A Short Interest Print Looks Live. It Is an Eleven-Day-Old Snapshot.

A short interest figure is not a reading of how many shares are sold short at the moment it appears on a screen. It is a snapshot of member firms' books taken on one designated settlement date, filed two business days later, and published on the seventh business day after that date. Across the twenty-four reporting cycles on FINRA's 2026 schedule, the distance between the date being measured and the date the figure reaches the public runs from 9 to 12 calendar days, with a median of 11.

The publication lag is the smaller half of the problem. Because the cycle runs only twice a month, each figure stays the newest one available for weeks after it appears. For the 23 cycles in the 2026 schedule that are followed by another cycle, the interval from settlement date to the moment a fresher number replaces it runs 24 to 28 calendar days, with a median of 26. A position that changed the day after a settlement date does not show up anywhere in the public record for most of a month.

Two separate limitations sit inside the same number, and they are worth keeping apart. One is age: the figure describes a date in the past, and the published schedule fixes exactly how far in the past. The other is scope: the reporting rule draws a perimeter, and a meaningful amount of short exposure can sit outside that perimeter without anything being hidden or improper. Neither is a flaw in the data. Both are documented properties of it.

How the Cycle Fixes the Age of the Number

FINRA Rule 4560 requires each member to maintain a record of total short positions in all customer and proprietary firm accounts in equity securities, other than Restricted Equity Securities as defined in Rule 6420, and to report that record twice a month. The deadline is in the rule text: reports "shall be received by FINRA no later than the second business day after the reporting settlement date designated by FINRA." FINRA's filing instructions set that cut-off at 6 p.m. Eastern Time.

The designated settlement dates are the 15th of the month, or the preceding business day when the 15th falls on a weekend, and the last business day of the month on which transactions settle. FINRA compiles what it receives and publishes on the seventh business day after the settlement date.

The mid-August 2026 cycle shows the whole chain in one line. The settlement date is Friday, August 14. Reports are due by 6 p.m. ET on Tuesday, August 18. FINRA publishes on Tuesday, August 25. Seven trading sessions open between the snapshot and its release, and none of what happens in them is in the number.

Eleven Days Is the Floor, Not the Ceiling

The publication lag is what most descriptions of this data stop at, and it is the part that matters least. What matters more is how long the released figure remains the freshest one on the board. The August 14 snapshot is published on August 25, and it is not replaced until the August 31 cycle publishes on September 10. From settlement date to replacement is 27 calendar days.

The chart below plots both segments for every 2026 cycle that has a successor. The gold segment is settlement date to publication. The full bar is settlement date to the moment the next cycle publishes and the figure stops being the newest available.

The distribution is narrow and predictable, which is the useful part. The shortest window is 24 days and the longest is 28. The 28-day cases are the March 13 and December 15 settlement dates, where a mid-month snapshot is followed by a month-end cycle whose own publication falls late. Nothing about this is irregular; it falls out of counting business days on a calendar that FINRA publishes a year ahead.

The practical reading is that a short interest level is a monthly-frequency variable dressed as a semi-monthly one. If a position thesis depends on the level changing within a two-week horizon, the data cadence cannot confirm or deny it inside that horizon. That is a statement about resolution, not about direction.

Under T+1, the Snapshot Is Older Than Its Label

There is one further day hidden in the label. The reported figure is a settlement date position, not a trade date position. Since the compliance date of May 28, 2024, the amended Rule 15c6-1 standard settlement cycle for most securities transactions is T+1 — one business day after the trade date rather than two.

A position that settles on August 14 therefore reflects executions through the prior business day. The date stamped on the number and the date of the trading that produced it are not the same date. One session is a rounding error next to a 27-day staleness window, but it matters for anyone trying to line a short interest change up against a specific session's price action. The alignment is off by one before any other adjustment.

What the Figure Counts, and What Sits Outside It

The second limitation is structural rather than temporal, and it is the one more likely to produce a wrong inference. Rule 4560 and FINRA's reporting FAQ define a specific perimeter. Positions inside it are reported in full; exposure outside it is not reported at all, and its absence is not visible in the published number.

Three items on that list do real damage to the intuitive reading of short interest as "bets against the company."

• Positions are gross, not net. FINRA's FAQ is explicit that a short position is reported in full and is not netted against a long position held in the same account. A short-against-the-box position — short and long at once — counts in the numerator exactly like an outright directional short.
• The derivative itself is not reportable. What the rule reaches is the stock position. A short position created by the exercise or assignment of an option is reportable; the open option is not. Economic short exposure held through derivatives does not appear as short interest unless and until it becomes a stock position on a member's books.
• Books-and-records is the boundary, not the market. Firms report what is on their own books. The FAQ addresses positions held overseas at a separate legal entity and not reflected on the reporting firm's books, and stock loan entries booked as "short," and fails to receive, and receipt-versus-payment transactions — none of them reportable short interest under Rule 4560.

The consequence is that the number measures a defined reporting perimeter accurately. It does not claim to measure aggregate short exposure in a security, and reading it that way is an error the data itself does not invite.

Days to Cover Has a Denominator That Moves

Days to cover is derived, and the derivation carries assumptions that get lost when the ratio is quoted on its own. FINRA defines it as short interest divided by average daily share volume, rounded to hundredths, with values of 1.00 or less displayed as 1.00 and a null average daily volume displayed as N/A.

The denominator is the part to check. FINRA computes average daily volume as total volume divided by the number of trade days from the day after the previous settlement date through the current settlement date. That window is the interval between two consecutive settlement dates — roughly ten or eleven sessions — and its length changes from cycle to cycle. Three consequences follow:

1. A heavy-volume stretch inside the measurement window pushes days to cover down without any change in the short position.
2. Both terms are dated to the settlement date. Volume traded after the settlement date is in neither the numerator nor the denominator, so the ratio ages exactly as fast as the raw figure does.
3. A publisher using a different volume window — a trailing thirty sessions, a three-month average — will report a different days to cover from the same short interest. Comparisons are only meaningful within one publisher's series.

Fails to Deliver and Threshold Lists Measure Something Else

Two adjacent public datasets get treated as proxies for short interest, and neither is one.

The SEC publishes fails-to-deliver data drawn from the aggregate fail positions recorded in NSCC's Continuous Net Settlement system. The publication cadence is also semi-monthly and also lagged: the first half of a month is published around month end, and the second half around the 15th of the following month. The Commission's own caveat on that page is direct — fails to deliver can occur for a number of reasons on both long and short sales, are not necessarily the result of short selling, and are not evidence of abusive or "naked" short selling.

The threshold securities list under Regulation SHO is narrower still. A threshold security is an equity security with an aggregate fail-to-deliver position at a registered clearing agency for five consecutive settlement days, totaling 10,000 shares or more, and equal to at least 0.5% of the issuer's total shares outstanding. The lists are disseminated by self-regulatory organizations. Where a fail position in a threshold security persists for 13 consecutive settlement days, the participant is required to close it out by purchasing shares.

Both datasets describe settlement failure. Settlement failure and short interest overlap, but the overlap is partial and the SEC says so on the page that distributes the data.

The Reporting Regime That Would Change This Is Not in Force

A more granular regime exists on paper. Exchange Act Rule 13f-2 and its Form SHO require an institutional investment manager crossing a threshold to file monthly, within 14 calendar days after the end of each calendar month. For a security of an issuer registered under Section 12, the trigger is a monthly average gross short position with a dollar value of $10 million or more, or a monthly average gross short position of 2.5% or more of shares outstanding. For securities of issuers that are not registered, the trigger is a gross short position of $500,000 or more at the close of regular trading on any settlement date in the month. The Commission would aggregate by security to protect the identity of filers and disseminate via EDGAR on a delayed basis. Unlike Rule 4560 reporting, Form SHO's activity data is defined to include activity in derivatives such as options.

It is not operating. On December 3, 2025, the Commission issued an exemptive order, Release No. 34-104303, granting temporary exemption from compliance with Rule 13f-2 and Form SHO reporting effective January 2, 2026 and ending January 2, 2028, with the first reports covering the January 2028 reporting period and due within 14 calendar days after month end. The same order pushed Rule 10c-1a securities lending reporting to September 28, 2028 and its public dissemination to March 29, 2029. The order ties the extension to a Fifth Circuit remand directing the Commission to consider and quantify the cumulative economic impact of the rules.

Whatever the merits of that sequence, the operating fact for a trader is simple: through at least the beginning of 2028, the twice-monthly FINRA figure is the public short interest data that exists. There is no fresher official series waiting behind it.

What Would Invalidate This

This frame is about resolution and scope, and there are conditions under which it stops being the binding constraint.

• If the variable is being used at quarterly frequency. A structural read — whether a security's short interest has been persistently elevated across many cycles — is barely affected by an 11-day lag. The lag argument applies to short-horizon inference, not to slow-moving classification.
• If the short exposure genuinely sits inside the perimeter. For a security where short exposure is overwhelmingly ordinary borrow-and-sell activity at US member firms, the scope objection is weak and the figure is close to complete.
• If the schedule changes. The 9-to-12 and 24-to-28 day ranges are computed from FINRA's published 2026 calendar. FINRA sets these dates annually, and holiday placement moves them. The mechanism holds; the specific day counts are a 2026 fact.
• If Form SHO compliance arrives or is accelerated. That would add a monthly series that includes derivative activity — but monthly is less frequent than twice-monthly, and the SEC's dissemination is itself delayed. A newer regime does not automatically mean a fresher number.
• Staleness is not the same as uselessness. Nothing here argues that the level carries no information. A lagged measurement of a slowly moving quantity can still be informative. The argument is only that the number cannot answer questions posed at a resolution finer than the one it is published at.

Concrete Framework

A checklist for handling the figure at its documented resolution rather than an imagined one.

1. Write down the settlement date, not the publication date. Every reference to a short interest figure should carry the date it measures. If a note says "short interest is X," it should say "as of the August 14 settlement date."
2. Count the staleness before using it. Take the settlement date, find the next publication date on FINRA's schedule, and subtract. That number — 24 to 28 calendar days in 2026 — is how long the figure will remain the newest one available.
3. Refuse to use it for questions inside that window. If the question is what a position looks like this week, the data cannot answer it. Deciding not to answer is a valid output.
4. Check whether the days-to-cover figure is FINRA's or a vendor's. If it is a vendor's, find the volume window before comparing it with anything. Different denominators are not comparable, and the difference is not disclosed in the ratio.
5. Ask where the exposure would sit. For any security where a large part of the short interest could be borrow against an offsetting long, or where the exposure would plausibly be expressed in options, treat the reported level as a lower bound on directional intent rather than a measure of it.
6. Keep fails-to-deliver and the threshold list in a separate column. They measure settlement failure, are published on their own lagged schedule, and carry the SEC's stated caveat. Do not blend them into a short interest read.
7. Re-check the regulatory state before relying on the timeline. The Rule 13f-2 exemption runs to January 2, 2028, and the 10c-1a dates to 2028 and 2029. These dates have been extended more than once. Confirm against the current SEC order rather than a remembered version of it.
8. Size the position to the data cadence. If a thesis needs confirmation the data can only deliver every two weeks at 11 days' remove, the holding period and the risk budget have to tolerate that gap, or the thesis needs a different input.