive finding: "Do Stocks Outperform Treasury Bills?"
From 1926 to 2016, roughly 4% of listed US companies account for the entire net wealth the stock market created above Treasury bills.
About 86 companies produced half of the $16 trillion in net wealth the market created over those 90 years.
The majority of individual stocks returned less over their lifetimes than money parked in Treasury bills.
Read that middle stat again, slowly. The market's famous 10% average is real, and at the same time it is wildly unrepresentative of the typical stock. The average gets hauled upward by a thin tail of colossal winners, while most individual companies fade, stall, get acquired cheaply, or die. Buying one stock and expecting "the market average" from it is like meeting one random person and expecting the average human income: the average includes a handful of billionaires who bend the math.
The quiet miracle of an index fund follows directly. Owning everything guarantees you own the 4%, whichever companies turn out to be in it, and that is why the foothills on Chapter 1's trail map climb so reliably while individual ridge paths so often dead-end. None of this forbids owning single stocks. It sets the base rate you size them against, and Chapters 8 and 13 build the sizing rules on exactly this foundation.
Expect a stock's long-run return to come from its earnings and its dividends. Treat any return that arrived through a rising multiple as borrowed from the future, and never plan on one company being "average": the average is dragged up by a tail of giant winners that most stocks never join.
Where people go wrong
- Extrapolating the mood. A stock that returned 25% a year while its multiple tripled did most of that with borrowed return. Decompose before you celebrate, and especially before you buy more.
- Judging by price charts. A flat price chart on a 4%-yielding stock hides a respectable return; Elaine almost sold a perfectly good holding because of one.
- Expecting the average from one ticket. The 10% historical figure belongs to the whole market, not to your single pick. Most picks did worse than T-bills; a few did spectacularly better. That skew is the entire game.
- Forgetting the inflation haircut. Long-range plans built on 10% instead of about 7% real quietly assume inflation went away. It has not.
Key takeaways
- Every stock return decomposes into three pipes: earnings growth, dividends, and the change in the multiple. Price equals EPS times the multiple.
- In the worked decade, Harbor Foods grew earnings 7.2% a year and paid $29.87 in dividends, yet returned 6.6% a year because the multiple slid from 20 to 16. The mood giveth and the mood taketh away.
- US large-cap stocks returned roughly 10% a year nominal, about 7% real, from 1926 through 2025. Historical, period-specific, and never a promise.
- Bessembinder's finding: 4% of companies created all net wealth above T-bills from 1926 to 2016, about 86 stocks produced half of it, and most stocks lost to T-bills over their lifetimes.
- Owning the whole market guarantees you own the winning tail. Owning one stock is a bet on joining it, and should be sized like one.
Sources: Bessembinder, "Do Stocks Outperform Treasury Bills?" (SSRN) · Bessembinder research summary (ASU W. P. Carey) · Investor.gov: Introduction to investing · Finvest Personal Finance Guide