Trust Architecture for European Digital Finance

The legal treatment The legal treatment should follow the economic function, rights and vulnerabilities of the asset, not the database on which it is recorded. A financial instrument should not become a different legal object merely because it moves onto a distributed ledger. Technology neutrality requires exactly this restraint. Where clarification is needed, it should focus on functional characteristics and the authority exercised by issuers and intermediaries rather than creating a second regime based on technical form. follow the economic function, rights and vulnerabilities of the asset, not the database on which it is recorded. A financial instrument should not become a different legal object merely because it moves onto a distributed ledger. Technology neutrality requires exactly this restraint. Where clarification is needed, it should focus on functional characteristics and the authority exercised by issuers and intermediaries rather than creating a second regime based on technical form.

The remaining difficulty is not solved by adding more labels. Hybrid and composable assets can change function depending on how they are used. Classification should therefore be grounded in enforceable rights, economic function, custody, control and the vulnerabilities accepted by users. A token that technically looks identical may create very different trust relationships depending on who can upgrade it, redeem it, custody it or change its economics.

Please explain your ratings in question 3 a), focusing on the issues you deem most significant regarding the overall balance between investor protection, market integrity and innovation:

The central problem with Title II is not simply whether each disclosure obligation is individually reasonable. It is the architecture created when disclosure, authorisation, marketing restrictions, issuer obligations and intermediary controls accumulate. Evidence is useful. Permission is different. A white paper can make risk inspectable. Liability can make representations accountable. Neither requires the regulator to become the authority that decides in advance which laI believe Title II requires substantial redesign rather than incremental adjustment.

Its fundamental problem is architectural.

Article 4 begins from the position that a person shall not make an offer to the public unless that person satisfies a prescribed regulatory structure. Freedom to experiment therefore becomes an exemption from a prohibition rather than the starting condition.

I believe that burden should be reversed.

The starting presumption should be that people are free to build, offer, experiment, invest and voluntarily accept risk unless a specific activity creates a specific vulnerability that justifies intervention.

Fraud should remain fraud. Materially false representations should create liability. Misappropriation should create liability. If an issuer promises particular rights, reserves, governance arrangements or use of proceeds, those representations should be enforceable.

None of that requires the state to become the gatekeeper of the activity itself.

This distinction matters because disclosure can be useful without becoming permission. A simple, proportionate requirement to disclose who controls the project, what rights the asset represents, material risks, insider allocations, upgrade authority, custody arrangements and use of proceeds can provide valuable evidence to a prospective participant.

But the current architecture goes considerably further. It prescribes who may make an offer, the form through which information must be produced, notification requirements, continuing obligations and conduct requirements, while allowing only defined exemptions from that framework.

That is the wrong direction.

I would significantly expand the exemptions, particularly for small offerings and experimental projects, and reconsider whether an arbitrary numerical threshold should determine when voluntary economic activity becomes presumptively subject to the full regulatory apparatus.

The present structure creates another problem: compliance costs are not neutral.

Large institutions can retain lawyers, compliance teams, consultants and regulatory specialists. Small builders cannot. Every additional mandatory process therefore creates a competitive advantage for actors already possessing capital and institutional access.

That is particularly difficult to justify when regulatory status itself has repeatedly failed to guarantee safety.

European supervisors continue to identify serious AML/CFT deficiencies, regulatory arbitrage and control failures within regulated financial institutions. The lesson is not that there should be no law. The lesson is that certification, licensing and compliance cannot be treated as substitutes for trustworthiness.

The relevant question should therefore be what vulnerability exists and what proportionate defence addresses it, not whether the participant has successfully passed through an institutional gate.

This is the difference between scaffolding and the cage.

Scaffolding would require enough truthful, inspectable information for another person to understand the vulnerability they are being asked to accept. It would establish consequences for deception, theft and breach of clearly assumed obligations. It would expose hidden authority and conflicts.

Then it would leave room for people to decide. The current approach increasingly starts from the opposite presumption: the activity is restricted unless the participant demonstrates compliance with the architecture the regulator has prescribed.

That is effectively guilty until proven acceptable. It also creates a dangerous ratchet. Once the principle is established that participation requires satisfying an institutional gatekeeper, the argument becomes only about how many gates should exist and how demanding wful innovation may exist. I would therefore strongly preserve notification without prior approval and resist any movement toward a permission-first model. MiCA should start from a presumption that lawful activity is permitted unless a specific vulnerability justifies a specific restriction. The opposite presumption, that a new asset, protocol or business model is dangerous until an institution certifies otherwise, is the beginning of the cage described in my Trust Architecture work: formal choice remains, while the practical routes through which that choice can be exercised progressively disappear.

Please explain your ratings in question 3 b), focusing on the issues you deem most significant regarding the overall balance between investor protection, market integrity and innovation:

I believe Title II requires substantial redesign rather than incremental adjustment.

Its fundamental problem is architectural.

Article 4 begins from the position that a person shall not make an offer to the public unless that person satisfies a prescribed regulatory structure. Freedom to experiment therefore becomes an exemption from a prohibition rather than the starting condition.

I believe that burden should be reversed.

The starting presumption should be that people are free to build, offer, experiment, invest and voluntarily accept risk unless a specific activity creates a specific vulnerability that justifies intervention.

Fraud should remain fraud. Materially false representations should create liability. Misappropriation should create liability. If an issuer promises particular rights, reserves, governance arrangements or use of proceeds, those representations should be enforceable.

None of that requires the state to become the gatekeeper of the activity itself.

This distinction matters because disclosure can be useful without becoming permission. A simple, proportionate requirement to disclose who controls the project, what rights the asset represents, material risks, insider allocations, upgrade authority, custody arrangements and use of proceeds can provide valuable evidence to a prospective participant.

But the current architecture goes considerably further. It prescribes who may make an offer, the form through which information must be produced, notification requirements, continuing obligations and conduct requirements, while allowing only defined exemptions from that framework.

That is the wrong direction.

I would significantly expand the exemptions, particularly for small offerings and experimental projects, and reconsider whether an arbitrary numerical threshold should determine when voluntary economic activity becomes presumptively subject to the full regulatory apparatus.

The present structure creates another problem: compliance costs are not neutral.

Large institutions can retain lawyers, compliance teams, consultants and regulatory specialists. Small builders cannot. Every additional mandatory process therefore creates a competitive advantage for actors already possessing capital and institutional access.

That is particularly difficult to justify when regulatory status itself has repeatedly failed to guarantee safety.

European supervisors continue to identify serious AML/CFT deficiencies, regulatory arbitrage and control failures within regulated financial institutions. The lesson is not that there should be no law. The lesson is that certification, licensing and compliance cannot be treated as substitutes for trustworthiness.

The relevant question should therefore be what vulnerability exists and what proportionate defence addresses it, not whether the participant has successfully passed through an institutional gate.

This is the difference between scaffolding and the cage.

Scaffolding would require enough truthful, inspectable information for another person to understand the vulnerability they are being asked to accept. It would establish consequences for deception, theft and breach of clearly assumed obligations. It would expose hidden authority and conflicts.

Then it would leave room for people to decide.

The current approach increasingly starts from the opposite presumption: the activity is restricted unless the participant demonstrates compliance with the architecture the regulator has prescribed.

That is effectively guilty until proven acceptable.

It also creates a dangerous ratchet. Once the principle is established that participation requires satisfying an institutional gatekeeper, the argument becomes only about how many gates should exist and how demanding they should become. The wider MiCA review demonstrates precisely why that matters: similar reasoning is now being considered around DeFi certification, CASP access to protocols and even non-custodial software.

The exemptions in Article 4 should therefore be significantly expanded, while the mandatory framework should be rebuilt around actual harm, material representations and actual authority.

I would preserve strong civil liability for materially false or misleading representations. I would preserve proportionate rules against deceptive marketing. I would support simple disclosure of facts necessary to understand an asset's actual trust architecture.

But I do not believe European citizens need a regulatory authority to determine in advance which disclosed risks they may voluntarily accept.

Trust is willingly accepted vulnerability in the presence of risk. The proper role of regulation here should be to make hidden vulnerability visible, make deception consequential, and intervene where one party possesses authority capable of imposing involuntary vulnerability

Please explain your ratings in question 4, focusing on the issues you deem most significant and add any additional issues that you deem highly important:

The most important risks are information asymmetry, hidden control and incentives that are not visible to the person accepting the risk. Disclosure should therefore reveal not only tokenomics but effective control: administrator keys, upgrade rights, insider allocations, vesting, governance concentration, treasury authority and material dependencies. A technically accurate white paper can still conceal the practical trust architecture if it does not show who can change what after the user has committed capital.

Please explain your ratings in question 5, and what (other) regulatory or supervisory actions you would consider useful either at the issuance stage and during marketing or distribution:

The burden of proof matters. Regulation should not begin by assuming that every new issuance or marketing technique is abusive and then require the issuer to prove innocence through institutional permission. Where there is deception, undisclosed insider advantage, market manipulation or a material conflict, regulate that conduct directly. Where founders promise vesting, make the promise inspectable and enforceable. But do not turn every legitimate concern into a universal restriction simply because restriction is administratively easier. Traditional finance already demonstrates why this distinction matters. Sophisticated actors have repeatedly found legal and regulatory seams unavailable to ordinary participants. Cum-Ex and Cum-Cum dividend- arbitrage schemes became serious enough that the EBA launched an inquiry and action plan, and later had to review how supervisors integrated tax-integrity risks. More rules do not automatically create equal protection when complexity itself becomes something well-resourced actors can navigate better than individuals. evidence by EU already available. EBA, inquiry into Cum-Ex/Cum-Cum dividend arbitrage schemes, 12 May 2020. EBA, Peer Review on Tax Integrity and Dividend Arbitrage Schemes, 6 February 2025.

MiCA should be simplified wherever compliance asks the same fact to be proved repeatedly to different institutional actors. Every additional collection point creates cost, delay, gatekeeping power and another repository of sensitive information. Verification does not require possession. If a fact has already been reliably established, privacy-preserving credentials, attestations and selective disclosure should be usable wherever the legal objective can be met without another complete copy of the underlying identity or evidence. This is not merely an efficiency concern. Financial institutions are high-value cyber targets precisely because they aggregate financial and personal data. ENISA's threat reporting continues to show significant attacks against the finance sector. A rule that creates another compulsory database should therefore be treated as creating a new vulnerability, not merely as creating compliance. ENISA, Threat Landscape: Finance Sector, 21 February 2025. ENISA Threat Landscape 2025, finance-sector cybercrime dataset: data breaches represented 64% of cybercrime incidents in that dataset.

The most difficult categories are those where the token's technical form says less than the legal and economic relationship behind it: wrapped assets, governance tokens, hybrid tokens and tokenised claims on off-chain assets. Classification should ask what right exists, against whom, who can alter it, what backs it, and what happens if the intermediary or bridge fails.

No Don’t know / no opinion / not applicable Please explain your answer to question 8:

Material changes to governance, tokenomics, vesting, upgrade authority or control rights can alter the vulnerability a holder originally accepted. Those changes should therefore trigger clear, durable and machine- readable disclosure. A user should not have to discover after the fact that the trust relationship changed while the token name remained the same.

Please explain your answers to question 9:

Stablecoins are not a temporary anomaly to be contained until institutionally preferred digital money replaces them. They are an important experiment in separating money's functions: issuance, reserve management, redemption, custody, programmability and settlement no longer have to be controlled by the same institution. That separation is precisely why overly restrictive stablecoin regulation is dangerous. If reserve rules, licensing costs, geographic constraints and distribution rules make only a narrow class of institutionally approved structures economically viable, Europe can eliminate meaningful monetary experimentation without ever formally banning it. The relevant question is not whether stablecoins should be unregulated. It is which vulnerability requires which authority. Regulate reserves as reserves, redemption as redemption, fraud as fraud and custody as custody. Do not use those legitimate concerns to justify control of every layer beneath them. ECB Appia roadmap states central-bank money is to remain the anchor of the tokenised financial system and that Appia will shape the European ecosystem. Direction is right the execution is disturbing when Pontus is asked to preform all layers. ECB comprehensive payments strategy discusses EU-governed, properly designed and regulated stablecoins as complementary private settlement assets. Properly designed, is where I take some issue. Again, Pontus being a private blockchain is exactly the wrong design. This design makes Appia already a failed trust mechanism since it is built upon a compelled single source of failure or undemocratic choice in selection of the controlling group.

Please explain shortly your assessment and expectations in question 10 as regards the benefits of stablecoins for EU users and, where relevant, any potential risks:

The largest benefit is optionality: stablecoins can provide interoperable settlement without requiring every user or application to belong to the same bank, platform or ledger. That optionality is itself a resilience property. Europe should be careful not to define 'safety' so narrowly that every safe route converges on the same licensed gateways and settlement anchor. A system with many compliant but genuinely independent routes is often more resilient than a system in which every route is supervised but shares the same failure domain.

Please explain your answers to question 11.1 and 11.2, including your reasoning and any suggestions for recalibration or alternative approaches (e.g. greater risk sensitivity, differentiation by business model, or interaction with reserve and liquidity requirements):

I do not have sufficient issuer-level prudential data to recommend numerical recalibration of these thresholds. My recommendation is that capital requirements remain risk-sensitive and tied to the actual liabilities and authorities of the issuer rather than being used indirectly to privilege one technical architecture over another.

Please explain your answers to question 12:

The absence of authorised ARTs should be treated as information rather than proof of one cause. Regulation may be one factor, but product-market demand and competition from EMT structures also matter. The Commission should examine whether the framework is protecting against a demonstrated vulnerability or simply making one architecture uneconomic before increasing or relaxing requirements.

Where an ART is designed to maintain relatively stable value against referenced assets, its clearest consumer function is likely to be exchange, payment or settlement rather than speculative investment. The legal treatment should nevertheless follow the actual rights, redemption mechanics and risk accepted by the holder rather than the label attached to the token.

b) Please rate each aspect below as regard to requirements for significant ARTs : Please explain your answers to question 14 a) and b), including your reasoning and any suggestions for recalibration or alternative approaches (e.g. greater risk sensitivity, differentiation by business model, or interaction with reserve and liquidity requirements):

I would not recommend broad numerical recalibration without prudential evidence. The important principle is diversification of dependency. Reserve rules should reduce redemption and credit risk without unintentionally concentrating reserves in a small number of banks and transferring systemic vulnerability there. Stress testing and transparent reserve composition are more defensible than architectural mandates that assume one custody or deposit structure is inherently safest. This also leaves open more diversified assets as backing.

Please explain your answers to question 15 a) and b), including your reasoning and any suggestions for recalibration or alternative approaches:

The same principle applies to EMTs: the objective should be reliable redemption and resilient reserves, not compulsory dependence on a particular institutional layer. Requirements should be assessed for the vulnerability they reduce and the vulnerability they create elsewhere, including concentration of deposits in banks.

Others / alternative approach Don’t know / no opinion / not applicable Please explain your answer to question 16, including your reasoning and any suggestions for other/alternative approaches:

I would favor functional ring-fencing of the assets and claims necessary to protect EMT holders rather than automatically requiring a separate legal entity or reproducing the non-bank reserve model inside a credit institution. The question should be whether holders have clear, enforceable and insolvency-resilient rights, not whether every issuer uses the same organizational form.

Please explain your answers to question 17 a) & b), including your reasoning and any suggestions for other/alternative approaches:

I do not have evidence supporting numerical recalibration. Significance should not rely on size alone. A token's systemic importance also depends on concentration, substitutability, redemption dependency, interconnectedness and the authority its issuer can exercise over users or markets.

Yes No Don’t know / no opinion / not applicable Please explain how you think the qualitative criteria for classifying EMTs or ARTs as “significant” should be adjusted:

Qualitative criteria should explicitly consider concentration of dependency and substitutability. A token can become systemically important before a single numerical threshold captures it if essential settlement, collateral or payment activity becomes dependent on one issuer or infrastructure.

Yes No Don’t know / no opinion / not applicable Please specify which criteria or metrics and explain why:

Add measures of substitutability, concentration of settlement activity, reserve/custodian concentration, interoperability dependencies and the share of critical applications for which no credible alternative exists. Systemic significance is ultimately about correlated vulnerability, not only market capitalization or transaction count.

Yes, interest or interest-equivalent remuneration should be allowed fully Don’t know / no opinion / not applicable Please explain why you think interest or interest-equivalent remuneration should be allowed for EMTs, (including any benefits and risks) and specify the conditions under which it should be permitted:

A blanket prohibition is another example of regulating the capability rather than the vulnerability. If remuneration creates maturity transformation, liquidity risk, misleading representations or an investment-like exposure, regulate those risks directly. The danger of prohibition-by-category is that it protects incumbent product boundaries from experimentation. The digital age should not reproduce a financial architecture in which sophisticated structures remain available to institutions and wealthy participants through complex wrappers while simpler programmable alternatives are prohibited for everyone else in the name of protection.

b) for ARTs: No, interest and interest-equivalent remuneration should continue to be prohibited Yes, interest or interest-equivalent remuneration should be allowed under defined conditions Yes, interest or interest-equivalent remuneration should be allowed fully Don’t know / no opinion / not applicable Please explain why you think interest or interest-equivalent remuneration should be allowed for ARTs, (including any benefits and risks) and specify the conditions under which it should be permitted:

Same reasoning as above. EBA's continuing work on tax-integrity and dividend-arbitrage schemes illustrates that complexity in traditional finance has itself been used to exploit regulatory and tax seams.

I do not have sufficiently direct issuer-level redemption experience to provide evidence on operational processing problems.

I would likely spend them (e.g. on goods or services) I would likely sell or exchange them with a broker or at a trading venue I would likely turn to the EMT issuer or his agent to exercise my statutory redemption right Don’t know / no opinion / not applicable b) In a stressed market situation I would likely spend them (e.g. on goods or services) I would likely sell or exchange them with a broker or at a trading venue I would likely turn to the EMT issuer or his agent to exercise my statutory redemption right Don’t know / no opinion / not applicable Please explain your answer to question 22 a) and b):

Normal market: exchange/spend. Stressed market: exercise the direct redemption right with the issuer. A redemption right matters most precisely when secondary-market liquidity and confidence are under stress.

Please explain your assessment to question 23, including if a different approach is warranted for significant vs non-significant tokens, or in crisis scenarios and share any suggestions for adjustments you may have:

The right itself should remain strong and predictable. The most useful improvements are operational: clear timing, transparent procedures and evidence that reserves and redemption mechanisms work under stress. A nominal right that cannot be exercised when vulnerability materialises is not meaningful protection.

b) Current MiCA requirements for EMTs Please explain your assessment to question 24 a) & b), including if a different approach is warranted for significant vs non-significant tokens, and share any suggestions for adjustments you may have:

Recovery and redemption plans should be proportional to the scale and dependency created by the token. For significant tokens, plans should be tested rather than merely documented. The goal is inspectable evidence that the issuer can honor the vulnerability it has asked holders to accept under realistic stress.

Please explain your ratings in question 25, including potential benefits, risks, or implementation challenges:

The strongest measure is clear insolvency-resilient ring-fencing for token holders. Direct central-bank reserve access may reduce intermediary concentration where appropriately governed. Emergency liquidity or resolution support should not socialise private risk without corresponding prudential obligations and accountability.

1 - Not at all appropriate / desirable 2 - Slightly appropriate / desirable 3 - Moderately appropriate / desirable 4 - Mostly appropriate / desirable 5 - Highly appropriate / desirable Don’t know / no opinion / not applicable Please explain your ratings in question 26 and add any suggestions for recalibration or alternative approaches:

I would not preserve a fixed 30–60% commercial-bank deposit requirement as the default architecture. This is the Trust Architecture problem in miniature. The rule is presented as reducing reserve risk, but it also compels stablecoin issuers to place a large share of their vulnerability inside the banking system. Regulation has not removed risk. It has selected the institution that will receive it. That choice requires evidence. It should not be presumed safe merely because the receiving institution is regulated. EU authorities themselves continue to report serious AML/CFT and operational weaknesses inside regulated financial institutions. From April 2024 to March 2025, the EBA's EuReCA reporting recorded 1,279 serious deficiencies or material weaknesses exposing financial institutions to ML/TF risk, most commonly involving customer due diligence, systems and controls, and suspicious transaction reporting. In the following reporting window, authorities still recorded 833 material weaknesses. The lesson is not that banks are uniquely unsafe. It is that regulated does not mean riskless. Stablecoin reserve regulation should therefore specify liquidity, quality, diversification and redemption outcomes rather than compel dependence on one institutional layer.

Please specify to what other potential use case(s) you refer in your answer to question 27:

Democratized global savings.

Please explain your answers to question 27:

Global stablecoins are most valuable where money must cross institutional or national boundaries without rebuilding the settlement stack for every jurisdiction. Their benefit comes from interoperability and shared settlement. Their risk comes from concentrating dependency on a small number of issuers and reserve structures. Both facts can be true at once.

Please explain your answers to question 28:

European users should not be placed behind a regulatory wall from global stablecoin liquidity. The EU can impose obligations on actors that actually provide regulated services to EU customers. It should not turn those intermediaries into universal permission gates controlling which lawful assets Europeans may hold or which public networks they may use. A CASP license should establish obligations for the CASP. It should not silently become a license for the customer's economic life.

Please explain the reasoning behind your ratings in question 29. Describe the nature and importance of these risks, the underlying assumptions, market dynamics, or regulatory factors informing your view:

The material risks arise where fungibility masks different legal claims or reserve pools. Users should be able to know which entity owes them what, where the backing sits, and whether redemption remains enforceable under stress. The solution should focus on those claims and dependencies rather than attempting to infer the geographic identity of every self-custodied token holder.

Yes, the multi-issuance model for stablecoins should be permitted under MiCA No, MiCA should disallow the multi-issuance model for stablecoins Don’t know / no opinion / not applicable Please explain the reasons for your response to question 30 and describe the potential implications for the EU market and relevant stakeholders:

Multi-issuance should remain possible if the legal claim, reserve responsibility and redemption obligation are clear to the holder. Prohibiting the model because cross-border trust is difficult would sacrifice interoperability rather than solve the underlying problem. The correct response is to make the trust boundaries inspectable.

No Don’t know / no opinion / not applicable b) Does it discourage the international role of the euro? Yes No Don’t know / no opinion / not applicable Please explain your answer to question 31 a) & b):

Denominating a token in euros should not by itself make every issuance and use anywhere in the world an EU- regulated offering. There is a deeper strategic contradiction here. Europe says it wants the euro to remain internationally relevant, while simultaneously risking an architecture in which digital use of the euro is acceptable only when it enters through European-approved structures. Monetary relevance comes from people choosing to use a currency, not from making the currency a regulatory perimeter.

The advantage is strong protection and a clear regulatory perimeter for EU users. The disadvantage is that an overly expansive perimeter can make euro-denominated digital money less attractive globally and push innovation toward other currencies. Europe should protect users without making use of the euro conditional on submission to an architecture designed for domestic distribution.

Please explain your answers to question 33:

Euro stablecoins can strengthen the international role of the euro if they remain usable across genuinely different infrastructures. Strategic autonomy becomes dangerous when it is allowed to mean institutional enclosure. The ECB openly describes Appia as shaping Europe's tokenised financial architecture, preserving central-bank money as the anchor and strengthening strategic autonomy. Pontes and Appia are described as one strategy, with Pontes ultimately becoming a core component of the Appia ecosystem. That may be a coherent institutional objective. It should not become the only economically viable architecture. Europe should compete by making the euro useful, interoperable and trustworthy, not by making alternative routes progressively harder to use. Pontes and Appia are the right idea and the wrong execution.

Please explain your answers to question 34:

Safeguards are strongest when they address a specific vulnerability such as redemption, reserve quality or misrepresentation. Broad intervention powers and usage restrictions should be used carefully because they can create concentrated discretionary authority. The more powerful the control, the clearer its trigger, scope, evidence and appeal should be.

Partially, but with some limitations (e.g., delays, estimation challenges) No, current methods are insufficient for reliable compliance Don’t know / no opinion / not applicable Please explain your answer to question 35, including any specific remarks with regard to tokens held in unhosted wallets:

Hosted accounts can provide jurisdictional information. A self-custodial bearer asset generally cannot provide reliable residence information without attaching identity or surveillance infrastructure to the holder. That limitation should be accepted where it is intrinsic to the architecture. The solution cannot be to make every wallet persistently identifiable simply because regulation wants a jurisdiction field. IP addresses, metadata, device identifiers and transaction graphs already make privacy fragile. Requiring additional identity linkage increases the consequences of breach, coercion and physical targeting. The fact that information can be collected does not establish that an authority should possess it.

No Don’t know / no opinion / not applicable Please explain your answer to question 36:

A fungible token should not lose a core redemption right because it crossed an invisible jurisdictional boundary, particularly where determining the holder's location would require intrusive identity infrastructure. If an issuer has created a redeemable claim, the cleanest trust relationship is that the claim follows the token, subject to lawful sanctions and other specific restrictions.

Valid-substantiated Not valid-not substantiated Don’t know / no opinion / not applicable Please explain your answer to question 37:

The concern is structurally valid. A multi-jurisdiction reserve creates dependencies on banks, legal entities, supervisors and transfer rules that may behave differently under stress. Those dependencies should be disclosed and stress-tested rather than assumed away.

Don't

advantage of MiCA protections (including redemption rights)

Please explain your answers to question 38:

I strongly oppose making a hosted wallet, CASP relationship or certified intermediary a condition for exercising a basic redemption right. That would turn self-custody from a legitimate ownership model into a second-class status. It would also create exactly the control loop Trust Architecture warns against: first impose identity and intermediation as a condition of access, then point to the resulting intermediated system as evidence that intermediation is necessary. If the concern is reserve distribution, disclose and stress-test reserves. If the concern is sanctions, enforce specific lawful sanctions. If the concern is fraud, prosecute fraud. Do not solve each problem by requiring every person to enter a permanent institutional relationship.

Differentiated redemption rights would create an economic penalty for refusing institutional custody. That is not neutral consumer protection. It is an incentive structure designed to move people into a more controllable architecture. The walk-away test matters here. A choice is not meaningful merely because the law says I may self-custody if exercising that choice causes me to lose basic rights attached to the asset. The door can remain technically unlocked while the cost of walking through it becomes prohibitive.

Yes, in addition to mitigation measures No Don’t know / no opinion / not applicable Please provide reasons for your response to question 40, including any perceived advantages or disadvantages of equivalence regimes:

Equivalence can reduce duplicated regulation where another jurisdiction achieves genuinely comparable outcomes. It should be outcome-based rather than requiring institutional or technical structures identical to the EU model. Mutual recognition is itself a trust architecture: evidence of equivalent protection should substitute for redundant control.

An equivalence regime should rely on third-country rules for prudential and conduct obligations where outcomes are demonstrably comparable, while preserving direct EU requirements for activities actually directed at EU users. Equivalence should be reviewable and evidence-based, not a political label or a demand that another jurisdiction copy the EU architecture.

Please explain your answers to question 42:

Reduce duplicate reporting and authorisation evidence, permit reusable verified information, and calibrate obligations to scale and actual authority. Simplification should not mean weaker accountability. It should mean asking for the same proof once rather than repeatedly collecting the same underlying data. This is the copy problem creating more attack vectors.

Please motivate your answers to question 43 and briefly explain what the problems are for your highest-rated issue(s) and suggest how they should be addressed (e.g. legislative clarification, supervisory guidance, industry standards):

I do not have sufficient issuer-level implementation evidence to score each RTS reliably. The general principle should be proportionality by actual risk and reuse of already-verified information rather than repeated documentary compliance.

Yes No Don’t know / no opinion / not applicable Please explain your answer to question 44:

Standardization is useful where it makes evidence portable and comparable. Reporting formats should be machine-readable and minimize duplicate collection. A common schema for reserve, redemption, control and material-risk information would improve accountability without requiring every participant to rebuild the same verification process.

Yes No Don’t know / no opinion / not applicable Please explain your answer to question 45:

The service list must not become a mechanism for progressively converting every useful interface into a regulated intermediary. The key test is authority. Does the actor custody assets, execute on behalf of the user, make discretionary decisions, control settlement, or make material representations on which the user is expected to rely? If yes, obligations can follow that authority. If the actor merely publishes software, writes open-source code or gives the user a tool through which the user exercises their own authority, regulation should not manufacture an intermediary because an intermediary is easier to supervise.

I would explicitly exclude neutral software development, protocol publication and non-custodial wallet creation from the CASP perimeter unless the provider separately exercises a regulated function. This is not a technicality. If regulation forces software developers to acquire control capabilities merely so that they can satisfy compliance obligations, regulation has changed the architecture. It has created the controller it then claims to regulate. Being identifiable is not authority. Being able to write code is not custody. Publishing an interface is not possession of the user's assets.

No Don’t know / no opinion / not applicable Please explain your answer to question 47:

Appropriateness tests should not become another permission gate for adults using their own assets. Where a provider recommends a complex product, manages discretion or creates leverage, stronger duties can be justified. Where a person simply instructs an intermediary to execute a lawful transaction, the provider should not become a paternal authority deciding whether the customer is sufficiently sophisticated to exercise their own economic agency. Traditional finance already contains extensive gated categories, professional-investor distinctions and complex exemptions that are far easier for well-resourced actors to navigate. Digital finance should not reproduce that asymmetry by default.

No Don’t know / no opinion / not applicable Please explain your answer to question 48, specifying what in your view could be a relevant alternative:

A globally accessible protocol or website should not be presumed to be unlawfully soliciting Europeans merely because Europeans can reach it. The alternative is a European internet in which every global developer must either submit to EU permission or technically exclude Europeans. That is not technology neutrality. It is jurisdiction by accessibility, and it predictably favors large firms capable of maintaining global compliance machinery.

It would be warranted to adjust the prudential stance Don’t know / no opinion / not applicable Please explain your answer to question 49:

Prudential requirements should follow the vulnerability created by the service, not the regulator's desire for a familiar institutional form. A custodian controlling customer assets, an exchange holding counterparty exposure and a non-custodial interface are not variations of the same risk. Treating them as though they are selects for large incumbents able to absorb compliance costs while excluding architectures that deliberately remove custody and counterparty dependence. That is the opposite of risk-sensitive regulation.

Please explain your answers to question 50, indicating, if necessary, how the amount of minimum capital should it be recalibrated:

Capital can protect customers where a firm actually creates balance-sheet, custody or operational exposure. But capital requirements also operate as barriers to entry. The Commission should therefore be able to show which customer vulnerability each capital requirement covers. A requirement that simply makes small entrants nonviable while leaving the underlying risk unchanged is not consumer protection. It is market structure.

The existing approach should be maintained There should be enhanced oversight and coordination in relation to such entities Don’t know / no opinion / not applicable Please explain your answer to question 51:

Large traditional financial groups are evidence that institutional scale and regulatory status do not eliminate conflicts, money-laundering risk or sophisticated regulatory arbitrage. The EBA has continued to identify serious AML/CFT weaknesses inside regulated institutions, and its work on dividend-arbitrage schemes exists because complex, highly regulated markets still produced structures capable of exploiting legal and tax seams. The correct lesson is to supervise the actual concentration of authority and conflicts, not to assume that putting more activity inside a regulated group makes the system inherently safer. EBA, Operational risks and resilience: 1,279 material AML/CFT weaknesses reported in the 12 months to March 2025. EBA, Cum-Ex/Cum-Cum inquiry and 2025 tax-integrity peer review.

Please specify what kind of reporting would be useful and how often? Please explain:

I answered yes, but. Supervisors should collect information that is necessary to observe a defined vulnerability. They should not collect everything that technology makes collectable. Every reporting mandate creates a dataset, a custodian and a future use case for that information. Financial and identity data are valuable attack targets. The presumption should therefore be data minimisation: collect what is necessary, retain it only as long as necessary, and require a clear purpose for expanding the dataset. ENISA finance-sector threat reporting; ENISA notes financial institutions process significant volumes of financial and personal data and remain high-value cyber targets. We do not want protections to increase blast radius.

Broadly appropriate, with limited areas for clarification Partially appropriate, but improvements could be considered Not appropriate Don’t know / no opinion / not applicable Please explain your answer to question 53:

Environmental disclosure is useful when comparable and decision-relevant. It should distinguish protocol-level consensus characteristics from the specific activities of issuers and service providers, and avoid repeatedly attributing the same network-wide footprint to every application using the network.

Cybersecurity cannot be separated from the amount of power and information regulation concentrates. The EU can require stronger security controls and should. But a secure central repository of identity and transaction data remains a repository whose compromise, misuse or compelled disclosure can harm millions of people at once. ENISA continues to document significant cyber threats against financial institutions. This should make the Commission cautious about regulatory designs that require more identity data, more transaction metadata and more persistent linkage to be held by more intermediaries. The privacy question is not only whether the database is well protected. It is whether the database needed to exist in that form at all. ENISA Threat Landscape 2025 and Threat Landscape: Finance Sector (2025).

Yes No Don’t know / no opinion / not applicable Please explain your answer to question 55:

The objective should be one clear obligation for one real function. Where the same activity is simultaneously captured by multiple frameworks, the result should not be duplicate authorization or data collection. Clarification should continue to follow economic function and actual authority.

Europe should not mistake fragmentation from the rest of the world for sovereignty. If EU rules make global liquidity reachable only through a shrinking set of highly regulated gateways, those gateways acquire enormous power over what Europeans may buy, sell and use. The fact that the gateway is supervised does not remove that vulnerability. The EBA's own work on de-risking shows the predictable institutional response when compliance risk becomes too costly: institutions exclude categories of customers rather than manage the underlying risk. The same incentive can operate at the asset and protocol level. EBA, Opinion on de-risking, 5 January 2022.

Please explain your ratings in question 57 and suggest how rules could be made simpler and more proportionate (e.g. legislative clarification, supervisory guidance, industry standards):

Proportionality should be measured against actual authority, not simply firm size. A small custodian can create custody risk. A large software publisher may create none. A protocol with an administrator key may contain concentrated authority. A widely used immutable contract may not. MiCA should stop treating institutional recognisability as a proxy for risk. The thing that can be licensed is not automatically the thing that should be controlled.

Please explain your ratings in question 58 and suggest how rules could be made simpler and more proportionate:

I do not have sufficient implementation evidence to score each technical standard reliably. Across Level 2, the priority should be interoperable reporting, avoidance of duplicate evidence and proportionality to the authority and vulnerability created by the service.

DeFi applications present important benefits for this service, they are already significantly used and present limited risks Don’t know / no opinion / not applicable b) Payments and transactions services DeFi applications present limited interest and are moderately used for this service DeFi applications present important potential benefits for this service but their use remains limited due to some non-mitigated weaknesses DeFi applications present important benefits for this service and are already significantly used , despite important non-mitigated weaknesses DeFi applications present important benefits for this service, they are already significantly used and present limited risks Don’t know / no opinion / not applicable c) Trading and investment services DeFi applications present limited interest and are moderately used for this service DeFi applications present important potential benefits for this service but their use remains limited due to some non-mitigated weaknesses DeFi applications present important benefits for this service and are already significantly used , despite important non-mitigated weaknesses DeFi applications present important benefits for this service, they are already significantly used and present limited risks Don’t know / no opinion / not applicable d) Lending and borrowing services DeFi applications present limited interest and are moderately used for this service DeFi applications present important potential benefits for this service but their use remains limited due to some non-mitigated weaknesses DeFi applications present important benefits for this service and are already significantly used , despite important non-mitigated weaknesses DeFi applications present important benefits for this service, they are already significantly used and present limited risks Don’t know / no opinion / not applicable e) Custody services DeFi applications present limited interest and are moderately used for this service DeFi applications present important potential benefits for this service but their use remains limited due to some non-mitigated weaknesses DeFi applications present important benefits for this service and are already significantly used , despite important non-mitigated weaknesses DeFi applications present important benefits for this service, they are already significantly used and present limited risks Don’t know / no opinion / not applicable Please detail the benefits or weaknesses you associate with the use of DeFi applications for these different services in question 59, and the actions you would consider appropriate to optimise these benefits or address the possible issues they raise:

DeFi's most important contribution is not yield. It is architectural competition. It demonstrates that custody, execution, settlement, liquidity provision, lending and governance can sometimes be separated rather than bundled inside one institution. That does not make DeFi trustless. Smart-contract bugs, admin keys, oracle dependence, governance capture and economic exploits are real vulnerabilities. But the existence of those vulnerabilities does not justify rebuilding the intermediary that the architecture removed. Traditional finance has not earned a presumption of safety simply because its failures happen inside licensed institutions. EU authorities continue to record serious AML/CFT deficiencies in regulated financial institutions, regulatory-arbitrage problems and cyber incidents. Europol reports that criminal networks exploit weaknesses in financial systems using cryptocurrencies, money-laundering techniques and legal business structures. Crime routes around architecture. The regulatory question should therefore be: where is the actual authority and vulnerability in this particular system? Not: where can we insert a gatekeeper? The best lock in the world is an annoyance to the criminal but an impossible barrier to the just.

Please explain your answers to question 60:

The risks are real. The asymmetry in how they are treated is the problem. A smart-contract exploit is treated as evidence that permissionless infrastructure is dangerous. A major failure inside a regulated institution is usually treated as evidence that the institution needs another control. MiCA should apply the same analytical standard to both. EU authorities recorded 1,279 serious AML/CFT deficiencies in financial institutions in the 12 months to March 2025. EBA reviews have continued to identify weaknesses in risk assessment and enforcement, and divergent supervisory approaches have created opportunities for regulatory arbitrage. None of this means regulated finance should be abolished. It means 'regulated' cannot be used as a synonym for 'safe'. DeFi should be judged against the risks it actually creates and the risks it removes, not against an imaginary traditional system in which licensing has eliminated crime, failure and corruption.

Please explain your answers to question 61:

Decentralization must be judged by effective authority. Who can seize assets? Who can change the rules? Who holds administrator keys? Who can upgrade the contract? Who controls governance? Who can prevent a transaction? Who can reverse one? Those are regulatory facts. The identity of a developer is not. The existence of a website is not. Marketing is not. If regulation treats identifiability as control, it will manufacture intermediaries by imposing duties on whoever can be found. That is administratively convenient and architecturally destructive.

No Don’t know / no opinion / not applicable Please explain your answer to question 62:

No. This proposal is the clearest example of control creep in the review. If a protocol is genuinely decentralised, MiCA should not solve the regulator's inability to control it by controlling every regulated gateway through which Europeans can reach it. That does not regulate the protocol's vulnerability. It creates a new authority: the CASP becomes the institution deciding which lawful protocols the user may access. Europe does not need to ban permissionless finance if it can make every practical gateway ask permission on the user's behalf. The Commission should recognize the cumulative architecture being created. CASPs are licensed. Stablecoin issuers are heavily constrained. The review then considers certification of DeFi and non-custodial software and restrictions on CASP connectivity. At the same time, the Eurosystem is openly shaping the architecture, standards and governance of Europe's tokenized financial ecosystem through Pontes and Appia while preserving central-bank money as its anchor. Taken together, this is not neutral scaffolding around an open system. It is an emerging permission architecture. The walk-away test is simple: can an individual reject the institutional gateway and still retain a practical lawful route to use their own assets? If the answer progressively becomes no, formal freedom remains while meaningful freedom disappears. ECB, Appia roadmap, 11 March 2026: Appia will 'shape' development of the European tokenised financial ecosystem and preserve central-bank money as anchor. ECB, 23 March 2026: Pontes and Appia 'form a single strategy'; Pontes is expected to become a core component of the Appia ecosystem.

No Don’t know / no opinion / not applicable Please explain your answer to question 63:

No. Mandatory certification reverses the presumption that should govern an open society. The protocol becomes guilty until an approved institution proves it innocent. Certification is not safety. Europe already has highly regulated financial institutions, extensive AML rules, prudential supervision and data-protection law, yet EU authorities continue to document money-laundering control failures, regulatory arbitrage, tax-integrity problems and major cyber threats. The lesson is not that certification is useless. It is that certification is evidence, not absolution. A voluntary certificate can tell me who inspected the code, what they inspected and what they found. The moment the certificate becomes the legal condition for access, the certifier becomes a gatekeeper. The vulnerability has moved from the protocol into the certification institution. That is precisely the kind of trust transfer MiCA should be required to analyze before imposing it, and then rightly reject. Here a standard can be beneficial, so long as it is open to change over time as evidence arises.

Please explain your answers to question 64:

I strongly oppose these proposals. A non-custodial wallet is a tool through which an individual exercises their own authority. Requiring institutional certification of that tool before it may be distributed or used reverses the relationship. The individual no longer begins with authority over their own asset. Their ability to exercise that authority becomes conditional on institutional approval of the software. Embedded compliance creates the same problem at the protocol layer. It can force neutral infrastructure to acquire surveillance, blocking or identity functions it did not previously possess. Regulation then points to those newly created control functions as the place where regulatory obligations should sit. This is not regulating an existing intermediary. It is manufacturing one. It also creates security consequences. Persistent wallet identity, IP metadata, location data and transaction linkage can make individuals easier to profile, coerce or physically target. Data minimization is therefore not merely a privacy preference. It is personal security architecture.

Please explain your answers to question 65:

Certification should be plural, voluntary and contestable. The proposal to prevent CASPs from connecting users to uncertified protocols is particularly dangerous because it converts an informational tool into access control. The sequence is predictable: First certification is useful evidence. Then certified systems are treated as safer. Then regulated institutions are discouraged from touching uncertified systems. Then access to uncertified systems becomes a compliance risk. Eventually the individual remains legally free to use the protocol while every practical regulated route to it has disappeared. That is the Double-Edged Sword problem. No single step needs to announce capture. The assembled architecture produces it. A strong trust architecture permits multiple independent assessments and lets evidence accumulate. It does not appoint one master authority whose absence of approval becomes proof of danger.

Yes No Don’t know / no opinion / not applicable Please explain your answer to question 66:

Staking should not become another excuse to regulate a protocol function as though it were automatically an intermediated financial service. Where a provider takes custody, pools assets, makes guarantees, exercises discretion or misrepresents risk, regulate those actions. Where a person directly participates in a protocol with their own assets, the absence of an intermediary should be respected rather than treated as a regulatory defect.

No Don’t know / no opinion / not applicable Please explain your answer to question 67:

Do not regulate 'crypto lending' as though the technical substrate itself creates the vulnerability. If an intermediary takes custody, rehypothecates assets, sets discretionary terms, obscures conflicts or makes misleading representations, regulate those acts. If autonomous code matches willing counterparties without custody or discretionary control, forcing a license onto an identifiable developer does not make the transaction safer. It merely gives the state someone to punish. This distinction matters because traditional regulated finance has not eliminated lending abuse, money laundering, tax arbitrage or unequal access. The EBA continues to report material AML/CFT weaknesses in regulated financial institutions, and Europol continues to document criminals exploiting legal business structures as well as cryptocurrencies. The answer to imperfect decentralized systems cannot be to reproduce the weaknesses of the incumbent system while removing the individual's ability to walk away from it.

No Don’t know / no opinion / not applicable Please explain your answer to question 68:

NFTs should not become a regulated financial category merely because they use the same technical substrate as financial crypto-assets. Where an NFT is functionally a financial instrument or a service involves custody, fraud or regulated financial activity, existing functional rules can apply. Regulation should follow the right and vulnerability, not the file format.

Yes No Don’t know / no opinion / not applicable Please elaborate on your answer to question 69:

Prediction markets can aggregate dispersed information and provide useful signals, but they also create manipulation, consumer-protection and market-integrity risks depending on the market and operator. The DLT implementation is not itself the defining risk. Regulation should focus on the economic activity and the authority exercised by the operator.

By MiFID By MiCA Don’t know / no opinion / not applicable Please explain your answer to question 70:

The governing regime should follow the economic nature of the contract. Using a smart contract does not turn a wager or derivative into a different economic activity. Technology neutrality requires the same activity to receive comparable treatment regardless of settlement technology.

Requirements should focus on market integrity, truthful resolution criteria, custody where applicable, conflicts of interest, manipulation and clear disclosure of how outcomes are determined. The resolver/oracle is a critical trust point and should be transparent and contestable where possible.

By MiFID By MiCA Don’t know / no opinion / not applicable Please explain your answer to question 72:

Perpetual futures are economically derivatives. Their use of crypto-assets as an underlying or DLT as infrastructure should not move them into a different regulatory logic. Technology neutrality argues for MiFID treatment where the product is functionally a derivative.

Apply requirements appropriate to leveraged derivatives: margin, liquidation transparency, conflicts, market integrity, custody and clear risk disclosure. Particular attention should be paid to oracle design and liquidation mechanisms because these are technical trust dependencies that can materially alter the user's exposure.

Please explain your answers to question 74:

Tokenized deposits can be useful, particularly for wholesale settlement. The danger is that their usefulness becomes an argument for rebuilding digital finance around bank membership. A tokenized deposit remains a bank liability. If it can only circulate among customers of the issuing bank or inside institutionally controlled infrastructure, tokenization has changed the technology without changing the trust architecture. The more interesting future is interoperability among genuinely different forms of money and settlement, where users and institutions can choose which vulnerabilities they are willing to accept.

Delivery-versus-payment (DvP) and atomic settlement Intraday liquidity management (real-time funding and liquidity optimisation) Treasury and liquidity management services for corporates and financial institutions Programmable payments and smart-contract-based use cases On-chain collateral or margining in regulated markets Improved settlement speed and operational efficiency Reduction of counterparty and settlement risk Please explain why the case(s) you selected in your answer to question 75 are the most promising in the EU:

Delivery-versus-payment and intraday liquidity management are the most promising because they directly reduce time-based counterparty vulnerability. Atomic settlement can replace a period of trust in future performance with verifiable simultaneous exchange. That is exactly where cryptographic infrastructure adds value.

The largest constraints are fragmentation, interoperability, uncertain legal treatment, limited common settlement infrastructure and the temptation for each institution or jurisdiction to build another closed network. Tokenization adds little if every tokenized deposit can only move inside its issuer's private domain.

The EU should provide legal certainty and interoperability without prescribing the topology of the network. This is especially important because the Eurosystem is simultaneously developing Pontes and Appia and openly describes Appia as shaping Europe's tokenized financial ecosystem. The ECB has said Pontes and Appia form a single strategy and that Pontes may evolve into a core component of the Appia ecosystem. That makes architectural neutrality more, not less, important. Public, private, permissioned and permissionless systems should be allowed to demonstrate equivalent outcomes. The institution designing one major settlement route should not, through regulation, also determine that competing routes are illegitimate.

No Don’t know / no opinion / not applicable Please explain your answer to question 78:

I do not have sufficient prudential data to recommend CRD/CRR recalibration. The core principle is that tokenization should not change the underlying prudential treatment merely because the record of the deposit moves to DLT, unless the new architecture creates a demonstrably different operational or liquidity vulnerability.

No Don’t know / no opinion / not applicable Please explain your answer to question 79:

Deposit insurance should attach clearly to the legal claim, not become ambiguous because the claim is represented by a token. Users should be able to know whether a tokenized deposit is insured, against which bank, and whether transferring the token changes that status. The ledger should clarify the trust relationship, not obscure it.

Don't

Issuing a

with national

law

Issuing a

with national

law

Transferring

ownership (acquisitions,

dispositions,

including

good faith

acquisition /

bona fide

ownership

Possibility to

tokens in

national law

property

Rights of

custody

chains

Enforcement

and exercise underlying

assets

Treatment of

intermediary

Please explain your answers to question 80.1 and provide examples from national law. In particular specify if any of the answers is specifically referring to native or non-native tokens and specify if any of the answers would be different depending on if the token is recording a tokenised financial instrument or crypto-assets:

There is material uncertainty because technical control of a token and legal ownership are not always the same thing. This is a fundamental trust problem: the ledger can prove who controls a key while the law may assign the underlying right elsewhere. Priority should therefore be given to ownership/entitlement, third-party effect, custody, insolvency and enforceability. Yes. The most significant uncertainty is the gap between a cryptographic fact and a legal fact.

Control of a private key can demonstrate the technical ability to transfer a token. It does not necessarily establish legal ownership, beneficial entitlement, authority to dispose of the asset, or the rights of third parties. This distinction becomes critical in theft, compromised keys, delegated authority, custody, collateral, inheritance and insolvency.

German law provides a useful comparison. The Electronic Securities Act (eWpG) deliberately creates legal certainty for electronic securities within its scope. An electronic security generally has the same legal effect as a certificated security and is treated as a “thing” for purposes of §90 BGB. The Act identifies the registered holder, provides rules for transfers and good-faith acquisition, and establishes an applicable-law rule for rights in electronic securities.

That is useful precisely because it demonstrates the remaining problem. The law has deliberately mapped particular tokenised financial instruments into a defined legal architecture. The same certainty does not automatically follow for every native crypto-asset simply because cryptographic ownership or transfer can be demonstrated. I therefore distinguish native and non-native tokens.

For a non-native token representing a security, deposit, claim, real-world asset or other off-chain right, the critical question is what legal right the token represents and whether transfer of the token actually transfers that underlying right. The token and the legal claim cannot simply be assumed to be identical.

For a native crypto-asset, there may be no separate issuer or underlying off-chain asset. The central questions instead concern whether the token itself is recognised as property, when ownership legally transfers, the effect of theft or unauthorised transfer, good-faith acquisition, collateral, competing third-party claims and insolvency.

Intermediated custody creates another layer. The law should distinguish clearly between ownership of an asset, a contractual claim against an intermediary and beneficial entitlement to assets held collectively. Insolvency is where ambiguity in that trust relationship becomes most dangerous.

The objective should not be to force every token into the legal model developed for traditional securities. It should be to make the legal consequences of the architecture predictable.

Cryptographic control, legal ownership, beneficial entitlement and authority to dispose should not be treated as automatically synonymous.

That distinction will become even more important as smart accounts, multisignature arrangements and autonomous agents increasingly exercise technically valid authority over assets without necessarily possessing beneficial ownership or unlimited legal authority.

Please explain your answers to question 80.2 and provide examples from national law. In particular specify if any of the answers is specifically referring to native or non-native tokens and specify if any of the answers would be different depending on if the token is recording a tokenised financial instrument or crypto-assets:

The highest urgency should be attached to the questions that determine what a person actually owns and what happens when something goes wrong: property status, transfer, third-party effect, custody, enforcement and insolvency.

These are more fundamental than adding further rules governing issuance.

Again, Germany provides a useful example. The eWpG creates explicit legal effects for electronic securities within its scope, including rules concerning registered ownership, transfer, good-faith acquisition and applicable law. That demonstrates that technological innovation does not require abandoning private-law certainty.

But the solution should not be to require every digital asset to enter an approved register or intermediary structure before receiving legal recognition.

That would solve legal uncertainty by creating another permission architecture.

Native crypto-assets require particular attention because there may be no issuer, register operator or underlying off-chain claim around which traditional private law can organise itself. The law should be capable of recognising ownership and resolving competing claims without manufacturing an intermediary merely because existing legal concepts find an intermediary convenient.

Non-native tokens present a different problem. Where a token represents a security or another off-chain asset or claim, legal certainty is required concerning the relationship between the token and the represented right. A technically valid token transfer should not leave the parties uncertain whether the underlying legal entitlement moved with it.

Custody and insolvency are immediately important because they determine whether someone actually owns an asset or merely holds a claim against an intermediary when that intermediary fails.

I would also treat the distinction between cryptographic control, ownership and delegated authority as an immediate issue.

Possession of a valid signing key establishes technical capability. It should not automatically establish beneficial ownership or unlimited legal authority.

This distinction already matters for custodians, multisignature arrangements, compromised wallets and delegated keys. It will become considerably more important as autonomous software and AI agents transact through smart accounts on behalf of human and institutional principals.

Europe has an opportunity here to provide genuine scaffolding: clear property rights, predictable transfer rules, enforceable ownership, clear treatment of custody and insolvency, and legal recognition of delegated authority. That increases freedom rather than restricting it.

Legal certainty should tell people what their rights are when they choose an architecture. It should not determine which architecture they are allowed to choose.

Ownership/property status and effect against third parties

EU law should clearly establish whether and when a token is an object of property, who legally owns it, and whether that ownership is enforceable against third parties.

Cryptographic control and legal ownership should not automatically be treated as synonymous. Possession of a private key demonstrates technical ability to transfer an asset, but it does not necessarily establish beneficial ownership or legal authority to dispose of it. This distinction matters for theft, compromised keys, multisignature arrangements, delegated authority and increasingly autonomous agents.

Germany's eWpG provides a useful example for electronic securities. Qualifying electronic securities are given the same legal effect as certificated securities and are treated as objects of property under §90 BGB. EU law should provide comparable certainty for digital assets across borders without requiring every token to adopt the institutional architecture of a traditional security.

For native crypto-assets, legal recognition should not depend on the existence of an issuer, custodian or authorised register operator. For non-native tokens representing an off-chain asset or claim, the law should clearly establish the relationship between ownership of the token and ownership or entitlement to the represented asset.

Legal recognition of ownership is scaffolding. It should protect the individual's rights, not become leverage for requiring an approved ledger, custodian or token architecture.

Most important element number 2:

Transfer, including good-faith acquisition and unauthorised transfers

EU law should establish predictable rules for when ownership of a token legally transfers, the consequences of an unauthorised transfer, competing claims, and whether and under what circumstances good-faith acquisition is possible.

A blockchain can establish with considerable certainty that a transaction occurred. It cannot by itself establish that the person controlling the signing key was the beneficial owner or possessed legal authority to make that transfer.

Germany's eWpG again provides a useful example. It expressly defines requirements for transferring electronic securities and provides rules for good-faith acquisition. Similar legal certainty is needed across the EU for digital assets.

The treatment should recognise the difference between native and non-native tokens. For a native crypto-asset, the principal question may be whether transfer of the token transfers the property itself. For a token representing a security, deposit or other off-chain claim, EU law must additionally establish whether transfer of the token transfers the represented legal right.

The objective should be to make the legal consequences of a voluntary transaction predictable, not to require transactions to occur through an approved intermediary simply because intermediated transactions are easier for regulators to categorise.

Most important element number 3:

Custody and insolvency treatment

EU law should clearly establish the rights of token holders when assets are held through an intermediary and what happens to those assets if the intermediary becomes insolvent.

A holder should be able to know whether they legally own the underlying token, possess a beneficial interest in segregated assets, or merely hold a contractual claim against the custodian. The law should also make clear whether assets may be rehypothecated, how shortfalls are allocated, and how additional intermediaries in a custody chain affect those rights.

Germany's eWpG demonstrates one approach by expressly addressing individual and collective registration of electronic securities and the legal position of holders within those structures. The broader EU problem is ensuring that custody does not make ownership ambiguous precisely when the intermediary fails.

This is especially important because self-custody and intermediated custody represent fundamentally different trust relationships. In self-custody, the individual retains technical control and accepts the vulnerabilities associated with that control. In intermediated custody, authority is deliberately transferred to another party and additional counterparty and insolvency risks are created.

EU law should protect people when they choose that intermediary relationship. It should not use those protections to make intermediation mandatory. The goal should be clear rights in failure, not regulation that eliminates the ability to avoid intermediary risk in the first place.

Please explain your answers to question 81.2 and provide examples, in particular from national law. Where you favour full or partial harmonisation or the introduction of a conflict of laws regime, please specify which legal instrument would be most appropriate for that purpose (a new Regulation or the use of existing regimes, such as MiFID or MiCA). Also consider if there are risks with establishing an ownership regime at EU level and how these risks could be mitigated:

I favor a functional EU layer and clear conflict-of-law rules over wholesale harmonization of Member State property law. The objective should be interoperability between legal systems: predictable consequences for ledger entries, transfers and third-party rights while preserving national law where it does not create cross- border uncertainty.

MODEL 3: Token is the asset by law. In this case the law does not actively regulate legal aspects such as what property is or how transfer of ownership occurs etc. but instructs on the legal consequences of entries in a DLT register (this is often called a functional approach). The law could at EU level stipulate who may exercise rights, i.e. the person recorded in the DLT register as holder of a tokenised asset, that these rights have full third-party effect and protects from competing claims. This also creates legal certainty for collateral use. The EU law would create a uniform rule of “digital entitlement” that functions across all Member States’ property law structures. MODEL 4: The token is the legal carrier of rights associated with a related (often off-chain) asset. The law defines tokens as legal objects that do not create new rights, but can –like a “container” –represent various kinds of rights, such as membership rights, ownership, intellectual property rights, usage rights or rights of lien etc. The token would represent the rights stemming from the underlying assets and the law would ensure that the transfer of a token on the ledger legally transfers ownership of the rights stemming from the underlying asset. (Inspired by Lichtenstein Law . Art 2 TVTG defines token as a piece of information on a TT System which can represent claims or rights of memberships against a person, rights to property, or other absolute or relative rights. Art 5 TVTG states that the TT Key holder has the power of disposal over the Token. It is further assumed that the person possessing the power of disposal over a Token also has the right to dispose of the Token. For every previous holder of the power of disposal, it is presumed that he was the person possessing the right of disposal at the time of this ownership). MODEL 5: Any other legal ownership structures, for example based on the approaches taken by different countries, including France, Germany, Luxembourg or adopted in other jurisdictions or internationally, that could be implemented in EU law. You selected: MODEL 3 - a) What would be the legal interoperability of any such EU level measure with national private laws of Member States, or at least laws of the Member State that you are familiar with?

I favor the functional approach in Model 3 for native on-chain assets and the rights-carrier approach in Model 4 where a token represents an off-chain asset or claim. The law should not pretend every token has the same ontology. What matters is that the holder can determine what right the token carries, against whom, and what legal consequence follows from transfer. Model 3 appears capable of interoperating with German private law and, in my view, offers a useful route to EU- wide legal certainty precisely because it need not replace national property law wholesale.

Germany's eWpG already demonstrates a functional bridge between an electronic register and existing private law. It gives electronic securities the same legal effect as certificated securities, treats them as objects of property under §90 BGB, identifies the registered holder and provides rules governing transfer and good-faith acquisition.

An EU rule of digital entitlement could perform a similar function across Member States: establish defined legal consequences of an on-chain record, including who may exercise rights, third-party effect and protection against competing claims, while leaving broader national property-law systems intact where possible.

The important limitation is that the EU rule should not require an authorised intermediary or centrally approved register merely to obtain legal recognition. A genuinely decentralised ledger must be capable of producing legally recognised evidence and legal effects without the law manufacturing a gatekeeper where the architecture itself has none.

EU law should create interoperability between technical fact and legal consequence, not use legal recognition as leverage to prescribe the technical architecture.

MODEL 3 - b) Is it preferable to establish constitutive rules of ownership based on the asset being recorded on chain or functional rules regulating the effects of recording an asset on chain? Constitutive rules of ownership based on the asset being recorded on chain Functional rules regulating the effects of recording an asset on chain Don’t know / no opinion / not applicable Please explain your answer to MODEL 3 b):

I prefer functional rules regulating the legal effects of recording an asset on-chain.

A constitutive rule stating that the blockchain record itself conclusively creates ownership risks collapsing several different concepts into one: cryptographic control, registration, beneficial ownership and legal authority.

They are not always the same.

A stolen private key can produce a technically valid transaction. A custodian can control keys without beneficially owning the assets. A multisig participant can possess signing authority without ownership. An autonomous agent may be authorised to execute transactions within a limited mandate without owning the assets it controls.

The law should therefore specify what legal consequences follow from an on-chain record while preserving mechanisms for dealing with theft, fraud, mistake, delegated authority, competing claims and good-faith acquisition.

The ledger can provide exceptionally strong evidence of what occurred technically. The law should determine the legal consequences of that fact rather than pretending the technical fact answers every legal question.

This is also more technologically neutral. It recognises the evidentiary and transactional properties of DLT without making a particular technical record the source of every underlying legal right.

MODEL 3 - c) Are the different ownership models better suited for native or non-native tokens? For native For non-native tokens Don’t know / no opinion / not applicable Please explain your answer to MODEL 3 c):

Model 3 is particularly well suited to native tokens.

For a genuinely native crypto-asset there may be no separate off-chain asset, issuer or legal claim represented by the token. The token is the asset.

In that situation, creating legal consequences around the on-chain record is considerably more natural than attempting to locate an underlying off-chain right that does not exist.

Bitcoin and Ether are obvious examples conceptually: the asset exists natively within the relevant network rather than functioning as a digital receipt for an external asset.

The law should therefore be capable of recognising digitally native property on its own terms.

That does not mean that the ledger should conclusively determine every legal question. Theft, compromised keys, inheritance, insolvency, delegated authority and competing claims can still require legal rules. But those rules should operate around the native asset rather than pretending it represents something outside the network.

MODEL 3 - d) Can the same ownership model be implemented for all types of tokens (crypto-assets and DLT financial instruments)? Yes No Don’t know / no opinion / not applicable MODEL 3 - d) What would need to be different depending on the type of the token?

The same ownership model should not be imposed identically on all tokens because native and non-native tokens represent fundamentally different legal relationships.

For a native crypto-asset, the token can itself be the object of property. There may be no external issuer, security, deposit, physical asset or contractual claim behind it.

For a tokenised financial instrument, the token generally represents or carries rights that already exist within a wider legal relationship: rights against an issuer, shareholder rights, repayment rights, claims to assets or other legally defined entitlements.

EU law should therefore establish common principles of digital entitlement, transfer, third-party effect, custody and insolvency while allowing the legal consequences to reflect what the token actually is.

Technological similarity should not erase legal difference.

Please explain your answers for MODEL 3 and provide examples:

Model 3 is strongest for native digital assets because it allows law to recognise legal consequences of a digital record without inventing an underlying off-chain asset.

Germany's eWpG demonstrates that national private law can successfully connect electronic registration to property-law consequences. It expressly recognises qualifying electronic securities as property and establishes rules for ownership transfer and good-faith acquisition.

I would extend the principle, but not necessarily the institutional architecture.

A public permissionless ledger should not require an authorised central register operator merely to receive legal recognition. The useful principle is that a reliable digital record can have predictable legal effects.

The law should then address the exceptional cases the ledger cannot resolve by itself: theft, fraud, compromised authority, good-faith acquisition, inheritance, insolvency and competing claims.

The objective should be legal recognition of digital property, not regulatory ownership of its infrastructure.

You selected: MODEL 4 - a) What would be the legal interoperability of any such EU level measure with national private laws of Member States, or at least laws of the Member State that you are familiar with?

Model 4 should be capable of interoperating with German private law, but it requires greater care because the token and the underlying right are legally distinct things unless legislation expressly connects them.

This model is particularly useful where a token represents a security, ownership interest, intellectual-property right, lien, claim against an issuer or another off-chain entitlement.

EU law could establish that a valid transfer of the token transfers the associated legal right, giving the token a legally recognised carrier function across Member States.

Germany's eWpG demonstrates a related approach for electronic securities: §25 provides that transfer of ownership of the electronic security also transfers the right arising from that security.

EU harmonisation could provide similar certainty across borders.

However, the underlying right still matters. A token representing shares, a debt claim, real estate, intellectual property or a bank deposit cannot automatically override every substantive rule governing those different assets.

The objective should therefore be legal interoperability between the token and the underlying right, not the fiction that tokenisation makes every underlying asset legally identical.

MODEL 4 - b) Is it preferable to establish constitutive rules of ownership based on the asset being recorded on chain or functional rules regulating the effects of recording an asset on chain? Constitutive rules of ownership based on the asset being recorded on chain Functional rules regulating the effects of recording an asset on chain Don’t know / no opinion / not applicable Please explain your answer to MODEL 4 b):

I prefer functional rules here even more strongly than under Model 3.

For a non-native token, recording something on-chain cannot by itself determine every aspect of ownership because an underlying legal right exists outside the ledger.

A token representing a share, bond, property interest, intellectual-property right or other claim should have clearly defined legal consequences when transferred. Ideally, the law should make the connection sufficiently strong that transferring the token reliably transfers the represented right.

But the ledger cannot determine whether the underlying right existed validly in the first place, whether the issuer possessed authority to tokenise it, or whether mandatory rules governing that asset have been satisfied.

Functional rules can create a strong legal bridge between token and underlying right without pretending that the digital representation creates reality simply by being written to a ledger.

MODEL 4 - c) Are the different ownership models better suited for native or non-native tokens? For native For non-native tokens Don’t know / no opinion / not applicable Please explain your answer to MODEL 4 c):

Model 4 is naturally suited to non-native tokens because its purpose is to make the token the legal carrier of a right or asset that exists outside the token itself.

Examples include tokenised bonds, shares, deposits, intellectual-property rights, claims against an issuer, liens or interests in physical assets.

The important legal function is to make the relationship between token and represented right unambiguous.

If I acquire the token, I should know whether I acquired the underlying right.

If I transfer the token, I should know whether that right transferred.

If the issuer becomes insolvent, I should know whether I own an asset, possess a secured claim or merely stand among unsecured creditors. That is fundamentally different from a native crypto-asset, where there may be no separate underlying right for the token to carry.

MODEL 4 - d) Can the same ownership model be implemented for all types of tokens (crypto-assets and DLT financial instruments)? Yes No Don’t know / no opinion / not applicable MODEL 4 - d) What would need to be different depending on the type of the token?

The same model should not be imposed on every token.

For a native crypto-asset, the token itself can be the property. There is no need to invent an underlying right for it to carry.

For a non-native token, the legal architecture must define the relationship between the token and the underlying asset or claim.

For tokenised financial instruments specifically, existing substantive rights and obligations under securities, corporate and financial law continue to matter. Tokenisation should provide a new means of recording, holding and transferring those rights rather than silently changing their substantive content.

The common EU layer should therefore concentrate on legal certainty around digital entitlement, transfer, third- party effect, custody and insolvency, while the specific legal consequences reflect the nature of the underlying asset.

The law should follow the actual relationship rather than forcing different relationships into one model merely because they use the same technology.

Please explain your answers for MODEL 4 and provide examples:

Model 4 provides a useful legal bridge between DLT and existing property and financial law for non-native assets.

A tokenized bond should reliably carry the rights of the bond. A tokenized share should reliably carry the relevant shareholder rights. A token representing another asset or claim should tell the holder precisely what legal entitlement travels with the token.

Germany's eWpG provides a useful example. For qualifying electronic securities, German law connects ownership of the electronic security with the right arising from that security and establishes rules governing transfer and good-faith acquisition.

The EU could create comparable cross-border certainty without requiring every token to pass through one prescribed institutional architecture.

This distinction between Model 3 and Model 4 is important.

For a native asset, the token may be the thing.

For a non-native asset, the token may carry rights to something else.

Those are different trust relationships and should not be collapsed simply because both are represented using DLT.

The strongest EU framework would recognize that distinction while providing interoperable rules for ownership, transfer, third-party effect, custody and insolvency.

Again, the purpose should be scaffolding: make rights clear and enforceable while preserving architectural choice.

The central problem is not that Member States must have identical property law, but that token ownership, transfer, custody and third-party effect can receive different legal treatment while the token itself moves seamlessly across borders. I would therefore prioritize EU rules on cross-border effect and legal interoperability rather than attempting an exhaustive replacement of national private law.

Yes. EU conflict-of-law rules should be introduced for the proprietary aspects of tokens because many traditional connecting factors assume a physical location, identifiable intermediary, central register or system operator.

Those assumptions become difficult to apply to genuinely distributed systems.

A native crypto-asset on a public permissionless blockchain may have no issuer, no central administrator, no single register operator and no meaningful geographic location of the ledger. The same ledger may be replicated and validated across many jurisdictions simultaneously.

Trying to determine the applicable property law from the physical location of nodes, validators, developers or other technical infrastructure would therefore produce uncertainty rather than resolve it.

The problem is different for tokenised financial instruments and other non-native tokens. Where there is an identifiable issuer, regulated register or underlying legal asset, existing connecting factors may remain meaningful. Germany's eWpG demonstrates this approach for electronic securities: §32 generally connects rights and dispositions concerning an electronic security to the law of the state supervising the relevant register operator, with fallback rules where that connection is unavailable.

That works because the architecture deliberately contains a legally identifiable register operator.

EU law should not extrapolate from that model and require every token architecture to create such an operator merely to make conflict-of-law analysis possible.

A conflict-of-law rule for tokens should therefore accommodate both architectures: systems with identifiable legal anchors and genuinely decentralised systems without them.

The purpose of the rule should be to provide predictable legal consequences for architectural choices, not to force architectural choices in order to obtain predictable legal consequences.

Please explain your answers to question 84, 84.1 and 84.2 and provide examples. In answering this question, please consider if the suggested scope of the conflict of law rule in your view is the right one, and whether there are aspects missing:

My preferred hierarchy begins with an expressly specified governing law where that choice is meaningful, transparent and accessible to participants.

1. Law expressly specified in the token or its governing principles.

2. Law specified in the rules or governing principles of the system on which the token is recorded.

These provide predictability without requiring the law to manufacture a geographic location for a distributed ledger.

For systems that intentionally use a regulated register, the law of the state supervising that register is an appropriate additional connecting factor. Germany's eWpG provides an existing example of this approach.

For non-native tokens with an identifiable issuer, the issuer's establishment and supervisory jurisdiction can provide further fallback connecting factors.

I would not use the location of technology providers, validators, developers or other infrastructure providers as a primary proprietary connecting factor. Providing technology does not necessarily confer ownership, custody or authority over the asset.

I would be particularly cautious about PREMA, based on the residence of a system-relevant master-key holder. A genuinely permissionless blockchain may have no such key. Creating legal advantages around the existence of one could perversely encourage systems to introduce precisely the centralized coercive authority that decentralized architectures were designed to remove.

The same concern applies to PROPA. It may be appropriate where a system genuinely has an operating authority or administrator. It should not become a reason to require every system to have one.

Conflict-of-law rules should recognize where authority actually exists rather than manufacture authority because existing law finds a central administrator easier to locate.

Three emerging legal-certainty issues deserve attention. First, autonomous agents. A valid signature can prove that a key authorized a transaction without proving that the human principal intended that particular action. Authentication, delegated authority and fidelity to the principal's mandate are different questions. Second, privacy and physical security. Persistent linkage of identity, wallet activity, IP metadata and location can create risks that extend beyond data protection, including coercion and physical targeting. Regulation should not create these linkages unless they are necessary for a specific vulnerability. Third, regulatory concentration itself. Where access to assets, protocols or settlement becomes dependent on a small number of licensed gateways or certifiers, the law has created a new systemic trust dependency. That dependency should be assessed with the same seriousness as technical concentration.

The greatest issue this review needs to address is not any single MiCA provision. It is the architecture created when they are assembled.

MiCA should be considered alongside Europe's wider direction in digital identity, tokenized finance and settlement, including Pontes and Appia. Licensing, certification, identity, stablecoin, protocol-access, custody and settlement rules can each be justified individually. Together they can create something very different.

A system does not need to prohibit alternatives to eliminate meaningful choice. It can make the cost of exercising them progressively prohibitive.

If participation increasingly requires an approved intermediary, identity, asset, protocol or settlement route, individuals may retain theoretical freedom while losing practical freedom.

The walk-away test has failed.

This is the Double-Edged Sword problem. Technologies that make relationships more verifiable can also make control extraordinarily efficient.

Cryptographic verification is not itself trust.

Trust is willingly accepted vulnerability in the presence of risk. Where vulnerability is compulsory and meaningful exit has disappeared, the relationship is not trust. It is control.

I am therefore particularly concerned by proposals involving certification of DeFi protocols and non-custodial wallets, embedded compliance mechanisms and restrictions on CASPs connecting customers to uncertified protocols.

Certification can produce useful evidence. It should not automatically become permission.

Non-custodial software does not possess the authority of a custodian. A genuinely decentralized protocol does not possess the authority of an intermediary. If nobody can seize assets, reverse transactions, unilaterally change the protocol, custody funds or decide participation, regulation should not require someone to acquire those powers merely to create an identifiable party capable of regulatory control.

That changes the trust architecture itself.

The same applies to money and settlement. Central-bank money for tokenized settlement and interoperability can provide value. But providing an option is different from designing the regulatory environment until it becomes the only economically viable destination.

Europe should not confuse strategic autonomy with architectural enclosure. Nor should institutional permission be confused with safety. European supervisors continue to document serious deficiencies and control failures within regulated financial institutions. Regulation can reduce particular risks. It does not eliminate vulnerability.

Risk moves.

If the state restricts an individual's ability to hold an asset, use open-source software, interact with a permissionless protocol or choose another settlement architecture, it should identify the specific vulnerability, justify the intervention, make its authority inspectable and preserve meaningful exit wherever possible.

That is the distinction between scaffolding and the cage.

Scaffolding establishes property rights, exposes material risks and hidden authority, protects against fraud and intermediary failure, makes legal relationships predictable, and allows identity to prove what needs proving without requiring every verifier to retain another copy of the person.

Then it leaves people room to choose.

The cage begins when protection becomes permission: disclosure becomes approval, certification becomes access, identity becomes persistent identification, intermediaries become mandatory, and legal recognition becomes conditional upon approved architecture.

Europe needs more legal certainty, not less. But it needs discipline about what legal certainty is for.

Regulate vulnerability where it exists: reserve risk at the reserve layer, issuer risk at the issuer layer, custody risk at the custody layer, fraud at the point of fraud, systemic exposure where it actually exists.

Do not manufacture intermediaries because they are easier to regulate. Do not manufacture control because it is easier to supervise. Do not make permissionless systems permissioned in practice while calling them permissionless in theory.

Specific vulnerability should justify specific authority. That authority should be bounded, inspectable and revocable. People should retain a meaningful ability to refuse it and choose another architecture.

That is Trust Architecture.

MiCA should be judged not only by whether each rule appears reasonable in isolation, but by where they lead when combined.

If every road ultimately leads through institutionally approved identity, intermediaries, assets, protocols and settlement infrastructure, Europe will not have built trustworthy digital finance.

It will have built a sophisticated system of capture and called it trust.