DappCon 2023: Building a Stablecoin Risk Framework - Tammy Yang

Transcript

uh hi everyone my name is Tammy and today I'm giving a presentation over a project that I'm currently working on is to build a um a risk assessment framework to assess stable coin risk from different angles and in the next 20 minutes I'm going to cover the motivation and then the project goal what are the different stable coin risk being identified in the risk assessment framework and what are the challenge of having a assigning a overall rating to a stable coin at the end

yeah wait few yeah well they're trying to fix the the the the machine issue but I can give a little bit background about myself so oh always working so about me I'm currently uh working in a crypto exchange specialized in blockchain analytics before that I was a credit risk modeler for wholesale Banking and if you are interested in Oracle design I also have

a research paper been published recently it's about relationship between onchain and offchain Market data and about this project it's a personal project that independent from my current employer about I started about 6 months ago after receiving a grant from eum foundation and it will be completely open source after it's done so before getting into all the details about the different risks um that exposed to stable coin holder I would like to review the two significant stable coin event that happened in the last two years because we're going to use them as an example when we analyze

the risk the first one is the famous Fall Of Terror so what happened is it started with USD slightly depack after the market saw a large amount of uh USD deposit being removed from the anchor protocol as well as a large amount of us been dumped into the curve three pool uh lfg which is the Luna Foundation G tried to recover the pack by selling BTC but um it wasn't successful um at the time there are two way to um exit for us holder one is to sell us to the curve us3 pool and then the other one is for is to burn USD by

maintain Luna then um if anyone has a little bit knowledge about how us work you know that it is backed by Luna which Luna does not has any other use cases outside the stable coin systems so when Us's demand is up lunus price goes up and when us uh when us demand goes down lunus price goes down as well in the crisis of confident more and more us holder want to exit and then you know um Luna and us went into the death spiral so you in just in a couple of days us and Luna went um us value dropped from one US dollar to below 10

cent and then the next significant stable event was triggered by the collapse of the Silicon Valley Bank um on in mid March this year then um the svb has been shut by the US regulator and then all deposit withdrawal has been HED some usdc holder got worried because um svb is one of the uh custodian for the usdc cat res so they sold their usdc as at the time it's not clear that what is the portion is stuck with svb um usdc when depack and followed by

other stable coin who has direct or indirect exposure to usdc like uh D is one example where um D enable one to one swap for usdc to die in its pack the blood te module PSM we also saw a run on the um so we also saw gusd and usdp Reserve in the PSM has been drained which is show in the diagram here and then the Dune diagram shows the token breakdown for the curve free pool during the dep weekend then you can see um large amount of die and usdc has been dump into the curve three pool so on Sunday Circle

revealed that only a very small portion of usd's cash Reserve is stuck with svb then um usdc value starts going back and then on the following Monday FDIC promised that all the um they will enable all the deposit withdrawal for um svb then the usdc is back to $1 from the two example we just went through we know that the two general concern arises from holding a stable coin one is the the risk that um stable coin significant deact and has a negative impact to The Wider ecosystem in this scenario the value of the stable

coin may go back to the pack like the usdc uh example we walk through or it can also can be the case that stable conin holder may suffer from the haircut and then a second concern is the risk that the entire stable conin project collapse in this case value of the St coin cannot be recovered and that's the US's case and then the concern hasn't been demonstrated in the two example but has been discussed quite frequently is the risk of user being sensored from using the stable con that can be done by a centralized party who control the stable Con in this case the value is there but user cannot assess to the value there's many in-depth stable con

risk analysis but for common stable Co holder these are the matx that they use the most uh and then the most easily accessible to compare stable coin project they use um markets capitalizations circulating Supply um and then you can have matric that are not as easily accessible but has been mentioned uh in many um stable coin related discussion like historical price volatilities value of collateral liquidity and the incentive to hold but if you think about the two Cas system for Terror the trigger of the terror uh the failure of the Tera is because there's um fun fundamental flaw in their

uh stable con system they are under calized by Design and they use uh Luna which has the circular value that can go down to zero and for usdc is a temporary pack um they are fully B by cash and cash equivalent that's what people know but um it's temporary depack was triggered by the custodian risk they are exposing to which uh the custodian risk um the cash Reserve exist as a bank note so with all this matric on this slides here even the combination of them are not enough uh for analyzing stable coin

risk so this is uh uh my project goal I call it SRA stable coin race assessment framework to build a multidim ditional uh stable coin risk assessment framework to measure stable coin risk from different angles and then it's easy for common stable coin holder to assess the information and then easy for them to compare stable coin risk and most importantly to encourage more risk-driven decisions rather than Trend driven or return driven the risk assessment framework um further break the risk assessment into eight different categories uh ass set backing stability mechanism governance insurance and Redemption liquidity smart

contract Oracle and uh regulatory risk we believe that stable coin which has been assigned with the same risk values exposed to similar type of risk and then um on a very high level asset backing and stability mechanism are the two main area have been used to describe how a stable coin work so in the rest of presentation I'm going to focus on these two category but other categories are also important the asset backing risk um is the risk that the stable core may become under calized when um the asset that you use the asset that you use to bet the

primary value uh decline so staple con project normally choose asset from one of the asset type here uh exogenous asset which is the asset that have other uses outside the stable coin system um examples are ethereum Bitcoin uh Fiat and other stable Cod and then the other one is the endogenous asset which the asset value arises from the stable con system itself like equity in the system then the third type um some stable con used to have um uh have a design that with no explicit backing and they rely on reward mechanism to adjust the supply

and then the stabil uh to stabilize the price but they are not here anymore so I'm not going to the details of this category for end dogenous um asset the risk is that it is its value is derived from the stable coin itself when the demand is high then the endogenous asset value is high when the stable con demand is low then the asset Val endogenous asset value is low it will work in when the market is good but when there's um crisis of confidence the sell pressure on the stable coin

will trigger uh will may lead to B value going down together um SOA is a is a example but I want to mention that it does not mean all the endogenous as set a bad so back in 2020 D did sell um maybe eight million of mkr to cover the uh the shot four the fundamental difference is like um in in tera's design they use um majority of its circul ating value major majority of its primary value are bet by the endogenous

Luna from the beginning so um when things happen that or we know they wenter into the that spiral so let's imagine like for for working stable con system for every one new stable con let's assume is a US dollar pack stable coin uh been new stable coin been minted then it's supposed to have one1 US dollar worth of um asset to be entered into the stable con system but for Us's design for every new uh USD design that USD is uh minted then $1 worth of Luna is burned in this case Supply down and

Luna's price goes up Luna hold a benefit from the newly mined uh USD so it is almost like a bank is using using uh client deposit to execute the stock buyback um this is more details on the exogenous asset backing mechanism they normally choose from uh this asset type here um so stable G choose trapi asset will expose to traditional Market risk as well as the uh custodian risk and sensor racek that we learn from the USD pack and then some may choose other stable coin to back their primary value so it will expose to the risk that the

other stable coin there might be bad thing happened to other stable coin like um other stable coin deack or other stable coin has been seized by centralized party that's what we learn from D um four there are other cryptocurrency being used in the system they are normally more volatile compared to the other two and then used in the um more often used in The Leverage component like um ethereum in the in in in makers MCD um or etherum in LUSD 12 this is example of how the throughout framework assets um asset backing risk

we have six um risk factor defined types whether this is a custodial or noncustodial stable coin uh asset type as I just discussed and then we also have asset breakdown validation of real world asset asset location and information quality uh please note that the usdc asset breakdown is not uh updated it's not it's kind of outdated here um for usdc how they validate their real world that is they have a monthly attestation and annual audit and for usdt they have a quarterly attestation and for D since they use tokenized real world asset I still haven't figured out a way to find

out how they validate it so more work need to be done there and then for asset location us DC's um uh Reserve is with us Banks and then for usdt it's unknown it's maybe Bahama Banks and then for die you have the asset on chain and then for chain is for the real world assets information quality this is depends on how valid the information provided by uh the stable coin project um for that it's high because some of the asset is on chain is all transparent and then um I haven't figured out for the tokenized rwa so I put as a question

mark so let's talk about the stability mechanism um it is important to know about the to to analyze the stting mechanism of stable coin project so you know how they may fail to stabilize their value um on a very high level uh stable con stable con project use the reserve mechanism or The Leverage mechanism to St uh in to stabilize their value they can choose either or use both stable con project that use the reserve component uh in their design exposed to a similar risk which is the risk that

the stable coin can be become under calized if the asset in the reserve component decline then the for the stable con that also have a leverage component the reserve component will help to mitigate the risk of short squeeze but they will still experience the risk where if the Cal that used in The Leverage post uh component decline too fast then the liquidation might become too challenge or too expensive to be executed and then you have another type where you only have The Leverage component then without the reserve

component or without a negative rate setting the stable col may have a risk to suffer from short squeeze like the original diet design before they have the PSM this is SL is about about the governance risk to understand what governance risk are um our stable cor holder exposed to you need to understand firstly who owns the governance process how fast is the decision turn over the level of decentralization and transparency and what safeguards are there against Bad governance it is for most of the stable coin smart contract available they are

either controlled by trusted and entity or they are upgradeable so um so that they can update their risk parameter later or they can fix the smart contract if a programming bug has been identified but it is also very important to have um resilient governance uh designed to ensure that um to prevent like a malicious player to have um power to influence the decision- making process uh for example they make decision to benefit themselves by Harting other stable coin holder so um the current governance design

normally uh adopt one of the uh design here one is um a trusted base that's for all the custodial stable con they have a trusted entity to control the decision- making process and then the other one is uh by majority vote like the Dow and then the third one is the interesting one where all the parameter are already preset in the smart contract so there's no room for them to change after smart contract being deployed to the main net in this case um it's good that there will be no decision Malayan Malaya player can't really influence the decision making because there's no decision to be made

however um it has the witness that if smart contract Vol ability have been identified later they can't change it um liquidity can exist both off uh offchain and onchain is very difficult to accurately assess the liquidity offchain without knowing all the relevant addresses but for onchain liquidity you can have code set up to monitor uh liquidity at centralized exchange and these are the two diagram I created for my dashboard to monitor um the curve tripool token holding and then the Unis swap usdc di balance the these are the other

categories that also in the uh uh Define the project so insurance and Redemption it determines the stable coin Supply smart contract risk uh programming bug and logical error in smart contract um I didn't identify a lot there there's not much exploit to stable coin smart contract as of today but this is very common for other defi protocol and you also have the Oracle risk which is the risk of incorrect of change information being imported and then at the end is a regulatory risk the risk of changes in laws and regulation that may have negative impact to the stable cor project these are the past failure and a quick recap and then most of them

adopted a similar um asset backing design the challenge um I think the biggest challenge when you talk about stable con risk is how to assign an overall rating to the stable con whether it's a A++ n out of 10 seven out of 10 the reason that is difficult because it's very difficult to judge the weight on different risk categories and um it can be biased depending on the person that the US as objective and then it's also there's not enough data to measure the the probability of stable con default so it's hard to assign score based on that my concern is like having an over score for a stable con it might

mislead uh stapor holder that when they're using the rating they're not aware of how the weight is assigned to each risk category so our approach is to rather than have to have a overall risk score have a risk score for each risk category and then the user themselves decides the importance of each category based on on their own objective and RIS appetite potential application can be uh help user decide which stable coins they want to use based on the risk they are most worried about help protocols decide which stable coin to incorporate as cultural based on the risk the protocol that want to take or they don't want to

take and most importantly stable con is one of the most important crypto like application that can help to promote the the adoption of uh The Wider adoption of cryptocurrency in the General Financial system so it's very important to let the regulator to know um that not everything is us there's a lot of good stable coin design uh in the market so this is uh what's more to come from my project I will have a paper to list all the details description of the chosen risk category risk factor and what metrics uh how we use to measure them um St key stable con project overview and then I'm doing dashboard which is already done to show live

analytic that I used to monitor latest stable corn market movement and onchain activities Ah that's the end thank you thank you Tammy uh we we have time for one or two questions really quickly all right if not then we'll get started in about two minutes with the next talk thank you so much Tammy Take Care Round of Applause for Tammy