What Is Short Selling? How to Bet on a Stock Falling

You can sell something you don't own. That sentence bothers most people the first time they meet it, and it should. It's the strange part, and everything else about short selling follows from it.

Here's the ordinary version first. You think a company is going up, so you buy its shares, wait, and sell them higher. Buy low, sell high, in that order.

Short selling does the same two things in the other order. Sell high first, buy low afterwards. The only problem is that selling requires shares, and you haven't got any. So you borrow them from someone who has.

That's the whole idea. What's left is what borrowing actually involves, what it costs, and why this trade can lose you more money than you put into it.

What short selling actually is

Four steps, in order.

1. 1Borrow. Your broker finds 100 shares belonging to somebody else and lends them to you. A fee starts running that day.
2. 2Sell. You sell all 100 at today's price. The cash lands in your account.
3. 3Buy back. Later — a day, a month, whenever you decide — you buy 100 shares at whatever they cost then. Traders call this covering.
4. 4Return. The 100 shares go back to the lender. Whatever cash is left over is yours.

Now notice what you owe in the middle of that. You don't owe money. You owe shares — 100 of them, and it stays 100 whatever they're worth. Your debt is measured in shares, and shares get repriced every second the market is open.

That's the difference between this and every other trade you've done. A normal debt sits still while you pay it off. This one moves.

Where the shares come from

Most explainers give this one line and move on. It's the part that decides how the trade actually goes.

The shares come from another investor. Usually they're a customer of your own broker — somebody who owns the stock and holds it in a margin account. The agreement they signed when they opened that account lets the broker lend their shares out, and the great majority of them have no idea it happens. They still own the position, and every cent it gains or loses is still theirs. The actual shares are just somewhere else for a while.

You pay rent on them. The borrow fee is quoted as an annual percentage of the borrowed stock's value and charged by the day, and the range is enormous. A large company with plenty of shares sitting in lendable accounts might cost a fraction of a percent a year. A stock that everyone wants to short and nobody wants to lend is called hard to borrow, and the rate on those runs into tens of percent. It isn't fixed, either — your broker can reprice it daily, and it tends to jump at exactly the moment your idea starts working and everybody else crowds in behind you.

Then there's the part with no equivalent anywhere on the long side. The lender can sell their shares whenever they feel like it. They never agreed to hold them for you, they don't know you exist, and they don't need your permission. When they sell, your broker has to find replacement shares from a different lender. Often it can. Sometimes it can't.

When it can't, you get a buy-in. The broker closes your position for you, at the market price, on its own schedule. Your view on the company doesn't come into it. You can be completely right, sitting on a position that's about to pay, and be bought in on a Tuesday morning because a stranger decided to sell.

You can be right and still be finished. The shares were never yours, and the person they belong to can ask for them back.

A worked example, all the way through

One stock, one set of numbers, carried to the end.

You borrow 100 shares of a stock trading at $80 and sell them. That raises $8,000, and it goes into your account. It isn't spending money — it's collateral against the 100 shares you now owe.

Your broker wants more than that. Under Reg T the account has to hold 150% of the short's value, and the sale proceeds cover the first 100%, so you put up the other 50% yourself: $4,000. The account now holds $12,000 against a debt of 100 shares.

Look at those two numbers side by side, because the whole trade lives in the gap between them. The $12,000 doesn't move. The debt does.

Here's every outcome that matters. Your equity is the $12,000 minus what the 100 shares cost that day, and your profit or loss is that equity against the $4,000 you started with.

The stock falls to $60 and it works. Buying back costs $6,000 against the $8,000 you sold for, so you're up $2,000. That's 50% on the $4,000 you put in, from a 25% fall in the share price. The gearing is doing that, and it works in both directions — our leverage guide has the machinery.

The stock does nothing and you're flat at $80. Except you're not, quite, because the borrow fee has been running the whole time.

The stock reaches $120 and your $4,000 is gone. Not damaged — gone, all of it, at a price the stock reached by rising 50%. A 50% rise is not a freak event. Plenty of ordinary companies do it in a year.

The stock reaches $160 and you're past zero. Buying the shares back costs $16,000 and the account holds $12,000, so you're $4,000 short and the broker wants it. The stock has doubled. That's all it did.

Drag the price below and watch the two bars move. The debt bar grows while the equity bar shrinks, which is the thing that makes a short different: the position gets bigger as it loses.

The margin, and the price that ends it

The $92.31 in that table isn't rounded off from anything and it isn't the broker's mood. Here's where it comes from.

FINRA Rule 4210 sets the maintenance requirement on a short position at 30% of what the stock is currently worth. (It's the greater of that or $5 a share; at $80 the 30% is the one that binds.) So the account has to keep equity of at least 30% of the debt, all day, every day. Drop below and you get a margin call: put in more money, or the position gets closed.

Write both sides down. Your equity is the $12,000 credit minus what the shares cost, which is 12,000 − 100P. The requirement is 30% of what they cost, which is 30P. The call fires when the first drops under the second.

12,000 − 100P < 30P. Add the 100P to both sides and you get 12,000 < 130P, so P > $92.31. From $80, that's a rise of 15.38%.

Now do it again without the $80 in it. The account credit is always the proceeds plus half of them, so it's 1.5 × shares × entry price. The requirement is always 1.3 × shares × current price. Set them equal and the share count cancels off both sides, and so do the dollars: current price = entry price × 1.5 ÷ 1.3.

Sit with that for a second, because it's the most useful thing on this page. A short doesn't get a wider berth because the stock is expensive, or a narrower one because it's cheap. The cushion is always the same 15.4%. That's one earnings report. It's one takeover rumour. On a volatile stock it's a Tuesday.

One exception, and it only bites down at the bottom of the market. The $5-a-share floor takes over from the 30% rule on very cheap stocks, and below about a $14.44 entry the call comes sooner than 15.4%, not later. Penny stocks are the place where shorting is most tempting and the margin rules are least forgiving.

Notice what the call is not, though. At $92.31 you've lost $1,231 of your $4,000 — under a third of it. The call arrives well before the real damage, which is the system working as designed. Whether you can find the cash to meet it is a separate question, and it's the question that decides how the next section goes.

Why the loss has no floor

Buy 100 shares at $80 and your worst case is a number you can write down on day one. The company fails, the shares go to zero, you lose $8,000. It's an awful outcome and it's a known one.

Sell the same 100 shares short and there's no equivalent number to write down. Your loss is 100 × (price − $80), and the price has no ceiling. At $160 you're down $8,000 — the buyer's entire worst case, reached by the stock merely doubling. At $240 you're down $16,000. Stocks double. Some of them double in a week.

So the two sides of the same trade aren't mirror images of each other. The buyer's loss stops at what they paid. The seller's loss keeps going for as long as the stock does.

There's a second thing tucked inside that, and it's the one people miss. As the stock rises, your debt gets bigger. A losing long position shrinks — the money at risk falls along with the price. A losing short grows. You end up with more exposure exactly when you can least afford it, and nobody had to increase your position for that to happen.

The squeeze

A short squeeze is the thing short sellers are actually afraid of, and it falls straight out of the margin arithmetic above. Nothing mysterious happens in it.

Take a stock that a lot of people are short. The price rises, for any reason at all. Equity in all those accounts falls at once, some of them cross the maintenance line, and their brokers demand cash or close them out. Closing a short means buying the stock. That buying pushes the price up. Which pushes more accounts over the line. Which produces more buying.

The crucial step is the third one, because it doesn't ask anybody's opinion. A forced buy-in is an order sent by a broker protecting itself, and it doesn't care what the shares are worth or what anyone thinks of the company.

This is a different machine from a stop-loss sweep, which our guide to stop hunting covers. A stop is an order you placed, and you can move it or cancel it. A margin call is a rule your broker enforces on your account, and there's no clever version of the position that avoids it.

The famous example is GameStop in January 2021, and it's worth telling accurately, because the popular version isn't what the regulators found when they looked.

Start with the facts. On 22 January 2021, short interest in GameStop was around 140% of the company's public float. The stock went from about $17 at the start of that month to an intraday high of $483 on 28 January.

That 140% looks impossible and isn't. A borrowed share, once sold, belongs to whoever bought it — and their broker can lend it out again to the next short seller. One share ends up supporting two short positions, so short interest can exceed the number of shares that exist.

Now the part that gets left out. The SEC's staff report on that period looked directly at whether short sellers buying to cover drove the price up, and found that it wasn't what sustained the rise. Covering was part of the early move. What kept the stock climbing after that was positive sentiment — people buying because they wanted to own it, not because a broker made them.

So hold two things at once, and don't collapse them. The loop above is real. It happens, and it's the mechanism that ends short positions badly. And the single most famous episode people cite for it turned out to be mostly something else. If you learn short selling from the GameStop story as it's usually told, you'll walk away with a correct mechanism attached to the wrong evidence.

The practical lesson from January 2021 is smaller and more useful anyway. A crowded short is one where short interest is large against the float. It's a position where a lot of people need to buy the same stock at the same time if it moves against them, and that's a fragile place to stand whatever eventually lights the fuse.

What it costs to keep the position on

A long position costs nothing to hold. Buy the shares and you can ignore them for a decade. A short position has a meter on it, and there are four of them.

Run the first one against your own stake. $240 a year on the $4,000 you put up is 6% a year, before the stock has done anything at all. That's the hurdle a flat share price sets you.

Put the four together and short selling has a clock on it that buying simply doesn't have. An investor who's early can be wrong for two years and then right, and the wait costs them nothing. A short seller pays rent for those two years, and can be recalled out of the trade halfway through them.

The capped-risk alternative

There's a way to bet on a fall where the worst case is a number you can write down in advance: buy a put option. A put gives you the right to sell the stock at a fixed price, and the most it can cost you is what you paid for it. No borrowing, no fee running daily, no recall, no margin call, no uncapped loss. What you give up is time — the put expires on a date, so a fall that arrives late pays you nothing. (A CFD is a third route to the same direction, with the broker as your counterparty rather than a lender; our CFD guide covers what that changes.) Our options guide has how puts are priced and what the real trade-offs are.

Who actually does this, and the rules

Short selling isn't mostly individuals betting against companies they dislike. It's hedge funds running long and short positions side by side, market makers hedging inventory they never chose to hold, and index arbitrage. A lot of it isn't a bet on a company failing at all — it's one leg of a position whose other leg is long something else.

There's also a case for it that rarely gets made. Short sellers are the only people in the market who get paid for finding out that something is worse than it looks. Several large frauds were named in public by short sellers before any regulator got there. That doesn't make it a public service. It does explain why regulators have kept declining to ban it, despite being asked to, repeatedly.

In the US the rules sit mostly in Regulation SHO, and three of them reach you directly.

• The locate. Before your broker executes a short sale, it has to have reasonable grounds to believe the shares can be borrowed and delivered on time. That's the rule that stops naked short selling — selling shares nobody has arranged to borrow.
• The close-out. If delivery fails anyway, the broker has to buy the shares in within a set deadline. Persistent fails get force-closed, and the stock gets restricted until they clear.
• The circuit breaker, Rule 201. If a stock falls 10% from the previous day's close, short sales in it can only execute above the current national best bid for the rest of that day and the next. You can still short it. You just can't hit the bid on the way down.

Outside the US the details differ and the shape doesn't: disclosure of large short positions to the regulator, some form of locate requirement, and the occasional emergency ban when a market is falling fast. Check your own regulator before assuming any of the above applies to you.