Short selling is the act of profiting from a price decline: a trader borrows shares, sells them at today's price, and aims to buy them back later at a lower price to return to the lender, pocketing the difference. It is the structural mirror image of going long — but it is not symmetric in its risks. A long position's maximum loss is the capital invested (the stock goes to zero); a short position's maximum gain is capped at 100% (the stock can only fall to zero) while its loss is theoretically unbounded, because price can rise without limit. Layered on top of that asymmetry is a second one that defines this entire section: the short seller does not own a self-contained position. The shares are a recallable loan from a third party, re-priced daily, and the right to keep the position open can be revoked at any moment. This section is the domain of that plumbing — the borrowing, the costs, the crowding signals, and the forced-buying dynamics — that separate "being right on direction" from "making money on a short."

The core mechanic and its asymmetries

The lifecycle of a short has four steps: (1) the broker locates and borrows shares; (2) the trader sells short, with the cash proceeds held as collateral by the broker; (3) the position carries a daily borrow fee plus margin obligations while open; (4) the trader buys to cover — buying shares in the market to return to the lender and close the position. Per Wikipedia (Short finance)) and SEC guidance, a US short sale cannot legally execute unless the broker has first satisfied the locate requirement of Regulation SHO. The three asymmetries every short carries:

• Loss is unbounded, gain is capped at 100%. A stock shorted at $50 that runs to $200 is a 300% loss on one leg; the most the short could ever have made is $50/share.
• Margin is heavier than for longs. Shorting requires a margin account. Under Federal Reserve Regulation T, the initial requirement is 150% of the position value (the 100% short proceeds plus 50% additional margin). Under FINRA Rule 4210, maintenance margin for a short in a stock priced ≥$5 is the greater of $5/share or 30% of market value — versus 25% for a long (FINRA Rule 4210). A rising stock erodes equity and raises the requirement at once.
• You don't control your own exit. Recall and buy-in can force you out involuntarily, at the market's price — covered in the Hard-to-Borrow & Buy-Ins node.

The regulatory frame

Short selling is one of the most heavily regulated activities in the equity market. The two anchors:

• Regulation SHO (eff. Jan 2005). Its locate requirement (Rule 203) prevents executing a short without a borrow arranged, and its close-out rules (Rule 204) force prompt resolution of fails-to-deliver — both aimed at curbing abusive naked shorting (SEC). US settlement moved to T+1 in May 2024, compressing these timelines versus older T+2/T+3 references.
• Rule 201 — the "alternative uptick rule" (eff. 2010). A price-test circuit breaker: if a stock drops 10% or more from the prior close, short sales are restricted to prices above the national best bid for the rest of that day and the next, dampening downward pressure on already-falling names (SEC, FAQ Rule 201; Nasdaq short-sale circuit breaker). This replaced the original 1938 tick-test uptick rule, which was repealed in 2007.

The four sub-topics (and what each owns)

This section decomposes short-selling mechanics into four child nodes; each is a self-contained expert doc — point to them rather than re-deriving:

• Locating & Borrowing Shares — the supply chain that makes a short possible: the Reg SHO locate, the agent-lender/prime-broker stock-loan market, collateral (~102% in the US), and how the borrow is priced (cash-collateral rebate for general-collateral names; a positive fee for "specials"). This is the supply side.
• Short Interest & Days to Cover — the standard crowding metrics: short interest as a % of float (squeeze fuel relative to supply) and days-to-cover (short interest ÷ average daily volume, how trapped shorts are relative to liquidity). Critically: the FINRA data is reported twice-monthly with a multi-day lag, so it is a stale snapshot, not live positioning.
• Short Squeezes — the reflexive, self-reinforcing upward spike when forced covering (margin calls, recalls) — sometimes compounded by dealer gamma hedging — drives price violently up, then almost always round-trips (Volkswagen 2008, GameStop 2021).
• Hard-to-Borrow & Buy-Ins — the cost-and-tail-risk side: punitive borrow fees on scarce names (accruing every calendar day) and forced liquidation via recall or Rule 204 close-out. This is where a directionally-correct short still loses to carry or gets bought-in.

The unifying tension — and the evidence

The thread running through all four children is that a short position has a negative cost of carry and an exit you don't fully control, so direction is necessary but not sufficient. The honest evidence picture (detailed in the children) is two-sided and worth stating once at the section level:

• On the bearish/quant side, the academic record genuinely supports short-side information: high short interest (Asquith-Pathak-Ritter 2005) and especially high borrow fees (Engelberg et al., "Loan Fee Anomaly"; Cohen-Diether-Malloy 2007) precede underperformance on average — short sellers are, in aggregate, informed.
• On the bullish/tactical side, the same crowding is squeeze fuel — but squeezes are rare relative to the universe of heavily-shorted names, untimeable, and round-trip. Do not let the documented quant anomaly (a cross-sectional, fee-based signal) lend false credibility to the idea that any individual retail short is a reliable edge, or that any heavily-shorted name is a reliable squeeze.

Strengths & limitations

When it works: Shorting enables hedging, relative-value and pairs trades, and expressing a negative thesis; borrow-fee and short-interest data are among the few hard-to-game, market-priced gauges of crowding. When it fails / #1 misuse: treating a short as the simple inverse of a long. It is not — the unbounded-loss asymmetry, the recurring borrow cost, recall/buy-in risk, the heavier margin, and the price-test restrictions all bite against the short and in favor of the long. The single most common error is modeling a price target while treating borrow cost and forced-cover risk as zero.

Sources

• SEC — Key Points About Regulation SHO (locate Rule 203, close-out Rule 204): https://www.sec.gov/about/divisions-offices/division-trading-markets/key-points-about-regulation-sho
• SEC — Responses to FAQs Concerning Rule 201 of Regulation SHO (alternative uptick rule / 10% circuit breaker): https://www.sec.gov/rules-regulations/staff-guidance/trading-markets-frequently-asked-questions-7
• Nasdaq — Short Sale Circuit Breaker (Rule 201 trigger and operation): https://www.nasdaqtrader.com/trader.aspx?id=shortsalecircuitbreaker
• FINRA — Rule 4210 Margin Requirements (short maintenance: greater of $5/share or 30% for stocks ≥$5; 25% for longs): https://www.finra.org/rules-guidance/rulebooks/finra-rules/4210
• Wikipedia — Short (finance) (lifecycle, unbounded-loss asymmetry, buy-to-cover): https://en.wikipedia.org/wiki/Short_(finance)
• Child nodes (this section): Locating & Borrowing Shares; Short Interest & Days to Cover; Short Squeezes; Hard-to-Borrow & Buy-Ins — and the academic citations therein (Asquith-Pathak-Ritter 2005; Cohen-Diether-Malloy 2007; Engelberg et al. "Loan Fee Anomaly"; D'Avolio 2002).

Flags: Reg T 150% initial and FINRA 30% short maintenance are baseline regulatory minimums — individual brokers commonly impose higher house requirements, and hard-to-borrow names carry special margin. The "informed short seller / loan-fee anomaly" evidence is cross-sectional and on-average, not a per-trade guarantee. This is a section-overview node; all numeric depth and the full evidence record live in the four child docs.