Which Settlement Date Does the Short Interest Number on Your Screen Describe

Three Clocks Sit Behind One Short Interest Field

Open a US equity on almost any broker page and you will find a line labeled short interest: a share count, often a percentage of float beside it, sometimes a days-to-cover figure. The field sits next to the last sale and the session volume, both of which update in seconds, and it quietly inherits their air of currency. It should not. The short interest field is a photograph of a settlement date that has already passed, developed and released on a calendar FINRA publishes a year in advance.

On September 22, 2026, the most recent FINRA short interest figure a US screen can be showing comes from the August 31 reporting settlement date. Member firms filed it by 6:00 p.m. Eastern on September 2. FINRA released it on September 10. The next figure, capturing September 15, does not reach the public until September 24. So for twelve calendar days, the field labeled short interest has been describing the last Monday in August.

That is not a data quality failure. It is the rule working as written. The interesting part is what happens once you can name which settlement date produced a number: several arguments that traders have about short interest stop being arguments and become calendar arithmetic.

The Reporting Settlement Date Is the Real As-Of Date

The obligation lives in FINRA Rule 4560 (Short-Interest Reporting). Members keep a record of total short positions in all customer and proprietary firm accounts in all equity securities other than Restricted Equity Securities and report them to FINRA on FINRA's designated schedule. The FINRA Short Interest Reporting schedule page states the deadline without ambiguity: "All short interest positions must be reported by 6 p.m. Eastern Time on the second business day after the reporting settlement date designated by FINRA."

Two words carry the weight there. Reporting settlement date. Not trade date. The FINRA Short Interest Reporting Instructions define the capture points: "Firms must report their mid-month short positions as of the close of business on the settlement date of the 15th of each month or, where the 15th is a non settlement date, on the preceding settlement date." The second capture is the last business day of the month on which transactions settle.

Under T+1 settlement, a settlement date of Monday, August 31 corresponds to trades executed on Friday, August 28. A position opened on Monday, August 31 itself settles Tuesday, September 1 and therefore does not appear in the August 31 snapshot at all. It waits for the September 15 capture, which publishes September 24. A trade executed on the last day of August can sit twenty-four calendar days behind the public record before anyone outside the clearing chain can see it.

Then there is the publication lag. The FINRA rule filing now pending at the Commission describes current practice directly: "FINRA publishes the aggregate short interest data seven business days after the reporting settlement date." FINRA's own 2026 calendar confirms it row by row. August 31 settlement, September 2 at 6:00 p.m. due, September 10 published. September 15 settlement, September 17 at 6:00 p.m. due, September 24 published. September 30 settlement, October 2 at 6:00 p.m. due, October 9 published. In calendar terms the August cycle ran ten days from settlement date to publication.

Daily Short Sale Volume Measures Something Else

The second clock is the one most often mistaken for the first. The FINRA Daily Short Sale Volume Files cover short sale trades executed and reported to a Trade Reporting Facility, the Alternative Display Facility, or the OTC Reporting Facility during normal market hours for public dissemination, and FINRA posts them no later than 6:00 p.m. Eastern on the trade date itself. Same day. No settlement lag, no seven business days.

Because that file is fast and free, aggregators republish it, and a meaningful number of them label the resulting column short interest. FINRA addressed this directly in Information Notice 05/10/19 (Understanding Short Sale Volume Data on FINRA's Website): "Although some websites redistribute the Daily File and refer to the data as 'short interest,' it is not, in fact, the equivalent of reported short interest information. Some market participants mistakenly conclude that the bi-monthly short interest data is understated because the Daily File reflects short sale volume that is much larger than what is reported as short interest. However, short interest data reflects short positions held by market participants at a specific moment in time on two discrete days each month, while the Daily File reflects the aggregate volume of trades executed as short sales on each trade date."

The same notice gives the arithmetic example that settles the point: "For example, if a firm sells short 1,000 shares of security ABCD, then purchases 1,000 shares of ABCD later the same day, the short sale volume in the Daily File will include the 1,000 shares that were sold short. Because the firm sold short and purchased an equivalent number of shares that day, it did not establish or accumulate a short position in ABCD; thus, its short sale has no impact on the reported short interest in ABCD."

One is a flow of executions with a sell-short mark on them. The other is a stock of settled positions on two calendar points. A daily short volume ratio of 55 percent and a short interest of 2 percent of float are not in conflict; they are answers to different questions.

Fails to Deliver Run on a Third Publication Clock

The SEC publishes its own settlement-date series in the SEC Fails-to-Deliver Data files, and its release cadence differs again: the first half of a given month is available at the end of the month, and the second half of a given month is available at about the 15th of the next month. So a fail recorded on the August 31 settlement date surfaces publicly around the middle of September, roughly fifteen calendar days after the settlement date and a few days behind the short interest figure for the same date.

The SEC attaches a caution that deserves quoting in full because it is routinely dropped when the file is redistributed: "Fails to deliver can occur for a number of reasons on both long and short sales. Therefore, fails-to-deliver are not necessarily the result of short selling, and are not evidence of abusive short selling or 'naked' short selling."

That is a third dataset, on a third clock, measuring a third thing: an aggregate net balance of undelivered shares on a settlement date, not positions and not executions.

Form SHO Adds a Fourth Clock, Dated 2028

A fourth series is coming, and its timing matters for anyone building a research process now. The SEC adopting release for Rule 13f-2 and Form SHO sets two thresholds. Under Threshold A, an institutional investment manager will report when it reaches a monthly average of daily gross short positions worth $10 million or more in a reporting-company equity security, or 2.5 percent or more of shares outstanding. Under Threshold B, covering non-reporting issuers, the trigger is a gross short position of $500,000 or more at the close of regular trading hours on any settlement date during the month.

Once the rule takes effect, managers will file within 14 calendar days after the end of the calendar month, and the Commission will publish aggregated information derived from those filings on EDGAR within one month after the end of the reporting calendar month, which would put the aggregate up to about thirty calendar days behind the month it describes. That would produce a manager-level series that FINRA's broker-dealer-level series does not contain.

The compliance date has moved more than once. As of the most recent extension, first filings on Form SHO are due February 14, 2028. Treating Form SHO aggregates as something you can backtest against today would be a dating error of roughly seventeen months.

The Weekly Proposal Now Sitting at the Commission

The seven-business-day lag is itself under review. In Federal Register notice of SR-FINRA-2026-012 (Release No. 34-105482), FINRA proposed to adopt Rule 4321 on allocations of fail-to-deliver positions and to amend Rule 4560. On short interest, the filing says FINRA is proposing "to increase the frequency of both short interest reporting under the rule and the subsequent public dissemination of short interest data by (1) requiring members to submit short interest reports on a weekly rather than a bi-monthly basis, and (2) reducing the two business-day reporting turnaround period to one business day to allow for a more streamlined and timely publication process for short interest reports."

The filing states the intended result: the modifications "would allow short interest data to be published weekly, five business days after the reporting settlement date." It also proposes to capture stock borrowed by customers through a member's arranged financing programs, and to require a final short interest report for a security as of the last settlement date before its symbol is deleted. For scale, the filing notes that a total of 86 firms reported short interest for at least one security in 2025.

It is a proposal, not a rule. The Commission published a Federal Register Notice of Designation of a Longer Period (Release No. 34-105824) that "designates August 14, 2026, as the date by which the Commission shall either approve or disapprove, or institute proceedings to determine whether to disapprove, the proposed rule change (File No. SR-FINRA-2026-012)." That date has passed with no order posted to the file. Anyone wiring a data pipeline to the twice-monthly cadence should read the docket before assuming the cadence is permanent.

A Concrete Framework for Dating Any Short Number on a Screen

Here is the Concrete Framework I use before a short-selling figure is allowed into a note or a model. It is four questions, in order.

One: what is the as-of date, stated as a settlement date? If the source cannot tell you, the figure is not usable for event work. Short interest carries a reporting settlement date. Fails-to-deliver carries a settlement date. Daily short sale volume carries a trade date. These are not interchangeable, and under T+1 the settlement date sits one business day after the trade that created the position.

Two: what is the gap between that date and the moment the number became public? For FINRA short interest in 2026 that gap is seven business days. For the daily file it is hours. For SEC fails it runs about two weeks. Write the gap down next to the number.

Three: is the denominator from the same vintage? Short interest as a percentage of float divides a settlement-date share count by a float figure that usually comes from a company filing with its own, much older, as-of date. Days to cover divides the same stale numerator by an average daily volume window that ends much later. A ratio built from two different dates is a hybrid, and the direction of its error depends on which leg moved.

Four: is this a stock or a flow? Positions outstanding on a date are a stock. Executions marked short during a session are a flow. Fails outstanding on a settlement date are a stock of a different thing entirely. Mixing them produces the apparent contradictions that fill message boards.

Run those four questions and most short-interest disputes resolve without anyone needing a view on the underlying company.

Where This Doesn't Apply

This framework is about dating published figures. It does not help with several adjacent problems.

It says little about securities lending. Loan balances, borrow rates and utilization come from vendor data drawn from lending desks, on cadences those vendors set. Some of them update daily. A borrow-rate spike can appear well before any FINRA settlement date captures the position that caused it, and a stock can carry a heavy borrow fee for reasons other than directional short demand, including corporate actions and index events.

It does not apply to non-equity instruments. Rule 4560 covers equity securities other than Restricted Equity Securities. Short exposure expressed through options, swaps or futures does not show up in the FINRA short interest series at all, and a position hedged by a market maker in the underlying may appear there for reasons that have nothing to do with anyone's directional view.

It also does not apply outside the US reporting perimeter. Other jurisdictions run their own regimes with their own thresholds and their own publication lags. A European net short position disclosure and a FINRA short interest report are not measuring the same quantity on the same clock, and reconciling them by eye is unlikely to work.

And it is not a trading edge on its own. Knowing that a number is twelve days old tells you the number is twelve days old. Whether that staleness is priced is a separate empirical question.

What Would Invalidate This

Several specific developments would change the arithmetic above, and it is worth naming them so the claims stay falsifiable.

First, Commission action on SR-FINRA-2026-012. If the proposal is approved as filed, the twice-monthly cadence becomes weekly, the member due date moves from two business days to one, and publication moves from seven business days to five. Every lag figure in this article would need to be rewritten, and any code that hard-codes 24 reporting cycles a year would break. The docket, including the designation of August 14, 2026 as a Commission action date, is the place to check rather than any secondary summary.

Second, a further change to the Form SHO compliance date. That date has already moved more than once; the current first-filing date of February 14, 2028 is a regulatory decision, not a law of nature.

Third, a settlement cycle change. The T+1 relationship between trade date and settlement date is what produces the twenty-four-day gap in the August cycle above. A move to same-day settlement for some instruments would compress that leg.

Fourth, a vendor changing what it labels. If a data provider silently switches its short interest column from the FINRA series to a modeled daily estimate, a series that looked stable will step without any rule changing. The defense is to check the as-of date field on every pull rather than only on the first one.

The first of these is a live docket right now. That is the argument for storing the as-of date alongside every short figure you save, rather than storing the figure alone.

Two earlier pieces on this site take apart other screen fields whose timestamps are not what they look like: An Odd Lot Prints, Adds to Volume, and Never Sets the Last Sale and Two Minutes for Crude, Thirty Seconds for the Index: What a Futures Close Measures.