Here are a few excerpts from Walch’s book chapter, and responses:
Angela Walch: The question becomes, then, whether the benefits of public blockchains serving as the underlying technology of FMIs (e.g., reduction of settlement risk) can justify the […] operational risk [of rough consensus].
CleanApp: Since inception, blockchains (including Bitcoin) were not “financial infrastructure.” ‘Finance’ & FMI is the problem that Bitcoin set out to solve. See Bitcoin WP, pg 1. So too with Ethereum, arguably. The original goal was arbitrarily complex blockchain relationships “for more than just money.” So the narrow question is: shouldn’t folks who choose to build their FMI stuff on top of Ethereum be responsible for risk of loss, not Ethereum?
Angela Walch: This paper contributes to the public-versus-private blockchain debate, explicating how the use of traditional grassroots open source software practices in public blockchains would expose any financial market infrastructures they undergird to new and potentially increased operational risks in exchange for the benefits they seductively promise. There are tradeoffs to all improvements we make, and in this case, the new operational risks seem quite hefty.
CleanApp: But “public blockchains” aren’t choosing to “undergird” FMI. Some blockchain actors think of blockchains as FMI, but definitely not everyone. Some core developers want blockchains to be “finance” or FMI but not everyone. So why should developers of a project that has multiple yet-unknown financial and non-financial uses be held to a heightened FMI governance standard? If anything, shouldn’t those governance standards apply to, say, L2 DeFi developers who want Eth recognized as money & are actually building dFMI?
Angela Walch: As my analysis will reveal, a lack of defined or accountable control over these processes generates operational risks for public blockchains, impacting their suitability to serve as financial market infrastructure (or other critical societal systems).
CleanApp: By definition, public blockchains are premised on tolerance for (& mitigation of) the operational risk that comes with governance by rough consensus. As Walch points out in the chapter, rough consensus works, evidenced by 99%+ uptime (including in crisis governance situations). Further, mitigation of ‘operational risk’ is an existential need for Bitcoin/Ethereum & progeny. This is why traditional financial institutions are so keen to latch onto (& build on top of), say, Ethereum.
Actors in the “broader financial system” are making those FMI decisions, and they should be accountable/responsible for those decisions if their decisions result in losses. The same is true for developers of “other critical societal systems” who want to graft their (say, decentralized identity) processes onto the Ethereum Merkle tree.
The key point here is that lots of different stakeholders are building financial and many non-financial applications (dID, games, dispute resolution processes, SCM processes, NFTs, DAOs, LAOs, etc.) on top of Ethereum and other blockchains. Stakeholders have baseline expectations that certain characteristics of their blockchains will remain largely unchanged. Forcing blockchains to conform to FMI standards uproots these expectations and arguably contradicts their very raison d’être, with net negative economic and social effects.
Angela Walch: It is worth thinking through the implications of this governance model, particularly in the context of a public blockchain supporting financial market infrastructure (or any other critical public system, really). In considering the governance implications I describe in the following paragraphs, I ask the reader to imagine using this type of governance model with our military defenses (e.g., nuclear weapons) or in an intensive care hospital unit, to concretize how ill-suited this model is for high-stakes matters.
CleanApp: Please see the first point about the original intent of blockchains. Public blockchains were NOT invented to play a supporting role to existing or new financial market infrastructure (or any other critical public system, really). Public blockchains were invented to solve a very simple set of transactional/relationship problems: how to have peer-to-peer interactions while minimizing reliance on a trusted third party. Blockchains solve the simple peer-to-peer value/data transfer problem and the solution set is now being scaled to far more complex transactional (value) and non-transactional relational (eg, pure data) settings.
Are there prominent folks who disagree with this characterization? Yes, many; but these are folks whose primary motivation may be to legitimate their vast financial gains, as opposed to, say, further development of blockchain non-financial use cases (such as tamper-resistant climate datasets, tamper-resistant global legal datasets, etc.).
Angela Walch: Third, this amorphous governance model can lead to unacknowledged centralization of power, resulting in unaccountable or unchecked power. The core developers of public blockchains are more powerful than the rank-and-file developers on these projects.
CleanApp: 100% agree.
Angela Walch: The balance between concentrated and distributed power is difficult to strike, but the standard grassroots open source software development process [eg: rough consensus] appears too far along the spectrum of (nominally) distributed power to govern critical systems like financial market infrastructure.
ClanApp: Pls see above.
Angela Walch: “governance practices in public blockchains [eg: rough consensus is] unlikely to satisfy [the standard that] ‘FMI[s] should have governance arrangements that are clear and transparent, promote the safety and efficiency of the FMI, and support the stability of the broader financial system[.]’”
CleanApp: Imposing FMI standards onto public blockchains à capture of public blockchains by “the broader financial system.”
Angela Walch: Analogously, volunteer fire departments are relatively common in small towns, but cities pay fire departments to fulfill this important public function — the scale of the systems seems to dictate a more formal structure being needed in the more populous cities.
CleanApp: How did hyper-formalism work out for Notre-Dame de Paris? The point is that decentralized systems and processes do certain things much better than centralized systems and processes. This is as true in biology, computer science, as it is in economics, the social sciences, academia, etc. In fact, the law itself if one of the best examples of a set of systems and processes that is, arguably, optimal in a decentralized form.