Borrow Rates Explained: Why Your Short Position Bleeds Overnight
A borrow rate is the annualized cost to borrow shares for shorting. Hard-to-borrow names can cost 50-200% per year. Daily carry is position-value times rate divided by 365 — and most paper trading platforms charge you nothing.
When you short a stock, you don't own the shares — you've borrowed them from someone who does, sold them, and now owe them back. The lender charges you for the privilege. That charge is the borrow rate, expressed as an annualized percentage, and it accrues every calendar day you hold the position open.
For easy-to-borrow names — large-cap stocks with deep stock-loan inventories at every prime broker — the borrow rate is small. Sub-1% annualized, often well below. The cost of holding a short for two weeks is rounding error.
For hard-to-borrow names — small-caps with tight floats, names in active squeeze setups, post-IPO lockup-expiry candidates — the borrow rate is enormous. Fifty percent annualized is common. Two hundred percent isn't unheard of. Five hundred percent has happened.
A paper-trading platform that doesn't model this teaches you nothing about short economics. You hold a paper short for two weeks, you congratulate yourself on the trade, you don't realize that the live equivalent of the same hold would have cost you more in borrow than the trade made in P&L.
What a borrow rate actually represents
The borrow rate is the price set by the supply and demand of lendable shares. On the supply side, long holders of the stock (retail accounts that opted into share-lending programs, ETF baskets, hedge funds with the position) make their shares available to be lent. On the demand side, short sellers need to borrow those shares to enter or maintain a short position. The clearing rate that matches supply to demand is the borrow rate.
For stocks where lendable supply vastly exceeds short demand (the median large-cap), the rate clears at near-zero — fractions of a percent. For stocks where short demand outstrips supply (the typical short-squeeze candidate), the rate clears at whatever level the marginal demanding short is willing to pay. In acute squeezes, that level can climb fast.
The rate is set by the broker's stock-loan desk in real time. Different brokers have different lendable inventories and different rate-setting policies, so the same stock can have meaningfully different borrow rates at different brokers on the same day. Interactive Brokers is the most-cited retail and small-prop benchmark because IBKR publishes its borrow availability and rates publicly through (iborrowdesk.com), which is where most retail short-research tools — including Tapeboard's squeeze score — get their numbers.
How IBKR sources the rate
IBKR runs its own stock-loan desk. The desk maintains an inventory of lendable shares accumulated from IBKR customers who've opted into the Yield Enhancement Program (or its equivalent), plus inventory borrowed from other lenders in the institutional stock-loan market. The desk posts available shares and the rate to borrow them, updated through the trading day as supply and demand shift.
Other brokers do the same with their own inventories. Goldman, JPM, and Morgan Stanley prime brokers post rates that institutional shorts see; those rates can differ from IBKR's by tens or hundreds of percentage points on hard-to-borrow names. For retail purposes, the IBKR rate is a reasonable proxy — most retail brokers either route to IBKR for their stock-loan or set rates that approximate IBKR's plus a markup.
Daily carry math
The actual cost charged against an open short position each day is straightforward:
dailyCarry = positionValue * annualizedRate / 365Where positionValue is the current mark-to-market value of the short position (share count times the current mark price), and annualizedRate is the published borrow rate expressed as a decimal.
Worked example. You short 1,000 shares of a $4 stock at an annualized borrow rate of 47.5%. The position value is $4,000. The daily carry is:
dailyCarry = 4000 * 0.475 / 365
= 4000 * 0.001301
= $5.21Hold the position for 14 calendar days and you've paid $72.94 in carry. If the stock didn't move at all, your "flat" trade actually cost you $73. If the stock dropped 2% (from $4.00 to $3.92), your gross gain was $80 — minus $73 in carry equals a net of $7 on a 1,000-share short held for two weeks. Two weeks of capital tied up to make $7.
Worked example, painful version. You short 1,000 shares of a small-cap in active squeeze, marked at $3.50, at an annualized borrow rate of 250%. The position value is $3,500. The daily carry is:
dailyCarry = 3500 * 2.50 / 365
= $23.97Three weeks of holding costs you $503 — about 14% of the position's value, charged in borrow alone, before any market move. For the trade to break even you need the stock to drop 14%. To make a meaningful profit you need it to drop 20%+. And the borrow rate may go up while you hold, not down.
This is why hard-to-borrow shorts are not patient trades. The clock is running against you in real dollars every day. A short that "doesn't work fast" is functionally a losing short even if the price hasn't moved.
Why paper platforms that don't model borrow are dangerous
A paper trader who shorts on a platform that charges no borrow cost learns three wrong things.
Wrong lesson 1: shorts are free to hold. Paper short for two weeks waiting for a thesis to play out costs nothing in the simulator. Live, that same hold could cost more than the eventual profit. The trader learns to be patient on shorts when patience is the most expensive choice available.
Wrong lesson 2: hard-to-borrow names look profitable. The names retail short-squeeze speculators are drawn to — high SI, high borrow fee, low float — are exactly the names where the borrow cost would have eaten most of the profit. A paper platform that ignores this makes squeeze-shorting look like a viable strategy when it usually isn't, except for very fast in-and-out trades.
Wrong lesson 3: position sizing on shorts is the same math as longs. It isn't. A 1% account position on a long stock has no daily carry; a 1% account position on a short with a 100% borrow rate accrues 0.27% per trading day in carry. Over a month, that's roughly 8% of the position's value — meaningfully different risk math than the long-side equivalent.
The Tapeboard paper-trading simulator charges daily borrow carry on every open short position using the same IBKR feed that powers the squeeze score. The methodology is documented at /methodology/short-borrow. Hard-to-borrow names cost what they cost in the sim. Easy-to-borrow names cost the median. Paper shorts feel the same time decay live shorts feel.
What the simulator can't replicate
Three honest limitations of borrow modeling in any simulator.
No intraday tightening. Real borrow rates can spike intraday during forced recalls or acute squeeze days. The published rate goes from 40% to 200% in the space of hours. Tapeboard refreshes the rate once nightly, so an intraday tightening won't show up until the next ingest. Paper traders won't feel the same recall pressure live traders feel during fast events.
No genuine hard-to-borrow recall risk. In real life, a hard-to-borrow short can be recalled — the lender demands the shares back, and the short must either find new shares at a (typically much higher) rate or close the position involuntarily. Simulators don't model the recall event. The position stays open; only the carry is charged.
No prime-broker rate variance. IBKR is one rate source. Real institutional shorts may see different rates at GS, JPM, MS. The simulator uses the IBKR rate exclusively; if you live-trade through a prime-broker relationship with different rates, the sim's carry math won't match your live experience exactly.
These are real limitations, documented in the methodology, and they're the gap between any simulator and the live tape. Modeling the median carry honestly is still meaningfully better than modeling no carry at all.
What this means for your trading
Three takeaways.
If you're new to shorting, paper-trade on a platform that charges borrow carry. Without it, you'll build habits that don't transfer to live shorting and you'll be surprised — expensively — when your first live short on a hard-to-borrow name doesn't behave like the paper version did.
If you trade short-squeeze setups, the borrow rate is your single most important variable after the price. The Tapeboard short-squeeze leaderboard ranks names by composite score that includes the borrow rate (25% of the weighting). High squeeze score with high borrow means structural shorting pressure with high carry cost — a setup that either resolves fast or eats the short alive.
If you size shorts the same way you size longs, recompute. A short position on a hard-to-borrow name needs to be sized smaller or held shorter than the equivalent long position, because the daily carry is a hidden tax that compounds against you. The math is simple, but the discipline of doing it before every short trade is the part most retail short-traders skip.
Trade shorts knowing what they cost. Trade them on a simulator that charges you that cost in paper. The transition to live will not surprise you.