CAT rounds brokers’ prospective fee down 21% at the sixth decimal

The prospective Consolidated Audit Trail fee charged to an executing broker on each side of a trade is $0.000001 per executed equivalent share. The unrounded formula produces $0.0000012664944145. Rounding to six decimal places therefore cuts the displayed per-side rate by about 21.0%. At one trillion executed equivalent shares, the mathematical one-third allocation would be approximately $1.266 million; the billed six-decimal rate produces $1 million. That difference is not a permanent subsidy or an accounting error. The funding model explicitly allows actual collections to differ from budget and carries surpluses or shortfalls into later fees.[2][1]

The result exposes an important mechanism behind a fee that looks almost too small to matter. CAT LLC first divided $15.149648 million of budgeted costs by a projected 3.987 trillion executed equivalent shares, producing a systemwide fee rate of $0.000003799483243631228. Buyers’ executing brokers, sellers’ executing brokers and market participants each bear one-third. The one-third result has nine meaningful digits after the decimal before it is rounded to the six-decimal rate supported by billing systems. At enormous transaction volumes, the discarded fraction becomes material.[2]

A six-decimal rate changes the arithmetic at scale

The word “share” also needs care. The fee uses executed equivalent shares, a funding-model measurement across eligible securities. It is not simply a count of ordinary shares bought by retail customers, and it is not assessed only once at the market level. The model identifies the CAT Executing Broker for the buyer and the CAT Executing Broker for the seller, while Participants receive their own invoices. A firm can therefore be affected through its execution role even if it ultimately chooses not to pass the expense to a customer as a separate line item.[2][1][4]

The fee alert says the first invoice was to be issued in June 2026 based on May transactions, with payment due in July. It also says the fee is separate from any Historical CAT Assessment. That matters because a broker seeing two CAT-related entries is not necessarily being billed twice for the same cost period. The prospective fee funds current or future operating costs; a historical assessment recovers earlier costs under a different layer of the model.[1][3]

The tiny rate rests on a $102.4m reserve offset

The most consequential denominator is not the decimal place but the cost pool. CAT LLC estimated $94.033 million of costs for May through December and added a 25% reserve of $23.508 million, producing $117.541 million. It then applied $102.391 million of existing reserve, leaving only $15.150 million to be recovered through CAT Fee 2026-1. The reserve offset removed about 87.1% of the cost-and-reserve total before the transaction-volume formula was applied.[2]

The existing reserve came partly from earlier collections exceeding budget because actual executed-equivalent-share volume exceeded projections, and partly from cost-saving measures. The fee alert says that without the reserve reduction the rate based solely on reasonably budgeted May–December costs would have been $0.000010 per executed equivalent share. That comparison is about ten times the published six-decimal rate, but it does not mean brokers received a permanent 90% discount. Prior overcollection is being returned economically through lower future fees, and any later shortfall can be reflected in subsequent budgets.[1][2]

This creates a timing feature that customers and brokerage analysts can miss. The current rate is partly financed by money collected during an earlier period. A low fee in 2026 does not establish that the audit trail’s underlying operating cost has fallen to the amount collected each month. Conversely, a future higher fee would not necessarily mean a sudden equivalent increase in current cost; it could reflect depletion of the reserve, a volume forecast error, a new cost period or reconciliation of earlier undercollection.[2][4]

Rounding can undercollect or overcollect without breaking the model

The filing acknowledges that actual volume can cause CAT Fee 2026-1 to collect more or less than the intended two-thirds share of budgeted costs assigned to executing brokers. Excess money is used to offset future fees or fund the reserve. A shortfall is reflected in later budgets. That reconciliation mechanism is why the $266,494 difference per trillion shares should not be presented as a loss to CAT LLC or a windfall that can never be recovered. It is a predictable consequence of applying a coarse rate to an uncertain volume denominator.[2]

The same logic prevents a simple revenue estimate from one month’s public market volume. A broker’s invoice depends on its own buyer-side and seller-side executed-equivalent shares as represented in CAT Data. Market-wide share counts may use other definitions, omit eligible products or count transactions differently. The invoice is supposed to include calculation detail so executing brokers can confirm it. A defensible customer-cost estimate therefore needs the broker’s applicable CAT volume, the current fee schedule and the firm’s pass-through policy.[2]

A regulatory fee is not automatically a retail surcharge

The funding model charges Industry Members. Whether and how a broker recovers that expense from clients is a separate commercial and regulatory question. A broker may absorb the fee, incorporate it into commissions or spreads, or disclose a direct pass-through where permitted. The CAT documents do not establish that every customer should pay $0.000001 per share, because the regulated invoice attaches to the executing broker’s role and the broker’s customer arrangement can differ.[1][4]

A direct pass-through also needs correct units. Applying the CAT rate to a customer’s ordinary share count can be a useful illustration only if those shares map to the executed-equivalent-share definition and the broker is the relevant executing broker. Applying both buyer and seller rates to one customer order without knowing the firm’s role could double the estimate. Adding Historical Assessment 1A requires another separate rate and period rather than relabelling the prospective fee.[3][2]

The fee’s stated end date is not a hard stop

CAT Fee 2026-1 was described as an eight-month recovery arrangement, with invoices initially contemplated through January 2027. The alert and filing also state that the rate continues after that month until a subsequent CAT fee becomes effective and CAT LLC gives notice. The stated schedule is therefore a planning period, not an automatic expiry. The broader funding authority contains a later outer limit, but the relevant operating instruction is to monitor the replacement notice rather than assume the charge disappears after the eighth invoice.[1][2]

The original finding is a mechanism, not an allegation. A six-decimal system rate is materially lower than the precise one-third formula, while a prior reserve finances most of the current cost pool. Both effects are openly documented. Together they explain why multiplying a tiny fee by market volume can fail to reproduce the budget: rounding, actual volume, three-way cost allocation and reconciliation all intervene. Brokers and customers can verify the arithmetic only by preserving those denominators instead of treating $0.000001 as a self-contained cost fact.[2][1][3]

Two-sided execution can magnify a micro-rate without creating duplicate billing

A single market transaction can generate buyer-side and seller-side executing-broker obligations because the funding model allocates cost to both roles. Where the same firm performs both roles in different capacities or across routed executions, its CAT Data determine the applicable volume. That structure is not duplicate billing in the ordinary sense; it is the approved allocation method. A customer-facing disclosure should nevertheless explain whether a charge represents one side, both sides or an internal recovery convention.[2][4]

The projected-volume denominator also embeds a forecast risk. If actual executed-equivalent-share volume is materially higher, the six-decimal rate can collect more money despite its downward rounding. If volume is lower, the same rate can undercollect. The reserve and later fee process absorb both outcomes, which means the economic burden should be evaluated across successive fee periods rather than from one static rate.[2]

For brokerage cost models, the right sensitivity table varies both volume and duration. Assuming the rate ends in January 2027 can understate expense if no successor fee is effective; assuming it continues indefinitely can overstate it. The documents provide a notice-based replacement mechanism, so the model should contain a review date rather than a fictional fixed expiry.[1][2]

Report an error or submit an update