Marty McFly Meets His Maker(DAO)

As I covered last year on Coppercast, there are three main models for stablecoin mechanisms, each with its own particular risks:

• tokenized cash (USDC, BUSD, Tether): fiat-collateralized 1:1 backing or a mix of relatively low-risk, cash-like assets minted as a token by a centralizer issuer; I would argue that JPM’s tokenized deposits fall into this category as well in that both represent a tokenized claim on bank deposits, albeit JPM expresses this relationship more directly; they represent yield-free liabilities with embedded credit risk for which the receiver gets zero compensation.
• algostables (UST, Frax): typically a pair of tokens, one pegged to the price of another asset, typically but not always USD, and the other free-floating, acting like a stabilizing arm operated via arbitrage: incentives built into the tokenomics encourage traders to sell the pegged token when it drifts about the peg, lowering the price, and to buy when it drifts below; typically there involves a mint/burn of the stabilizer, in the case of Terra/LUNA this pair was UST and LUNA. In some cases, including in the latter days of UST, there is a “reserve” of some kind intended to be deployed to defend the peg in emergencies. Both smart contract risks in the form of flaws or exploits and tokenomic mechanism design errors can lead to tragic outcomes here even with billions in reserves. If you think the reserves are enough to protect the peg, ask ChatGPT who is “the man who broke the Bank of England” and see how you feel. When eye-watering profits lie in wait for the arbitrageur who can kick the asset of its peg, the cost of introducing enough doubt to succeed is often within reach.
• tokenized collateralized debt positions (DAI): an organization, typically a DAO, manages a treasury of crypto collateral well in excess of the issued stablecoin’s notional value, sometimes reinvesting assets in yield-generating strategies to further supplement the reserves. Where there is a secondary token, more often it is a governance token which controls the issuer. There are operational risks, smart contract risks but most importantly liquidity risks at play here. I am somewhat troubled by the recent excitement around the DJED stablecoin on Cardano as it emphasizes the 600% - 800% collateralization without acknowledging that much of this collateral is in the ADA token, which is intertwined with DJED itself via technical linkages. Its price is easy to manipulate with enough capital and issues with one part of this duo can overspill on the other, e.g. a Cardano network problem might slash the value of the ADA position, and if this were to be combined with a lot of selling pressure in ADA — likely — this could further erode that overcollateralization, especially as it would not take much selling to move the market in an illiquid token.

So: what happened in the last week?

• The SEC forced Paxos to abandon its support for Binance USD (BUSD), in part supported by a claim that BUSD is an unregistered security.
• The SEC filed an enforcement action against Terraform Labs over Terra/LUNA, again based on a claim that the stabilizing token in that pair is a security.

That a regulator acting in the name of investor protection triggered a $1B move from BUSD to Tether should give us all pause. Tether is backed by the full faith and credit of a small bank in the Bahamas and a former child actor from The Mighty Ducks, and has on the best of the days a less than ideal record of disclosure and auditing; I have yet to meet an institutional crypto trader who is willing to sit on a large USDT position any longer than absolutely necessary. Recent enforcement is very likely to create a classic market for lemons: by acting after-the-fact without a clear framework, the good actors will proceed with caution and the worst players will charge ahead.

For a long while in DeFi, the reaction has been to shrug, say that there is no alternative, and move on:

But in the face of the decentralized dumpster fire that is the current situation, I think it has to be said that this is not fine. Two of the three dominant models for stablecoins are in the eyes of the SEC based on unregistered securities, and the one that was most likely to get regulatory clarity in the U.S. — tokenized cash — was further undermined by a Fed Board of Governors statement that same week coming out against banks holding crypto assets on balance sheets. That means Circle may not be able to get an OCC bank charter and the associated regulatory clarity — bank accounts have a carve-out under securities law — and bank supervision unless Congressional action makes very clear that this is OK. The anti-flight-to-quality that resulted was a reaction, essentially the market saying that sketchy offshore tokenized cash is more likely to survive, whatever its demerits.

But as I was on my way to Singapore, I started thinking about this from a different angle. What follows is not a solution, but a way of reframing the problem, and effective problem-finding is often more important than problem-solving. In Agile product development, there is in my view an over-emphasis on “listening to the customer,” and I started to wonder if stablecoins were a case of listening to what the customer says they want rather than listening for what the customer needs. What if that TINA assumption that stablecoins are needed and even desirable is wrong? What if, where we are going, we don’t need roads?

When you are managing risky assets, you often want to have an option to trade in and out of risk to keep the total portfolio risk within your mandates or at least your comfort level: it’s a safe place to go when the winds are howling. Furthermore, most of us have liabilities denominated in fiat currency, whether tax payments due or mortgages; a store of value that is highly-volatile could leave us insolvent, unable to deliver cash flows to our creditors — and the more volatile, the more likely we are to go suddenly bankrupt. This is not fine; we don’t want this. You can summarize the properties of an instrument that satisfies these needs as:

• Liquid with respect to transfers in and out of fiat
• Low volatility relative to fiat — but not necessarily zero
• Offering adequate compensation for any assumed credit risk
• Offering an inflation rate sufficiently low and stable to absorb price level adjustments and high enough to allow for adjustment across the economic cycle

The U.S. dollar has these properties by construction if you have U.S. dollar liabilities, and the full faith and credit of the U.S. government, whether directly deployed or indirectly through mechanisms like FDIC protection for bank accounts makes the credit spread component stable and low so long as Congress doesn’t do something batshit crazy like not raising the debt ceiling. (Note: not a 100% failsafe assumption.) Speaking to the inflation assumption will probably get me something very smelly in my mailbox from goldbugs and Bitcoin maxis, especially after recent increases in inflation, but let’s just run with this one for a while: hyperinflation, as LUNA holders discovered, is not cool, and we want to avoid it.

The credit risk is an interesting dimension. Financial models may have a concept of a risk-free rate, but any time you have a liability there is some credit risk, and every fiat currency, bank account, bond, LP token, CDP, etc. is a liability. As an EM CDS trader once said to me, you always have to price in both can’t-pay and won’t-pay risks. Yield, whether fixed or floating, is the liability-holder’s compensation for assuming these risks. If someone is paying you 20% to hold a USD-pegged asset, they are basically jumping around and shouting at you that there is a relatively high probability that your Benjamins may not come back to you. As noted above, if you have a mortgage payment tomorrow, this is not good. We’d prefer to at least have the option of holding an instrument which is much less racy but at least gives us a high degree of comfort that the money we need is there when we require it.

Which brings us to MakerDAO, and its roadmap to stop pegging to the U.S. dollar. I have to be honest, my initial reaction here was that this was ideologically motivated and kind of missing the point: what good is a stablecoin that’s not pegged to fiat? But I think that’s the wrong way of thinking about it. The better question to ask is: could you construct a tokenized financial instrument without a fiat currency peg which nevertheless has the desirable properties outlined above? And you can. What we have described is a tokenized debt instrument: a bond, issued by a DAO.

It is not even necessary to lose USD price quote. Even in trading, we already have this concept of separating the price something is quoted in and the settlement asset delivered to you by a derivative contract or in an outright sale: AAPL is quoted in dollars, but you get a security delivered into a custodian. Similarly, an order book that quotes in USD but is financed with crypto-assets can be made: in fact, Deribit has had one for quite some time for its BTC-USD option contracts. These positions are funded with Bitcoin but quoted in USD, even though no USD ever moves in or out of that exchange’s custody. All you need is an exchange rate to translate from your USD quote to the delivery asset. If this delivery asset is far less volatile than Bitcoin and offers some yield to compensate for its credit risk, it does not have to be based on tokenized cash or deposits. If constructed properly, there should be a highly-liquid market for exchanging this semi-stable asset for USD, in the same way that you can easily trade in and out of money-market funds in the U.S. today.

What Marty McFly needs is not a stablecoin, but rather a fixed income instrument. In fact this is better, because one of the most troubling things about the various collateral-based stablecoins is the fact that they offer zero compensation for credit risk: Tether collects the interest on its billions of reserves but does not pass this income on to its token-holders, creating perverse incentives for them to take higher risks on reserve management because the ultimate bag-holder here is the person who minted the Tether token, not the management team. But wait! There’s more! Finance offers all sorts of wonderful examples of stripping out risks and directing those risks to those better able to manage them, e.g. the classic fixed-float interest rate swap. So if you are willing to trade your variable credit spread on that tokenized instrument to someone who might better better able to understand and manage the evolution of those credit spreads and other interest rate risks over time, you may well be cool with giving them some compensation for it: take a lower, fixed rate in exchange for certainty. So you start to not only get a solution for stablecoins, but you start to see a term structure of risk evolve. (The bonkers crypto yield curve is its own topic, so let’s just leave it at this for now: having a liquid rates market and a yield curve is useful.)

MakerDAO may or may not be the right vehicle for this, and the fiat-to-token exchange does not go away; in many ways it becomes more complex, because what we just described above is most certainly a security in the eyes of the SEC. But traditional finance offers models here too for actors who make fiat-to-security conversions with regulatory clarity: securities exchanges. The natural on-ramp from fiat-to-crypto is not necessarily someone with a money transmitter license, but NASDAQ and the brokerages who have licenses to interact with it, who among other things have these nice rules about segregating client funds which FTX, um, did not. You thus have a liquid marketplace with well-understood interchanges and regulation, including appropriate exchange supervision. All you need is a way for those brokerages to then coordinate the transfer to on-chain custody, and thereafter we are in the Web3 and DeFi world again, but this time with a portable instrument that’s not pegged — in fact, it floats as market views on credit risk of its issuer evolves — but offers a safe harbor for funds during periods of high volatility and a store of value. You have also created a chain of custody: from the bank to the brokerage to the transparent blockchain ledger, which could go a long way toward reassuring KYC/AML concerns.

Like I said, this is not a solution, but a way of thinking about the problem. However DeFi evolves over time, Cloudwall will be here to help market participants understand the manifold risks. But the starting point should be risk, and not magical thinking. Zero-yield instruments with underlying credit exposures are dangerous. Waving your hands at high yields and telling customers it’s because DeFi is so amazingly efficient is intellectually dishonest; risks should be disclosed and understood, not hidden. There are better ways, and they don’t even have to be new ways; finding new applications of old ideas is an honorable way to innovate. In the wake of the FTX meltdown, Terra/LUNA and their aftershocks, something better may arise.

Note: this essay has been written from an American perspective. Most of the principles here could be applied in other regulatory contexts, e.g. the euro. But as it’s being written from an airport, you are getting the shorthand.