“You’d have to be a Lunatic…”
The Actual Economics of the UST Stablecoin
Disclaimer #1: This is not financial advice. Although if you’re the kind of person who takes financial advice from a random person on the internet, you’re probably not the type to read disclaimers.
Disclaimer #2: I am not a neutral source of information. I have spent more or less of the last 6 years working full time with the MakerDAO team and in the Maker ecosystem. I can fairly be accused of being biased towards Dai.
If I’ve gotten something wrong in this post please reach out so that I can correct it.
Introduction
Over the past 18 months a once-obscure algorithmic stablecoin, UST (Terra USD), has increased its market share by over 1,000x to become the 11th largest cryptoasset and the 3rd largest stablecoin. This post will analyze UST in an effort to understand its meteoric rise. I’ll examine how the coin functions underneath the hood, compare it to other assets with similar utility, and attempt to predict what may happen to UST in the future.
What is the purpose of a stablecoin and why do people want to hold them?
A stablecoin is intended to be used as everyday currency in a blockchain ecosystem. It must be able to provide a unit of account in decentralized applications, a mechanism by which users can store value without being concerned with price volatility, and a fair way to denominate debts.
The demand for a stablecoin will depend on what I like to call endogenous and exogenous yield, which are the core drivers of demand for any currency. Endogenous yield is the yield that comes “risk free” to the holder of the stablecoin directly from its issuer, whether that issuer be a centralized corporate entity, a DAO, or an L1. A TradFi example of endogenous yield is the interest paid on US Treasuries (the “risk free rate”). An example of endogenous yield in DeFi is the Dai Savings Rate, which is claimable by all Dai holders. The yield paid by Anchor to UST depositors is also an example of endogenous yield, as 85% of it is provided directly by Terraform Lab’s sales of LUNA and does not come from organic borrowing on the platform. Exogenous yield is what a stablecoin holder can earn by using the coin elsewhere in the economy. For example, depositing USDC on Compound or providing liquidity on Uniswap, a centralized exchange, or in a yield farming pool are activities which give USDC an exogenous yield.
Endogenous yield comes from the issuer of the stablecoin via the management of its assets, while exogenous yield comes from economic utility (as a store of value/unit of account/medium of exchange/standard of deferred payment) in the ecosystem.
Let’s quickly look at the current endogenous yield on the top 5 stablecoins:
USDC: 0%
USDT: 0%
BUSD: 0%
DAI: 0%
UST: 16.61%
What is the role of a stablecoin issuer?
As stated above, stablecoin issuers can come in all shapes and sizes. They can be centralized corporate entities, like Tether, Circle, and Binance/Paxos, decentralized entities like MakerDAO, or a hybrid like Terra/Terraform Labs/Luna Foundation Guard (LFG). The role of an issuer, however, is uniform — to manage supply and demand of the stablecoin in a way which maximizes market price stability (1 stable = 1 dollar), and to protect against insolvency by managing a pool of underlying assets (asset/liability management).
An issuer must be sufficiently prepared to face significant inflows and outflows of capital into/out of its stablecoin, it must also ensure that the assets purchased/sold or created/destroyed as a result of these inflows and outflows are sufficiently liquid and valuable to match or exceed the value held in the stablecoin. Put simply, a stablecoin is a liability of its issuer and the underlying reserves or loans are its assets. The market price of a stablecoin will first depend on the issuer’s ability to match inflows/outflows, and will subsequently depend on the confidence of holders that there are sufficient assets backing the stablecoin should all holders wish to redeem simultaneously. When you hear about “death spirals” it’s the process by which mismanagement of the latter leads to the inability to manage the former, in a vicious cycle.
Review of the top stablecoin issuers and their management procedures
Issuer: Circle
Asset: USDC
Model: Centralized
Behind the scenes Circle is just a regular bank, and a very conservative one at that. It runs a simple asset/liability management scheme to ensure that it has sufficient cash reserves to redeem USDC at will, and also ensures that any of its assets not held in cash are sufficiently liquid to be sold quickly under stressful market conditions for total losses that should not be in excess of its equity. Inflows/outflows to USDC are just inflows/outflows from its balance sheet, as it is redeemable at will for $1 of cash. Circle makes a profit by investing the capital used to create USDC into these TradFi assets and keeps the net interest margin for itself. Due to this conservative operating model, Circle has thin profit margins and cannot offer a very high endogenous yield. Rather, it relies on the confidence, stability and liquidity brought about by its centralized and regulated business model to encourage other economic actors (e.g. market makers) to utilize USDC elsewhere in the blockchain economy, providing a strong exogenous yield. The primary risk of holding USDC is censorship, as Circle is a prime target of regulatory capture.
Issuer: Tether
Asset: USDT
Model: Centralized
Honestly I have no idea what Tether’s balance sheet looks like or what goes on underneath the hood. There’s speculation that some of the balance sheet is in Chinese commercial paper tied to real estate developers, which should be rather concerning to most holders if true. At one point, USDT was partially backed by an unsecured loan to its then-parent company, Bitfinex. The one thing you can say about Tether is that it’s beaten the odds to get to where it is and it’s clearly kept its core holders (market makers) confident enough to prevent significant outflows. I would not be surprised if Tether ends up insolvent, but I also won’t be surprised if it doesn’t. Either way, I don’t have enough information to feel comfortable holding it on my own balance sheet. Its business model is otherwise the same as Circle’s. The primary risk of holding USDT is not knowing what the primary risk of holding USDT actually is.
Issuer: Binance (via Paxos)
Asset: BUSD
Model: Centralized
This is essentially identical to USDC.
Issuer: MakerDAO
Asset: DAI
Model: Decentralized
MakerDAO’s assets are crypto loans (mainly against ETH), “real-world” loans (e.g. mortgages) and other stablecoins (mainly USDC). As a decentralized issuer, all of these loans are actually self-issued by the borrowers and the collateral can’t be used in Maker’s own asset/liability management (with the exception of the USDC). Due to this reality, Maker needs to manage its inflows/outflows using an endogenous interest rate called the Dai Savings Rate (DSR), which it takes from the net interest margin on the loans. By raising rates when the market price of Dai is falling, Maker can increase the incentive to pay back loans and therefore decrease the overall Dai supply, while simultaneously increasing Dai’s endogenous yield by passing these increases to Dai holders and increasing demand. MakerDAO makes a profit by collecting the net interest margin on the loans (what it collects minus what it spends on labor minus and it pays out via the DSR). The smart contracts that power MakerDAO are designed to ensure that the value of its assets, vis a vis the underlying collateralization of its loans, is always sufficient to back the full value of its liabilities (i.e. the Dai outstanding) — in a tail event, these smart contracts have a function called emergency shutdown which settles all outstanding Dai into $1 of underlying collateral. It’s via this mechanism that Maker can credibly maintain confidence in the Dai peg when its interest rate mechanism isn’t reactive enough to manage inflows/outflows. The primary risk of holding Dai is MakerDAO’s potential mismanagement of the loan portfolio (e.g. lending against bad assets or getting its USDC frozen).
Issuer: Terra Blockchain/Terraform Labs/Luna Foundation Guard (LFG)
Asset: UST
Model: Hybrid
UST is a hybrid stablecoin with a mixture of decentralized and centralized entities fulfilling the various functions of an issuer. The three main entities are (1) the Terra Blockchain (Terra), which is its own L1 with a native token called LUNA, (2) Terraform Labs, a centralized corporate entity, and (3) the Luna Foundation Guard (LFG) which appears to be a centralized Singaporean limited company. The creation/destruction of UST is handled by Terra, endogenous yield is provided by Terraform Labs, and reserves are maintained by the LFG. For convenience I’ll refer to this collective of entities as “Terra.” Terra manages inflows/outflows by paying an above-market endogenous yield on UST (currently 19.54%). This yield is achieved by monetizing its L1 token, LUNA. LUNA must be destroyed in order to generate new UST, and minted in order to destroy it. For example, if you want to create 1 UST, you will need to purchase and burn $1 of LUNA and vice versa. Inflows to LUNA must exceed outflows of UST in order for the system to remain solvent — solvency being the ability to fit the entirety of the value of outstanding UST into the value of LUNA, less reserves (LFG has been trading LUNA for BTC recently, in order to act as a less volatile form of reserve). If the value of LUNA is below the value of UST less reserves, it is likely that a bank run will take place unless Terra puts some kind of capital controls in place which inhibit outflows. I intend to go into further detail regarding these mechanics in the second half of this post.
Other Issuers: Reflexer, Liquidity, Fei DAO, Frax
Assets: RAI, LUSD, FEI, FRAX
Model: Decentralized
Other notable mentions in the stablecoin issuer space are Reflexer, Liquidity, Fei DAO and Frax. Reflexer and Liquity are basically MakerDAO with a few different trade-offs. RAI (Reflexer’s stablecoin) is probably the safest on the market from tail risk, but its trade-off is that it punishes holders with negative interest rates if there’s excess demand for the stablecoin. LUSD (Liquity’s stablecoin) has some tail risk but is overall a decent ETH-only design. It occasionally forcibly sells its user’s collateral in order to manage its outflows, which is its primary trade-off. Fei DAO and Frax are more or less decentralized versions of UST with much higher/more sustainable levels of reserves. They do not provide an endogenous yield, which gives them a bit of breathing room at the expense of growth. Both of these stablecoins still carry high tail-risk.
Diving further into UST
The best analogy for the Terra ecosystem may be to think of it as a startup with a high burn rate that’s running a blitzkrieg on the stablecoin space. Terra sells its “equity,” LUNA, in order to create a product, UST, which is actually just a repackaging of its equity with a fixed price and a stable (but technically variable) yield. It pays people a subsidy to use this product via Anchor, until its eventual scale and liquidity can motivate unrelated ecosystem participants to offer UST exogenous yield. If and when this happens, they can stop paying people to use it. It’s kind of like the way WeWork operated, except in this analogy the office buildings would burn down if WeWork’s stock tanked. While this strategy sounds exciting on paper, perhaps even something you’d like to invest in, it creates a binary outcome for Terra — total domination or total collapse. Unfortunately, since the equity is inextricably linked to the product, UST is also in this binary position. Startups trying to shoot the moon usually pay more than 20% to their investors, just saying…
Let’s crunch the numbers to see how this strategy is playing out…
Current UST Market Cap: $18.12B
Source: Messari
The UST market cap has grown approximately 1000x in 18 months. It should be noted that roughly 4–5B UST was created by Terraform labs by monetizing LUNA and sent to the community fund. It’s hard for me to understand exactly what happened/is happening to this 4B (to be conservative we’ll call it 4 instead of 5), but for the remainder of this analysis I’ll be subtracting it from market share calculations.
Using current info from Anchor…
Current Anchor Deposits: $13.022B
Current Anchor Borrows: $3.052B
Source: app.anchorprotocol.com
Revenue (annualized, not compounded): $343.04M
Liabilities (annualized, not compounded): $2.541B
Net Income (annualized, not compounded): -$2.198B
Anchor is costing LUNA holders $2.198B/yr in outflows. This is, however, not a compounded figure. Taking compounding into consideration, this liability will grow exponentially and will require equally exponential inflows to LUNA in order to keep the system fully collateralized.
Current LUNA Market Cap: $31.725B
Outflows as % of LUNA Market Cap: 6.93%
Anchor Net Rate (adjusted for borrows): 16.88%
Let’s extrapolate out 5 years to demonstrate the compounding nature of Terra’s liabilities. In the following projection, I am assuming no additional inflows or outflows from UST or LUNA, only the compounding liabilities from the Anchor protocol under the conditions that exist today.
Over the course of 5 years, accounting for nothing else aside from a static Anchor protocol needing to monetize LUNA, it will require over $17B in inflows in order to continue to pay the Anchor yield and keep the system solvent. Short of that, the system will be approximately 50% undercollateralized. Note that this is just to keep the system at 100% collateralization, and that these numbers are conservative as they are rounded down and annually (rather than continuously) compounded. If UST were to grow to the size of USDT, it would at current metrics require $13.672B in continuous, annual and compounding inflows, to maintain the Anchor yield. UST will reach this level by the end of the year given its growth trajectory. I am pointing this out to emphasize that while growth is a solution it’s also the problem.
What about the reserves?
Recently Terra has been monetizing LUNA by converting it into bitcoin, USDC and USDT. It’s calling these assets “reserves” and claiming that they’ll be used to purchase UST on the market if outflows exceed inflows and the peg begins to break. I think this shift in narrative is potentially counterproductive to UST finding product-market fit, but I’ll save that for later. For now, let’s just examine the economics.
Source: dashboard.lfg.org
Currently the LFG has $2.240B in assets standing ready to defend the peg. This amounts to approximately 12.36% of the overall UST supply, and just about one year of Anchor interest payments. This, assuming the LFG can access the market efficiently, gives UST a current intrinsic value of $0.12. Of course, the future value of these reserves is almost entirely dependent on the price of bitcoin and the inflows/outflows of UST. It’s a big bet that could cause UST’s reserves to plummet right when they need them the most. It could also pay off spectacularly — a similar outcome happened to FEI, which now has an enormous amount of reserves relative to its outstanding supply. It should probably be pointed out that if you’re buying an asset that gives you a fraction of the overall yield of bitcoin and collapses if bitcoin underperforms, you should probably just buy bitcoin. Either way, UST’s reserve balance will be an important metric to watch as it continues to grow.
Product-Market Fit
This all leads us to the most important point. As demonstrated, UST will eventually collapse under its own gravity if it does not find demand external to yield on Anchor. So let’s see how that’s going. Currently, if we back out the 4B UST that was sent to the community fund, it would appear that 92% of the actual floating UST supply is sitting in Anchor. Given that I’ve been conservative in my numbers, I am fairly confident in saying that virtually all of the floating UST is in Anchor.
So right now, endogenous yield from Anchor Protocol is the only driver of demand for UST. What about the broader narrative and role it’s trying to fill? From day one, UST has been marketed as a “truly decentralized” stablecoin, criticizing DAI as impure for having USDC reserves. Putting aside the fact that this is effectively an admission that the system doesn’t function as intended, Terra is now openly claiming that it’s keeping BTC and USDC reserves in the LFG — which is a completely centralized company presumably being secured by a multisig. Is this the game-changing decentralized stablecoin we’ve all been waiting for? It sounds more like a hedge fund to me. Continuing down the path of backing UST with the balance sheet of a centralized entity may improve stability, but I can’t imagine how it materially helps on the business development front. Major dapps already have stablecoins with a similar operating model, why would users adopt a riskier alternative absent some kind of excess reward? Whether or not anyone wants to admit it, UST has thus far utterly failed to gain any kind of product-market fit outside of providing a high endogenous yield.
Looking to the future
So what’s in store for UST? Let’s play out a few ideas and scenarios.
Cut the Anchor yield to a sustainable level that comes from organic borrowing?
Yeah, this just isn’t going to work. As I’ve shown above, the only reason anyone holds UST in the first place is because of this yield. Every point it drops will begin to have a large impact on demand for UST, causing outflows that decrease the LUNA market cap and cause liquidations of LUNA borrowers (who are the vast majority of borrowers, as it’s not really attractive to pay 11% APY for a loan against cross-chain ETH) on Anchor. The entire system is in a tight feedback loop around the Anchor yield and lowering it will be a major blow to confidence. If a lower yield causes outflows, which causes LUNA price decreases, which causes liquidations of LUNA loans on Anchor, which causes the sustainable interest rate to drop further, which causes further outflows…yeah that’s a death spiral. Ultimately the laws of gravity are going to force them to lower this rate, I imagine they’ll do it slowly and paired with big marketing announcements so as to not spook the market.
Bring UST to the current major stablecoin demand centers on Ethereum?
I think it’s going to be hard to convince market makers or blue-chip DeFi protocols to hold UST on their balance sheet without getting forced back into the same vicious LUNA monetization cycle that they’re in right now. Market makers and DeFi protocols have existential risk if the assets on the balance sheet suddenly go to zero. In the presence of competing, far more economically sound stablecoins, why would they risk holding UST? If it’s supposed to be because of the reserves being primarily in decentralized assets, I’d imagine they’d prefer a more economically sound alternative like Fei.
Continue to drum up retail demand for UST and try to become too big to fail?
I think this is the only actual strategy they have. They’ll buy their way into choice integrations, get a bunch of non-crypto consumers invested in UST and pray that this liquidity and scale is enough to overcome any reservations about the operating model. In this world, users would not know or care about the value of LUNA and would just trust that 1 UST = 1 USD. In short, they’re shooting for “Lindy.” From here, UST would actually replace cash in the “real” economy and exogenous yield would build. I think this is extremely unlikely, but it’s possible.
Conclusions
All this is to say that the current trajectory of UST, LUNA, and the Terra ecosystem is going to collapse under its own weight if they ever find themselves unable to monetize LUNA to match UST outflows. Because of this, you should expect them to take a blitzkrieg/shoot the moon strategy to grow as fast as possible, regardless of whether it’s sustainable. While this strategy could work, UST holders are not being properly compensated for taking the risk. There are many investments with a higher ROE than 20% and a much lower chance of going to $0. In short, can this all work out? Maybe. But to hold UST, you’d have to be a lunatic.
