Stablecoin: a coin that exhibits low or zero volatility to some reference asset, typically USD.
The newest rage in crypto is here, with a recent study reporting that there are more than 50 stablecoins in development. Understandably, there is a lot of confusion about how trustworthy, safe, and decentralized they are. Critically, one has to understand that not all stablecoins are built the same. The two most prominent ones, Tether and DAI, are worlds apart in form and function despite providing the same utility. In this piece I will go over a general taxonomy of stablecoins, followed by an analysis of DAI’s stability mechanism.
Broadly speaking, to properly evaluate a stablecoin one must understand the underlying risks in the proposed stability mechanism. A stablecoin that’s worth $1 when you go to sleep and anything else in the morning serves no purpose. Worse, this risk is often obfuscated (intentionally or not), which can lead to a very false sense of security.
The binary risk profile of these products can’t be stressed enough. At equilibrium, a stablecoin is worth either $1 or $0. There is really no in-between here. A prime example of this was Liberty Reserve, a company founded in 2006 which issued e-dollars in exchange for real dollars. These e-dollars could be transacted with electronically outside of the typical banking system, until one day in 2013 the government stepped in and shut the entire operation down. The value of this e-dollar instantly went from $1 to $0.
Another example that can be seen unfolding is Tether, a USD-backed stablecoin. Tether has always been intentionally opaque about its affairs, and it’s very possible that government agencies can once again swoop in and shut the whole operation down. Because Tether is a centralized entity, there is inevitably counterparty risk. And so even though Tether has been perfectly stable up until now, there is no reassurance that its peg wouldn’t suddenly break one day if they were to run into legal issues. No one wants to be the final bagholder of Tether when the cat and mouse game they are playing with regulators comes to a halt.
To make matters worse, government intervention isn’t the only risk factor that could cause Tether’s value to plummet. If it’s ever found out that Tether is not backed 1-to-1 by dollars in a bank somewhere, then all confidence will be lost, and 1 Tether will almost certainly cease to be worth 1 dollar. Again, there’s really no in-between here. The money is either there or it isn’t.
This context provides the beginning of a rough sketch for what qualities a hypothetical ‘ideal’ stablecoin should have:
- Censorship-resistance / decentralization: it should be clear by now that if a government agency can target a specific bank or company, the stablecoin is not very robust. Ideally, it should match the censorship-resistant qualities of native cryptocurrencies such as Ether or Bitcoin.
- Publicly auditable stability mechanism: if a stablecoin relies on reserves (either off-chain or on-chain), the user-base should be able to independently verify its legitimacy. Furthermore, the stability mechanism should have failsafes in case of unpredictable tail-risk events.
Under this framework an investor can begin to evaluate competing stablecoins by challenging their levels of censorship-resistance and the robustness of their stability mechanisms. Let’s explore a number of different stablecoin approaches.
Building a stablecoin from scratch
Naive approach: The simplest stablecoin uses a pure faith-based solution. Create a token and declare it to be worth $1. The creator’s word is the stability mechanism. Of course, this sounds ludicrous, and it is. However, it’s worth understanding specifically why this is such a weak system. The answer is that this proposed stability mechanism fails to inspire any confidence in the underlying architecture. The central challenge for all stablecoins is convincing the user-base that the system maintains stability.
In fact, confidence is the key ingredient in maintaining stability. This is most apparent in the centralized realm where a fiat currency often loses purchasing power due to diminished confidence in the government’s ability to service its debts. Most recently, this has occurred in Venezuela and Turkey where the value of the Bolívar and Lira have tanked as investor confidence has waned. So how can confidence be built in a decentralized system where there is no government to rely on? How can people be made to believe that a stablecoin will be worth $1 for the foreseeable future?
The answer is collateral.
Slightly less naive approach: Create a token and declare it to be worth $1. ‘Collateralize’ the token, meaning that every unit of this stablecoin is backed by $1 worth of assets. In a hypothetical situation, these assets could be anything from commodities to real estate or even real dollars sitting in a bank somewhere. Does this collateral have to provably exist? No! This is the Tether approach. They claim that each Tether is backed by a real dollar sitting in a bank, yet they refuse to provide proof. Therefore, all investors have to go on is their word. This is arguably better than the naive approach, but not by much.
Okay, but let’s say the collateral provably existed. What does this accomplish? How does collateral provide confidence to users that the stablecoin should carry exactly $1 worth of purchasing power? It does so by allowing the user to trade in their stablecoin, no matter what its price may be, for $1 worth of collateral. If the stablecoin crashes to 0, that’s not a problem, as one should still theoretically be able to trade it in for $1 worth of collateral. In fact, if a user knew the collateral provably existed in reserve, and the redeem mechanism still functioned, then they might even buy more of this stablecoin at lower prices and immediately trade it in for the collateral. This convertibility is designed to help users sleep well at night.
Legitimate but centralized approach: Create a token and declare it to be worth $1. Declare that the token is collateralized, yet also comply with regulators and provide regular audits. This is the approach taken by TrueUSD, Gemini Dollar (GUSD), and Circle (USDC). Each unit of these stablecoins is provably backed by dollars sitting in a bank account, and they have the audits to prove it. If you can scoop up some of these coins for less than one dollar, you should, and then trade them in to the parent organization for a real dollar. After all, you know they have it due to their audits, and they’re obligated to redeem it because they’re regulated. It should be free money, and since everyone likes free money, this stablecoin should stay priced close to $1 at all times.
There is one critical drawback to this approach, however, and that is obviously the level of centralization. There are two subtle but distinct sources of centralization here. First, there exists a parent organization which, although currently complying with regulations, can theoretically be shut down if the government has a change of heart. Second, the collateral reserves are in the form of dollars sitting in the parent organization’s bank, meaning that even if the parent organization could somehow evade capture, the government could merely target the reserves, thereby ruining confidence in the entire system. Despite providing great utility, this level of centralization renders these stablecoins uninteresting from a cryptocurrency perspective, which prioritizes censorship-resistance.
Back in 2014, a visionary team called MakerDAO (‘Maker’) began working on a decentralized solution to the above problem. Founded by Rune Christensen and Nikolai Mushegian, Maker’s DAI is the most progressive iteration in this sequence of stablecoin approaches. It is an attempt at providing a collateral-backed stablecoin without the existence of a parent organization, nor the confiscatable reserves held in a bank.
Legitimate and decentralized approach: Create a token and declare it to be worth $1. Declare that the token is collateralized. There is no need to provide audits, as the collateral can be independently verified on the blockchain. For the collateral, instead of dollars in a bank (remember, there is no bank), use native cryptocurrencies such as Ether, along with a mixture of other crypto-native tokens and securities (it is important to have diversified collateral, as will be seen shortly). In this model, there is no parent organization. There are no confiscatable reserves, except for the caveat of security tokens (we will also revisit this risk shortly). The goal is that if cryptocurrencies are used as collateral, the stablecoin built on top can also maintain censorship-resistance. Let’s take a deeper look into the mechanics.
MakerDAO’s DAI operates on a collateral-based stability mechanism. As mentioned earlier, collateralization is effective because it gives users confidence that they can trade in their stablecoin for $1 worth of collateral, despite the market price of the stablecoin. With centralized services, this swap is quite easy. Gemini will be happy to redeem your GUSD for real dollars if you request it (after completing a KYC/AML check, of course). As a reminder, this convertibility is the primary confidence backstop necessary for maintaining stability.
How can this same concept be achieved in a decentralized setting, though? In the Maker system, there is no one who will redeem your DAI. Instead, the ‘ultimate backstop’ necessary is a process called ‘global settlement.’ Global settlement is essentially a graceful unwinding of the entire MakerDAO system, where everyone can trade in each DAI for $1 worth of collateral. The threat of global settlement is the crux of DAI stability. If an attacker is short-selling DAI and keeping the price below $1, then global settlement can be triggered, leading to a full recapitalization for all DAI holders (to the detriment of the short-seller).
To recap, there is no government risk, as there is no way for them to ‘shut down’ Maker (Maker is simply a set of smart contracts secured by the Ethereum blockchain). There is also no way for the government to confiscate reserves (since they are locked in the Maker smart contracts). Speculative attacks will also fail and be costly under the threat of global settlement.
In addition, there are a number of secondary stabilizers that help keep prices in line.
- Market makers / arbitrageurs known as ‘keepers’ are incentivized to keep the DAI price close to $1. They would not arbitrarily do this. Instead, they make tight markets because they know if something serious were to occur, global settlement would be triggered.
- If DAI drops significantly below $1, users are incentivized to ‘buy back’ cheap DAI and retrieve their locked collateral for a discount.
- DAI issuers are charged interest (called a ‘stability fee’) on any DAI borrowed. This parameter can be tinkered with to help smooth out supply/demand imbalances that might lead to a mispricing of DAI.
- DAI holders earn an annual savings rate, currently set to 2%. This parameter can also be tuned to help smooth out supply/demand imbalances.
The astute reader will have noticed that one crucial, system-breaking scenario has been overlooked: what if the value of the collateral falls? With Gemini Dollars, for example, the collateral backing it is simply real dollars. There will never be a mismatch in total marketcap of the stablecoin and total value of collateral. They are identical. With crypto-collateral, however, you have to ask the critical question: “If 10,000 DAI is backed by $10,000 worth of Ether, what happens if the price of Ether drops?” Restated another way: if the value of the collateral falls, then how effective will the threat of global settlement be for maintaining confidence in the system?
Let’s consider two separate ‘falling collateral value’ scenarios: one in which the value falls gradually, and another where it falls precipitously (in some type of ‘black swan’ event).
Let’s take a simple numerical calculation. A user locks up 100 ETH as collateral worth $25,000 (assuming an ETH/USD price of $250). From this 100 ETH the user issues 15,000 DAI in what’s known as a CDP (collateralized debt position). The current collateralization ratio for this CDP is $25,000 / 15,000 DAI = 166.67%, which is dangerously close to the 150% minimum threshold required by the system (a 150% buffer instead of the minimum 100% is a risk parameter subject to change). Now let’s assume the ETH price drops to $200. The new collateralization ratio is 133.33%. It’s important to note that the system is still fully collateralized (and thus, global settlement would still be effective), but it is beneath the specified threshold. The Maker system can now seize the 100 ETH of collateral (which is still worth $20,000 at this time), and auction it off to the market in exchange for DAI. 15,000 DAI needs to be soaked up from the public supply to recapitalize the system (with the balance of the collateral going to the original owner). Once this DAI is raised, it is burned, after which the system is now collateralized above the 150% threshold again.
To recap, a user who is no longer sufficiently collateralized has his collateral confiscated (via smart contract) and auctioned off to the public in exchange for DAI. The DAI proceeds of this auction are then burned, thereby decreasing the amount of DAI in circulation.
In terms of risk factors one must also consider the scenario where the value of the collateral drops so quickly that there may not be enough proceeds raised from the seized collateral to adequately reduce the supply.
Continuing the numerical example above, assume a scenario where the price of Ether suddenly drops from $250 to $100. Selling off the 100 ETH now only raises 10,000 DAI, which leaves a shortfall of 5,000 DAI. How can Maker account for this un-backed 5,000 DAI, which is still floating around in the public space? More importantly, how can this be accounted for in a decentralized fashion?
Enter the MKR token, the core innovation in the MakerDAO platform. MKR is a governance token that functions as a recapitalization resource of last resort. In the above scenario, the MKR token smart contract will auto-inflate and auction off the necessary amount of MKR in order to soak up the shortfall. Why would a MKR token holder subject himself to this dilution risk? Obviously, there is an upside to token ownership as well. MKR holders are the beneficiaries of the fees that accrue in the system (such as the stability fee).
Essentially, MKR holders are tasked with running an efficient operation with proper risk management. If they perform well, they will be rewarded. If they mismanage the parameters of the system, they will be penalized. You can read through the white paper to see the full spectrum of governance responsibilities that MKR holders are entrusted with. The most important one in the context of this discussion, however, is management of the collateral pool. In the next iteration of the Maker system (scheduled to launch by end of year), Ether won’t be the only collateral type. Additional forms of less correlated asset types will be added in order to avoid a specific type of risk known as the ‘death spiral’.
The theory of the death spiral is as follows. If the Ether price keeps plummeting, too much dilution of MKR might spook current holders, leading them to sell their MKR stash, which would necessitate further dilution in order to cover collateral shortfall. Why would investors buy MKR if it’s going to perpetually dilute? There needs to be an incentive here for investors to maintain confidence.
Now imagine a diversified collateral pool of 20 uncorrelated asset types (such as REP, tokenized gold, or security tokens). If one of these collateral types suddenly fails, MKR dilution would occur. Critically, however, investors should theoretically step in to buy this newly minted MKR, as there is still a healthy revenue-generating system in operation. A diversified collateral pool insulates the Maker system from a death spiral. Therefore, one of the most important governance decisions MKR holders need to make is allowing only the highest quality assets to be used.
Several governance questions remain in the Maker system regarding what these high quality assets will look like. As of today, there is no silver bullet on tokenizing physical assets onto the blockchain. While there are mitigations being explored, this could end up becoming be a central point of failure. Details on the intricacies of this system will likely become available shortly.
As with any investment, one must always weigh the trade-offs when choosing which stablecoin to use as a shelter against volatility. While centralized approaches might seem attractive because of their apparent lack of complexity, one shouldn’t forget that they reserve the right to freeze your money as part of their regulatory compliance. For a bit of complexity, however, a DAI user can enjoy the benefits of a decentralized censorship-resistant stablecoin.
*Disclaimer: Cyrus Younessi is Director of Research and Trading of Scalar Capital, which holds a position in MKR. This post is not investment or legal advice.
Thank you to Linda Xie, Jordan Clifford and Rich Brown for providing feedback on this article.