For many observers, the collapse of the crypto exchange FTX understandably raises questions about the future of the crypto economy, or even of public blockchains as a technology. The topic is high on the agenda of the U.S. Congress this week, with the House Financial Services Committee set for a Dec. 13 hearing with FTX CEO John J. Ray III and founder and former CEO Sam Bankman-Fried, followed by a Dec. 14 hearing of the Senate Banking Committee on “Crypto Crash: Why the FTX Bubble Burst and the Harm to Consumers.”
To some extent, the significance of the FTX case is likely to be exaggerated due to the outsized media attention that Bankman-Fried was able to generate. Nevertheless, many retail and institutional cryptocurrency holders were harmed by FTX and thus both users and policymakers will likely respond to what happened. In this post, I will contrast three perspectives on what may and should happen next for crypto.
‘Centralization Caused the FTX Fiasco’
The first perspective—likely the prevailing view in the crypto community—is that the FTX collapse was a failure of a centralized service, which should be emphatically distinguished from “true” or “crypto-native” decentralized services. The distinction between centralized and decentralized services is sharper in theory than in practice, and it should be seen as a spectrum of decentralization, rather than a simple binary distinction. There is, however, little doubt that crypto-asset exchanges like FTX, which predominantly operate “off-chain” (i.e., on their own servers, not on a public blockchain network) are the paradigmatic case of centralization in the crypto space. They are thus not “decentralized finance” (DeFi), even though much of DeFi today does rely on centralized services—e.g., for price discovery.
As Vivek Ramaswamy and Mark Lurie argued in their Wall Street Journal op-ed, the key feature of a centralized exchange (a “CEX”) “is that somebody (…) takes custody of user funds.” Even when custody is subject to government regulation—as in traditional stock exchanges—custody creates a risk that funds will be misappropriated or otherwise lost by the custodian, as reportedly happened at FTX.
By contrast, no single actor takes custody of customer funds on a decentralized exchange (DEX); these function as smart contracts, self-executing code run on a blockchain like Ethereum. DEX users do, however, face other risks, such as hacks, market manipulation, bugs in code, and situations that combine features of all three. Some of these risks are also present in traditional stock exchanges, but as crypto insiders recognize (see below), the scale and unpredictability of risks like bugs in smart contracts is potentially significant. But as Ramaswamy and Lurie observe, the largest DeFi protocols like “MakerDAO, Compound and Clipper hold more than $15 billion, and their user funds have never been hacked.”
Aside from the lack of custody, DeFi also offers public transparency in two key respects: transparency of the self-executing code powering the DEX and transparency of completed transactions. In contrast, part of what enabled the FTX debacle is that external observers were not able to monitor the financial situation of the centralized exchange. The solution commonly put forward for CEX services on the blockchain—proof of reserves—may not match the transparency that DEX services can offer. Even if a proof-of-reserves requirement provided a reliable, real-time view of an exchange’s assets, it is unlikely to be able to do so for its liabilities. Because it is a business, a CEX always may incur liabilities that are not visible—or not easily visible—on the blockchain, such as liability to pay damages.
Some have proposed that a CEX could establish trust by offering to each user legally binding “proof of insurance” from a reputable insurer. But this simply moves the locus of trust to the insurer, which may or may not be acceptable to users, depending on the circumstances.
‘The Ecosystem Needs Time to Mature Before We Get Even More Attention’
As a critique of today’s centralized crypto services, the first perspective is persuasive. The implication that decentralized solutions offer a fully ready alternative has been called into question, however, both within the crypto space and from the outside. One internal voice of caution has been Ethereum founder Vitalik Buterin, one of crypto’s key thought leaders. Writing shortly before the FTX collapse, Buterin said:
… I don’t think we should be enthusiastically pursuing large institutional capital at full speed. I’m actually kinda happy a lot of the ETFs are getting delayed. The ecosystem needs time to mature before we get even more attention.
… regulation that leaves the crypto space free to act internally but makes it harder for crypto projects to reach the mainstream is much less bad than regulation that intrudes on how crypto works internally.
Following the FTX collapse, Buterin elaborated on the risks he sees for decentralized crypto services, singling out vulnerabilities in smart-contract code as a major concern.
Buterin’s vision is one of a de facto regulatory sandbox, allowing experimentation and technological development, but combined with restrictions on the expanding integration of crypto with the broader economy.
Centralization Will Stay, but with Heavier Regulation
It is even more understandable that observers who come from traditional finance have reservations about the potential of decentralized services to replace the centralized ones, at least in the near term. One example is JPMorgan’s recent research report. The report predicts that institutional crypto custodians, not DeFi, will benefit the most from FTX’s collapse. According to JPMorgan, this will happen due to, among other factors:
- Regulatory pressure to unbundle various roles in crypto-finance, such as brokerage-trading, lending, clearing, and custody. The argument is that—by combining trading, clearing, and settlement—DeFi solutions operate more efficiently than centralized services and will thus “face greater scrutiny.”
- DeFi services being unattractive to large institutional investors because of lower transaction speeds and the public nature of blockchain transaction, both of which run counter to trading history and strategies.
The report listed several other concerns, including smart-contract risks (which Buterin also singled out) and front-running of trades (part of the wider “MEV” extraction phenomenon), which may lead to worse execution prices for a trader.
Those concerns do refer to real issues in DeFi although, as the report notes, there are solutions to address them under active development. But it is also important, when comparing the current state of DeFi to custodial finance, to assess the relative benefits of the latter realistically. For example, the risk of market manipulation in DeFi needs to be contrasted with how opaque custodial services are, creating opportunities for rent extraction at customer expense.
JPMorgan stressed that the likely reaction to the FTX collapse will be increased pressure for heavier regulation of custody of customer funds, transparency requirements and, as noted earlier, unbundling of various roles in crypto-finance. The report’s prediction that, in doing so, policymakers will not be inclined to distinguish between centralized and decentralized services may be accurate, but that would be an unfortunate and unwarranted outcome.
The risks that centralized services pose—due to their lack of transparency and their taking custody of customer funds—do not translate straightforwardly to decentralized services. Regarding unbundling, it should be noted that a key reason for this regulatory solution is to prevent conflicts of interests. But a DEX that operates autonomously according to publicly shared logic (open source code) does not pose the same conflict-of-interest risks that a CEX faces. Decentralized services do face risks and there may be good reasons to seek policy responses to those risks. But the unique features of decentralized services should be appropriately accommodated. Nevertheless, it is admittedly a challenging task, partially due to the difficulty of defining decentralization in the law.
The collapse of FTX was a failure of a centralized model of crypto-asset services. This does not mean that centralized services do not have a future, but more work will need to be done to build stakeholder trust. Moreover, the FTX affair clearly increased the pressure for additional regulation of centralized services, although it is unclear whether it will prompt certain specific regulatory responses.
Just before the FTX collapse, the EU had nearly finalized its Markets in Crypto-Assets (“MiCA”) Regulation that was intended to regulate centralized “crypto-assets service providers.” There is an argument to be made that MiCA might have stopped a situation like that at FTX, but—given the vague general language used in MiCA—whether this would happen in future cases depends chiefly on how regulators implement prudential oversight.
Given the well-known cases of sophisticated regulators failing to prevent harm—e.g., in MF Global and Wirecard—the mere existence of prudential oversight may be insufficient to ground trust in centralized services. Thus, JPMorgan’s thesis that centralized services will benefit from the FTX affair lacks sufficient justification. Perhaps, even without the involvement of regulators, centralized providers will develop mechanisms for reliable transparency—such as “proof of reserves”—although there is a significant risk here of mere “transparency theatre.”
As to decentralized crypto services, the FTX collapse may be a chance for broader adoption, but Buterin’s words of caution should not be dismissed. JPMorgan may also be right to suggest that policymakers will not be inclined to distinguish between centralized and decentralized services and that the pressure for increased regulation will spill over to DeFi. As I noted earlier, however, policymakers would do well to be attentive to the relevant differences. For example, centralized services pose risks due to lack of transparency and their control of customer funds—two significant risks do not necessarily apply to decentralized services. Hence, unbundling of the kind that could be beneficial for centralized services may bring little of value to a DEX, while risking giving up some core benefits of decentralized solutions.