Fiat-Backed Stablecoins and Mehen’s USDM

Background

First, none of this is financial advice. If you are reading this you may be using DeFi applications which are extremely risky and subject to many different types of risk, beyond what is discussed below. Please do not take advice from random anonymous people on the Internet.

My view on a Stablecoin protocol is that they are effectively a simple bank. Stablecoin protocols have both assets and liabilities. The liabilities of the protocol is the stablecoin, like USDC, USDT, or USDM. The assets of the protocol are the reserves being held to back the liabilities. For fiat-backed stablecoins, often the reserves/assets are US Dollars, money market funds, treasury bills, reverse repurchase agreements, short term commercial paper, etc. There can be many different assets in the reserves and its important to review the risk profile of all the assets in totality, to measure both liquidity and solvency risks. I am going to focus on fiat-backed stablecoins and specifically on Mehen’s USDM token.

Mehen

Mehen the protocol is issuing USDM when users deposit US Dollars into the protocol. The US Dollars become Mehen’s assets and the USDM is the liability of Mehen. Initially Mehen is going to take the US Dollars and directly purchase 2 money market mutual funds. The assets of Mehen will be a combination (at least initially) of 2 money market funds and uninvested US Dollars temporarily in custody of a transaction bank. Maybe over time as Mehen grows they will directly access the underlying securities of the money market funds (usually treasury bills, agency debt, reverse repurchase agreements, etc.) either directly managing these assets or through a separately managed account, managed by a professional asset manager.

My understanding of the asset flow of Mehen protocol is:

1. Users transfer US Dollars from their banks directly into 1 of Mehen’s bank accounts, Mehen’s clearing accounts.
2. At the end of the day the final amount of US Dollars is tallied up and some portion of that is moved through their introducing broker directly into 1 of 2 Money Market Funds.
3. Mehen directly purchases the funds, using an introductory broker, so that Mehen is actually on the transfer agency record/registry for the funds. These funds are not held by a broker in Mehen’s name, they are directly owned by Mehen.

Both Money Market Funds use BNY Mellon as their custodian and both Money Market Funds are managed by extremely well respected asset managers. I don’t know if the actual tickers are public but below is some summary data on the funds. Their average maturity and weighted average life are a bit longer than Circle’s backing USDC managed by Blackrock (USDXX) and a bit longer than some popular federal money market funds like Vanguards VMFXX. Overall I still think they are well within the normal range of federal money market funds duration exposures, even if they are a bit on the longer side.

Money Market Funds used by Mehen to back USDM

Major Risks

There are 3 major risks of fiat-backed stablecoin protocols.

1. Peg Stability — how close the token stays to the asset its pegging (i.e. how close USDM stays to 1 US Dollar)
2. Liquidity — when people redeem USDM for US Dollars, how quickly can they get back 1 US Dollar for 1 USDM (excluding fees)
3. Solvency — are the assets greater than or equal to the liabilities, i.e. is there at least 1 US Dollar (or equivalent in assets) backing 1 USDM

Peg stability

Peg stability is intrinsically tied to both liquidity and solvency (see below). Peg stability really relies on the fact that there are enough assets in the reserve to back 1 USDM with at least 1 dollar of assets, and that these assets can be accessed in an appropriate amount of time. These are both the solvency and liquidity risks.

I fully expect a large majority of the time that USDM will trade between 0.97ish and 1.02ish on decentralized exchanges, over longer periods of time as liquidity on chain grows. It’s always possible in the short term for some instability before users arbitrage the DEXes, especially initially with low liquidity. As with any asset with low liquidity, large price swings are possible. Also note that the arbitrage mechanism requires bank transfers, which can take days. Given the time delay of the arbitrage (it’s not all within DeFi), this soft peg window may be wider. We may need to increase our window to account for larger discounts (wider depegs for larger arbitrage opportunities) given the time value of money.

The reason the upper bound is closer to 1 is because there is no minting fee, if there was a minting fee it would be closer to 1 + minting fee. This means that if USDM trades above 1, there is an arbitrage opportunity. Someone could take US Dollars, mint say 100 USDM for 100 US Dollars, then trade that 100 USDM for over 100 US Dollars worth or ADA or another stablecoin, then trade that other ADA or stablecoin back into US Dollars. So the arbitrage opportunity depends on the fees charged by DEXes. Most stableswap pairs will have extremely low fees so its likely this upper bound could be close to 1. The reason it’s not closer to 1 is there is a requirement to KYC to off ramp through the primary market (minting and burning), and it requires bank transfers which takes time. It’s possible there may not be enough users willing to KYC to close the arbitrage opportunity.

The lower bound is going to be close to 1 minus the fee to burn (currently set to 1.5%). The reason its going to be closer to 0.985 is because if USDM trades below 0.985 then there is an arbitrage opportunity. Some could buy say 100 USDM using less than 100 US Dollars worth of ADA, then burn that 100 USDM worth of ADA for 100 * (1- 0.015) = 98.5 US Dollars, then use that 98.5 US Dollars to buy back ADA and end with more ADA then they started. This could also be done with other stablecoins. The reason it’s not closer to 0.985 are the same reasons stated above for the upper side of the peg.

Liquidity

Liquidity risk is the risk that the protocol is still solvent (there are enough assets to cover all the liabilities, i.e. the reserve is greater than the outstanding stablecoins), but the assets cannot be sold/ turned into US Dollars fast enough to satisfy the demand to burn or mint the stablecoin. There is an asset / liability duration mismatch. Liquidity usually becomes an issue when the liabilities can theoretically be redeemed at any time, but the assets may take some time to turn into uninvested US Dollars.

The money market funds that will be purchased by Mehen have a 1 to 2 day settlement time, I believe will be 1 day at the end of May 2024. However, if USDM is theoretically immediately redeemable and there are not enough US Dollars in the bank account, then there could be an illiquidity event that could cause some peg instability on exchanges. This is the reason many protocols have uninvested cash to offset unexpected outflows. there is also intra-bank transfers that need to happen between Mehen’s banks and users’ bank, which could take up to 5 to 7 days. I do not know how long it will take for someone to burn USDM and when the US Dollars will settle in their bank account.

Note that the release valve of liquidity events tends to be the stablecoin price on chain. Stablecoins tend to trade below peg when there is demand to redeem but actual burning through the primary market will take too long, so people sell on the secondary market (DEXes/CEXes). Stablecoins will tend to trade above peg when there is excess demand to buy the stablecoin and minting/arbitrage takes too long in the primary market, so people buy in the secondary market (DEXes/CEXes).

Maker DAO has the PSM (peg stability mechanism) for this reason, which is filled with USDC. Circle at one point kept 15% of their assets in bank accounts for this. Unfortunately they had 15% in Silicon Valley Bank that failed and caused an illiquidity event for USDC and caused USDC to trade below 1 US Dollar as the illiquidity event turned into solvency fears. At the end of the day Circle remained solvent and USDC remains backed by at least 1 US Dollar worth of assets. Maker recently had to increase their DAI Savings Rate and interest rates on their Collateralized Debt Positions when there were large demands for redeeming DAI and their PSM was drawn down to the lowest level in years. The point here is even the most well known and largest stablecoin protocols can sometimes face liquidity events.

Solvency

Solvency is strictly the fact that the protocol has enough assets to cover all liabilities at the peg rate. So for Mehen, solvency would be having asset values in US Dollars greater than the amount of USDM outstanding. The current Oracle for USDM will be reporting assets once a day, so solvency risk can be estimated on chain once a day.

I see 3 main solvency risks, beyond the normal risk of poor internal controls of any protocol with some centralized functions. The first risk is the money market funds that Mehen invests in ‘break the buck’ (i.e. have some losses on their underlying investments). The investments of both money market funds are as risk free as possible within the US financial system. They are all guaranteed by the US government. So, to me, this risk is extremely low.

The second risk would be whatever reserves Mehen keeps in their bank accounts (to offset liquidity risks) become tied up in bank bankruptcies, or is permanently frozen through regulation or enforcement by a US Agency. This risk is low but not as low as the previous risk. Before Silicon Valley Bank and Signature Bank collapsing in 2023, I would have rated this risk lower, but bank failures can happen, as we have recently seen. In the US, first insured depositors are paid back, then uninsured depositors are treated as general creditors. So the first 250K US Dollars would be insured, after that, any other deposits would be treated like other general creditors of the bank. And US Regulation is an unknown risk. Hopefully the fact that Mehen has a FinCEN registration should help. I view this risk as low but not extremely low.

The final risk comes from the fact that intra-bank settlement can take 5 to 7 days. So if a user tries to buy USDM using US Dollars by transferring US Dollars from their bank account to Mehen’s bank account, and USDM is issued before the US Dollars settle, it’s possible the bank transfer is reversed and USDM is issued by unbacked US Dollars. This risk is tough to measure so we will have to trust intra-bank transactions. I do not know how long it will take for USDM to be accessible by the minter. Mehen has been given assurances that this should only take 1 to 2 days, but banks do have the ability to take up to the maximum amount of time to settle intra-bank transfers.

Conclusion

Overall, USDM seems like a safe fiat-backed stablecoin for Cardano. FinCEN registration is great to see as well. There are definitely risks, as there are with all stablecoins and tokens in general, but I do view these risks as low. If laws ever change, I would love to see Mehen pass along some of their interest earned on the money market funds back to USDM holders. First, I would like to see them increase their assets with the interest earned to build an asset buffer against solvency (like an insurance fund). It may be smart to open multiple bank accounts under the 250K FDIC insurance limit to act as a ‘PSM’ of sorts for liquidity events and to diversify bank bankruptcy risk. It also may make sense to hold some USDC on separate chains to act as a secondary PSM.

I think a strong fiat-backed stablecoin is needed for Cardano and I am happy to see Minswap vote on the proposal to mint some USDM in their Treasury. USDM passes my risk assessment. My main concern would be liquidity risk, leading to some peg instability. If USDM is stated to be immediately redeemable, but there aren’t enough assets in the bank to cover 1 days worth of redemptions/burns, it may take another day or two to get your US Dollars, and the price on chain may vary. There is a balance between keeping too much money in the bank and the risks of the banks that currently are willing to bank crypto projects. I believe the risks to the money market funds are actually lower than the risks of holding US Dollars in a bank account, so I would be willing to take some small amount of illiquidity risk if I was Mehen. The other main risk to me is making sure that USDM is only issued once there is confirmation that the intra-bank settlement of US Dollars takes place.

Overall, this is an amazing opportunity to move Cardano forward and get us a native fiat-backed stablecoin. I hope all projects across the Cardano ecosystem will support Mehen. I will be a proponent for them in both Minswap Labs and Optim Labs (which I am actively involved as a community and DAO member).

Congratulations to Mehen.

Thank you for the comments and suggestions Vanessa Harris and Blockjock.

About Marco

I am in the Kitty Farmer Committee for Minswap. I am a DAO Working Group member and Iporian at IPOR. I am also a member of Optim’s ODAO.

This is not Financial Advice!

I would love to hear your feedback/questions/comments. Reach out to me on Twitter… Marco_112358, or Discord… marco_112358.