We wanted to share our thoughts on the recent report prepared by the LlamaRisk team.
While we viewed the process as generally collaborative and agree that there’s work for us to do in a number of areas as our protocol evolves, we disagree with some of the findings.
USDR integrates a novel approach to the integration of tokenized RWAs. At this point in time, any steps related to the purchase, registration and management of real estate occur in centralized, legacy systems. Taking a risk-based approach on managing RWAs requires that some of our operations are correspondingly centralized. As legal and regulatory frameworks adapt to include web3 governance structures, these operations can be progressively decentralized.
The goal of everyone in this space should be to support builders and the development of protocols that seek to move the category forward, expanding the use case for blockchain/tokenization. What we’re doing is hard, we’re creating something entirely different from the ground up. It’s not a fork and there’s no past example to build from; Real USD is a meaningful evolution in the integration of RWAs on-chain and wealth preservation for users world-wide.
We hoped for support in our efforts, at a minimum we expected a balanced read-out of USDR’s benefits and areas for improvement when the assessment was distributed on Twitter. LlamaRisk presents itself as an impartial body, so we were surprised to see clearly inaccurate comparisons drawn between USDR and notable scams such as the Mochi/USDM incident.
Given the importance placed on these reports — produced by an anonymous team of volunteers — we’d like to request a list detailing the various members of the LlamaRisk team, places of employment and major token holdings. We believe this is a fair request to help guarantee the impartiality and authority of this team’s valuable work in the future, for protocols and users alike.
In the following response, we’ll look at each of the major points of LlamaRisk critique individually.
TNGBL BACKING/MINTING WITH TNGBL
“Is 100% CR the optimal parameter for USDR, considering the risks that come with RE as the main type of collateral?”
“There are concerning and potentially unsustainable strategies that can be used by having a partially endogenous collateral type. For instance, the bribe manager regularly deposits TNGBL to mint USDR for its incentive scheme. This gives Tangible a similar power to mint unbacked stablecoins that prompted an emergency action against Mochi’s USDM.”
“The image below displays the current collateral structure…TNGBL’s share on the other hand is over 15%. A difference of plus 5% from the intended allocation.”
We agree that 100% CR is not the optimal parameter for USDR and have built it to be overcollateralized, however the design choices we’ve used have either been misunderstood or misrepresented in the report.
A real estate backed stablecoin cannot be built without an asset like TNGBL in the backing. Given the potential for real estate to lose value over the short term, stablecoins require a buffer like TNGBL or an insurance fund to maintain collateralization and capital efficiency in the early stages of growth. Without these types of assets — that we’ve integrated by design — the required value of real estate to back each $1 of the stablecoin would leave it capital inefficient, lacking enough native yield to maintain purchasing power and meet consumer needs.
The total amount of Real USD that can be minted with TNGBL cannot exceed 10% of the amount of USDR minted, minus USDR redeemed. The report even acknowledges this in the section “Minting Real USD.” So under normal circumstances, TNGBL will constitute no more than $0.10 of every $1.00 of circulating USDR. As such, the claims of “mint[ing] unbacked stablecoins” with TNGBL are a misrepresentation of protocol functionality. And the comparisons to projects without a minting cap or those backed exclusively by endogenous assets are without merit.
The team does account for the majority of the minting with TNGBL at the moment, however we never exceed the limits imposed by the system. Minted funds are ultimately used for bribes or to purchase other assets within the Insurance Fund i.e. VELO, BAL, CRV, etc. and flow into the bribe and insurance wallets.
While our goal for TNGBL backing is ~5–10%, in actuality the backing is substantially less than that. Currently, loose TNGBL backs only 1% of USDR in circulation. Accounting for the TNGBL in the BAL LP (~$610k) the TNGBL backing is still significantly under our target at ~3.8%, contrary to statements in the report. Any remaining TNGBL in the treasury is simply overcollateralization, a buffer that is not accounted for in the USDR market cap.
As the report notes, the price of TNGBL went up which resulted in an expansion to this component of the backing. While new USDR was minted against these gains, the team determined that the most responsible action was to burn this new USDR, preventing future undercollaterlization if TNGBL retraced. Minting against gains to TNGBL will be deprecated in USDR v3.
**NOTE: At the time of response publication, TNGBL backing to circulating tokens is ~2.9%, encompassed entirely in the Insurance fund. See site for latest stats.
REAL ESTATE, PRICING ORACLES AND PROOF OF RESERVES
“A collaboration with Chainlink and an independent auditor would be a substantial improvement. This would remove concerns surrounding Tangible’s conflict of interest. However, using Hometrack as the sole price authority moves the question of reliability to another single entity. While it’s definitely an improvement, it can’t guarantee that the oracle quotes a reliable liquidation price.”
“Using RE as collateral comes with some benefits but also risks. While its yield and price stability are very attractive, its valuation and liquidation can pose an issue.”
“Should USDR have a built-in risk management and liquidation system?”
The report flagged issues with RWA pricing, verification and ownership over 10 times. Those concerns were made explicitly clear, as was our commitment and ongoing work with reliable third parties to address these concerns. Of course the issue needed to be noted in the report and we don’t expect the team to give us credit for new oracles or PoR until they’ve been integrated. However, at a certain point, the repeated mentions and suggestions of a conflict of interest/potential for impropriety around asset valuation and verification start to feel like FUD.
Proof of Reserves will provide a verified, trusted third party to confirm real estate ownership is being accurately presented in the protocol. Oracles will integrate timely, third-party data as a source of real estate valuation. Our goal will be to integrate as many reliable sources of property valuation as possible. What’s available now is Hometrack, as other providers come online, we’ll explore integration.
When it comes to liquidation, two points:
- The report does not take into account real estate asset appreciation in the criticism of valuation and liquidations. While the liquidation value may not match the market value provided by the oracle, the market value doesn’t necessarily reflect the purchase price of the property. Given real estate’s reliable track record of appreciation vs fiat, a liquidation price a few percentage points below the current market value may well be a net gain for the sale. The report incorrectly assumes market value is always equal to purchase price.
- The recent announcement of Baskets gives Tangible an opportunity to build an automated liquidation system for tokenized real estate in the USDR treasury.
DECENTRALIZATION
“Tangible is prioritizing growth and fast releases of new features over the decentralization and sustainability of the existing infrastructure.”
“The protocol and its stability strategy are not “battle-tested” and it is fully dependent on the founding team. Custody and management are still centralized both on and offchain”
“As mentioned above, the Tangible platform and the USDR smart contracts involve a role-based access control system owned by a few multi-sig wallets. The custody risk thus lies in the hands of the Tangible: Deployer and in these signers. They basically control the entire project, making it a fully centralized project.”
We’re a small, self-funded/bootstrapped team, building something completely different from anything else on the market. As soon as we got traction, we immediately invested in additional audits, dev resources and began work integrating custom solutions for Chainlink Proof of Reserves and price oracles. Our ability to grow is our path to greater decentralization and sustainability. We’re investing in the protocol as every responsible project should.
In our view, the access rights concerns are unfounded. The blockchain allows for transparent tracking of which addresses possess which access rights. Initially, the Deployer has complete administrative privileges immediately following deployment. However, typically, the multisig performs the next transaction, which removes the Deployer’s administrative rights.
Further, while the deployer EOA may not have complete control over deployed contracts, it still has the ability to deploy new ones. However, activation or replacement of existing contracts requires a transaction from the multisig, which provides an added layer of security. As previously pointed out, the multisig’s first transaction on a new contract is usually one that revokes the deployer’s admin role. At this time, we believe our multi-sig strategy is sufficient for a project of our size and relative maturity in the market
Lastly, we believe that human intervention and responsiveness is critical to maintaining stability at this time. While increasing decentralization is the goal, until we’re fully “battle-tested” as called out, we need to remain nimble and make the right decisions based on the info we have. While contract immutability is often highlighted as important, it may not always be necessary in practice. Without mutability, code can’t be improved, bugs can’t be fixed and we can’t remain responsive while building something entirely new from scratch. Of course mutability requires trust, which we continue to build within our community on a daily basis.
MISCELLANEOUS
“Tangible currently owns ~42% of the Curve pool, indicating that there is relatively low organic demand to provide liquidity”
The Llama team is aware that we have pools on other chains that exist without any POL, indicating a high organic demand to provide liquidity. For example, our pool on Velodrome (OP) is $18M and the fourth largest stablecoin pool on the chain. In fact, it’s larger than our pool on Curve and entirely organic.
“It’s also worth pointing out that there is a high concentration in terms of wUSDR token holders. On all chains outside of Polygon, almost all wUSDR tokens are deposited into the above-mentioned DEXs”
Like Curve on Polygon, we’ve identified the top liquidity layers on other chains (BNB, OP, ETH) and concentrated liquidity on the most trusted protocols. We can’t be criticized for not having organic demand to provide liquidity and then be chastised for responsibly building organic liquidity outside Polygon.
“With a constantly increasing price, these liquidity pools are not “pure stable pools”. They can not be considered to meaningfully contribute to USDR’s pegging mechanism”
This is false as you can buy USDR, wrap it and sell into a wUSDR pool, pools which have incredibly deep liquidity as noted above. This does, in fact, help support the peg.
“USDR has grown to a market cap of $11.5M before receiving a Curve gauge.”
If this section is meant to highlight the impact of CRV incentives, we reached a market cap of $15.5M prior to the distribution of any CRV emissions, a meaningful misrepresentation of ~35%.
ASSESSMENT OBJECTIVITY
Looking through past reports, it’s impossible not to notice a stark difference in tone between the USDR report and language used in the assessments of Frax and USDD. By Frax’s own admission, they’re ~21% decentralized in backing and nearly 6% algorithmic. USDD is a Luna fork with some additional BTC and USDC, primarily backed by endogenous assets. Yet neither faced the same level of scrutiny or criticism as USDR or calls to not support a gauge.
Remove TNGBL completely from our backing, inclusive of the insurance fund, and we’re over 96% backed. Further, using Frax’s methodology to calculate our decentralization ratio, USDR is 16.3% decentralized, net 4% points less decentralized than Frax (which is currently 21.4% decentralized.)
Lastly, a quick look at the top stablecoin pools on Curve reveals a 78.5% centralization ratio. Unaudited attestations from USDT are widely accepted, while our motives are continuously questioned in the report despite an ongoing push for greater decentralization and transparency.
So what is really going on here? At best there’s a notable degree of hypocrisy and at worst some serious intellectual dishonesty and deliberate bias at play. Different protocols are being evaluated with different sets of criteria for unknown reasons.
A stablecoin that delivers a scalable, native yield is highly disruptive to the stablecoin landscape and to many vested interests. For this reason, we think it’s essential that the LlamaRsk team reveal their employment details and token holdings to give their report a sense of impartiality. These types of disclosures are typical in traditional finance and should also be utilized in our space. When LlamaRisk reporting has the potential to cause serious harm to a protocol’s future, all potential for bias must be removed. Afterall, this is simply the same level of transparency the team is requesting from the projects they audit.
CONCLUSION
“USDR is not a trustless or decentralized stablecoin. It relies on the protocol’s own RWA on-ramp service and management by the team.”
“RWA custody, governance, and collateral structure is insufficient and merits caution”
We are not attempting to build a trustless, fully decentralized stablecoin. RWAs and real estate are, by nature, centralized. Let’s acknowledge the incremental risk this presents, but tokenization protocols for physical assets simply cannot be held to the same standard as fully decentralized digital assets or stablecoins backed by those assets. Of course the project can’t continue if the team vanishes, but that’s an unrealistic expectation to set at this time. And by this measure, if the team at Circle walked away, 78% of current stablecoin TVL on Curve would also vanish.
We recommend the LlamaRisk team explore evaluation criteria that’s reflexive to projects like ours. DeFi remains risky, users must do their own research and proceed with caution. However, if Curve protocol support necessitates that every project fit into the same tiny box, how do we expect the category to evolve and flourish? This rigid system is stifling innovation. There’s widespread belief that crypto projects offer little to no real world value. We’re working hard to change that perception and hope to have the category’s support in that work.
