The short workflow: (1) your broker LOCATES shares available to borrow, typically from another customer's margin account or an institutional lending pool; (2) the shares are borrowed and sold into the market at the current bid; (3) you pay a daily BORROW RATE -- a fee for the duration you keep the position; (4) when you close, you buy shares back in the open market and return them to the lender. If the lender recalls their shares before you are ready, you have to cover whether the thesis has played out or not.
| Term | What it is | Why it matters |
|---|---|---|
| Locate | The broker's confirmation that shares are available to borrow | Required under Reg SHO before a short sale can be entered |
| Borrow rate | Annualized fee paid daily to the share lender | Under 1% on easy names; 20-100%+ on hard-to-borrow names eats short returns |
| Hard to borrow | Names where lendable inventory is thin | Borrow rates spike; locates can be denied; recalls are common |
| Failure to deliver (FTD) | A trade that did not settle on the standard timeline (T+1 for US equities) | Persistent FTDs trigger Reg SHO close-out requirements |
| Buy-in | Forced repurchase of borrowed shares when the lender recalls or FTD persists | Buy-ins happen at market price, not your chosen exit -- often near the worst possible price |
Regulation SHO is the rulebook that governs short selling on US equity markets. The two pieces most relevant to retail: (1) Rule 203 requires a locate before any short sale, with limited exceptions for bona fide market making; (2) Rule 204 requires brokers to close out failures to deliver within a defined window (typically T+3 after the original settlement date). Persistent FTDs land a stock on the Reg SHO threshold list, which itself signals to the market that borrow has become scarce.
Settlement matters. US equity trades settle T+1 (one business day after trade date) as of May 2024. Short sales must DELIVER actual shares by settlement -- the borrowed shares are how that obligation is met. When the locate was sloppy or the borrow disappeared mid-trade, the seller's broker cannot deliver on time and the trade fails. Settlement failures are not theoretical: persistent FTD volumes on certain hard-to-borrow names have triggered multiple SEC enforcement actions over the past two decades.
The cost stack of shorting is asymmetric vs going long. A long position can lose at most 100% of its capital and has no carry cost beyond the time-value of money. A short position has UNBOUNDED upside loss (a stock can theoretically rise without limit), pays borrow daily, and can be force-closed via recall or buy-in at the worst possible moment. The mechanical setup favors the long side; the case for any short has to clear that headwind.
Sit with the ideas.
Saltmarsh Capital wants to short 5,000 shares of a small-cap company they believe is overvalued. The broker confirms a locate at a 28% annualized borrow rate and the trade fills at $40 per share. Their thesis assumes the stock will drop 25% over the next four months. Twelve weeks in, the stock is flat at $40 -- thesis has not played out -- and the lender recalls the shares because the institutional holder is unwinding their position. Saltmarsh has 24 hours to either find a new locate or buy in. New locates on the name are unavailable. What does the disciplined post-mortem identify as the structural cost the trade was always carrying, before any thesis question?