BitMEX founder: What is the best way to design a stablecoin?

We have little control over why humans exist in this universe. As a civilization, we spend an inordinate amount of energy trying to bring calm and stagnation to a turbulent planet. Just turning the temperature up and down in our homes and workplaces takes a lot of energy.

Because we instinctively understand that humans are reeds blown by the wind, which gives individuals and institutions tremendous power. Politicians tell us there is a plan, business leaders map out a path for the future, and we hope to always succeed. But time and again, reality throws unexpected problems, and the plans made by leaders often fall short. But what can we do, try again and again?

Just as civil society exhibits composure, the money that powers civilization must also remain stable. Fiat currencies are designed to slowly depreciate over time. Humans cannot comprehend decades or centuries of lost purchasing power when referencing the purchasing power of fiat currencies in the past. We are conditioned to believe that today’s dollar, euro, yen, etc. will buy the same amount of energy tomorrow.

The behavior of Bitcoin and the crypto movement it inspired is extremely tragic. Satoshi Nakamoto was essentially a revolutionary, and the love and anger of crypto believers has led to Bitcoin's price volatility relative to fiat currencies and pure energy itself. While believers claim to accept this volatility with unwavering faith, we are only human and sometimes abandon the gold standard. Stablecoins have shown us their sweet tune in difficult times, but many people fail to recognize that they are fundamentally incompatible with the financial world we hope to create.

Many people have asked me about my thoughts on stablecoins. The recent volatility surrounding Terra’s USD stablecoin UST, which is pegged to $1, has prompted me to start a series of articles on stablecoins and central bank digital currencies (CDBCs). Both concepts are interrelated to the fundamental nature of the debt-based, partially banking system that dominates the global financial system.

This article will explore the broad categories of blockchain stablecoins - including fiat-backed, overcollateralized crypto, algorithmic, and Bitcoin-backed stablecoins. While there is no proven solution yet, the last part of this article will cover my current view that a stablecoin based on Bitcoin and pegged to the US dollar that is also an Ethereum Virtual Machine (EVM) compatible ERC-20 asset is the best way to combine these two incompatible systems.

In this time of heightened downward market volatility, the only solace is our breath. Our actions and thoughts are not completely controlled by reason, so we must breathe in and out slowly, methodically, and consciously. Only in this way can we faithfully preach the good word of the Lord.

"The Fremen have a saying: 'God created Arrakis to train believers.' Man cannot disobey God's word." - Paul Atreides, 'Dune'

Fiat-backed stablecoins

As explained in the previous article "I Still Can't Draw A Line", banks are utilities that operate fiat value. They help individuals and organizations conduct business activities. Before the advent of the Bitcoin blockchain, banks were the only trusted intermediaries that could perform these functions. Even after the advent of Bitcoin, banks are still the most popular intermediaries, which has led some banks to even do reckless behavior because they believe that the government can print money to provide bailouts for their actions.

Banks charge a very significant tax on the time and fees users spend transferring value. Given that we now have instant and nearly free encrypted communication methods, there is no reason to continue paying so many fees and wasting so much time with traditional banks.

Bitcoin has created a competitive peer-to-peer payment system with low time and money costs. For many people, used to using fiat currencies and energy (i.e. a barrel of oil) as a benchmark, Bitcoin is extremely volatile. To solve this problem, Tether created the first stablecoin pegged to the US dollar using the Omni smart contract protocol built on Bitcoin.

Tether created a new digital asset class on a public blockchain, backed 1:1 by fiat assets held by banking institutions, which we now call fiat-backed stablecoins. After Tether (also known as USDT) and USDC, various other fiat-backed stablecoins have sprung up, and the fiat custody assets (AUC) held by each project have surged as the tokens have grown. Currently, USDT and USDC together have more than $100 billion in fiat AUC.

Because the Bitcoin economy lacks infrastructure, our means of payment are still in US dollars or other fiat currencies. Since traditional fiat currency sending and receiving methods are very expensive and complicated, bypassing the bank payment system and sending fiat currency instantly at low cost is a very valuable thing. I would rather send USDT or USDC to someone than use the expensive global fiat bank payment system.

The fundamental problem with this type of stablecoin is that it requires willing banks to accept the fiat assets that back the token. The transaction fees of stablecoins do not go into the pockets of bankers, but there is a cost for banks to hold these huge fiat assets. It is well known that central banks have destroyed the lending business model of commercial banks, which makes it impossible for them to agree to agreements for commercial banks to hold billions of dollars to achieve decentralization.

Fiat-backed stablecoins want to use the bank’s storage facilities but pay nothing for it. To me, this strategy will not survive. A few billion might be fine, but expecting commercial banks to allow a fiat-backed stablecoin’s AUC to reach trillions is impossible.

Fiat-backed stablecoins will not be the payment solution that supports Web3 or a truly decentralized global economy. They cannot be the digital payment service that connects the physical world quickly, cheaply, and securely. We saw this dissonance when the Federal Reserve banned Silvergate Bank from being a partner in Facebook's Diem stablecoin. Due to Facebook's large user base, Diem would immediately become one of the largest currencies in circulation in the world if it launched. It would be in direct competition with traditional fiat currencies, which cannot be allowed to happen.

The next iteration of stablecoins is a wave of projects that overcollateralize major cryptocurrencies to maintain a peg to fiat value.

Overcollateralized stablecoins

In short, these stablecoins allow participants to mint a pegged fiat token in exchange for crypto collateral, the most successful of which is MakerDAO.

MakerDAO has two currencies. Maker (MKR) is the token that governs the system. It is similar to a share in a bank, but the bank's goal is to have more assets than liabilities. These assets are mainstream cryptocurrencies such as Bitcoin and Ethereum, and MakerDAO promises to create a token pegged to the US dollar, DAI, after receiving the crypto assets.

1 DAI = 1 USD

Users can borrow DAI from MakerDAO with a certain amount of crypto collateral. Since the price of crypto collateral may fall in USD value, Maker will programmatically liquidate the pledged collateral to satisfy the DAI loan. This is done on the Ethereum blockchain and the operation process is very transparent. Therefore, the level price of Maker's liquidation can be calculated.

This is a chart of the percentage deviation of DAI from its $1 peg, a reading of 0% means DAI is perfectly holding the peg. Maker is doing a great job of keeping the USD peg.

The system is very robust as it has survived several price crashes in Bitcoin and Ethereum, with its DAI token still maintaining a value close to $1 on the open market. The downside of this system is that it is overcollateralized. It effectively eliminates liquidity from crypto capital markets in exchange for the stability of a pegged fiat asset. As we all know, stagnation is expensive, while volatility is free.

f(x) for MakerDAO and other over-collateralized stablecoins: Completely draining the ecosystem of liquidity and collateral. Maker token holders can choose to introduce risk into the business model by lending out idle collateral in exchange for greater income. However, this introduces credit risk into the system. Who are the reliable borrowers paying positive interest rates in crypto, and what collateral are they pledging? Is that collateral the original asset?
The benefit of fiat fractional banking is that the system can grow exponentially without running out of collateral for all currencies. These over-collateralized stablecoins fill a very important niche, but they will always be a niche for the fundamental reasons outlined above.
The next iteration of stablecoins aims to completely remove the link to any “hard” collateral and be backed solely by fancy algorithmic minting and burning schemes. In theory, these algorithmic stablecoins could scale to meet the needs of a global decentralized economy.

Algorithmic Stablecoins

The stated goal of these stablecoins is to create a pegged asset with less than 1:1 crypto or fiat collateral. Often the goal is to use assets other than “hard” collateral to back the pegged stablecoin. Given that Terra is the subject at hand, I will use LUNA and UST as examples to explain the mechanics of algorithmic stablecoins.

LUNA is the governance token of the Terra ecosystem.

UST is a stablecoin pegged to 1 USD, whose “assets” are simply the LUNA tokens in circulation.

Here’s how UST’s peg to $1 works:

Inflation: If 1 UST = $1.01, then the value to which UST is pegged is overvalued. In this case, the protocol allows LUNA holders to exchange $1 worth of LUNA for 1 UST. LUNA is burned or removed from circulation, and UST is minted or put into circulation. Assuming 1 UST = $1.01, the trader makes a profit of $0.01. This drives up the price of LUNA because its supply is reduced.

Deflation (where we are now): If 1 UST = $0.99, then UST is undervalued relative to its peg. In this case, the protocol allows UST holders to exchange 1 UST for $1 worth of LUNA. Assuming you can buy 1 UST for $0.99 and exchange it for $1 of LUNA, you will make a profit of $0.01. UST is burned and LUNA is minted. This causes the price of LUNA to go down because its supply increases on the way down. The biggest problem is that investors who now own newly minted LUNA will decide to sell it immediately instead of holding it in the hope of a price increase. This is why LUNA faces constant selling pressure when UST is trading at a price that is significantly decoupled from its peg.

The more UST is used in commerce in the Web3 decentralized economy, the more valuable LUNA becomes. This minting and burning mechanism is very useful on the way up. But if UST cannot reverse its downward trend, then a death spiral may begin with LUNA minted indefinitely in an attempt to get UST back to its peg.

All algorithmic stablecoins have some kind of minting/burning interaction between the governance token and the pegged stablecoin. All of these protocols have the same problem: how to increase people’s confidence in restoring the peg when the pegged stablecoin is trading below the fiat peg .

Almost all algorithmic stablecoins have failed due to the death spiral phenomenon. If the price of the governance token drops, then the governance token asset that backs the pegged token is viewed as untrustworthy by the market. At that point, participants begin to sell off their pegged tokens and governance tokens. Once the spiral begins, it is very expensive and difficult to restore confidence in the market.

The death spiral is no joke, it is a confidence game based on a debt-based banking system. However, this game has no government that can force users to use the system.

In theory, profit seekers should be willing to ignore the decline in collateral to save the algorithmic protocol in order to reap the huge profits of governance tokens created out of thin air. But this is just hypothetical.

Here is a chart of the percentage deviation of UST from its $1 peg. As with MakerDAO, 0% means the peg is rock solid. As you can see, everything was fine until UST broke free.

Many similar projects have failed or are failing. This is not to say that this model can't work, at least for a period of time. I am working on a specific algorithmic stablecoin project and holding its governance token. The project is currently profitable at the protocol level, which makes them attractive. The protocol has a similar structure to Terra, but in addition to its governance token, it also accepts other mainstream collateral to support its anchor stability.

In theory, this model resembles parts of a bank and could scale to meet the needs of a decentralized Web3 economy, but would require near-perfect design and execution.

Bitcoin-backed stablecoins

The only laudable goal of stablecoins is to allow tokens pegged to fiat currencies to be issued on public blockchains. This has practical uses until true Bitcoin economics arrives. So let’s try to make the best of a fundamentally flawed premise.

The original crypto collateral is Bitcoin. How can we transform Bitcoin, which has a 1:1 USD value ratio, into a USD-pegged stablecoin that is difficult to break?

Various top cryptocurrency derivatives exchanges offer inverse perpetual swaps and futures contracts. These derivatives contracts are based on BTC/USD but are margined in BTC. This means that profits, losses, and margin are denominated in Bitcoin, while quotes are in USD.

I’ll hold your hand while we do some math — I know it’s hard on your TikTok-damaged head.

Each derivative contract is worth $1 of Bitcoin at any price.

Contract Value Bitcoin Value = [$1 / BTC Price] * Contract Quantity

If BTC/USD is $1, the contract is worth 1 BTC. If BTC/USD is $10, the contract is worth 0.1 BTC.

Now let’s create $100 using a combination of BTC and short derivative contracts.

Assume BTC/USD = 100 USD.

At a BTC/USD price of $100, how many contracts or $100 worth of BTC are there?

[1 USD / 100 USD] * 100 USD = 1 Bitcoin

Intuitively, this should make sense.

100 synthetic USD: 1 BTC + 100 short derivative contracts

If the price of Bitcoin goes to infinity, the value of a short derivative contract denominated in Bitcoin approaches the limit of 0. Let’s demonstrate this with a BTC/USD price that is larger but less than infinity.

Suppose the price of Bitcoin rises to $200.

What is the value of our derivatives contract?

[1 USD / 200 USD] * 100 USD = 0.5 BTC

Therefore, our unrealized loss is 0.5 BTC. If we subtract the unrealized loss of 0.5 BTC from our 1 BTC collateral, we now have a net balance of 0.5 BTC. But at the new BTC/USD price of $200, 0.5 BTC is still equal to $100. Therefore, even if the price of Bitcoin rises and causes an unrealized loss on our derivatives position, we still have $100 in synthetic USD. In fact, it is mathematically impossible for this position to be liquidated upwards.

The first fundamental flaw in this system occurs when the price of BTC/USD approaches 0. As the price approaches zero, the value of the contract becomes greater than all the Bitcoins that will ever exist - making it impossible for the short seller to pay you back in Bitcoin.

This is math.

Suppose the price of Bitcoin drops to $1.

What is the value of our derivatives contract?

[$1 / $1] * $100 = 100 Bitcoin

Our unrealized gains are 99 BTC. If we add the unrealized gains to the initial 1 BTC collateral, we arrive at a total balance of 100 BTC. At a price of $1,100, Bitcoin is equivalent to $100. Therefore, our synthetic peg of $100 is still valid. However, notice how a 99% drop in the price of Bitcoin increases the value of the Bitcoin in the contract by 100x. This is the definition of negative convexity, and shows how this peg breaks when the price of Bitcoin approaches 0.

The reason I am ignoring this scenario is that if Bitcoin goes to zero, the entire system will cease to exist. At that point, there will no longer be a public blockchain capable of transferring value, because miners will not expend the sheer energy of maintaining a system where the native token is worthless. If you are worried about this being a real possibility, just continue to use the fiat banking rails — no need to try something that might be cheaper and faster.

Now, we have to introduce some centralization, which brings a host of other problems to this design. The only place where these inverse contracts are traded at a scale large enough to accommodate a Bitcoin-backed stablecoin that can serve the current ecosystem is on centralized exchanges (CEXs).

The first point of centralization is the creation and redemption process.

Creative process:

Send BTC to the foundation.
The foundation pledges BTC on one or more CEXs and sells inverse derivative contracts to create sUSD, or synthetic dollars.
The foundation issued a sUSD token based on a public blockchain. For ease of use, I recommend creating an ERC-20 asset.
In order to trade these derivatives, the foundation must create an account on one or more CEXs. The BTC collateral is not held with the foundation but now stays with the CEX itself.

Redemption process:

Send sUSD to the foundation.
The Foundation buys back one or more CEX short reverse derivative contracts and then destroys sUSD.
The Foundation withdraws the net BTC collateral and returns it to redeemers.
There are two problems with this process. First, the CEX (for whatever reason) may not be able to return all of the BTC collateral entrusted to it. Second, the CEX must collect margin from losers. In the case of this project, the project's derivatives are profitable when the BTC price falls. If the price falls too far, too fast, the CEX will not have enough long margin to cover it. This is where various socialized loss mechanisms come into play. TL;DR, we cannot assume that if the BTC price falls, the project will receive all of the BTC profits it deserves.

set up

The foundation needs to raise funds for the development of the project. The biggest need for funds is a general fund that covers counterparty risk on the exchange. The governance tokens must initially be sold in exchange for Bitcoin. This Bitcoin is specifically for situations where the CEX does not pay out as expected. Obviously, this fund is not inexhaustible, but it will give confidence that the $1 peg can be maintained if the CEX returns less than it should.

The next step is to determine how the protocol will earn revenue. There are two sources of revenue:

• The protocol will charge a fee for each creation and redemption.

• The protocol will gain a natural positive basis for the derivative contract with the underlying spot value. Let me explain.

The stated policy of the Fed (and most other major central banks) is to inflate their currency at 2% per year. In fact, since 1913 (the year the Fed was founded), the dollar has lost more than 90% of its purchasing power when pegged to the CPI basket.

BTC has a fixed supply. As the value of the denominator (USD) grows, the numerator (BTC) remains constant. This means that we should always assign a higher value to the future value of the BTC/USD exchange rate than the spot value. Therefore, fundamentally, the contango (futures price > spot) or funding rate (perpetual swaps) should be positive - meaning income for those who are short these inverse derivative contracts.

One could counter that US Treasuries have a positive nominal yield and there are no risk-free instruments nominally priced in Bitcoin - so it is incorrect to assume that the dollar will depreciate relative to Bitcoin over the long term. While this is true, as I and many others have written, negative real interest rates (i.e. when the nominal risk-free Treasury rate is less than the GDP growth rate) are the only mathematical way for the US to nominally repay debt holders.

The alternative is to increase population growth above 2% per year, which would require couples to collectively avoid contraception and other family planning methods. In 2021, the population growth rate is 0.1%, according to the U.S. Census Bureau. If you exclude immigration, the tax rate would be negative.

The final option is to discover some new amazing energy conversion technology that greatly reduces the energy cost per dollar of economic activity. Neither of these alternative solutions seems likely to be realized anytime soon.

Long Bitcoin vs. short inverse derivative contracts should have positive returns year after year. Therefore, the more sUSD there is in circulation, the more Bitcoin is held in custody compared to short derivative contracts, generating a large compounding interest income stream. This provides a large pool of capital for governance token holders.

Perfection is impossible

It is not possible to create a stablecoin pegged to a fiat currency on a public blockchain without many compromises. It is up to the users of the solution in question to determine whether the compromises are worth the goal of making fiat faster and cheaper on a public blockchain than on a centralized payment network controlled by banks.

Of the four options presented, I like Bitcoin and derivatives-backed stablecoins best, followed by overcollateralized crypto-backed stablecoins. However, each of these solutions pins cryptocurrencies in large pools of money. As I mentioned in The Doom Loop, the problem is that these public networks require assets to move between parties in order to generate transaction fees that pay for network maintenance. Holding is toxic in the long run. So let’s not get complacent, but keep working to create a farm-to-table Bitcoin economy.

Will Terra/UST survive?

Terra is currently in the deepest of a death spiral. Reading this tweet thread from founder Do Kwon, what is happening right now is completely by design. The protocol is working fine and the fact that people are surprised by what is happening means they haven’t read the whitepaper properly. Luna-tics also didn’t question hard enough where Anchor’s 20% UST yield is coming from.

The spiral stops when UST market cap equals LUNA. If left unchecked, the protocol will find market cap equilibrium. The question then becomes what the final resting market cap is. Most importantly, when arbitrageurs buy cheap UST to create fresh LUNA, who will buy that LUNA? When you know there are billions of dollars of LUNA selling pressure, as long as UST < $1, why would you buy LUNA from those who are selling it.

Even if LUNA and UST survive this event, some genius protocol change would have to happen in the long run to bolster market confidence that LUNA’s market cap will always exceed UST’s float. I don’t know how to do this. That’s why I only LARP. Many have highlighted this fundamental issue — check out this post by Dr. Clements for a more detailed discussion.

Here is a graph of [UST Market Cap – LUNA Market Cap]. When this value < $0, the system is healthy. A spike upwards means that UST must be destroyed and LUNA issued to re-peg UST.

Algorithmic stablecoins are not much different from fiat debt-backed currencies, except for one crucial factor: Terra and others like them cannot force anyone to use UST at any cost. They must convince the market with their fancy designs that the governance tokens backing the protocol will have a non-zero value that grows faster over time than the number of fiat-pegged tokens issued. However, a government can always eventually force its citizens to use its currency at gunpoint. Therefore, there is always an inherent demand for fiat currency, even if everyone knows that the "assets" backing the currency are worth less than the currency in circulation.

Another victim is a group of investors who shouted “Yeah, Yeah, Hooray!” to DeFi because of their enthusiasm for Terra. Now these investors will be busy repairing their balance sheets instead of buying Bitcoin and Ethereum when they resume their downward trajectory.

Adjournment

This debacle isn't over yet...

During a proper crash, the market seeks out those indiscriminate sellers and forces their hand. This week’s crash was exacerbated by the forced sale of all Luna Foundation Bitcoins to defend the UST:USD peg. As usual, they still failed to defend the peg. This is how all anchors fail in the face of cosmic entropy.

I dutifully sold $30,000 of Bitcoin and $2,500 of June puts. I have not changed my structural long crypto positions, even though they are losing “value” in fiat terms. If anything, I am evaluating the various altcoins I own and adding exposure.

I did not expect the market to move through these levels so quickly. This collapse came less than a week after the Fed raised rates by the expected 50 basis points. This market cannot handle rising nominal rates. It is shocking to me that anyone can believe that long-term risk assets at historically high price multiples will not succumb to rising nominal rates.

US CPI rose 8.3% YoY in April, down from 8.5% YoY previously. 8.3% is still too hot to handle, and the Fed in firefighter mode cannot give up its quixotic pursuit of fighting inflation. A 50bp rate hike is expected in June, which will continue to undermine long-term risk assets.

The crypto capital markets must now determine who is overly exposed to the risk of anything Terra related. Any service offering above average returns that is perceived to have any exposure to this melodrama will experience rapid outflows. Given that most people have never read how any of these protocols actually work in a distress situation, it will be an exercise in sell first, read later. This will continue to weigh on all crypto assets as all investors lose confidence and would rather grab the safety rug and hold fiat cash.

Crypto capital markets must be given time to recover after the bloodletting is over. So it would be foolish to try to understand a reasonable price target. But I will say that given my macro view that eventual money printing is inevitable, I will close my eyes and trust the Lord.

Therefore, I am a buyer at $20,000 for Bitcoin and $1,300 for Ethereum. These levels roughly correspond to the all-time highs for each asset during the 2017/18 bull run.

We do not know who among us is in league with the devil yet pledges allegiance to God. So keep yourself away from the "buy" and "sell" buttons and wait for the dust to settle. All will be revealed to the believers in due time.