Bootstrapping an Ecosystem

As time progresses rollups will look more like public goods than they do today. The question of long-term economic sustainability is contingent on the demand for blockspace.

Rollups can contribute to this by funding public goods and using governance as a tool to distribute ecosystem funding. Users and developers will want to go where the action and attention is. We have seen this play out in the last economic cycle with the persistence of the bridge, farm & dump routine. Its reasonable to expect the next economic cycle to bring “L2 season” or “rollup wars” or w/e meme sticks to describe the activity.

The new activity brings not only retail users but in terms of volume and tx value, highly sophisticated actors with a lower latency view of the world and the skill-sets to express their granular preferences. These actors may be independent searchers or institutional cross-domain MEV experts, Mafia EV. This economic activity comes with potential benefits and undesirable trade-offs.

Random Half-baked Thought

Before we get into the tradeoffs let’s think about MEV. I’ll direct the reader to a quote from ‘Sexy Sun’

Random half-baked thought:

A relatively standard train of crypto is that MEV exist because the ordering protocol allows MEV to be extracted. But one could flip this argument on its head: perhaps one could argue that protocol emerge due to the existence of MEV to redistribute

I want to linger on this for moment. Protocols emerge due to the existence of MEV to redistribute. Is the claim testable? Would it pass Popper’s criterion of demarcation? Probably, not. It is though good science in that we are starting with an intuition (deduction) a priori and testing it for flaws. Maybe the claim is a philosophical wrapper that can help us see the state of the world as it us, not how we want it to be. Maybe this is why protocol’s exist.

One of crypto’s core use cases is financial disintermediation. Some have described DeFi as un-bundling an investment bank into its core primitives. Technically this might be a stretch but the analogy is still useful. Banks have historically occupied the privileged position to extract value from retail, institutional customers, and each other. Then there is also the risk free IOER-EFFR arb. As more fiat is converted to cryptocurrency more value can be captured on chain putting blockchains in the privileged position to capture value.

Unfortunately banks typically redistribute the MEV captured in the form of share buybacks, executive compensation, and acquisition targets. One could easily argue that stock buybacks are an efficient way to share profits with investors and helpful for the firm’s stock price. Financial firms collectively compose ~ 11.5% of the S&P 500 by weighting. Therefore by operating profitably and buying back shares firms here are lifting the overall index which is tracked by and sits in the 401k accounts of 58% of Americans, a benchmark for retirement savings. Stock buybacks can contribute to the overall welfare of the system. Is this optimal way to maximize welfare redistribution though?

During the financial crisis, politicians signed off on the largest fiscal spending package since the great depression. They did this because they knew without fiscal stimulus quantitative easing would take to long to ‘trickle down’ into the real economy. It is also likely that they were aware that poorest segment of the population that owned x% of equity would be unable to perform retail / service sector / working class jobs in a lock down / post lock-down environment. This also might not be the best way to maximize welfare redistribution.

Sequencer EV

If we accept this as a starting point then it becomes clear that L2s have an opportunity to redistribute MEV in a optimal way. This is a way that monolithic tradefi, national governments, and even Ethereum L1 protocol do not posses. Why? The former must strive to maintain credible neutrality at all times. Having an opinion that deviates significantly from public perception becomes untenable and risks the survival of the monolith. L2s are intrinsically biased because they pick the L1 they want to build on and typically start as a private organization, not a public good.

Additionally, L2s are building blockchains from scratch. There are centralized sequencers who user transactions must flow through. As an example here is Arbitrum’s design.

The Sequencer receives user transactions and orders them and then publishes the sequence. The Sequencer every several minutes will take a sub sequence put them into a batch using a general purpose compression algorithm that writes batches to L1. Therefore the sequencer is in a privileged position to order or insert their own transactions to capture any available MEV.

Already today Searchers who co-locate in the same Data Center as the permissioned sequencer have a significant latency advantage over avg users. These searchers can easily extract top of block back-run opportunities generated from users. Also, a variety of cross chain MEV is available; eg. arbitrage ~ CEX-DEX, L2-L1, etc.

Schools of Thought

Generally there are two schools of thoughts.

(1) Granular Fair ordering (FCFS → Aequitas → Themis)

Themis is a scheme for introducing fair ordering of transactions into (permissioned) Byzantine consensus protocols with at most f faulty nodes among n ≥ 4f+1.

~ Kelkar, Deb, Long, Juels, Kannan

Themis has the same communication complexity of the underlying protocol. It is a minimum modification to any existing partially synchronous leader-based protocol

Existing fair ordering protocols require replicas to gossip tx orderings to everyone else as their first step. This results in n^2 communication complexity. Themis requires replicas only to send fair ordering to current leader. The Leader uses n-f of these to compute a fair proposal. The Leader computes a SNARK and provides this with a block proposal. Replicas check proof to ensure fairness. Everything else can be taken as is from the underlying protocol.

Due to granular fair ordering users face less adverse selection eliminating some forms of MEV like sandwiching, without sacrificing on user experience.

Though it should be noted that sandwiching actually improves order routes and leads to better fills with less slippage than without. The axiom of Fair ordering relies on some assumptions like this one. There is also the existential question, Is it really fair?

(2) MEV Auctions

MEV Auctions is an idea put forth originally by the Optimism team. Specifically, MEVA is the idea that the Protocol would host an auction for the rights to order the transactions in n amount of blocks. These rights would be purchased by Alice in advance for n amount of blocks. The Optimism Protocol would capture the revenue from sales of future blockspace. This revenue - expenses = profit. The Profit could then be redistributed through the ecosystem via retroactive public goods funding voted on by the Optimism DAO.

The Scheduler concept from the ATOM 2.0 white paper is loosely based on this idea as well. The core concept is efficiently selling blockspace to the highest bidder making MEV extraction maximally efficient and allowing the protocol to capture the value rather than searchers/builders. The revenue would then be used to fund an On chain Treasury for the Cosmos Hub which would be controlled by an agent, The Allocater. This type of revenue model leads to recycling MEV back into the ecosystem via venture funding, protocol owned liquidity, over-collateralized lending, and retroactive public goods funding.

Conclusion

The First approach while very useful does not address the issue of how most efficiently to capture MEV and redistribute it. Fair ordering protocols by themselves treats MEV as if it is a Mole to be whacked rather than a revenue engine for resource allocation. I haven’t run simulations so I’m just speculating. However hard on-chain data is needed before writing fair ordering protocols off.

Historically, off-chain auctions have proven to be most effective way for users to express their preferences with respect to the ordering of transactions in their bundles and where their bundles lands within a block. This approach lends itself to selling blockspace in advance allowing protocols to capture MEV and redistribute to their communities through mechanism like retroactive public goods funding.

Contributions

Thank you to @vote_escrow for reviewing and providing suggestions. Conversation with Tom Walpo inspired some of the writing as well.