Public Comments

Status: Original and derived artifacts are available for this document.

Download: pdf

Comment on the Renewal of Information Collection
3038-0076, Requirements for Derivatives Clearing
Organizations
Submitted to: Commodity Futures Trading Commission, via www.regulations.gov,
specifically referencing “OMB Control No. 3038-0076”
Docket: CFTC-2026-1685 (Regulations.gov document CFTC-2026-1685-0001)
Federal Register citation: 91 Fed. Reg. 50816 (Aug. 6, 2026)
Comment period closes: October 5, 2026
Submitted by: Tokenization Systems (www.Tokenization.Systems)
Date: October 5, 2026

Introduction and summary
Tokenization Systems is an independent research program studying the operational control
layer of tokenized financial systems: the technical and organizational functions through
which an issuer or platform actually exercises control over issuance, custody, transfer, and
the execution of lawful orders. It develops measurement instruments intended to make
those functions observable to supervisors, including the Minimum Viable Equivalence Pack
(MVEP), a ten-category structural diagnostic applied at decision points such as onboarding
and collateral acceptance, and the Control Layer Intensity Index (CLII), a continuous
concentration measure.
Tokenization Systems has submitted more than a dozen prior federal comments across
eight federal agencies. This is its first comment to the Commission, and it continues a
position taken before a sibling regulator rather than opening a new one. In an April 2026
comment to the Office of the Comptroller of the Currency on its rule for permitted
payment stablecoin issuers (Docket OCC-2025-0372), Tokenization Systems observed that
tokenized instruments originating in different regulatory regimes were receiving comparable
collateral treatment, and that “[e]ach regulator supervises the instruments within its
perimeter. Nobody supervises the seam where they are combined.” That comment named
the Commission as one of the regulators whose perimeter meets that seam, and
recommended a standardized equivalence assessment as a supervisory tool.
This comment is directed to two of the four matters on which the Commission invites
public comment: whether the collection is “necessary for the proper performance of the
functions of the Commission, including whether the information will have a practical use,”
and “[w]ays to enhance the quality, usefulness, and clarity of the information to be
collected.” It does not contest the Commission’s burden estimate. It takes no position on
any pending rulemaking, and none of its recommendations requires one.
The four recommendations are:

1

Regulation
inside this
collection

#

Recommendation

1.

Define the evidentiary form of the
tokenized-collateral equivalence
determination
Record the liquidation-path analysis
behind the liquidity finding

39.13(g);
39.27(b)

Record whether concentration was
assessed at the redemption path, not
only at the asset
Record when the equivalence
determination was last revisited

39.13(g)

2.

3.

4.

39.13(g)

39.13(g)

What it would make visible
Whether an accepted instrument
conveys the rights the margin
treatment assumes
Whether quick liquidation
survives a named stress, not only
the base case
Concentration that sits below
the asset level and is invisible
when measured at it
Staleness, without the
Commission having to set a
review interval

Recommendations 1 and 4 are near-zero burden. Recommendations 2 and 3 are records of
analysis a registrant performing the 39.13(g) test in good faith has already done.

A note on scope and evidence
Tokenization Systems’ empirical work studies tokenized funds, stablecoins, and the gateway
infrastructure through which they move. It does not study derivatives clearing
organizations, and Tokenization Systems has examined no DCO’s collateral practices.
Where measured findings appear below, they are offered to illustrate a measurement
method and a failure mode, not as evidence about any registrant subject to this collection.
Each point of use is flagged. Where this comment describes what Staff Letter 25-39 says, it
quotes it.

The collection now supports a determination it does not capture
The notice states that this collection covers the reporting requirements established by
Commission Regulations 39.10 through 39.19, 39.21, 39.24 and 39.27, including Regulation
39.13, and that the obligations “are necessary to implement certain provisions of the CEA,
including ensuring that registrants exercise effective risk management.”
Since this collection was last renewed, Commission staff have addressed a new form in
which initial margin assets may be held and transferred. On December 8, 2025, the Market
Participants Division, the Division of Market Oversight, and the Division of Clearing and
Risk issued CFTC Staff Letter No. 25-39, setting out those Divisions’ views on the use of
tokenized assets as collateral. That guidance is staff-level, states that it does not create
rights and may not be relied upon, and represents the views of the three Divisions rather
than the Commission. Tokenization Systems does not treat it as binding and does not ask
the Commission to treat it as binding here. It is relevant to this collection for a narrower

2

reason: it describes determinations that registrants subject to this collection are now
expected to make.
The guidance also places itself in a sequence that leads to this collection. It describes the
CFTC’s tokenized collateral initiative as “part of the CFTC’s efforts to implement the
recommendations in the President’s Working Group on Digital Asset Markets report.”
That report, issued in July 2025, recommended that the Commission consider using its
rulemaking, interpretive, and exemptive authority to “[p]rovide guidance for DCO
acceptance of digital asset collateral (including payment stablecoins),” addressing financial
resources, haircuts, settlement finality, custody, system safeguards, reporting, and legal risk
“under CFTC Regulations 39.11, 39.13, 39.14, 39.15, 39.18, 39.19, and 39.27.” Each of
those seven regulations is among the thirteen whose reporting requirements this collection
covers. The information that shows how DCOs apply any such guidance is therefore
information this collection covers.
Three features of Staff Letter 25-39 bear on the practical use of the information collected
under 3038-0076.
First, it routes tokenized collateral to the standard this collection already supports. The
guidance notes the standard in Regulation 39.13(g), under which, in the guidance’s words,
a DCO “must limit the assets it accepts as initial margin to those that have minimal credit,
market, and liquidity risks,” apply appropriate haircuts, and “apply appropriate limitations
or charges on the concentration of assets they accept as initial margin to ensure quick
liquidation.”
Second, it identifies a threshold question the traditional-asset case does not present. The
guidance states that “[m]arket participants should consider whether a tokenized form of an
asset provides a holder with legal and economic rights that are the same or functionally
equivalent to the rights of the asset in traditional form.” It supplies no standard for making
that determination. On legal enforceability it is firmer: “[r]egistered entities or registrants
are required to demonstrate that non-cash assets collected as regulatory margin meet legal
enforceability requirements,” and for DCOs it cites the legal-framework requirement of
Regulation 39.27(b), which is also within this collection. A requirement to demonstrate
presupposes a record of the demonstration.
The Divisions have since put the equivalence test itself in those terms. In frequently asked
questions updated on September 24, 2026, they answered that an FCM or DCO may invest
customer funds in tokenized forms of permitted investments “so long as the FCM or DCO
can demonstrate” that, among other things, “the tokenized form of the asset grants the
holder legal and economic rights that are the same or functionally equivalent to the rights
received by holders of the asset in its traditional form.” That answer concerns the
investment of customer funds under Regulation 1.25, not initial margin. It confirms,
however, that staff now treat equivalence as something a DCO must be able to show, and
in more than one setting. What that showing looks like in the record is the question this
comment raises.
Third, it forecloses treating the question once and generally. The guidance reminds
registrants, including DCOs, that “any tokenized asset or tokenization structure must be
3

analyzed on an individual basis to ensure it meets all requisite regulatory requirements and
registrant policies and procedures,” and observes that while tokenization “need not change
the fundamental characteristics of that asset,” different methods “may provide different
rights to token-holders or different levels of protection.”
The result is a per-instrument determination that registrants are expected to make, that
varies by tokenization structure rather than by asset class, and that conditions whether the
39.13(g) analysis rests on a sound premise. The information collected under 3038-0076 does
not capture it in any defined form. A supervisor reading the collection can observe that a
haircut was applied. The collection does not show whether the instrument the haircut was
applied to conveys the rights the haircut assumes.
That is a limit on practical use rather than a defect in the collection as designed. The
collection predates the circumstance. But the standard the Office of Management and
Budget applies to every collection asks for more than availability: practical utility means
“the actual, not merely the theoretical or potential, usefulness of information to or for an
agency” (5 CFR 1320.3(l)). A determination recorded, where it is recorded at all, in
whatever form each of 204 respondents chooses has potential usefulness to the Commission.
A defined form would make it actual.

Recommendation 1: define the evidentiary form of the
equivalence determination
Tokenization Systems recommends that, in renewing this collection, the Commission
specify the form in which a registrant records the equivalence determination for any
tokenized asset accepted as initial margin, rather than leaving the form to each registrant.
The recommendation is deliberately narrow. It does not ask the Commission to adopt a
substantive standard for equivalence, which is a policy question beyond a PRA renewal. It
asks that where the determination is made, the record of it be comparable across the 204
respondents this collection covers. Comparability is what converts an individually
reasonable judgment into information with practical use for the Commission.
A defined form would record (i) the decision date; (ii) an instrument description sufficient
to identify the tokenized collateral and its redemption or liquidation pathway; (iii) the
equivalence framework the registrant applied; (iv) the outcome recorded for each category
in that framework; and (v) the system-of-record reference that preserves the decision
package. Each is a record of the registrant’s own analysis rather than a Commission verdict
on substantive equivalence.
The guidance itself supplies a natural first set of categories: the five areas of regulatory
concern it names, namely eligible tokenized assets, legal enforceability, segregation, custody,
and control arrangements, haircuts and valuation, and operational risks. A form organized
around them would require no framework beyond what Commission staff have already
published.
MVEP is offered as one more detailed worked model, not the only possible one. Each

4

category names a required artifact and a pass-or-fail test, applied by the party taking the
risk rather than by the issuer of the instrument: Rights Parity; System of Record; Records
and Reconciliation; Custody and Customer Property; Insolvency Posture; Settlement
Finality; Reorg and Exception Handling; Governance and Change Control; Disclosures;
Collateral Operability.
Four map onto regulations already inside this collection:
MVEP category

Question it forces

Custody and Customer
Property

Has the segregation model
been verified under stress, and
does it match the disclosures
Is the holder’s claim
enforceable against the issuer
in a receivership, and is claim
priority accessible before it is
needed
What is “settled,” and does
value integrity survive an
oracle or pricing disruption
Can the asset be liquidated
within the assumed close-out
horizon under each named
stress

Insolvency Posture

Settlement Finality

Collateral Operability

Nearest provision in this
collection
39.15

39.15; 39.27(b)(3), (4)

39.14(d); 39.27(b)(5)

39.13(g); 39.27(b)(4)

Tokenization Systems applied MVEP across a published matrix of fourteen tokenized
products, nine in the core sample and five comparators. Every product returned
Non-Equivalent at the headline level, separating only on a subordinate strength score.
Those products are tokenized funds and stablecoins, not DCO margin assets, so the result
is not a finding about anything a DCO holds. It is offered for a narrower purpose: it
indicates that the determination is not trivially satisfied and that its outcome is not
predictable from the asset class alone. That is the condition under which a defined
evidentiary form is worth more than a free-text one.

Recommendation 2: record the liquidation-path analysis behind
the liquidity finding
Regulation 39.13(g) limits the assets a DCO accepts as initial margin to those that have
“minimal credit, market, and liquidity risks,” requires haircuts set “taking into
consideration stressed market conditions,” and requires concentration limits or charges that
protect the DCO’s “ability to liquidate such assets quickly with minimal adverse price
effects.” Regulation 39.27(b)(4) separately requires the DCO’s legal framework to provide
for “the unimpeded ability to liquidate collateral and close out or transfer positions in a
timely manner.” For a tokenized asset, the liquidity leg of that test has a structure that the
5

traditional case does not, and recording only its conclusion discards what makes it
checkable.
The characteristic risk is that tokenized collateral posts instantly and converts to cash
slowly, and that conversion depends on redemption paths that can constrict together.
Tokenization Systems models this in MVEP’s tenth category as three paths, four
cross-cutting constraints, and a set of named stresses:
• Paths. An instant-redemption pool, available continuously but of finite capacity;
secondary-market sale, which requires market makers to be present at the moment of
need; and standard fund redemption, which runs on business hours and settles on a
T+1 or longer cycle.
• Constraints that cut across all three. Pool capacity, whitelist or eligible-holder
requirements, cross-chain bridge dependency, and the transfer-restriction model.
• Stresses. Pool depletion; market-maker withdrawal; fund-administrator
unavailability on a weekend or holiday; a transfer-restriction upgrade; an oracle or
net-asset-value disruption; sanctions or blocked-property status; and a
composed-product cascade, in which an underlying pool is depleted while a wrapper
still presents instant redemption.
The pass test Tokenization Systems applies is that the asset remains liquidatable through
at least one path under each named stress within the applicable deadline, or that loss
allocation is explicit. The test is deliberately not “the base case works,” and it is not “two
paths work,” because the paths share constraints and can fail together. Two of those
constraints, eligible-holder requirements and the transfer-restriction model, bear directly on
whether liquidation is unimpeded in the sense of Regulation 39.27(b)(4): either can limit
who may receive the asset at the moment the DCO needs to sell it.
Recommendation: the collection should capture, for each tokenized asset accepted as
initial margin, the liquidation paths the registrant relied on, the stresses considered, and
whether at least one path survives each. This is a record of the analysis the 39.13(g)
liquidity finding already rests on. It converts a conclusion into a reviewable one.

Recommendation 3: record whether concentration was assessed at
the redemption path
Regulation 39.13(g) requires a DCO to apply concentration limits or charges “in order to
ensure its ability to liquidate such assets quickly with minimal adverse price effects.”
Concentration is conventionally measured at the level of the asset: how much of one
instrument a portfolio holds. For a tokenized asset, a second concentration sits underneath
the asset, in the infrastructure through which it is redeemed or moved, and it is invisible to
a measure taken at the asset level.
Tokenization Systems measures this directly for stablecoin transfer infrastructure through
the Control Layer Intensity Index. In a sample of nineteen gateway entities across fifty-one
addresses, covering 27.7 percent of Ethereum USDC and USDT transfer value (a coverage
figure for the sample, not a market share), tier-level concentration measured by the
6

Herfindahl-Hirschman Index is approximately 5,000, against the 1,800 level conventionally
used to mark a highly concentrated market. Within the top tier the split is approximately
56 and 44 between two entities. Over the observed period, transfer volume roughly doubled
while unique counterparties fell about 25 percent and cross-gateway counterparties fell
from 5.8 to 2.6 percent.
Two scope limits apply and Tokenization Systems states them rather than leaving them to
be inferred. This measures gateway infrastructure for two stablecoins on one chain; it is
not a measurement of any DCO’s collateral, and no DCO’s practices were examined. And
the index of approximately 5,000 is computed across the three tiers rather than across
individual entities: computed across all nineteen entities it averages approximately 2,700,
still above the 1,800 level, and the largest single entity holds approximately 37 percent of
gross flow and sits outside the top tier.
Independent research finds the same structure at the point of redemption itself. A National
Bureau of Economic Research study reports that the largest stablecoin issuer “only allows
6 agents in an average month to redeem stablecoins for cash,” that the largest of those
agents accounts for 66 percent of redemption activity, and that a second large stablecoin
had 521 redeeming agents in an average month; the authors treat their concentration
figures as a lower bound. Two instruments of the same asset class can therefore differ by a
factor of more than eighty in how many parties stand on their redemption path, and no
measure taken at the asset level shows which is which. Stablecoins fall within the CFTC’s
tokenized collateral initiative, but that study, like the program’s, concerns stablecoin
markets, not any DCO’s collateral.
The relevance to this collection is methodological. If a registrant accepts a tokenized asset
whose liquidation depends on redemption or transfer infrastructure, an asset-level
concentration limit can be satisfied while the path that limit exists to protect is
concentrated. The measurement is feasible from public on-chain data, which is why both
that study and Tokenization Systems can report it.
Recommendation: the collection should record whether the registrant’s concentration
analysis was performed at the asset level only, or also at the redemption and transfer path.
This is a single categorical field. It does not require the Commission to set a threshold, and
it does not require the registrant to adopt any particular index.

Recommendation 4: record when the determination was last
revisited
The equivalence determination is made when an asset is accepted. Nothing in the current
collection records when it was last revisited. The parameters that rest on it are treated
differently: Regulation 39.13(g)(12) and (13) require a DCO to evaluate the
appropriateness of its haircuts, and of its concentration limits or charges, “on at least a
monthly basis.”
This matters because the facts that determine equivalence can change without producing
any signal in the variables a registrant monitors continuously. A custody amendment, a
7

change to a redemption gate, or a change in administrative-key policy can alter the rights
an instrument conveys while its price, market depth, and attestation cadence remain
unchanged. Continuous monitoring can therefore remain undisturbed after the basis for the
original acceptance has changed, and a monthly haircut review can reconfirm a haircut
whose premise no longer holds.
Recommendation: capture the date of the most recent equivalence determination for
each accepted tokenized asset and the event class that prompted any redetermination. This
is a date and a category. It makes staleness visible without requiring the Commission to
specify a review interval, which would be a policy choice beyond this renewal.

Burden
Tokenization Systems does not contest the estimate of 204 respondents, 40 average burden
hours per respondent, or 8,160 total annual burden hours.
Recommendations 1, 2 and 3 are intended to be burden-neutral in substance. The
equivalence analysis in Recommendation 1 is analysis Staff Letter 25-39 already directs
registrants to perform on an individual basis. The liquidation-path and concentration
analyses in Recommendations 2 and 3 are the reasoning a registrant performing the
39.13(g) test in good faith has already done in order to reach the finding the collection
records. In each case the recommendation addresses the form of the record, not whether
the work is done. To the extent that structuring existing analysis imposes any incremental
burden, it should fall on presentation rather than on substance, and a defined form should
reduce burden over time by removing the need for each registrant to design its own.
Recommendation 4 adds two fields per accepted tokenized asset.
All four apply only to registrants that actually accept tokenized assets as initial margin.
Registrants that accept none should record nothing further, and the burden estimate for
them is unchanged. Tokenization Systems notes that this scoping is what keeps the
recommendations proportionate: the incremental burden is borne by the registrants for
whom the information has the most supervisory value.

Conclusion
This collection exists in part to support the Commission’s ability to see that registrants
exercise effective risk management. For a tokenized asset, the soundness of that judgment
rests on a prior determination about what the instrument conveys, which registrants are
now directed to make individually and for which no evidentiary form is specified, and on a
liquidity finding whose supporting analysis the collection does not capture.
Defining the form of those records, and recording when the equivalence determination was
last revisited, would enhance the quality and usefulness of the information collected
without expanding its substantive scope, without setting a threshold, and without
requiring a rulemaking.

8

Tokenization Systems appreciates the opportunity to comment and is available to provide
further detail on the category structure and the measurement methods described above.
Respectfully submitted,
Zach Zukowski
Tokenization Systems
[email protected]
www.Tokenization.Systems
ORCID: 0009-0006-3642-2450
SSRN Author ID: 10386216

9

Appendix: Selected Sources
The notice.
• Commodity Futures Trading Commission, Agency Information Collection Activities:
Notice of Intent to Extend Collection 3038-0076: Requirements for Derivatives
Clearing Organizations, 91 Fed. Reg. 50816 (Aug. 6, 2026) (FR Doc. 2026-16048).
Abstract, burden statement, and the four matters on which comment is invited.
• 5 CFR 1320.3(l) (Office of Management and Budget, Controlling Paperwork Burdens
on the Public): the definition of practical utility.
Commission regulations and staff guidance.
• 17 CFR 39.13(g), including paragraphs (g)(10) (types of assets accepted as initial
margin), (g)(12) (haircuts, set taking stressed market conditions into account and
evaluated at least monthly), and (g)(13) (concentration limits or charges, also
evaluated at least monthly), each also cited in Staff Letter No. 25-39.
• 17 CFR 39.27(b), including paragraphs (b)(3) (the DCO’s interest in collateral),
(b)(4) (the unimpeded ability to liquidate collateral on a clearing member’s default),
and (b)(5) (finality of settlement), cited in Staff Letter No. 25-39 for legal
enforceability; and 17 CFR 39.14(d) (settlement finality).
• CFTC Market Participants Division, Division of Market Oversight, and Division of
Clearing and Risk, Tokenized Collateral Guidance, CFTC Staff Letter No. 25-39
(Dec. 8, 2025). Staff guidance of the three Divisions, not Commission action.
• The same three Divisions, FAQ: Frequently Asked Questions Concerning Registrant
and Registered Entity Activities Relating to Crypto Assets and Blockchain
Technologies (updated September 24, 2026), Q12 (investment of customer funds in
tokenized forms of permitted investments under Regulation 1.25). Staff views, not
Commission action.
Official-sector report.
• President’s Working Group on Digital Asset Markets, Strengthening American
Leadership in Digital Financial Technology (July 30, 2025), pp. 52 and 53 (the
recommendation that the CFTC consider using its authority to provide guidance for
DCO acceptance of digital asset collateral, including payment stablecoins, under
Regulations 39.11, 39.13, 39.14, 39.15, 39.18, 39.19, and 39.27).
Independent research.
• Yiming Ma, Yao Zeng, and Anthony Lee Zhang, Stablecoin Runs and the
Centralization of Arbitrage, National Bureau of Economic Research Working Paper
No. 33882 (May 2025). Concentration of stablecoin redemption among a small
number of agents, measured from public blockchain data.
Concentration benchmark.
• U.S. Department of Justice and Federal Trade Commission, Merger Guidelines
(2023). Markets with a Herfindahl-Hirschman Index above 1,800 are treated as highly

10

concentrated.
The program’s published first-party work (method only; not a finding about
any derivatives clearing organization).
• Zachary Zukowski, Minimum Viable Equivalence Pack: An Institutional Diligence
Framework for Tokenized Products, SSRN abstract 6363138 (2026), doi
10.2139/ssrn.6363138. The ten MVEP categories, including Collateral Operability,
and the fourteen-product assessment matrix.
• Zachary Zukowski, Routing the Dollar: Gateway Infrastructure, Monetary Policy
Transmission, and the Dollar’s International Functions in Digital Markets, SSRN
abstract 6267698 (2026), doi 10.2139/ssrn.6267698. The Control Layer Intensity
Index gateway sample (nineteen entities, fifty-one addresses, 27.7 percent coverage);
Exhibit 15 and the accompanying text (tier-level and entity-level concentration);
Exhibit 17 (counterparty network).
The program’s prior comment, referenced and not restated.
• Tokenization Systems, comment on the Office of the Comptroller of the Currency’s
proposed rule on permitted payment stablecoin issuer activities, RIN 1557-AF41,
Docket OCC-2025-0372, comment OCC-2025-0372-0057 (April 10, 2026). Section
III.B (coordination across federal regulators on the collateral eligibility boundary)
and Section III.C (a standardized equivalence assessment as a supervisory tool).

11