"Hard-to-borrow" (HTB) and "buy-in" describe the two ways the plumbing of the stock-loan market punishes short sellers when borrowable supply runs thin. Hard-to-borrow is a pricing/availability condition: a stock is in such short supply at lending desks that the borrow carries a meaningful (sometimes punitive) fee, or cannot be located at all. A buy-in is a forced liquidation: a short seller's position is closed out involuntarily — at the market's price, not the trader's — because the borrowed shares were recalled or a fail-to-deliver had to be cured. The core tension is that a short seller does not own a fully self-controlled position; it is a leveraged loan from a third party that can become more expensive, or be terminated, without warning, often precisely when the trade is going against you.

How it's formed

When you sell a stock short, your broker must first borrow the shares to deliver to the buyer. The loan terms depend on supply and demand in the securities-lending market, which splits names into two buckets:

• Easy-to-borrow (ETB) / "general collateral" (GC): ample supply. The borrow is essentially free or near-free; the lender even rebates most of the interest earned on the cash collateral back to the borrower. In D'Avolio's (2002) sample ~91% of loaned stocks were GC, with fees averaging only ~0.17%/yr.
• Hard-to-borrow / "special": scarce supply. The lender keeps more (or all) of the rebate and charges an explicit fee. In D'Avolio's sample ~9% of loaned stocks were specials, averaging ~4.30%/yr; in acute scarcity (squeezes, tiny floats) brokers have disclosed fees of 100%+ annualized.

Brokers publish an HTB list pre-market with an indicative borrow rate, then update it intraday as supply tightens or frees up. The fee is quoted as an annualized percentage of the position's market value and accrues every calendar day the short is held (not just trading days). Brokers such as Interactive Brokers and tastytrade document this daily accrual on settled positions.

A stock becomes HTB for predictable reasons: small float, heavy existing short interest, an active squeeze, an event (M&A, index rebalance, dividend record date), or recent IPO/SPAC mechanics. Two structurally related concepts live on sibling nodes — Short Interest & Days-to-Cover and Short Squeezes.

How buy-ins work

A buy-in arises from one of two triggers:

1. Recall. The beneficial owner who lent the shares sells them (or wants them back to vote/collect a dividend). The lending agent recalls the loan; if your broker cannot re-source the shares from another lender, your short is closed out — you are bought in at prevailing market prices. 2. Fail-to-deliver close-out under Reg SHO Rule 204. If shares are not delivered by settlement, the clearing participant must purchase or borrow to cure the fail. Per 17 CFR 242.204, the deadlines (in settlement days) are: a short-sale fail must be closed out by the open of the settlement day following settlement date; a long-sale fail by the third consecutive settlement day after settlement; and certain deemed-to-own / restriction-removal sales by the 35th consecutive calendar day after trade date. (Older trader shorthand of "T+3/T+4" predates the May 2024 move to T+1 settlement, which compressed these windows.)

The teeth of Rule 204 are in 204(b): a participant that fails to close out on time — and the brokers that clear through it — may not effect further short sales in that security without a pre-borrow (bona-fide borrow secured up front) until the fail is purchased and that purchase clears and settles. This "penalty box" is why brokers aggressively cure fails and pass buy-in risk to clients. The first step in the chain is covered on the Locating & Borrowing Shares sibling node.

How it's used in practice

For a short seller, HTB status and buy-in risk are cost-of-carry and tail-risk inputs, not signals to be traded on directly:

• Carry math. A 60% annualized borrow fee is ~0.16%/day. Hold a hard-to-borrow short for two weeks and the fee alone can eat several percent — before any adverse price move. The position must clear this hurdle to be profitable.
• Indicative ≠ guaranteed. The pre-market rate can multiply intraday. Squeeze names have been reported at 100%+ and, in extreme scarcity, several-hundred-percent annualized rates (per IBKR and broker disclosures). A short can be "right" on direction and still lose to carry.
• Buy-in defense. Experienced shorts avoid maximally crowded HTB names for size, keep dry powder for recalls, and treat a sudden borrow-rate spike or "no shares available" message as a warning that a forced cover may be near.

A high and rising borrow fee is itself information: it confirms heavy short crowding and thin float, the same preconditions a squeeze needs.

Adoption, debate & evidence

The mechanics are not contested — they are codified in Reg SHO and the daily operations of every prime broker and clearing firm. What's debated is the predictive value of borrow data. D'Avolio (2002), "The Market for Borrowing Stock" (JFE), is the foundational descriptive study: using ~18 months of one intermediary's data it found ~91% of loaned stocks are general collateral (fees averaging ~0.17%/yr) while ~9% are "specials" (averaging ~4.30%/yr), with scarcity rising in divergence of opinion — but it is not the source of the predictive-return claim. The predictive finding comes from later work: Cohen, Diether & Malloy (2007) show that increases in shorting demand precede negative abnormal returns, and Engelberg, Evans, Leonard, Reed & Ringgenberg ("The Loan Fee Anomaly," Management Science) report that the loan-fee signal is among the strongest cross-sectional return predictors tested. So high borrow fees tend to precede underperformance — a reasonably robust short-side anomaly — though magnitude estimates vary by study and sample. The folklore that retail can reliably trade off broker HTB lists is far weaker: retail lists are delayed, indicative, and the actionable cost (the fee) often already reflects the crowding. So the signal has academic support; the easy retail edge does not.

Strengths & limitations

Strengths. Borrow fee and availability are a real, market-priced gauge of short crowding and float scarcity, and the academic link between high loan fees and subsequent underperformance (Cohen-Diether-Malloy; the Loan Fee Anomaly literature) gives the data genuine analytical weight. Reg SHO close-out rules also meaningfully reduce naked-short fails versus the pre-2008 regime.

Limitations. Borrow is a moving cost — the largest hidden risk in shorting. A position can be involuntarily closed via recall or buy-in at the worst moment, removing the trader's exit control. The #1 misuse is ignoring carry: traders model price targets while treating a 30–100%+ annualized borrow as negligible, then bleed out on a thesis that was directionally fine. A close second is overconfidence in indicative rates that reprice violently intraday.

Sources

• SEC, Key Points About Regulation SHO — sec.gov/investor/pubs/regsho.htm (locate rule 203, close-out framework).
• 17 CFR § 242.204 (Cornell LII) — law.cornell.edu/cfr/text/17/242.204 (exact close-out deadlines and 204(b) pre-borrow penalty).
• SEC, Shortening the Securities Transaction Settlement Cycle (T+1, eff. May 2024) — sec.gov (Rule 204 timeframe compression).
• FINRA Rule 4320 — finra.org (non-reporting threshold-security close-out: 13 / 35 settlement days).
• Interactive Brokers, "The Risks of Shorting Series, Part II: Borrow Fees" and Short Sale Cost pages — interactivebrokers.com (fee accrual, rate ranges).
• tastytrade, "Hard-to-Borrow (HTB) Fees and Share Availability" — support.tastytrade.com (daily charging, indicative lists).
• General-collateral vs. special / negative rebate mechanics — securities-lending practitioner descriptions.
• D'Avolio (2002), "The Market for Borrowing Stock," Journal of Financial Economics 66 — descriptive lending-market facts (~91% GC at ~0.17%/yr; ~9% specials at ~4.30%/yr; recalls/specialness rise with divergence of opinion).
• Cohen, Diether & Malloy (2007), "Supply and Demand Shifts in the Shorting Market," Journal of Finance — shorting-demand increases precede negative abnormal returns.
• Engelberg, Evans, Leonard, Reed & Ringgenberg, "The Loan Fee Anomaly: A Short Seller's Best Ideas," Management Science — equity loan fees among the strongest cross-sectional return predictors. (Edge applies to the fee signal, not to trading delayed retail HTB lists.)

Dispute flag: close-out timeline shorthand varies in older sources (T+3/T+4) because they predate T+1 settlement; the regulatory text is framed in settlement days, used above. The "borrow fee predicts returns" finding is well-supported academically but is a cross-sectional anomaly, not a guaranteed per-trade edge.