Let me illustrate my concerns with reference to Taylor Swift, who is, incidentally a fan of very different cats.[2] Recently, I was discussing with my nieces and a friend of theirs, all around thirteen years old, what they would pay to see Taylor Swift in concert. All three girls insisted they would pay ticket prices of $3000 to go to the concert. “OK, but what if your parents weren’t going to pay? What if you had to earn the money yourself to pay for the ticket?” After some reflection, my oldest niece conceded that if she had to pay, she would still go to the concert, but she would sit high up in the nosebleed section.

The funding model we are considering today ensures that none of my niece’s pragmatism enters the conversation about CAT. The Commission is the tween Taylor Swift fan who knows she is not footing the bill and so insists on premium seats. The incentives aggravated by this funding plan make it even harder for real concerns about monetary, security, and privacy costs to be taken seriously.[3]

The funding proposal we are considering today—the latest effort by the CAT LLC to fund past and future expenses—employs an apparently plausible approach to allocating CAT costs. For each transaction, the executing broker for the buyer, the executing broker for the seller, and the relevant exchange or FINRA will split the CAT fee three ways. CAT LLC will adjust the applicable fee rate twice a year. FINRA and the brokerage industry raise meaningful objections to this allocation of expenses, including its operational complexity, the relative burden borne by brokers, and the disparate effect on different firms within the brokerage industry. These concerns are understandable, but with a price tag of more than $500 million to date and billions to come, discontent would be inevitable under almost any method of allocation, as reaction to previous fee filing models has shown.

So, if this approach is apparently plausible, why can I not support it? The CAT funding model we are considering is the product of the CAT LLC, the decisions of which are made by an operating committee consisting of members appointed by each of the self-regulatory organizations (“SROs”)—national securities exchanges and FINRA—who are participants in the CAT NMS Plan (“plan participants”). While broker-dealers and other interested parties have participated on the CAT Advisory Committee and weighed in during the Commission approval process, their ability to shape CAT, its budget, and its fee allocation is limited. They do not have a vote. Without a seat at the official table, these firms have little control over CAT costs. These costs have risen rapidly. When the CAT NMS Plan was approved in 2016, plan participants estimated the average annual costs of operating the plan at a little more than $51 million; in March of this year, the CAT published a 2023 budget of $223 million.[4]

The allocation of fees under these amendments will widen the misalignment of incentives under the plan to control costs. It ensures that most, if not all, of the CAT’s costs are borne by parties that have little or no influence over how the CAT is implemented or how its requirements are interpreted or applied. The Commission oversees the plan participants and so theoretically could act as a moderating force. The rule that launched the CAT set its parameters, but subsequent implementation choices made by the plan participants and the Commission added to CAT’s costs. Negotiations between the SROs and the Commission about CAT design and implementation will continue to affect costs throughout the CAT’s life. Some debate about CAT functionality and coverage has spilled into public, but many discretionary choices are the product of private discussions between the Commission and the SROs. The Commission does not foot any of the CAT bill (aside from the considerable staff time devoted to implementation), so while the Commission is instinctively inclined to expand its data demands, it has little if any incentive to take CAT costs into account. Asking for more data points, requiring them in an aggressive timeframe, or demanding them in a particular form might be costly to market participants, but the Commission—enthused by the prospect of real-time, granular data at no cost to itself—is unlikely to carefully balance the real costs against the perceived benefits. The SEC has little incentive to agree to tweaks that would only minimally affect the quantity, form, or timeliness of CAT data but would substantially reduce costs to the firms providing the data. Similarly, the SEC and the other regulatory users of CAT have little incentive to be judicious in their use of CAT queries.[5] The SROs, which in the funding model before us today bear a third of the cost, do care about constraining costs.[6] They have pushed back on certain elements of the plan and are even litigating with us on some features of CAT. But they are likely to pass some or all of their portion of the cost on to their customers in the form of higher fees. Moreover, the SROs, in their regulatory function, are users of the CAT, so they, like the Commission, have less interest in constraining the CAT’s power as a regulatory tool.

The process for approving semi-annual adjustments to the fee rate and any fee filings by which plan participants recoup costs for their share of the CAT affords some opportunity for external constraint on CAT costs. These filings, however, are a very indirect way to police CAT costs. Today’s order suggests that the Commission will exercise oversight in its review of the plan participants’ fee rate filing that will be required to impose CAT fees on plan participants and industry members,[7] but the critical analysis of CAT costs needs to happen before the budget is finalized.

Although the amendments to the plan would require CAT LLC annually to propose a “reasonable” budget, neither the proposed amendment nor the Commission’s order explains how this provision may, as a practical matter, be enforced.[8] What would be required to find the budget unreasonable? In practice, only the plan participants have much insight into, let alone say over, the budget. Frustrating the Commission’s or industry’s ability to question the budget is its opacity. The vast majority of the budget is one giant, unelaborated item—“cloud hosting services,”[9] which amounted to $99.6 million in 2021 and $135.7 million in 2022.[10] Moreover, the Commission’s role in interpreting the scope of the plan’s requirements—and, according to some commenters, interpreting it so broadly as to drive up the costs of the CAT—may make it more difficult for the Commission to subject these budgets to scrutiny. How can the Commission argue that the CAT budget and the resulting fees are unreasonable if the Commission is a significant cause of the bloated budget?

The allocation required under this plan, with industry members bearing directly two-thirds of the costs of CAT, will reduce the incentives of plan participants and the Commission to subject the budget and resulting fees to rigorous scrutiny. If plan participants are able to pass their costs through to industry members, a possibility the Approval Order recognizes,[11] nobody with financial skin in the game will be among those setting or reviewing the budget, or among those able to propose amendments to the plan or involved in day-to-day discussions about interpretation or implementation of the plan. To modify my Taylor Swift analogy, my nieces might avoid premium tickets if they had to pay out of their own allowance, but they would have no incentive to buy cheaper tickets if their parents tied their allowance to the girls’ expenses, and the girls successfully characterized their $3,000 tickets as expenses justifying a higher allowance. The allocation required by these amendments is not fair given the further misalignment of incentives that it encourages.

The costs of CAT go far beyond the historical building and future maintenance costs that today’s funding mechanism will allocate. In addition to the direct costs of building and maintaining the system, providers of CAT data incur the costs of modifying their systems so that they can meet the data demands. Although these costs will come up in Advisory Committee meetings, they likely will not be central to discussions between the Commission and plan participants about CAT implementation.

When it approved the CAT, the Commission stated that plan participants could “recoup their regulatory costs . . . through the collection of fees from their members, as long as such fees are reasonable, equitably allocated, and not unfairly discriminatory.”[12] Although the plan participants maintain that the funding model before us meets these conditions, the absence of a braking mechanism means that we are asking broker-dealers and likely their customers to pick up the tab for a project that has no reasonable limits. That strikes me as unfair and unreasonable. When CAT budgets are being drawn up and when decisions are made about how CAT should work, who will speak up for the end investor who will foot the bill, give up her right to financial privacy, and perhaps see her data compromised? Until we answer that question, I am not able to support a CAT funding model.

Although the plan does not directly allocate costs to investors,[13] investors will certainly bear the brunt of the financial and non-financial costs of the CAT. Yet their interests have been and will continue to be given short shrift under the plan. Evidence of this can be seen in the Commission’s delay in addressing less visible—but critically important—costs to investors such as the invasion of investor privacy, particularly of retail investor privacy. Although we have taken steps to obscure certain personally identifiable information, such as social security numbers, CAT collects and stores other personally identifiable information. Recognizing the vulnerability of the CAT and hence the retail investors whose information it collects, in August 2020 the Commission proposed amendments to the national market system plan governing the CAT to bolster data security. We have not yet adopted these amendments, even though the CAT is vacuuming mountains of investor information and thus putting Americans at risk to malicious hackers that might seek to mine the data. More fundamentally, we have completely ignored the more difficult, but essential discussion about whether a comprehensive surveillance system designed to monitor Americans suspected of no wrongdoing is consistent with our American ideals.[14] Brokers are more sensitive to these costs than the plan participants because they have direct relationships with retail investors, but they, too, as I have already discussed, have no seat at the table.

Before closing, I want to thank the staff. The CAT funding problem eludes simple solutions. The staff in Trading and Markets, the Division of Economic and Risk Analysis, and the Office of General Counsel have worked extremely hard through many difficult issues over the past decade. The CAT team has logged countless hours on the project and has been patiently good-natured in their interactions with this especially skeptical commissioner. I am sorry that I cannot support the funding proposal before us today.

I have several questions.

Questions:

1. One commenter characterized the proposed funding model as imposing “industry-wide invoicing inefficiencies.”[15] Are we putting the fees on the market participants best able to pay them and pass them on appropriately and inexpensively?
2. Will smaller broker-dealers be burdened disproportionately? If so, is there anything we can do to mitigate costs for small firms?
3. One commenter called for “an independent cost review mechanism to help ensure that future CAT Fees are fair and reasonable and to ensure that controls are put into place to guard against unchecked spending.”[16] In addition to the Section 19(b) process for reviewing fees, could we establish an independent review mechanism, such as a committee made up of representatives of retail investors, broker-dealers, and SROs to advise on security and privacy issues, the CAT budget, its fee allocations, and any proposed adjustments to how CAT works or what it collects?
4. Why not simply impose 50% of CAT costs on the exchanges and 50% on Industry Members? Wouldn’t that be a less costly approach that would roughly match costs to the entities responsible for them?
5. Why does it make sense for FINRA to bear any of the costs of the CAT? After all, FINRA does not run a market, does not operate for profit, and is funded by member firms. Given FINRA’s plans to increase member fees to absorb CAT costs imposed on it, why not just impose FINRA’s portion directly on its member firms or on the exchanges?
6. When do you anticipate presenting a recommendation to adopt the data security amendments?
7. Does the Commission have the authority to fund its primary market surveillance tool with money that does not run through the Congressional appropriation process?