What a Stock Actually Is (and Why the Market Keeps Mispricing It)

The Question Almost Nobody Can Answer

In January 2021, shares of a shrinking mall retailer went from $17.25 to an intraday high of $483 in eighteen trading days. The company sold used video game discs. Its industry had moved to digital downloads. Nothing about the business changed that month.

Ask most people who own stocks to explain how that's possible, and you'll get a shrug. Which is worth pausing on: millions of people own an asset they can't define, priced by a mechanism they can't describe.

So this is the first lesson, and we're going to do it properly. What a stock actually is, in plain numbers. And then the one framework that explains the market's apparent insanity - Benjamin Graham's voting machine and weighing machine - tested against the historical record instead of asserted.

We used Barebone to pull the price and earnings history behind every example in this piece: GameStop's squeeze, Meta's round trip, Cisco's quarter-century, and 98 years of S&P 500 holding periods. The receipts: +1,625% in one month for a business in decline. -64% in one year for a business whose profits fell 38%. And 25 years and 8 months for the price of a genuinely great company to get back to even.

All three numbers are correct. By the end you'll see why none of them contradict each other.

What One Share Actually Buys You

A stock is a fractional claim on a real business - its profits, its assets, its future. Not a ticker, not a line on a chart. A piece of the company.

Make it concrete with the largest example available. Apple has roughly 15 billion shares outstanding. Buy one and you own about one fifteen-billionth of the entire enterprise. That sounds homeopathic until you look at what the enterprise produced in fiscal 2025, per Apple's own Form 10-K:

What you hold	What it was worth in fiscal 2025
Ownership of the business	~1/15,000,000,000 of Apple
Your slice of $416.2B in sales	about $27.74
Your slice of $112.0B in net income	$7.46 (the diluted EPS)
Cash actually paid to you	$1.04 per year at the current $0.26 quarterly dividend
Your say in how it's run	one vote per share

That $7.46 is just net income divided by share count - which is all "earnings per share" means. Apple sends you $1.04 of it as dividends and reinvests or uses the rest (largely to buy back stock, which quietly grows your fractional slice each year without you doing anything).

This is the part of investing that is genuinely simple. The share is a claim on cash the business generates. If the business earns more over time, your claim is worth more over time.

The confusing part was never the stock. It's the market that prices it.

The Voting Machine

In 1934, in Security Analysis, Benjamin Graham and David Dodd wrote down the single most useful sentence about market behavior:

"...the market is not a weighing machine, on which the value of each issue is recorded by an exact and impersonal mechanism... Rather should we say that the market is a voting machine, whereon countless individuals register choices which are the product partly of reason and partly of emotion."

A stock's price on any given day is not a measurement. It's an election. Every buy order is a vote up, every sell order a vote down, and the voters include rational analysts, bored gamblers, panicking retirees, and algorithms reacting to other algorithms.

GameStop in January 2021 is the cleanest specimen of a pure vote on record. The setup: short sellers - investors who borrow shares and sell them, betting the price falls - had positions totaling roughly 140% of GameStop's available shares. A crowd organized on Reddit realized that if they bought aggressively, the shorts would be forced to buy back shares at any price to exit. The votes started stacking.

The tape, as reported at the time (GameStop later split its stock 4-for-1, so historical feeds show a quarter of these prices):

Date (2021)	Close	What happened
Jan 4	$17.25	A struggling retailer, heavily shorted
Jan 13	$31.40	The squeeze thesis spreads
Jan 26	$147.98	Volume in the hundreds of millions of shares
Jan 27	$347.51	+134.8% in a single day
Jan 28	$193.60	Intraday high $483; brokers restrict buying; -44.3%
Jan 29	$325.00	+67.9%; the month ends +1,625%
Feb 2	$90.00	-60% in a day, -81% from the intraday peak

At the top, the market briefly valued GameStop at $33.7 billion - about thirty times where it started the month - for a company whose core product was being replaced by downloads.

Here's the correction to the version of this story you usually hear: the price didn't move "for no reason." It moved for a very specific reason - a structural short squeeze meeting a coordinated crowd. What it didn't move on, in either direction, at any point, was the business. The voting machine ran a full election cycle, mania to crash, in five weeks, and the company's earnings power was never on the ballot.

That's the signature of voting-machine markets, and it's not limited to meme stocks. Narrative rallies, panic selloffs, sector manias - same mechanism, slower clock. In the short run, prices measure feelings.

The Weighing Machine

The script most people learn attributes the famous quote to Warren Buffett. The record is more precise, and the precision matters. Graham wrote the voting machine metaphor in 1934 - note that his original sentence says the market is not a weighing machine. The optimistic half, that it eventually becomes one, is Buffett's distillation, from his 1987 letter to Berkshire Hathaway shareholders:

"The market may ignore business success for a while, but eventually will confirm it. As Ben said: 'In the short run, the market is a voting machine but in the long run it is a weighing machine.'"

Strong claim. Does it survive contact with data?

Meta is the best modern stress test, because the market voted it almost dead. The company entered 2022 at $338.54. Apple's privacy changes were strangling its ad targeting, TikTok was taking attention share, and it was burning billions a year on the metaverse. By November 3, 2022, the stock closed at $88.91 - down 74% in ten months.

Now look at what the vote was supposedly measuring. Meta's diluted earnings per share fell from $13.77 in 2021 to $8.59 in 2022. A real deterioration: -38%. The price fell -64%. The vote didn't track the damage - it nearly doubled it, then kept going.

Then the weighing began. Meta cut costs hard through 2023, the ad business recovered, and earnings re-accelerated: $14.87 in 2023, $23.86 in 2024. The stock followed - +194% in 2023, +66% more in 2024.

Index both lines to 100 at the end of 2021 and the punchline writes itself. By the end of 2024, Meta's earnings index stood at 173.3. Its price index: 174.1. Three years of wild voting - a 74% crash, a 194% rebound - and the price landed within one point of exactly tracking the earnings.

Year	Year-end close	Diluted EPS	Price (2021 = 100)	EPS (2021 = 100)
2021	$336.35	$13.77	100.0	100.0
2022	$120.34	$8.59	35.8	62.4
2023	$353.96	$14.87	105.2	108.0
2024	$585.51	$23.86	174.1	173.3

The scale doesn't care how the election went. It weighs the cash.

The Scale Weighs Both Ways

Before you conclude "great company, hold forever, the weighing machine will pay me" - the machine has a second mode, and it's the one that ruins decades.

On March 27, 2000, Cisco closed at $80.06 and became the most valuable public company in the world, passing Microsoft. It was arguably the best business of its era - it sold the plumbing of the internet - and it traded at a price-to-earnings multiple above 200. Investors weren't just voting for Cisco's future; they were prepaying for several decades of it.

What happened next to the business: revenues roughly quintupled from 1999 levels, earnings per share grew about eightfold. The weighing machine fully confirmed Cisco's success.

What happened to the stock: it did not close above that March 2000 record until December 10, 2025 - $80.25, a round trip of 25 years and 8 months.

Cisco the company succeeded. Cisco the stock spent a quarter century paying down what people voted for it in March 2000. The weighing machine doesn't promise your stock goes up. It promises price converges toward business value - and if the vote put the price miles above the value, convergence is a long walk down. What you pay is part of the machine.

How Long Does a Vote Last?

So the framework has two gears. Reasonable question: where's the crossover? When does voting stop and weighing start?

The honest answer is that it's a dial, not a switch - and the dial is your holding period. Across the past 98 years of S&P 500 history (through the end of 2025), the share of rolling periods with positive returns climbs steadily as the horizon stretches:

Over any single year, the market was up only 74% of the time - one year in four, the vote went against you regardless of fundamentals. At five years, 88%. At ten years, 94%.

Jack Bogle, Vanguard's founder, had a clean way of explaining why. Decompose any stock return into three parts: the dividend yield, the growth in earnings, and the change in what investors will pay per dollar of earnings (the P/E multiple). He called the first two the investment return - the business itself. The third he called the speculative return - the vote. Over months, the speculative term dominates everything; over decades, it tends to wash out, leaving you with roughly what the businesses actually earned. Meta's three-year round trip to a one-point gap is Bogle's arithmetic in miniature.

The vote never fully disappears. It just stops being the main character.

Where the Framework Misleads

A framework you can't break is a slogan. Here is where this one fails, and each failure mode has a body count.

Value traps: the weight itself can shrink. "Price converges to value" cuts both ways when the value is falling. GameStop looked statistically cheap for years before 2021 - the scale kept reading lower because the business kept shrinking. Buying something cheap and waiting for the weighing machine only works if the earnings power survives the wait. The machine weighs what the company will earn, not what it used to.

Reflexivity: the vote can change the weight. The framework treats price and value as independent. George Soros built a career on the observation that they aren't. A soaring stock can sell new shares, raise cheap capital, attract talent, and acquire competitors - manufacturing real value out of enthusiasm. GameStop did precisely this, selling stock into the squeeze to rebuild its balance sheet. A collapsing price can equally force the cuts and panic that destroy the value it was supposedly mismeasuring. Sometimes the election rigs the scale.

There is no clock. "Markets can stay irrational longer than you can stay solvent" - a line usually pinned on John Maynard Keynes, though no one has found it in his writing (this lesson's second misattributed quote; check your citations). Cisco's weighing took 25 years. Anyone who needed that money in 2010 didn't experience a framework. They experienced a 60%-ish loss with a story attached.

You are the voting machine. Morningstar's Mind the Gap study measured what fund investors actually earned versus what their own funds returned: 6.3% per year versus 7.3% over the decade through 2023. The missing 1.1 points a year - roughly 15% of the total return - was lost to timing: buying after rallies, selling after crashes. Most people hold weighing-machine assets and trade them like ballots. The machine that costs you the most money is the one running between your ears.

What This Means

This is a framework for reading the market, not a buy list. Four things to take from the data:

Know which machine priced the stock you're looking at. When a price moves violently on no change in earnings power - GameStop in January 2021, Meta in late 2022 - that's a vote, and votes reverse. When price grinds along with earnings for years, that's the scale talking. The question to ask of any move is the one almost nobody asks: did the cash this business generates actually change, or just the mood?

If you're buying because the price is moving, you're voting - and your exit depends on the next voter showing up. Someone owns the right side of the GameStop chart. The crowd that piles in late and never knows when to leave is exactly who funds the people who did their homework.

The entry price is part of the weighing. Cisco is the permanent reminder that a great business at 200 times earnings can be a generation-long mistake. The scale eventually weighs both the company and what you paid for it.

Horizon is the one lever a beginner actually controls. You can't out-analyze the voting machine on day one. You can choose which game you're playing: 74% odds over one year, 94% over ten - before any skill at all.

Every trading day, the market runs both machines on the same screen and quotes you a single number per company. The number never tells you which machine produced it.

Knowing the difference is the entire game.

---

Data: Barebone | Sources: Graham & Dodd, Security Analysis (1934), Berkshire Hathaway 1987 shareholder letter, Apple FY2025 Form 10-K, Morningstar Mind the Gap (2024) | Data as of April 13, 2026