The Stock Market Isn't Up. It's Skewed.

An investor bought 10 shares of AMD in 2004 for $150. The stock rose 33% in six months. He sold at $20 per share, pocketed a neat profit, and learned a confident lesson: he picked a winner and took his money off the table.

The wrong part was not the pick. It was the exit.

Those same 10 shares, held through a decade when the stock traded underwater, a financial crisis, multiple turnarounds, and at least one period where nobody in the office would defend it at a dinner party — would be worth roughly $5,000 today. The $50 profit he celebrated in 2004 was 1% of what the stock eventually became. He sold the one thing that was doing the job and convinced himself he'd beaten the market.

He didn't sell the market. He sold the rare animal that is the market's return.

Here is the picture most investors carry around — and the part it deletes.

You've heard "stocks go up over time." You interpret that as: pick any stock, hold it long enough, and it appreciates. It feels true because the S&P 500 chart slopes up. The Dow has never permanently returned to a prior low. Your 401(k) statement gets bigger.

But the S&P 500 is not 500 stocks each climbing steadily. It is 10 or 20 stocks carrying the other 480. And if you look at individual stocks across their entire lifetimes — not indices, not hand-picked samples, every public company that ever traded — the picture is the opposite of what that upward slope suggests.

Over 100 years of U.S. stock market data, studied by finance professor Hendrik Bessembinder, the median individual stock didn't beat Treasury bills. It didn't beat a shoebox of cash under the mattress. Nearly 60% of all stocks underperformed the risk-free rate. Roughly 96% of all stocks combined barely matched T-bills.

And yet the market returned about 10% per year. How?

Because 4% of all stocks created every dollar of net market wealth. Just 86 stocks accounted for half of all wealth the U.S. market generated from 1926 to 2019.

The market doesn't go up because stocks appreciate. It goes up because a tiny handful of companies explode, and the index you own is structured to capture them.

Put away the acronym for thirty seconds. Think of it as a casino.

A casino has thousands of slot machines. Most players lose. Some break even. A few hit jackpots so large they cover every loss from every other machine and still leave the casino rich. The casino's profit doesn't come from most machines paying out. It comes from owning the entire floor so that when the jackpot hits, the casino keeps the spread.

Now you're a guest at the casino. You have two choices:

Choice A: Walk in, sit at one machine, play for a while, cash out when you're up $50, and leave feeling like a winner. You just sold your exposure before the machine could pay out big — or before it lost everything.

Choice B: Buy a piece of the casino itself. You now own a claim on every machine on the floor. You don't know which one will hit. Most won't. But when one does, your piece of the building captures it.

Choice A is picking individual stocks and selling them. Choice B is owning a broad index fund and holding it.

The casino works only if you stay in the building when the jackpot hits. You can't know which machine it'll be, and the machine that looks ordinary today might be the one that makes the casino profitable for the next decade.

Now label the props.

• The casino floor = the broad stock market, all public companies
• Each slot machine = one individual stock
• The jackpots = the 4% of companies that generate all net market wealth
• The player who walks away at $50 = the investor who sells a stock that's up a "good" amount
• Owning a piece of the building = owning a broad, low-cost index fund long-term

The mechanism that makes equity investing profitable is not that most companies grow your money. It's that you own a wide enough pool of companies for long enough that when the rare mega-winner appears, you're holding it. The index captures the winners because it includes every company. Individual stock-picking misses them because you can't know which 4% will be the 4%.

In the toy version, there are only three people and ten dollars.

Imagine 100 companies. Over 20 years: - 60 companies lose value. You put in $10 each, get back $4. - 35 companies return roughly what Treasury bills would have paid. You put in $10, get back $10. - 5 companies are extraordinary. Each turns $10 into $100,000.

If you picked 20 stocks at random, you'd likely own 12 losers, 7 break-evens, and maybe 1 or 2 winners. But if you sold that winner early — when it went from $10 to $50 instead of $100,000 — you just turned the best investment in the room into a pleasant anecdote.

Now replace the toy numbers with real data. The Bessembinder study looked at every U.S. stock from 1926 through the 2010s. The result: the median stock lifetime return was about 0.4% per year — worse than the 0.46% per year that one-month Treasury bills paid during that same period. But the average return was positive because a handful of mega-winners pulled the whole system up.

That analogy has now done its job. Here is where it breaks.

Casino jackpots are pure luck. Stock winners aren't. The companies that create outsized returns tend to have identifiable characteristics: they compound profits, reinvest at high returns, and survive competitive attacks that destroy their peers. That said, you still can't know which ones they'll be. AMD spent roughly a decade below its 2000 peak after the dot-com bubble. Anyone who sold during that decade wasn't wrong on any rational short-term basis. They were wrong only because they sold the machine that eventually paid out.

Also, real indices have fees, taxes, and concentration risk that the toy casino doesn't. And unlike a casino, you can lose money in the stock market over multi-year periods. The "stay in the building" rule requires that you survive the drawdowns without selling — which is the psychological part that separates the mechanism from the practice.

Bring the model back to the market.

Right now the top 10 stocks in the S&P 500 account for more than 40% of the entire index — higher concentration than during the dot-com bubble of 2000. That's not a separate story from the power law. It's the power law working in real time. A few companies are doing so much of the heavy lifting that "500 companies" is a label, not a description.

This changes how you should think about three common moves:

Selling winners early. The 33% gain on AMD in six months looked like a great result in isolation. But the mechanism of equity returns rewards the people who hold through the years when the stock does nothing, then explodes. If you've been trained to "take profits," ask yourself what profit level would have been high enough to sell Apple, Microsoft, Amazon, or Alphabet at any point in their history. The answer is almost always higher than what feels comfortable.

Picking individual stocks. The math doesn't forbid it. It just tells you that the vast majority of individual stocks underperform safe bonds over their lifetimes. If you pick stocks, you need a process that is better than randomly owning the casino. Most people don't have one, and most don't know they don't have one until they've been doing it for 15 years.

Owning a broad index. This is not a compromise strategy for people who "can't pick stocks." It is the strategy that explicitly exploits the power law. You own every machine on the floor. You can't miss the winner. You don't need to predict it. The cost is that you also own every loser — but the winners more than cover the losses.

The one number to remember: over the last century, the median stock underperformed Treasury bills. The market returned roughly 10% anyway. That gap — between what one stock does and what the market does — is the entire reason the index exists. It's not a backup plan. It's the mechanism.

If you remember one test, use this one: before you sell a stock that's up, ask whether you're selling your exposure to one of the rare companies that creates market wealth, or whether you're just collecting a fee for holding a position that the next owner will do the same math on. The stock doesn't know which question you're asking. The math doesn't care.