What the Rule 611 Comment File Actually Cites

The Concordance

The comment file on the SEC's proposal to rescind Rule 611 and Rule 610(e) closed on August 17, 2026. It is a large record: 174 items in the SEC index, plus the two form-letter templates the docket hosts separately, on top of a campaign of more than a thousand individual signatures the Commission counts on its own.

Rather than characterise it, we counted it. We downloaded every record the SEC hosts, extracted the text, and tallied references to each provision of Regulation NMS and the rules that sit next to it. 172 of those 176 documents parsed. The four that did not are named in the method section. The result is a concordance: not what the file argues, but which rules it touches.

The percentages below are taken against the 102 substantive letters, meaning public comments of 500 words or more. A one-paragraph submission has no room to cite a rulebook, and including those would understate every provision equally. The nine meeting memoranda and the one staff study in the index are parsed and published in the table but excluded from the percentages, since they are not comment letters.

Two things stand out. The file engages Regulation NMS in detail, so this is not a record of people who did not read the rulebook: three quarters of substantive letters reach Rule 610(e), just under a quarter reach the access fee cap, a third reach Rule 605, and nearly half reach FINRA's best-execution rule. And within Rule 610, engagement is uneven in a specific way. Paragraph (c) appears in 23 letters. Paragraph (e) appears in 77. Paragraph (d) appears in one.

The full table, every record with its counts and a link to the source document, is published as a CSV at the end of this piece. The counts here can be reproduced from it.

Reading Across Groups

A single number can hide a split. If the fee-transparency provision were being discussed by exchanges and ignored by everyone else, that would be a different finding than a flat zero. So the same count, broken out by commenter type, for every group with at least three substantive letters.

The pattern holds across every group. Exchanges, broker-dealers, asset managers, academics, digital-asset firms, trade associations: each engages 611 and 610(e) heavily, each reaches the access fee cap occasionally, and none reaches paragraph (d). The single citation sits in the public-interest row, and section 03 explains why it does not actually count.

Named letters, for readers who want to check a specific firm. Numbers are citation counts, not mentions of the underlying concept.

Positions are as coded in the underlying classification and are included for context only. Nothing in this piece turns on whether a commenter supports or opposes rescission. The concordance measures citation, not agreement.

The One Citation

Rule 610(d) appears once in the parsed corpus, in a 23,370-word letter from the Healthy Markets Association that cites Rule 611 by name 229 times. Read in place, the sentence is a quotation of a third party recommending that the Commission repeal the prohibition of locked markets under Rule 610(d) and lower the access fee cap under Rule 610(c).

The prohibition of locked markets is not Rule 610(d). It is Rule 610(e). It was Rule 610(d) until the 2024 Regulation NMS amendments, which added a new paragraph (d), titled Transparency of fees, and redesignated the locking and crossing provision as paragraph (e). The quoted recommendation predates that renumbering.

So the count is not one letter engaging fee transparency. It is one letter quoting older material that uses the old numbering, and pointing at the provision the file already discusses seventy-seven times under its current name. The substantive count for paragraph (d) as it exists today is zero.

This is worth stating without any edge to it, because the renumbering is a real trap and the Commission fell into it too. After the 2024 amendments, Rule 610(e)(3) still contained a cross-reference to paragraph (d)(1) of the same section, which by then pointed at the fee provision rather than at the locked-markets exceptions it was meant to reference. The SEC corrected it by technical amendment effective May 5, 2026. When the agency's own rulebook carries a stale (d) reference for eighteen months, a citation in a comment letter carrying one is a hazard of the drafting, not a lapse by the drafter.

We also checked that the zero is not an artifact of how we searched. Every record was scanned for the phrasings a regular expression on 610(d) would miss: paragraph (d) of Rule 610, the unparenthesised 610 d, and constructions of the form fees ... determinable. All returned nothing.

The Dates Diverged

The 2024 amendments arrived as a package and are usually discussed as one. They no longer share a schedule. The adopting release set a single compliance date, the first business day of November 2025, for the Rule 610 amendments. Two subsequent exemptive orders pulled the pieces apart.

The October 2025 order granted paragraph (d) a short extension, to the first business day of February 2026, while giving the fee cap and the tick regime a year. The June 2026 order extended those three again. Its title enumerates them: 600(b)(89)(i)(F), 610(c) and 612. Paragraph (d) is described in the order's background and is not in its relief. It has bound national securities exchanges since February 2, 2026, roughly six months before the comment file closed.

One detail worth stating precisely, because it is easy to get backwards. The codified text of Rule 610(c) already reads one tenth of a cent. Every US exchange currently charges up to three tenths of a cent to remove a displayed protected quotation, and that is lawful because the compliance date is deferred, not because the old figure is still in the CFR. The number in the rulebook and the number on the invoice have been different since 2024.

What 610(d) Requires

The operative text is one sentence. A national securities exchange shall not impose any fee, or provide any rebate or other remuneration, for the execution of an order in an NMS stock that cannot be determined at the time of execution.

Three features of that sentence do most of the work, and each is easy to misstate.

It binds exchanges, not trading centers. Paragraph (c), the fee cap, uses the phrase trading center, which reaches alternative trading systems. Paragraph (d) says national securities exchange. ATSs are outside it. Given that a large share of US volume prints away from exchanges, the transparency obligation covers a narrower slice of the tape than the fee cap immediately above it.

It does not prohibit volume tiers. This is the most common misunderstanding and the adopting release is explicit. Tiers survive, but they have to key off historical rather than future volume, so that the applicable rate is ascertainable before or contemporaneously with the fill. The Commission's stated problem was that many fees and rebates were calculated at month end, which means the cost of a trade is unknown at the moment the trade happens.

It prescribes no format. The release says so directly: the rule enhances transparency regarding exchange fees and rebates but does not require a specific format for publication. Exchanges must post fees on their websites and file changes under section 19(b). Beyond that, how a schedule is laid out, how a rate is keyed to an order, and what a participant has to do to compute an expected fill cost are all left open.

That last point is where a measurement is possible. Determinable in principle and determinable in practice are different claims, and only one of them can be checked against published documents.

What the Schedules Show

We built an access fee dataset for the registered US exchanges by reading each venue's own published schedule, and separately mapped the native liquidity codes those venues use to identify how an order was filled. Both efforts are downstream of the same question paragraph (d) asks: given an order and a venue, what does the fill cost.

The record from that work, stated as findings rather than complaints:

• Seventeen of the nineteen registered exchanges publish a standard rate for removing a displayed protected quotation. Thirteen charge exactly the capped amount. The two absences are 24X National and Green Impact, neither of which is trading yet, so neither has a transaction schedule to read.
• Six of those venues publish no standard add rate at all. Their add economics appear only as tier ladders. A rate exists for every participant, but there is no single published number that describes the venue.
• Rebates are signalled two different ways across the industry, by parentheses at some venues and by a trailing word such as credit at others. Neither is a minus sign, and reading one convention with the other's rules inverts a venue's economics. Both conventions produced wrong output in our own extraction before we caught them.
• Venues encode the fill in five structurally different ways, from a flat enumeration of codes to a bitset to a parametric field. We wrote a separate parser per architecture.
• Of 732 native liquidity codes across sixteen venues, we can currently attach a verified published rate to 217. The rest sit on routing, auction, sub-dollar, non-displayed and modified-order lines that are published inside tiered sections rather than as standard rows.

None of that is a violation of anything. Paragraph (d) requires determinability and prescribes no format, and a venue whose rates live in a tier ladder keyed to last month's volume can satisfy it completely. What the numbers describe is the distance between a rule that is satisfied venue by venue and a market in which fill economics can be compared across venues. Those are different properties, and only the first one is required.

Whether the second one matters is a judgement, and it depends on what happens to Rule 611. If the protected quotation stops being an entitlement, the question of what a fill actually cost carries more weight than it does today, and the answer to that question currently lives in nineteen separately formatted documents. That is an argument, and it is ours rather than the record's. The concordance above is the part that stands on its own.

Where We Might Be Wrong

The counts are reproducible from the published table. The reading of them is not, so here are its strongest objections.

The absence may be exactly right. A comment file responds to a proposal, and this proposal concerns order protection and locked markets. Paragraph (d) is not in it. A commenter writing about whether rescission widens spreads has no reason to raise fee transparency, and treating that as a gap assumes a connection between the two that the Commission did not draw. This is the objection we take most seriously.

Citation counting is a blunt instrument. A letter can engage the substance of determinability while discussing tiered pricing or fee complexity without ever writing the subsection number. Our concept search found the underlying language in a handful of records, and thirty of the substantive letters discuss rebates. Counting citations measures what the record names, which is narrower than what it thinks about.

The existing disclosure regime may be sufficient. FINRA Rule 5310 has coexisted with Rule 611 for two decades with a methodology built largely on Rule 605 and 606 statistics rather than on per-fill economics. Removing 611 may change that emphasis far less than we suggest.

Rescission may not happen, or not in this form. The file is genuinely split, several exchanges support it only with conditions, and at least one large wholesaler opposes it outright. A threshold-based or phased version is a realistic outcome and would change the stakes considerably.

The concordance survives all four. It is a count of what a public record contains, and it is useful to anyone reading that record regardless of which way they think the proposal should go.

Method and Sources

The table. Every parsed record with its per-provision counts, commenter group, coded position, word count and a link to the SEC-hosted source document, as a CSV.

Download the concordance (172 records, CSV)

How the counts were produced. We downloaded every item in the SEC comment file index for S7-2026-20, extracted text from each PDF and HTML record, normalised whitespace so that citations broken across lines still match, and counted regular expression hits per provision. A subsection letter is required for every Rule 610 match: a bare 610 collides with page numbers, release numbers and dollar figures. Percentages are taken against the 102 public comments of 500 words or more, since shorter submissions have no room for a rulebook discussion and would deflate every provision equally. Both scripts and the full output are reproducible from the CSV.

What did not parse. Of the 176 documents, 172 parsed. Two SEC-hosted links return 404 and could not be retrieved: Summer K. Mersinger of the Blockchain Association, and Travis Meredith. Two further records yielded no extractable text, being scanned images: submissions from James Valmond and from a commenter listed only as John. We exclude those four rather than count them as zeroes, since a failed extraction is not an absent citation. The Commission separately reports a form-letter campaign of 1,063 submissions plus six compilation submissions, which are counted once each as texts here rather than once per signature.

On venue coverage in section 06. Our fee dataset holds verified standard rates for seventeen of the nineteen registered national securities exchanges. The two absences are 24X National Exchange and Green Impact Exchange, and neither is a failure of extraction. 24X received temporary conditional relief in August 2026 to trade overnight from January 24, 2027 but is not yet operating. Green Impact was not trading at the snapshot. Neither has published a transaction fee schedule, which is what a venue that is not yet live would be expected to do. Green Impact did file a comment letter on this rescission, so it is present in the regulatory record while absent from the fee record.

• SEC proposal to rescind Rule 611 and Rule 610(e), Release 34-105655, File No. S7-2026-20 (Jun 11 2026)
• Comment file index, File No. S7-2026-20
• 17 CFR 242.610, current text (eCFR)
• SEC Release 34-101070: 2024 Regulation NMS amendments adopting Rule 610(d) (Sep 18 2024)
• SEC Release 34-104172: exemptive relief, Rule 610(d) to first business day of Feb 2026 (Oct 31 2025)
• SEC Release 34-105656: relief for Rules 600(b)(89)(i)(F), 610(c) and 612 to Nov 2027 (Jun 11 2026)
• Technical amendments correcting the Rule 610(e)(3) cross-reference, 91 FR 24116 (May 5 2026)
• FINRA Rule 5310, Best Execution and Interpositioning
• Sapinover: The Access Fee Cap Is Not a Ceiling. It Is the Price.
• Sapinover: Five Ways US Exchanges Encode Execution Economics