Short Selling Explained: Borrow, Locate, Squeeze and Unlimited Risk

Short sellers talk about theses. The desk talks about the loan. Everything difficult in a short position comes out of the fact that you sold something you had to borrow first: the fee that accrues every night, the lender who can ask for the stock back on a morning that suits them and not you, the regulator’s locate requirement, the squeeze. Get the loan right and the rest of the trade is ordinary. Get it wrong and the direction of the stock stops being the thing that decides your result.

Market microstructure applies to shorts the way it applies to anything else, with one addition. Here your counterparty list includes the person who lent you the shares.

The loan is the trade

Your broker locates stock to borrow, usually out of other customers’ margin accounts or from an institutional lending desk. Those shares are delivered to whoever bought from you. The sale proceeds stay in your account as collateral against the loan.

You now owe shares, not dollars. That distinction does the work. If the company pays a dividend while you are short, you owe it to the lender, because the person holding your borrowed shares is collecting the real one. You close by buying the stock back and handing it over, and the loan disappears.

Two features of the arrangement deserve more respect than they get. The borrow is callable, so the lender can demand the shares whenever they want, and a recall forces a buy in at whatever price the market happens to be showing that afternoon. And the fee floats. A position that cost 1 percent annualized to carry in one quarter can cost 40 percent in the next, with no notice, no negotiation, and no option other than returning the stock.

Locates, Reg SHO and the price you are allowed to sell at

Regulation SHO governs short sales in US equities. Before your broker accepts the order it must have reasonable grounds to believe the stock can be borrowed and delivered by settlement. That is the locate, and it is why a platform sometimes refuses a short in a small cap and tells you nothing beyond “not available.”

The same rule set forces closeout of failures to deliver once they persist past a defined window, which is the machinery that stops undelivered short sales from piling up. It also carries an alternative uptick restriction. Once a stock falls more than 10 percent from the prior close, short sales in that name are restricted to prices above the national best bid for the rest of that session and the next one.

The cost stack, and two outcomes worked

Cost	How it is charged	Rough magnitude
Borrow fee	Annualized percent of market value, accrued daily	Under 1 percent easy to borrow, 10 to 100 percent hard to borrow
Dividends	Paid through to the lender on the ex date	The full declared amount
Margin interest	On any debit balance in the account	Your broker’s schedule
Short rebate	Interest credited on the cash collateral, offsetting the above	Varies with the policy rate and with your agreement

Assume the rebate is zero unless your agreement says otherwise, because in a retail account it usually is.

Now put numbers on it. Short 1,000 shares at $50, so proceeds of $50,000. The borrow is quoted at 12 percent annualized and you hold for 60 days.

Borrow cost = 50,000 x 0.12 x 60/360 = $1,000

The thesis works and you cover at $42.

Gross gain = (50 - 42) x 1,000 = $8,000 Net of borrow = 8,000 - 1,000 = $7,000

A quarterly dividend of $0.30 a share went ex while you were short, so you owe 0.30 x 1,000 = $300 and keep $6,700. The borrow fee took an eighth of the gross gain in a trade that worked quickly, and holding the same short for a year at the same rate puts the fee at 50,000 x 0.12 = $6,000, which is most of the move you were hoping for.

The other direction. The stock rises and you cover at $65.

Gross loss = (65 - 50) x 1,000 = $15,000 With borrow = $16,000

Under Reg T you posted 50 percent initial margin, so $25,000 of your own money is in this, and a $16,000 loss is 64 percent of the capital you committed on a stock that rose 30 percent. Run your own combinations through the short selling calculator, and the account side through the margin calculator.

The margin call does the deciding

This is where shorts actually end, and it rarely feels like a decision.

Your account holds the $25,000 you deposited plus $50,000 of sale proceeds, a credit of $75,000. At $65 the liability is $65,000, so equity is 75,000 - 65,000 = $10,000. A typical maintenance requirement of 30 percent of market value demands 0.30 x 65,000 = $19,500.

You are $9,500 short of the requirement. Meet the call or the broker buys the stock back at the market, and your forced buying joins whatever was already pushing the price up. Nobody asks whether you still believe the thesis.

Short interest, and why the number is stale

FINRA collects short interest twice a month, at the mid month and end of month settlement dates, and publishes it after a lag. Two numbers come out of it.

Short interest against float measures crowding. An 18 million share short position against a 60 million share float is 18 / 60 = 30 percent, which is very high, and it tells you that nearly a third of the tradeable supply has already been sold by people who are contractually obliged to buy it back.

Days to cover divides short interest by average daily volume. With 3 million shares of daily turnover, 18,000,000 / 3,000,000 = 6 days. Six full sessions of normal volume consumed entirely by covering. That is the fuel load.

Both figures describe positions from roughly two weeks ago. Anyone quoting short interest as a read on what is happening today is quoting history, and the gap has been wide enough in fast moving names to reverse the conclusion entirely. Use it as structural background.

The squeeze, mechanically

Three inputs and a loop. Crowded positioning supplies buyers who have no choice. A limited float supplies scarcity. A catalyst starts the thing moving up.

Then the sequence runs on its own. The price rises, equity in short accounts falls, brokers issue calls, shorts buy to close, the buying lifts the price, and the next tier of accounts gets a call. Meanwhile the borrow fee spikes, because demand for the loan is rising while supply falls, so carrying the position gets most expensive exactly when it is going wrong. Lenders recall, forcing buy ins with no warning at all.

Call buying adds a fourth accelerant, since dealers hedging short calls have to buy stock, which options flow and dealer gamma covers in detail.

None of the four inputs requires anything to have changed about the company. That is the part traders refuse to accept, and it is why being right about the business is not a defense.

Ways to be bearish that cannot end you

Buying puts caps the loss at the premium and removes borrow and recall risk completely, at the cost of theta and the need to be right about timing as well as direction, which the options introduction covers.

Inverse ETFs give short exposure without a margin account, and they carry the daily reset problem set out in the leveraged and inverse ETF guide, which makes them a few day instrument at most.

A short inside a pair trade strips out market direction and leaves the relative call, which is a survivable bet in a way an outright short in a rising market is not.

And the outright short remains available to people who have accepted the distribution: stocks drift upward over long horizons, so you pay a structural headwind before anything else happens. Your best case is a hundred percent gain on a company going to zero, and your worst case has no number attached to it. A business can be deteriorating precisely as you diagnosed in the earnings quality guide while the stock doubles on a financing announcement and your borrow goes to 60 percent. The thesis, the timing and the funding all have to hold at the same time.

So keep every short small, write down the price at which you are wrong before you enter, and check the borrow rate weekly. Risk management for traders has the sizing framework that stops one short from ending an account.