We spoke to a consultant who reviewed Facebook's Libra about its stability, legality, and future

Anonymous Consultant: There is a lot of work that goes into mapping Virtual Machine state to the underlying Blockchain. Telegram’s TON spends 150 pages describing this mapping and it can be hard to implement correctly. I think it ends up being an advantage if your consensus protocol, VM, and language are co-designed. It becomes hard to extricate one part and inject it elsewhere without a performance or security hit.

The Block: The Libra white papers were noticeably lacking from an economics specification standpoint – expected gas fees, validator rewards, etc. What kind of economic descriptions would you have liked to have seen? Presumably this is more relevant to the ‘permissionless’ stage of Libra’s life cycle?

Anonymous Consultant: I think Basis, Celo, Terra, and Reserve did much more to instill some base confidence in their reserve and MKR does a really good job with their audits.

The Block: On the topic of underspecification, there was no formal description regarding the composition of Libra’s reserves and its rebalancing mechanisms. How can the Libra Association reconcile composition and rebalancing transparency while avoiding front running from hedge funds? What kind of max volatility vs. USD do you think is acceptable? Is taking a Special Drawing Rights-esque approach sensible?

Anonymous Consultant: Proof of Reserves for exchanges is a hot topic in centralized exchange land and can be employed in a smart contract.

Money markets are less complicated than SDRs — after all, the U.S. economy has < 25% of dollars in the “real” economy. Most money market reserves are in repo/bonds and the shadow banking of overnight lending, especially post crisis, has increased a lot.

I do think that you'll end up holding swap-like instruments (e.g. MakerDAO achieves stability as a swap between long volatility, short duration traders and short volatility, long duration users) that swap with SDRs, but I'm not sure you need dollars directly to hedge out volatility.

The Block: Libra Association has pursued the Swiss foundation model, initially popularized by Ethereum and widely pursued throughout 2017/2018. Is there a reason to be skeptical about this Foundation model?

Anonymous Consultant: I think the skepticism is that this model ends up having governance problems that we don't seem to have with U.S. C-corps (e.g. Tezos).

The Block: What are the major considerations asset managers will be thinking about when determining how to approach Libra? What are the major risks? What kind of guarantees will they need to participate as liquidity providers? Does Libra qualify as a swap under Dodd-Frank?

Anonymous Consultant: Liquidity, counterparty risk, margin. 

Listing is easy, but liquidity, derivatives, and tools for traders (margin, etc.) are harder. You also need to provide proof that you're fair — which is hard when bootstrapping and you're forced to give early market makers advantages. Other infrastructure includes exchange technology with performance guarantees (this is hard! It took CME forever) and standardized products.

As for other considerations: what if the association doesn’t allow members to trade more than x% of the float and this is too low for many [market makers] to participate? It can be confusing, given that the membership agreement with regard to trading is opaque.

I suspect that the Libra Association does not want transactions with the reserves to be considered a swap, so one would need to tread carefully here. Delivering this mechanism in a mathematical formalism and description that makes sense to traders and regulators is likely of the utmost importance for the Libra Association.

The Block: Part of Libra’s mission statement is to provide financial services for the 1.7 billion unbanked. A large driver of financial services exclusion is a lack of state issued identification. How will Libra address this?

Anonymous Consultant: I'm unconvinced by most ID initiatives that I've seen to date.

The Block: Libra’s gas model was largely underspecified, although it seems as if they will be using a similar first-price auction structure as Bitcoin and Ethereum. As demonstrated across Bitcoin and Ethereum, a bad gas model can have unpredictable, high-variance costs, aggressive front-running, and difficult optimizations. What does this mean for average developers? Could a Vickrey Clarke Grove auction structure be a solution?

Anonymous Consultant: The average developer works in a high-level language and doesn't think about op codes, register files, caches, or memory hierarchies. No web developer is going to analyze transaction fees and ordering in the Web Assembly.

Batch auctions prevent front-running in that they give a “slack” that allows formally incentivized users to post their true valuations. But VCG is exponential in basket size, so you do need heuristics — akin to those used in Google’s ad auctions.

The Block: There seems to be a lot of confusion around the ability to run non-validator nodes and discussion that third party clients will really just be connected to a Libra API. Are these third-party clients analogous to full nodes in Bitcoin and Ethereum or are they more similar to non-verifying light clients?

Anonymous Consultant: The nodes that Bison Trails will be running for members are full nodes, but I'm not sure how the interaction will work externally.

The Block: How will the Libra Association manage the Libra Core repository? What kind of Pull Request policy will it use? And how viable would it be to fork Libra considering the native digital asset is backed by real world currencies and bonds?

Anonymous Consultant: Intuitively, a massive multisig might be good enough to secure the reserve – unlocking will be hard. In practice, they might end up using aggregated signatures like BLS.

As for Pull Request policy, it might behoove the Association to specify its current thoughts.

The Block: What role does the derivatives market have to play in relation to Libra and what concerns might the CFTC have when it comes to approving a derivatives products tied to Libra?

Anonymous Consultant: The CFTC’s hesitation to admit an unrestricted license for an Ethereum futures product (as expressed via LabCFTC’s Request for Information) is direct evidence that regulators believe that derivatives markets for cryptocurrencies, especially when switching from Permissioned and/or PoW to PoS, can dramatically affect market behavior and protocol security.

Specifically, the attack vectors include front-running of rebalances and synthetic stake attacks, where users can use the derivatives markets to “coordinate” bribes – e.g. accumulate stake by buying token settled futures and take delivery of a sizeable portion at expiry and over take a shard with lower than 51% safety.