Musing on UMA: DeFi Infrastructure
Honestly, looking at the UMA website right now filled mainly with explainers and docs extolling the broad design space enabled by their open synthetic asset infrastructure, it’d be easy to think they are a young DeFi upstart hopping on the train. There’s no product interface, no dashboard of market metrics, nor even demo of functionality. The irony, of course, is that UMA (Universal Market Access) has been in DeFi far longer than anyone having such a thought — so why no dapp?
It’s a valid question given the established DeFi playbook pioneered by the likes of Compound, Synthetix, Uniswap, Aave — all projects conceived in the same pre-Cambrian DeFi Era as UMA itself. For those projects, the roadmap was basically some variation of:
1. Pick a function of the traditional financial system
2. Code up a set of smart contracts that serve that function in a trustless, decentralized way
3. Build an interface for users to access that functionality
4. Encourage developers to build on top of and integrate that functionality
It’s clearly an effective strategy: be a product, then a money lego.
UMA has opted to exist only as lego, building open financial infrastructure for the largest financial primitive, accounting for over $540 trillion dollars in often custom OTC deals: the synthetic contract.
There are plenty of resources out there to understand the basics of the UMA protocol, so I’ll summarize extremely briefly: the UMA protocol consists of an open framework for synthetic asset creation that tracks counterparties, margin accounts, economic terms, contract interactions, and an oracle mechanism. That oracle mechanism is UMA’s DVM, which is powered by $UMA holders and includes economic guarantees against corruption.
To understand UMA’s go-to-market strategy, a good analogy would be if Compound or Aave, rather than creating a full-fledged lending protocol and app, instead created the rails for various other parties to create lending apps with customizable interest rate curves and supported assets. One could imagine the likes of Argent and Zerion deploying their own set of lending contracts, conforming to Compound’s specs and likely tying back in some way, perhaps a unified insurance pool. But there’s one obvious reason for lending protocols not go that route: fractured liquidity.
All DeFi money protocols are in a competition to lock up as much liquidity in their contracts as possible. If a lending infra protocol allowed for many iterations of itself for various niche use cases, that liquidity would be split into each use case, rather than amassed in a few large pools. The question is, would that be so bad?
For Compound, the answer is likely, yes, though this is probably its own hypothetical debate. The question truly at hand is what will be the implication for UMA’s synthetic contracts?
There are two key implications that are immediately apparent: limited exposure and limited liquidity. The idea of limited exposure is specifically in contrast to Synthetix, which has one staking pool of SNX liquidity that funds their entire synthetic ecosystem. This single pool system exposes minters to a complex debt risk in which your debt from minting fluctuates in relation to the entire pool’s debt balance, ie. if you mint and hold sUSD, and sETH and sBTC rise in value, your debt owed increases. In UMA’s contract-specific system, when you mint a synthetic token, your exposure and debt are limited only to that token and contract, making for far simpler hedging and collateralization maintenance.
The more obvious implication of limited liquidity also creates a contrast to Synthetix as it relates to trading. Where Synthetix offers a native exchange offering zero slippage with prices set by Chainlink oracle, UMA’s synthetic tokens are “priceless,” which really just means it is up to the market to find equilibrium with the pegged asset price. As we’ve seen with DAI over the past few months, a synthetically pegged asset does not necessarily retain its peg in the open market. For UMA, this means that the dawn of synth liquidity mining in a couple weeks might drive demand to depeg the synthetic assets.
Ultimately, this isn’t a major issue as market makers are incentivized to mint more of the synthetic to capture the arbitrage, and the UMA contracts have a token expiry at which time they will be redeemable for their true value, enforced by the DVM. This fractionalized liquidity does make these synthetic tokens more susceptible to manipulation, however, which might limit synth composability in systems like Aave. This risk is obviously mitigated in synths that manage to garner enough liquidity.
Clearly UMA offers a synthetic system that has some advantages over Synthetix for certain use cases, but it’s hard to beat those SNX pumpamentals. Luckily UMA has a couple tricks up its sleeves here.
UMA’s oracle system, powered by $UMA holders, not only serves as arbiter for contract settlement, but has built-in economic guarantees against manipulation. The DVM’s contracts measure the Profit from Corruption (PfC), calculated by summing the value at stake in the contracts, and the Cost of Corruption (CoC), calculated by the cost of garnering 51% of $UMA tokens for voting. If the PfC ever exceeds the CoC, the protocol charges the underlying financial contracts fees in order to buy and burn $UMA to raise $UMA’s price and CoC.
While yes of course we love an economically secure model, DeFi has shown itself a bit more concerned with tokenomics — and this system greatly benefits $UMA holders. The DVM’s logic mandates that the price of $UMA rises with the amount of value locked in their financial contracts. This means investors can calculate a price floor for $UMA based on the amount of value currently locked in the contracts, as well as estimate future value based on a growth of that value locked. With a current $UMA price of $2.20 and circulating supply of 53M, the current CoC is about $59MM, which certainly outstrips the existing $250k locked in UMA’s yCOMP and yETHBTC synths. But with the recent liquidity mining explosion, it is not hard to imagine a scenario in which $UMA rewards incentivize the creation of additional synths and locked value overtakes the current CoC, which would necessarily lead to an increase in the price of $UMA.
While taking such an open infrastructure approach and creating a bespoke oracle are unique strategies, one of the biggest gambles of UMA’s strategy seemed to be the lack of a user interface. It pretty much guarantees a slow ramp up of users and TVL and limits exposure. For what must be regulatory reasons, you could only really interact with UMA financial contracts through Etherscan for weeks following the initial launch of their first synth, yETHBTC.
It would seem, however, that the community is rising to the occasion. Between the time I started this piece and posting it, a well-designed, minimalist UI has been created at usynth.xyz Additionally, two talented teams are currently building custom products using the UMA protocol. Potion Labs will use an AMM system and Black-Scholes pricing to build an options protocol that allows users to create custom strike prices on a wide variety of assets — offering the flexibility of Hegic with a wider variety of assets and a more rigorous technical system, courtesy of UMA’s long development and extensive audits.
Additionally, the well-hyped Jarvis Network is currently building their synthetics exchange on the UMA protocol, part of their developing DeFi ecosystem, which includes margin trading and wallet interface in addition to their synthetic exchange.
To sum up, UMA has built an incredibly flexible protocol that can be customized to meet the needs of the estimated $540 trillion synthetic derivative market, created a unique oracle mechanism with strong economic guarantees governed by $UMA holders, has a community of active builders, and is set to launch liquidity mining in the next couple weeks. Despite their unconventional route, their near term prospects are extremely bright, but what does the future hold?
While the openness of UMA’s financial contracts is a great feature, the openness of the DVM will allow UMA to bring remarkable tokenizations to DeFi. One way to think of it is Augur but linked to more complex financial contracts than prediction markets. This unlocks near limitless possibilities, such as a Celebrity Futures protocol, built by a Hack Money team, in which a user can long and short the popularity of celebrities.
As long as there is an ability for $UMA holders to ascertain the pegged value, the DVM can become an oracle for anything.
For instance, forget replicating Synthetix, let’s fork CeFi. FTX has some highly interesting products. Some enterprising individual with an FTX account could submit UMIPs (UMA Improvement Proposals) to have FTX-replicating synthetic financial contracts added to the DVM, mint FTX synthetic tokens, hedge on FTX itself, and offer users the same exposure one gets through FTX through Uniswap.
Ultimately, this is where UMA beats Synthetix in the long run. While you cannot deny the SNX pumpamentals, the single collateral pool and reliance on Chainlink oracles simply limit design space in a market that is based around customized, OTC deals in the traditional system. Much of the synthetic asset market requires this flexibility.
Not to worry, SNX Spartans, there will be space for more than one player here, but the UMA experiment to build the infrastructure for most of those players will use is certainly one to keep an eye on.