UMA makes it possible for you and me to invest in derivatives i.e. it allows us to gain exposure to an asset without actually owning the asset.
Let’s say you like Tesla but the country in which you reside does not allow you to buy US stocks. UMA facilitates the creation of synthetic Tesla stocks that you can trade irrespective of the jurisdiction in which you are based. Or let’s say you like gold but instead of buying gold outright, you want to use the more capital efficient “Total Return Swaps” to gain exposure. Brokers however offer Total Return Swaps only to institutions and accredited investors. UMA makes it possible for you to invest in Total Return Swaps irrespective of your investor status.
=> Net-net, UMA enables a single global platform on which any individual irrespective of nationality, wealth, status, caste, class, gender can buy and sell any form of financial risk.
How do such synthetic assets work in traditional finance?
Let’s say that Susan likes gold and believes its value will go up but doesn’t want to pay the full price of gold. Alex owns some gold. He wants a return of 5% on the gold he owns i.e. he is willing to give up the upside from appreciation in the value of gold as long as he gets his 5%.
Susan and Alex, therefore, enter into a Total Return Swap Agreement where Susan pays Alex 5% on the value of gold, and in return, any capital appreciation (or depreciation) of gold accrues to Susan. To secure the contract, both parties deposit say a 7% margin (or $70k). At the end of the year, Susan will owe Alex $50k (5% interest on $1million) and Alex will owe Susan the capital appreciation in the value of gold. So if gold climbs 9% then he will owe Susan $90k. The contract will settle with Alex paying Susan $40k (90k-50k) from the margin already deposited.
This works for both parties. Susan earns the upside from appreciation in the price of gold. Alex can sell the gold he owns to cover the amount he owes Susan and still be left with 5% that he initially wanted.
During the one-year period, appropriate collateral/margin amount must be maintained i.e. if say the value of gold climbs 20%, then Alex must post additional collateral of $200k. If sufficient collateral is not posted then the contract is liquidated with Alex’s $70k initial collateral being taken by Susan with no possibility of Alex recovering the money even if the value of gold drops in the future.
So where does decentralization and UMA come in?
The key issue with this type of instrument is counter-party risk- the luxury to trust that the counterparties to the contract will honor their commitment i.e. pay their dues. Traditionally one needs centralized agencies to do “due diligence” and rely on expensive legal procedures to enforce such contracts i.e. enforce the right to liquidate the contract if sufficient collateral is not posted with the assurance that the contract will not be liquidated when sufficient collateral is posted. The complexity and cost of such a transaction can therefore only be handled by large institutions with teams of legal experts.
Of course, even large institutions have a tough time ensuring that the 2 parties pay their dues. The recent Archegos fiasco highlights this risk where Archegos Capital defaulted on margin calls and caused Credit Suisse and Nomura to take losses of billions of dollars.
This is where UMA comes in. It provides a template for the creation of financial smart contracts on the Ethereum blockchain on any underlying asset and uses economic incentives to get the counterparties to always post sufficient collateral without the need of a central authority policing their actions.
Let’s get into the details
Susan and Alex enter into the same contract described above where Susan pays Alex 5% on $1mn and Alex pays Susan the capital appreciation on $1mn of gold. The margin requirement is 7% but both deposit the margin of $100k (10%) in a smart contract to provide an additional buffer for themselves.
The contract has the following details:
* Susan and Alex’s public addresses.
* Susan and Alex’s margin sub-accounts in which both parties deposit the margin.
* Economic logic to determine who owes what i.e. Alex owes Susan: (%change in the price of gold-5%[annualized])*1mn.
* The oracle of choice from where pricing data will be gathered.
* Termination terms like expiry date, settlement procedure, default procedure, default penalties (say 3%), early termination conditions, etc.
This agreement will work exactly like a traditional Total Return Swap Agreement except it will execute automatically through a smart contract.
Now let's consider the constraints placed on such a derivative contract by the architecture of the blockchain
Usually, smart contracts that require the price of an outside asset (like gold), use oracles that update prices every 5 seconds or so, to check if sufficient collateral has been posted. Such price updates are expensive and can clog the blockchain pipes. Also, there is a possibility of a corrupt oracle providing incorrect price information and causing an unfair liquidation with the counterparties having no mechanism to dispute the liquidation.
UMA realized that the cost of on-chain computation is expensive but the cost of off-chain computation is cheap. So it transferred the responsibility of monitoring margins to the counter-parties. Alex and Susan, therefore, monitor the price of gold themselves and run a remargin() function only if the change in gold price requires a change in the margin. For example, if gold price increases by 5%, it is in Susan’s interest to run the remargin() function. Knowing that Susan is likely to run the remargin() function, Alex will have already posted additional collateral because the smart contract will automatically liquidate in Susan’s favor if and when she runs the remargin() function if Alex is below the required 7% margin. The contract will terminate with him losing 5% which is the drop in value of gold + the 3% default penalty.
=> Another way of explaining the UMA mechanism is to say that UMA uses “priceless” financial contracts i.e. contracts that use on-chain price feeds only in case of disputes. Otherwise, they depend on the counter-parties to monitor prices and maintain sufficient collateral on their own.
So what happens if there is a dispute?
In case of dispute, a dispute resolution protocol called Data Verification Mechanism (DVM) is called into action.
Liquidators constantly monitor if a position is properly collateralized. They are rewarded for identifying under-collateralized contracts and liquidating the same. Incorrect liquidations attract penalties. The liquidation mechanism ensures that the parties are always collateralized correctly even without oracles regularly feeding price data to the contract. The fear of liquidation keeps them honest.
Disputers exist to remove the possibility of incorrect liquidations. Those who believe that a contract is being unfairly liquidated can dispute the liquidation. They will be rewarded if they block an incorrect liquidation. In case they’re proved wrong, they are penalized.
=> Frivolous liquidations and disputes are unlikely given the penalties for the same. Liquidators/Disputers have to stake a bond when they propose a liquidation/dispute and the party that is proved wrong loses its bond.
Data Verification Mechanism (DVM) is UMAs oracle. DVM resolves the dispute by getting UMA token holders to obtain the price of the underlying asset at a given timestamp. Token holders are rewarded for providing correct information. Of course, the oracle isn’t actually needed unless there is a dispute and contracts don’t generally go into dispute.
Avoiding Oracle corruption
UMA’s oracle system is designed such that:
Cost of Corruption (CoC) > Profit from Corruption (PfC)
=> Therefore, it would never be worthwhile to corrupt the system for profit
Cost of Corruption (CoC) - In order to influence the oracle vote, one must control 51% of the UMA token holders that vote in the DVM. The CoC, therefore, is the market value of 51% of UMA token holders that vote.
Profit from Corruption (PfC) - If one gains control of 51% of the UMA tokens that vote then one can potentially drain all the collateral posted in various UMA financial contracts. The sum of all collateral posted across all registered UMA contracts, therefore, is the PfC.
=> The way to maintain CoC>PfC is to ensure that the price of voting tokens is higher than 200% of collateral in the system. If the CoC falls closer to PfC, the Risk Labs Foundation draws funds from registered UMA contracts and uses them to buy UMA tokens raising UMA token price and hence CoC.
UMA Tokenomics
UMA is an ERC20 token with an initial supply of 100 million tokens. They launched the token through an Initial DEX Offering (IDO) on Uniswap where they sold 2mn tokens at a price of 26 cents each. Of the remaining tokens, 14.5mn are for future token sales, 35mn for developers and users, and 48.5mn to founders.
UMA is a utility token used for:
Governance - Token holders vote on UMA Improvement Protocols (UIPs) for which they receive an inflation reward.
Dispute Resolution - Token holders settle disputes and hence protect from oracle manipulation.
Problems with UMA
* The role of oracles is performed by UMA token holders who may mostly be investors without the expertise to determine correct data. UMA token holders are rewarded for voting honestly and correctly but I’ve been unable to find data on what percentage of UMA token holders participate in such disputes. If the number is small then there is a possibility of oracle manipulation.
* It is in the interest of liquidators to liquidate under-collateralized positions as long as the position has some collateral left. For example, if gold price suddenly rises by 5% and Alex has not added collateral to the $100k already deposited then he is under-collateralized as he needs to have a margin of at least 7%. The liquidator in this situation can claim the 5% remaining collateral. On the other hand, if the price of gold rises 20% and the collateral deposited is still only $100k, then there is no incentive for liquidators to liquidate as there is no collateral left in the contract to collect.
* With 48% of UMA tokens with founders, the protocol‘s decentralization is questionable. Maybe if they could code in that founders won’t vote in oracle votes, the DVM would become more trustworthy.
Final Words
Like the rest of DeFi (Decentralized Finance), UMA is motivated by the unfairness that has crept into traditional finance and aims to democratize access to financial assets. I believe that this is a genius product (issues identified above notwithstanding) with a lot of thought having been put behind the smart contract template, economic incentives, and dispute resolution mechanism. I prefer UMA to the Synthetix model.