each crypto asset, blockchain, and project needs to be assessed on its own merits. Talking about crypto as if it is a monolith obscures important differences and makes it easier for scammers to hide in the crowd.

Commissioner Hester M. Peirce

We suggest here a nomenclature of crypto-finance project. Any endeavor to create a nomenclature is nevertheless doomed to fail, as it is a multidimensional continuum more that a few categories. This categorization is therefore more defined to elaborate on qualities of those crypto-finance projects and less about providing a definitive framework.

Crypto-currencies

The first flavor of crypto-finance is crypto-currencies. This is the ultra-sound money meme in action or the Austrian economic dream of Bitcoiners. It is explored in the Synthetic community money paper by George Selgin.

One can use BTC or ETH in this category. Both are very trustless, immutable, composable, transparent, open-source and accessible. They aim to be a generally accepted form of money, which they are not yet, but have strong arguments.

The current problem is their lack of good payment system (which seems basic for a form of money). It seems it was easier to make (digital) purchases in 2015/2016 than in 2023.

Beyond this flavor, the rest of this article is more about providing finance on top of money and is indifferent to the form of money is is based on (which can be a fiat currency of a crypto-currency). Crypto-finance is orthogonal to crypto-currencies.

PureFi

The PureFi flavor is what people usually expect (but don’t get) when they speak about Decentralized Finance (DeFi). Those protocols are governed by their smart contracts and there is no external dependencies.

Any feature change (including bug fixes) requires a new smart contract deployment. Users will also need to migrate to the new version. The older version will remains forever on the blockchain for anyone to use.

Building a business in PureFi is challenging. As the protocol is only based on open-source code, any tentative to extract fees can be countered by anyone redeploying the contracts.

The most famous example here is Uniswap which allows to trade tokens. At least up to version 3, all contracts are immutable. The first version is still alive and used. They are censorship resistant as no one can do anything about it. For instance, Uniswap Labs removed the security tokens from its interface. Yet, they are still traded on Uniswap v2 (here for example). The price of the swap is derived internally and there can’t be any so called “Oracle attack”. There is a fee switch which would divert some transaction fees to UNI holders, but it wasn’t activated yet and it seems unlikely to be activated.

PureFi protocols are of paramount importance as those are bricks on which you can safely build. Yet the trustless constraint is difficult to integrate. For instance, currently there is no PureFi lending protocols (also this should change in the coming months).

DaoFi

While PureFi represents the gold standard of DeFi, the limitation of smart contracts makes it sometimes necessary to add two external components: oracles and governance. Oracles deliver exogenous information to the blockchain like the price of a crypto-currency or a commodity. Governance allows people, usually governance tokenholders, to make changes to the protocol like changing fees, setting new interest rates, updating the protocol, … .

Both of those external dependencies are risky, Oracle manipulation are one of the biggest reasons for hacks and the example on Mango shows a case where an attacker used a oracle price manipulation to steal funds, then used those stolen fund to vote a governance action to whitewash his actions. Chainlink, biggest Oracle protocol is for instance managed with a 4 out of 9 multisig with unknown signers, which is not considered strong on a decentralization spectrum.

Moreover, governance is usually working like a group of shareholders incentivize to maximize their value. There is very limited, if any, minority tokenholders protection or creditor protection. It is the case that a majority of MKR tokenholders could collude to steal all the collateral, not only of MakerDAO but also from Compound v2 and Aave v2. The fact that such attack has almost no chance of happening shouldn’t be taken as an excuse to not create new standards of safety. Governance can also vanish as in the case of Fei Protocol that was shut down.

Those shortcomings shouldn’t discard DAOs. In my view the main difference between a DAO and a corporation is mainly a cultural one by working the the open. We should move from the cultural part (and to some extend regulatory arbitrage) and improve the smart contract part (and I couldn’t care less if it’s a DAO or a regular LLC).

BridgeFi

BridgeFi encompass projects that allows to bring on-chain assets that aren’t endogenous to it. The biggest examples are the fiat-backed stablecoins, where a company issued 1 token (e.g. USDC) for each US Dollar it has under management.

The primary market (creation and redemption) is usually limited to KYC people only, but the secondary market is wide open, meaning anyone can transfer the tokens freely. There is a blacklist to freeze accounts (which is not great) but so far it has worked well. This at least enable composability within the space.

Those projects are based on trust. In that light, it should be mandatory for them to provide audited financial statements and use narrow purpose entities (bankruptcy remote SPVs from the rest of the company).

So far, we are not there yet. Lindy effect is more prevalent that transparency as a metric. Despite the lack of clarity of Tether, USDT remains one of the most used stablecoin in the crypto-economy. Details on the legal structure are usually not disclosed and we have a best the virtual safety of regulation under money transmitter or limited chartered trust.

A good example is WBTC, which bridge Bitcoin to Ethereum. While you have both an on-chain link on the asset and liabilities, there is no trace whatsoever of any legal agreement between BitGo and a WBTC DAO. MakerDAO for instance is mentioned as a member but never was (probably Maker Foundation was, but it is now defunct and no one seem to remember where the key is). This obviously lead to trust being shaken every now and then.

This is quite an uncomforting situation as we can’t build a trillion dollars ecosystem on top of black boxes. Without the much needed transparency and openness, the future is not bright if we can’t solve that.

RegFi

One type of BridgeFi project are similar but qualitatively different due to the use of a whitelist, a list of accounts allowed to hold the token. This limit composability a lot making any product an order of magnitude, at least, less used. Some solutions to create bridges are explored but they obviously are less than ideal.

This state of affair is solely due to the SEC and securities laws in the US which is quite unfortunate as the discussion moves to regulation compliance instead of the actual safety of a token (full disclosure, I lost all hope on the securities laws and protecting the retail investor with the launch of APE).

Quite unexpectedly, this space is showing quite significant progress towards a better safety and transparency. For one, all users need to sign some kind of agreement so they actually have some legal documents with is already a progress. But, moreover, . You can for instance find all the legal documentation of the MakerDAO HVB deal on the permaweb. This should be the case every time, just like the code of smart contract is published, so should be the legal documents.

CeFi

Finally, we arrive at CeFi, for Centralized Finance. Those are the FTX, BlockFi and Genesis of the world. They attract a lot of attention and negative press around crypto but are quite orthogonal to DeFi. They live in the shadows and provide no transparency whatsoever. Neither are they composable. If they blow up, they can’t damage DeFi as they have no link to it.

Terra/Luca would also be listed there. If you just reduce Terra/Luna to the on-chain activities, then it would be a DaoFi, and it was for all to see the shortcomings (as noted in this tweet, Terra was clearly insolvent by design). But part of the Terra narrative was also reserves (which weren’t enough anyway) hold at the Luna Foundation Guard. And they obviously disappeared without a clear trail (there was an audit it seems but it is no longer online so who knows).

One should still mention that Coinbase is publicly listed which makes it somewhat transparent. They also issued cbETH which would be in the BridgeFi category.

Conclusion

As we have seen, the crypto universe is filled with projects that are on different point in the “decentralization” spectrum (if that means something at all). It is unclear what will be the state of the industry in 5 years but here are some key learnings I made along the way:

• Crypto-currencies may or may not be more money-like in the future. That doesn’t have any meaningful impact on the rest of the ecosystem.
• The jury is still open on the extend in which PureFi can grow. It is the gold standard of DeFi and we should devote more resources on this area. It doesn’t mean that PureFi should be against tokenized off-chain assets, they are a lower level building bloc, real public goods.
• BridgeFi, bringing off-chain assets in DeFi should be significantly improved in term of transparency (audited financial audit) and all legal docs tied in the off-chain world should be open-source.
• DAOs should also improve minority tokenholders and creditor protections. The key shouldn’t to recreate organizational framework but to migrate what already exist in a smart-contract way.

We are early, but the next steps will be significant to define where we will land. I’m sympathetic to those claiming that anything beyond PureFi is evil. Personally, I want to minimize the amount of RegFi and CeFi. Finance should be let to walled garden only for the big financial institutions like it is today. We need to level the playing field and allow everyone to participate. A good mix of PureFi (maximized), BridgeFi (with transparence) and DaoFi (for complex protocols) is needed to create a sound Crypto-Banking System.

And I have no doubt that ETH will become international money in that case.

Following last week discussion on credit vs currency, this article dig into currency boards and what we could apply to stablecoins. This will also lead to an improved value proposition for MakerDAO and its stablecoin DAI.

Stablecoins and currency board canonical link

Currency boards

A currency board is a system where a local currency keep a fixed exchange rate with a foreign currency. A great starting point is the Currency Boards Phd thesis of Kurt Schuller. Currency Boards allow the creation a currency by being pegged to a foreign currency like the US$. In it’s simplest form, it could be seen as a government money market fund (GMMF) which shares would be a currency. A more crypto way to represent it is fiat-backed stablecoins as USDC. It has foreign safe assets and you issue the local currency. Conversion between the two (US$ and USDC) is possible at the board (Circle) or a branch (Coinbase).

Having their assets mainly invested in the foreign currency credit instrument, they take a risk on the future of the foreign currency.

A currency board is generating revenue by having a small maturity mismatch. The local currency is redeemable for the foreign one on demand, but part of the reserve is used in foreign debt instrument that are producing interest revenues. Those are called seigniorage. There is also sometimes seignorage for minting and redeeming operations.

The purpose of the currency board system is not to maximize seigniorage, but to capture seigniorage subject to the condition that local notes and coins always be convertible into the reserve currency
Currency Boards — Kurt Schuler — 1992

Most British colonial boards adopted a two buckets balance sheet. 30 to 50% where defined as a liquidity reserve and were hold in species and high-quality securities (British government bonds) with maturities shorter than 1 year. This is somewhat close to the idea of High Quality Liquid Asset in the Basel Framework or what GMMF or fiat-backed stablecoin are holding. It is interesting to note that local currencies was usually non yielding (just like USDC currently), therefore the currency boards was already generate a profit on this part of the balance sheet.

The rest of the balance sheet was called investment reserve, and was devoted to invest in longer term securities and possibly of lower quality. It is supposed that an “hard core” of local currency in circulation will never be redeemed and that not investing corresponding assets would be a economic waste.

In order to remain solvent, a piece of equity (i.e. more assets than issued currency) was needed, around 5 to 10%. This piece was called the “surplus account” not far from the similar Surplus Buffer of MakerDAO.

To ensure that capital losses on securities would not reduce reserves below 100 percent, many boards also held a “surplus account” of 5 or 10 percent
Currency Boards — Kurt Schuler — 1992

Quite interestingly, there was some rules on how to manage the surplus account.

Many boards had a small margin for discretion with the surplus account: within the range of 100 to 105 or 110 percent reserves, they could either retain the seigniorage accumulated in the previous year or pass it on to the government. Below 100 percent they had to retain seigniorage and above 105 or 110 percent they had to pass it on.
Currency Boards — Kurt Schuler — 1992

This is again not very different from MakerDAO where all profits are going to fill the Surplus Buffer up to a threshold. From there, profits are given back to Maker Governance by buying and burning MKR token in the market.

The split between liquidity and investment reserves was also sometime rule based. For instance, Mauritius board was targeting 30–50% liquidity reserves and was supposed to refill (by selling securities) to 50% when the liquidity part was going below 30%.

The Currency Boards were more diverse that what exposed here. Some where more “pure” and hold all their asset at the Foreign Central Bank (like Faroe Islands), therefore removing the need of the surplus account. Some invested part of their reserve in lending to local government.

One interesting feature of currency board is the ability to change the peg. The Mauritius Currency Board was pegged to the Indian rupee but later changed to the sterling (at a rate of 15 Mauritius rupees = £1 sterling). Other switched from £ sterling to US$. Some used gold as backing. The British Honduras held both gold, £ sterling and US$.

Most of the Currency Board ended up becoming Central Bank in order to provide monetary sovereignty and economic flexibility. Politics (looking more independent from the former colonial power) was a big factor as well.

Comparison with a Central bank

The main rule that is broken with a central bank is the ability to hold local currency assets which, usually, end up in lending to the Government. This lending is quite a risk and you end up stopping convertibility to the higher level money (i.e. say gold during the Gold Standard) and becoming a fiat currency.

The figure below show the balance sheet representation of Currency Boards and Central Banks. The key differentiator is not riskiness of assets but the recursivity of being able to transform local currency debt, especially government debt, into local currency money.

In Perry Mehrling taxonomy, Central Bank, under the influence of the government, tends to favor too much elasticity in the local currency. The UK recent drama, wanting to bring discipline but switching to more elasticity at the first issue, is a prime exemple. Currency Boards lean toward more discipline, even if it is probably too much under the Pure Currency Board displayed above.

The Modern Monetary Theory says that you can be elastic until you get inflation and then you need to raise taxes and reduce spending (in order to repay or at least stabilize debt issuance). As inflation is very high, we will see how that will play out.

This makes me quite bearish on fiat currencies like the US$. While the dollars is rising (against other currencies, but against gold as well), I believe this strength is more due to too much $-denominated debt to be repaid than a rational decision to hold the asset. We already seen a similar pattern with DAI being valued above $1 when debt was contracting and it wasn’t possible to create DAI without borrowing.

In the Pozsar vs Mehrling debate, I feel more on the side of Pozsar seeing the prelude to a birth of a new world monetary order. Nevertheless, being more digital oriented, I would lean more towards cryptocurrencies than traditional communities.

With those macro-economics concerns and what I’ve learned studying those Currency Boards, my view of the value proposition of MakerDAO changed somewhat.

A path for MakerDAO?

I tend to agree on Ohno view that a monetary system should seek short-term stability and long-term flexibility against US$. The US$ is the main currency for now, therefore having a 1:1 strong peg is condition to be relevant. Following the article discussing credit vs currency for stablecoins, this would make DAI being a currency, but behaving like credit.

Yet, we need to keep the long-term flexibility and provide an added value to DAI. When (and if) the US$ fails we should have the flexibility to switch to another unit of account system to safeguard DAI value as a store of value. Transition will years or even decades but so far my best bet would be Ethereum (Bitcoin as close second choice). Ethereum (ETH) is currently too speculative and not used in international trade. Yet, within a decade the situation might change.

Therefore, I would suggest the following structure for MakerDAO balance sheet (which is mostly a representation change from the ALM forecast I published this summer).

Following the Currency Board model, most of the DAI backing would be in US$ assets. A position in ETH would be started. As defined in the ALM model, this would consume a lot of capital but could constitute the kernel of the foreign currency switch. Another key, more innovative, element would be the holding of DAI assets, i.e. crypto-backed loans. This element is important as a local currency asset doesn’t create a foreign exchange risk when you change your referential. Rules must be in place to avoid the self referential issue of central banks (having only crypto-backed loans would probably remove the problem). The Ethereum network state doesn’t have a government to start with.

At this stage, it is just an idea. There are probably many second order effects to consider. Creating a currency instead of a credit instrument is also not easy, you need to create a currency ecosystem.

Nevertheless, it would, I hope, retains most of the benefits of having DAI as credit with the additional value of safeguarding DAI store of value capacity in case the US$ start to fall.

In this article, we will analyze the difference between money-like credit and a currency. Then we will try to understand how such a dichotomy related to the stablecoin construct.

Canonical version of Stablecoin: currency of credit?

Money-like credit versus currency

In a poll on MakerDAO forum, I’ve asked the community if they see DAI as credit or as a currency. Results where unexpecting quite uniformly distributed.

First, let’s define what we mean by currency and credit.

A currency is defined here as a system of money. It is commonly defined as store of value, medium of exchange and unit of account. A more detailed presentation is given in The 7th Property: Bitcoin and the Monetary Revolution.

In our case, the most import point of a currency is the unit of account. Indeed, we are using a more broader definition of currency as a system of money. You can also read my article How Money Works for a simple explanation and Crypto-Banking 101 where I details more in depth hierarchies of money.

For the scope of this article, I will refer to the following diagram representing the US Dollar currency / system of money. At the top there is the Fed Note (and Fed reserves not displayed for simplicity). This is the monetary base, or layer 1, of this system. It is the most pristine US Dollar. But the money supply (everything that we might consider US Dollar cash or cash equivalent) is broader and encompass banks deposits, Treasury Bills and stablecoins like USDC. Those are not cash but credit instruments with high moneyness (claims on higher level money).

Therefore, the original question is to know if DAI is part of the US Dollar currency tree (like USDC, at the inside money level), or if DAI (the token) is the monetary base of the Dai currency tree (at the outside money level, i.e. first layer).

The case of DAI

While the “contractual engagement” (if any) is only to allow redeem of DAI for $1 at emergency shutdown, those last years, MakerDAO always provided ways to go from/to DAI to/from USDC. In a way, it is commercially expected.

The MakerDAO Whitepaper assume DAI as a unit of account. Nevertheless, the paper recognize that it is not widely used beyond the protocol itself. It feels akin saying the Citigroup IT system express everything in Citi $.

Dai is not used as a standard measurement of value in the off-chain world, it functions as a unit of account within the Maker Protocol and some blockchain dapps

Currently the stablecoin ecosystem is mainly driven by $-linked stablecoins. While there is more than $100 billions of $-linked stablecoins, the next represented currency, the Euro, is far below 1 billions €. The crypto-banking system is a US dollar world. Some experiments are trying to create a stable currency using a new unit of account, like RAI, are limited to niche markets. Even speculative crypto-currencies (BTC, ETH) have less daily volume than the $-stablecoins.

We are in a world of intense competition and, to thrive, a new currency should provide a new value. Fiat currencies around the world have the avantage of being legal tender in their country. The US dollars is currently the unit of account of the international trade. That might change in the future for the benefit of another national currency or a return to hard money in the form of a commodity standard (e.g. gold) or based on a crypto-currency (BTC, ETH, …).

Therefore, my view is that DAI is clearly in the credit side in the US Dollar system. Nevertheless, DAI could keep the optionality to migrate to a new unit of account system (or a basket, like the SDR).Maybe to ETH when it will be recognized as the international unit of account?

Liquidity as the systemic risk and the death of the dealers

One issue during the Global Financial Crisis was the refinancing of long term debt instruments with short term funding. This is shadow banking. In case of stress, there is a liquidity crunch and the whole system stall. One option, proposed in The Money Problem, is to forbid non-regulated entities from issuing short-term, money-like, liabilities.

The Crypto-Banking System is no stranger to issuing short-term liabilities, from stablecoins to aggressive use of repurchase agreements (repo, i.e. secured short-term lending). If we zoom in on the full-reserve crypto banking part, there is a maturity transformation by funding longer maturity assets (e.g. bonds), by the mean of open repurchase agreements. We identified clearly that a key assumption was the “existence of a liquid market for the bond”.

Such a liquidity is needed to enable orderly liquidations. If $1B of bonds are repo with a 10% haircut (the overcollateralization), the system stays safe, if, and only if, those bonds could be sell in a short period without moving the price by more than 10%. Otherwise, there will be a liquidation loss.

Concerns about the liquidity of markets for the bonds used as collateral led to increases in repo “haircuts”: the amount of collateral required for any given transaction. With declining asset values and increasing haircuts, the U.S. banking system was effectively insolvent for the first time since the Great Depression.
Gary Gortno and Andrew Metrick, Securitized Banking and the Run on Repo, 2009

Liquidity is a key driver for asset prices. The existence of abundant liquidity is, therefore, key for orderly markets.

Not only does the demand for liquidity move prices, but the breakdown of liquidity is one of the primary drivers of crashes, as well.
Richard Bookstaber, The End of Theory: Financial Crises, the Failure of Economics, and the Sweep of Human Interaction, 2017

In current markets, the liquidity is provided by dealers that are making markets. They leverage their balance sheet and provide a bid/ask spread so sellers don’t have to wait for a buyer and the opposite. Yet, while it works well in normal markets, it tends to fail in stressed market. Dealer reduced their risk in the 2013 sell-off and the reduction of dealer balance sheet capacity accounted for 25% of the sell-off during the Covid crash.

Other kind of market participants should obviously be incentivized to buy and asset when there is a price discrepancy due to market stress. In his article Asset Price Dynamics with Slow-Moving Capital, Darrell Duffie call those market participants “inattentive investors”. They tend to come too late to the party (after a liquidity crunch spiral is already in motion).

What if those inattentives investor could be active and get rewarded for it?

Crypto-banking and the abondance of liquidity

Significant research advocate for diversified and automatically rebalanced portfolios, like the Dragon Portfolio from Artemis or the All-Weather Portfolio. At the same time, we live in a rise of passive investing macro trend. Playing the market is no longer seen as better than just holding the market.

As displayed in the chart below, holding a 50% USDC / 50% ETH portfolio in a Liquidity Pool (LP) would have been granted a 50% overperformance versus balancing every hours (not taking in account cost). The providing of liquidity in Uniswap is akin to rebalancing by selling options (and collecting the premium) instead of rebalancing by market orders.

Moreover, performance and efficiency can again be improved by depositing the underlying assets of the Liqudiity Pool to a money market like Aave and Compound. While this was started with iearn Curve pool, it proved too gas intensive. The Balancer Boosted Aave USD is a more recent iteration on the concept (keeping some assets uninvested for easier trading).

It is still early days on research in this area. For one, we are still remote from the days where a portfolio allocation can be done fully on-chain. While we can expect stocks and bonds to come soon on the blockchain (using Backed Finance for instance), so far the investment world is limited to natives crypto-assets, gold and stablecoins. Not exactly the basis of a sound allocation.

There is also a need to understand, in a theoretical way, what drives the performance of a liquidity pool. For pure monotonic assets (assets going only up or down), in the absence of organic trades, such a strategy would result in a loss. Volatility and some mean reverting around a trend is much needed. Having a way to infer to profitability (if any) of an LP from the volatility and correlation of the underlying assets would be a great contribution.

Nevertheless, the fact that most assets would be provided as deep liquidity would result as a stabilization mechanism for the financial system. In Three Principles for Market-based Credit Regulation, Perry Merhling insist on the need of the Central Bank to be a dealer of last resort when commercial dealers are stopping doing their job. In this article, we show that passive investors could play this role and have an incentive to do so. This is made possible by the composability of Decentralized Finance.

Link to the canonical version of The Crypto Banking System

The chart below, based on the work of Zoltan for the Shadow Banking System, provide an high level overview of the Crypto Banking System. It is also an extension to the work done in Crypto Banking 101 and Crypto-banking vs shadow banking.

On the right, we have ultimate lenders (households with excess savings, financial institutions, sovereign wealth funds, …), agents with excess cash that could be used (for a cost, the interest rate) to finance ultimate borrowers (households with a mortgage, small and medium enterprises, big corporations, governments) on the left.

As you can see in the chart, while native crypto-assets (think ETH, BTC, …) are in the legend, they appear nowhere. It is by no means complete and further work will analyze the place of crypto-assets. You can also consider Cash being a crypto-currency and, assuming real-world assets are using the same unit of account, it would work the same.

As we will see, the keys components of the Crypto Banking System are the followings:

• creation of deeply liquid tokenized bonds representing real-world credit (private credit and public markets);
• core infrastructure of a market-based economy with exchanges and repo markets, the crypto market;
• intermediation of saving and lending with crypto-banks that operate in maturity transformation.

The need for new primitives

DeFi being build on a trustless basis, most lending is against a collateral. The key component of a good collateral is deep liquidity. One should be able to liquidate in size and quickly without affecting the price too much.

The main collaterals used in DeFi currently are ETH and WBTC. Both are highly volatile (hence asking for a high haircut), quite limited in supply, and are speculative. They might well be the collateral of the future but, currently, they are not convenient.

Therefore, we need to introduce a new type of collateral. Gold could be an example, it is already tokenized (PAXG) but it didn’t get much traction.

If we look at TradFi, the collateral of choice migrated from mercantile bills to government bonds.

Public credit tokenization

The first area to source a liquid collateral is to onboard public markets on-chain. Corporates are issuing bond that are traded and rated by credit rating agencies. Governments issues bonds as well that are highly liquid and rated. Contrary to crypto-assets, those asset are more stable meaning more efficiency for collateral usage. In TradFi those instrument are considered safe and liquid enough to be included in High Quality Liquid Assets.

Public market tokenization would simply have some public securities on the asset side and issue token (with probably 1:1 relationship). To achieve greater liquidity securities of the same kind can be pooled or ETFs can be onboarded directly (until the underlying are on-chain and can be pooled on-chain).

One example is Backed Finance that is working with MakerDAO to tokenize the iShares $ Treasury Bond 1–3yr UCITS ETF.

Private credit tokenization

If we stay at the public credit stage, there would be plenty to use as collateral for a few years, but this isn’t creating a better system as it would exclude households and small and medium enterprises (SME) from funding.

The solution lies in securitization in which an intermediary pool those illiquid assets (loans are not tradable) and issue two class of tokens, a senior piece and a junior piece. The credit enhancement of the junior tranche provide safety and easier price discovery for the senior tranche as well as force the intermediary to have “skin in the game”. Ideally this senior piece would have to be rated by credit rating agencies.

The pooling should also be big enough and transparent enough to encourage a strong liquid market for the senior part.

Two examples of such tokenization are New Silver, which is in the fix-and-flip loan business (technically not for households yet) and FortunaFi which is in the revenue-based finance space aggregating loans from many asset originators.

Scale and liquidity trumps everything

One key point, both on the public part and the private one, is to target such a scale that a liquid market forms. Fragmentation of liquidity is to be avoided.

As a first data point, during March 2020, bonds ETF have shown to be way more liquid that their underlying as showed in the chart below. Some research also suggest that such pooling reduce the impact of fire sales.

One issue during the GFC is that illiquid securitization was used as collateral, sometimes without a haircut. When the crisis got serious, higher haircut were applied or such collateral just excluded from repo markets.

It is therefore of paramount importance to have a few of very liquid and very transparent instrument to support the crypto-banking system instead of an endless list of less liquid, harder to understand instruments.

Not displayed in the Crypto-Banking System are the bonds issued by on-chain protocols and DAOs, either secured by crypto-collateral or not (see the Capital Structure for DAOs article). Pooling those will indeed allow the creation of another liquid and transparent bond instrument.

With those deep and liquid primitives, we would be able to construct a sounds crypto-market.

The crypto-market

At the heart of the Crypto-Banking System, lies the crypto-market which provide deeps pools of liquidity (both for trading and short-term funding).

The crypto-market itself is composed by two sub-markets, one for trading (decentralized exchanges or Dex), and one for short-term funding (money market). Both of those markets should be governance-minimized and to the best extend possible immutable contracts. It should be trustless, permissionless and non-custodial.

There are three kinds of operators of the crypto-market: the value investors, the speculators and the arbitragers.

The whole stability of the system is provided by value investors. Those can be individual, DeFi institution (like cryptobank or insurance protocol) or TradFi investors. For simplicity, let’s assume they allocate their assets to a predefined asset allocation (e.g. 50% Cash, 50% Bonds). They will deposit some of those assets into a decentralized exchange to passively let the market arbitrage their position to keep the exposure constant. They will also get trading fees on top (contrary to TradFi where keeping a constant allocation is a service for a fee). They would also be able to deposit their unused liquidity and some bonds into money markets to let other actors borrow them for a fee (again improving value investors returns while TradFi custody is at a cost).

Arbitragers (hedged dealer books) are here to provide efficient markets by providing more liquidity for small spread. If the price for a investment grade bond move significantly, they can borrow some in the money market swap it for some treasury bond of similar maturity (to hedge their interest risk) and some higher credit risk bonds (to hedge their credit risk). Those bonds will be used as collateral on the money market to finance their borrowing of bonds. Those arbitragers will also be able to provide liquidity on options protocols and provide an orderly market for on-chain ETF (by arbitraging the ETF price with the underlying).

Finally, the speculators will be actors willing to take more risk (non-hedged) than arbitragers when a speculative profit opportunity arise by being long one asset and short another. An example is to be long treasury bonds and refinancing those position on the money market to get leverage when an speculator think the yield curve is too steep (or the reverse just as well) . By speculating, they provide a price discovery mechanism and try to achieve efficient markets.

By centralizing (pun intended) the liquidity to those crypto-markets instead of sitting idly in wallets, the Crypto-Banking System can achieve more fluidity, at a lower cost and lower complexity, than traditional markets.

The Bond primitive was already designed to be highly liquid by being an aggregation of lower liquidity instruments (i.e. corporate bonds and mortgages), the crypto markets are providing a way to use most of the market cap of those primitives as trading liquidity or funding facilitators (and nothing prevents Dex to use Money Markets in the background for their liquidity).

To add even more efficiency, we need to add a new player, the cryptobanks.

Towards fractional-reserve crypto-banks

As defined currently, the Crypto-Banking System is composed of Bonds instruments and Cash instruments. We have seen that Bonds are composed on tokenized pooled private credit and public markets instruments. But what is Cash?

Defining Cash in the Crypto Banking System

In a simplified version, fiat-backed stablecoins are simply $1 on-demand liabilities backed by Treasury bills (or bank deposits). Such a backing allows stablecoins to be redeemed at any time (on-demand) without any liquidity issue. Currently the spread between those two (which is the t-bill rate) goes fully to the stablecoin issuer (or distributors), nothing goes to the stablecoin holder. That is probably going to change but, by construction, the interest rate given to stablecoin holders will be limited by the t-bill rate.

At scale such a system might be adverse for credit intermediation. Indeed, if stablecoins are the new bank deposits, it will reduce the scale of the latter and diminish credit creation by traditional banks. This is where cryptobanks are coming into play.

Solving the liquidity preference and expanding the money supply

There is a fundamental mismatch between ultimate borrowers that need to borrow long term and ultimate lender who have a liquidity preference. As you can see below, most ultimate lenders are holding cash, possibly large amount of it. While it is unlikely to get a shortage of state deficits, it would not allow to finance economic activities enough and cause economics limitations.

To solve this problem, fractional-reserve banking was introduced as a way to extend the supply of cash based on long-term lending (see my article How Money Works).

As depicted in the chart below, a small amount of fiat-backed stablecoin can be expanded to a larger amount with fractional-reserve banking thanks to an intermediary, the cryptobank. The cryptobank issue a stablecoin considered money-like by creditor as convertibility to a fiat-backed one (the reserve) is possible. But, and that’s the catch, there is not enough reserves for redeem all stablecoins at one. History shows that it leads to bank run when trust is reduced but that the banking system can operate for many years without runs on the banking system, even in recessions. You can also check the article Crypto-banking 101 for more details.

Crypto-banking is neither traditional banking (bank holding illiquid loans), nor market-based banking (no maturity transformation there), nor shadow banking (illusion of market-based maturity transformation). Crypto-banks hold highly liquid assets as a mean of defense against bank runs.

Conclusion

As we have seen, creating a robust crypto financial system requires three components: tokenized real-world credit (Bonds), robust crypto markets, trading and lending (to provide deep liquidity and price discovery) and crypto-banks to provide maturity transformation.

At the time of writing, crypto markets are still imperfect but working (Uniswap for lending and Aave as a money market). Yet, we are critically missing the Bonds part. Rails are there, but they are mainly used for speculation. Crypto-banks like MakerDAO are already integrating with money markets with tools like the D3M. MakerDAO is also helping creating the Bonds primitives both on the private credit side with Centrifuge than of the public market side with Backed and the proposal to onboard up to 1 billion $ of short-term bonds.

We have never been so close to get a robust Crypto Banking System.

Link to canonical version

Banking is the alchemy of bridging ultimate lenders that want safety and liquidity to ultimate borrowers that want long dated maturities and are making risky bets. It is both an art as a science as achieving this goal is impossible and bring fragility. Yet, it is needed to create grow for our economies. Using learnings from the real-world, we will think on this can inform us for crypto-banking.

Real-world banking

Traditional banking is a model used for centuries. It is one intermediary that provides credit, maturity and liquidity transformation. The lender gets an on-demand saving deposit while the borrower get long term loan.

The mismatch between the two sides is bridged by adding some forms of credit enhancements (the equity of the bank) and a liquidity buffer (reserves). On top of that, we can have a central bank providing liquidity of last resort and a government insurance to provide trust.

Due in part to the regulation on bank, a chunk of the lending/borrowing mismatch management was transferred to the so called Shadow Banking. The article Shadow Banking Economics and Policy is an excellent starting point. For this paper, suffice to say that shadow banking is splitting the transformations in subparts. Loans are bundled and sliced to provide many kind of credit risk profiles (ABS for Asset Backed Securities). Those are refinanced by short-term commercial paper (ABCP — Asset Backed Commercial Paper) with a bank providing a liquidity put as guarantee. Those ABCP are then pooled in Money Market Funds (MMF) as safe short-term investments. The MMF shares are considered as money-like instruments.

The Global Financial Crisis showed the fragility of Shadow Banking as complexity, mark-to-market valuations and lack of liquidity broke the system.

Enter crypto-banking

Crypto-banking is still young and it unclear how it will evolve. In DeFi (Decentralized Finance), the prime importance is given tokens (securitization), composability and liquidity. One can make the case that DeFi is inherently suited for market-based finance.

In such setup, a pool of private credits (loans) are pooled together to create securities/tokens. There can be many tranches for different risk-profiles. This is quite similar to the Shadow Banking system. The pooling and tokenization provide a market for such assets and help with price discovery. Being a tokenized assets with a price feed and liquid markets, it becomes a collateral that can be used for repo (composability). Therefore, one main difference with Shadow Banking is the size of such securitization. It should be big enough to have strong liquidity (i.e. billions at least).

Initially, crypto-banks where involved in crypto-backed loans. Those are still private and illiquid credit even if quite secured and short-term. Nevertheless, evolution of the crypto-bank is towards market-based activities where assets are sourced and put on the balance sheet. Same as traditional banking, there is a credit enhancement (equity) and some liquidity buffer. This is the model I analyze in depth in Crypto-banking 101. DAI, FEI, FRAX, agEUR are based on such model.

Another view of crypto-banking is a more shadow banking approach. In such view, stablecoin are fully-reserved, i.e. backed only by cash or cash equivalent (like USDC, BUSD, …). To make the connection with the borrowers, stablecoin holders should leverage other protocols as showed below.

The tokenized is the same as already discussed. Then comes the maturity transformation. This would be done by using the existence of a liquid market for the bond to be able to repo it on a money market (e.g. Aave, Euler, …) displayed on the right side (will come back to it soon). The bond maturity could be years while the repo is open term. As there is an haircut on the repo market, there is a need for equity which is adding a credit enhancement as well.

The money market, on the right, provides some liquidity transformation by keeping a buffer of liquidity and adjusting in real-time the interest rate on repo.

There is no liquidity put so the expected scenario in case of stress is interest rate going through the roof on the money market, value of the collateral being depressed and liquidations occurring.

If the market of such a bond is no longer liquid it will end up wiping the middle intermediary and accruing losses to the ultimate lender. Notice that such a loss is due to market illiquidity and doesn’t mean the loans are defaulting in the end.

This last item makes me feel fractional-reserve crypto-banking is a better alternative. It is still unclear how to address systemic issues like a generalized bank run. Using mark-to-market valuation in an illiquid market could trigger a crypto-bank insolvency (at least from accounting) and therefore a bank run.

Yet, at the same time, a fully market-based would provide a significant amount of collateral and required transparency. Maybe part of the answer is in the Option Clause like the one used by Royal Bank of Scotland during Free Banking Era?

Moreover, while we can self-regulate to be on the safe side in term of allocation (equity buffer, liquidity buffer), it is impossible to ensure liquidity in a market.

For those reasons, I believe fractional-reserve crypt-banking is the future. We take the best from traditional banking and shadow banking.

Link to canonical version

Decentralized finance promised a Cambrian explosion of nature of tokens. As we will see, we are still far behind TradFi. In this paper, the different source of on-chain funding will be presented with an analysis of pro and cons.

When researching for the balance sheet view of stablecoin protocols, I produced the following illustration showing the balance sheet of stablecoin protocol and some real world banks like Bank of America and the Fed. We will focus on the liability side.

Key takeaways:

• Stablecoins are a great source of funding for DAOs able to create stablecoin utility.
• Secured bond/repo should see boarder adoption going forward as it is a flexible a cheap source of funding. It remains to be seen what the main collateral will be for this system (ETH, stETH, tokenized Treasuries?).
• Securing the lender is still work in progress. I believe we will see legal contracts and lender representatives having an impact on DAOs governance.

Equity: the governance token

The most common source of funding is issuing a governance token and allowing people to buy it with a money-like instrument (stablecoin, ETH or more complex tokens).

An first exemple is the issuance of TRIBE at genesis. Most DAOs are raising capital this way with such IDO (Initial Dex Offering) or by selling on the private markets to VCs.

Another exemple is backstop mechanisms that protocol like Maker offers. When the system detect too much bad debt, it will automatically issue some MKR token (governance token) for cash (DAI, the MakerDAO stablecoin, therefore not getting funding per se but reducing debt).

A last exemple is how stablecoins like FRAX and UST are constantly issuing (or buying back, but not the subject for today) governance token when there is a stablecoin minting. Yet, as seen in the case of UST downfall, you can easily be mistaken in your ability to raise capital by this mean. As in the inelastic markets hypothesis, flows are big drivers of price. Having a $10B market cap, doesn’t mean you can raise up to $10B, it’s at least an order of magnitude less.

Stablecoin issuance

A stablecoin is a on-demandish (real or expected) liability that provide money-like benefits. Customer finding utility value in this liability, the cost is usually lower than another kind of debt.

This liability is on demand, but due to the actual or perceived utility of the stablecoin, the behavioral maturity is split between short and long term.

The key for stablecoin is to boostrap an organic demand (people using and holding the stablecoin as a liquidity preference).

While other kind of debts have mainly a cost that scale with the amount (the interest rate), stablecoin demand upfront fixed costs (marketing, integration, …).

As described in Crypto-Banking 101, one key component to drive stablecoin demand will be providing an interest rate on the stablecoin. It will still be lower than using straight debt as there is a utility yield for the customer/lender.

debt interest rate = risk-free rate + credit risk of the issuer

stablecoin interest rate = risk-free rate + credit risk of the issuer - utility yield

Therefore, while difficult to get running and more difficult to manage, having a stablecoin provide great benefits at scale.

Repo markets

If the protocol own some valuable assets (like ETH), those can be pledged at lending protocols like Aave, Fuse and MakerDAO as a source of funding. This funding is described as a repurchase agreement (repo in short) but it is actually a secured loan (“a repo is a secured loan disguised as a sale”). We are considering here a on-demand repo where the repo last as long as no party want to close it (it’s a bilateral repo in the case of MakerDAO but an aggregated tri-party repo for Aave). When the lender wants to close the repo, he will increase interest rate (or the minimum collateral ratio) to a penalty rate, incentivizing the borrower to repay the loan to get his collateral back (or ending up liquidated).

Repo market can rehypoticate the collateral or not. For instance, if you borrow stablecoins on Aave with ETH, part of your ETH will be lent to other borrowers and secured by stETH. Using MakerDAO, the collateral is just escrowed and not used. The latter is therefore more secured but more expensive as well (as the collateral on Aave is providing some yield).

In both cases, there is a trust assumption on the repo market protocol. Governance usually has access to the collateral in one way or another (updating the contracts, messing with oracles, …). While expected to be safe, repo market have seen issues with oracles or liquidations.

Another risk with repo funding is the volatility of the interest rate which can fluctuate widely.

We can expect repo market rate to end up at SOFR + xxbps (SOFR being the repo rate of a Treasury bond). For new protocols or in period of stress, the repo market might provide the lowest cost source of funding. But, obviously, it demand to hold collateral to be pledged.

Secured bonds / term repo

The repo market described previously is open repo. While usually liquid, it can still dry up quickly (and/or get very expensive). For financing longer dated asset, it make sense to increase the maturity of the liabilities as well. A secured bonds (or term repo, but we will focus on a tokenized and fungible version here) is a zero-coupon bond with a maturity date that keep some collateral in escrow. The collateral is sold at maturity or if the collateral ratio falls too low. Therefore, the lender don’t have to trust the borrower.

Example of such bond protocols are Yield and Notional. Currently maturities are going up to 1 year.

Just like open repo, being secured, the interest rate doesn’t take in account the credit risk of the borrower but just a small spread for the secured bond protocol risk. It also demand collateral that not all DAO would have. Borrowing at a fixed interest rate is usually more expensive but provide visibility.

The secured bond market is quite young, but should be quite promising as a source of funding.

Corporate bonds

Corporates bonds is a source of funding where the lender doesn’t get a strong security. While the stablecoin is the on-demand debt of the DAO, corporate bonds are all debt instruments.

Governance token collateral-backed

They can be seen as secured by the governance token, but that is only correct when 2 conditions are met:

• The protocol can’t mint it’s governance token
• A collateral liquidation would be orderly (limiting scale of borrowing).

Those conditions are usually not met and therefore they shouldn’t count as a secured bond. If the protocol can mint the governance token, it can always find enough collateral to avoid being liquidated as long as the governance token price is not zero.

As we have seen with UST/LUNA debacle, relying too much on the governance token as collateral doesn’t work as liquidation demand a buy demand for the governance token. Such a demand doesn’t exist at scale when things are bad.

While stablecoin holders can redeem quickly and front-run a malicious DAO decision (like stealing the collateral), corporate bond holders are locked. The solution of using a timelock is not enough to bring safety on this front.

Example of in this direction is Porter (bonds for DAOs). Yearn, Polygon and Mana have (or had) loans taken on MakerDAO.

Revenue-based finance

Revenue-based finance is a kind of debt where the repayment is taken directly from revenues. This sounds like a good fit for an on-chain protocol as such a mechanism can be implemented directly on-chain. In a way, Uniswap Liquidity Position are a kind of revenue-based instrument (and smart contract being immutable, it is very safe).

Angle protocol also provide such instrument with it’s SLP token. While in such a case it is a on-demand liability, actual redeemability depend on some conditions (contingent to protocol safety levels).

Overall, the corporate bond space is very early and complex to implement in a trustless way. I would argue that, currently, such corporate bonds are more based on trust than legal obligations. DAOs usually have a lot of control on their protocol and can change the conditions as they want (directly or indirectly). Either the space will move towards immutable smart contracts (which is risky and not very flexible) or will see addition of legal contracts. On the later, it could the addition of a trustee (kind of a security agent) representing the bond holders and having a veto rights on the decision of the DAO if it breaches some covenants defined in the bond issuance. It remains to be seen if a respectable corporate trustee would be able to take such a role (and the corresponding liabilities).

Intra-stablecoin issuers money markets

This funding is more detailed in the ClearingDAO paper. It is really just a way to solve the supply/demand temporary imbalance between stablecoin issuers (some stablecoin issuers having an excess of funding from their stablecoin while other having a funding deficit).

Link to canonical version

We will model a simple economy at the aggregate level. In our simple system, there is only 2 aggregate of money:

• the monetary base which includes central bank notes and reserves. This is the top of the US Dollar hierarchy (level 1) from my Crypto Banking 101 article.
• the money supply represents what the non-bank actors can hold and consider as money (some level 1 plus the other levels). In this system, we will define it as central banks notes and bank deposits.

Our little economy

The diagram below shows an example of such a financial system.

The central bank hold treasuries (issued by the government and not represented here) and issue notes (including coins) and reserves. Our central bank will have no equity part. We also assume that the central bank can convert reserves to notes on demand.

Banks are holding some reserves to provide liquidity but are mainly extending loans to businesses. Banks are financed by equity (contributed by households) and bank deposits.

Households and businesses are holding the bank deposits and the notes. Households have financial investments in the form of treasuries and equity while businesses have productive assets. Businesses are financed mainly by banks loans. We use the term net worth instead of equity for households.

As you can see in the left part, the money supply is more important than the monetary base. This is called the money multiplier. Regulation can be used to limit or influence this money multiplier, for instance by making a legal reserve requirement (forcing the banks to hold a percentage of reserves).

Expansion of the money supply

In the diagram below, the households are selling their treasuries for bank deposits. As we can see the money supply increased while the monetary base remained the same. This illustrates the view that banks are creating money by expanding their balance sheet.

But what happens if the central bank expands its balance sheet?

Money printer go Brrrr

The central bank is a bank and it can extend its balance sheet quite easily. Let’s assume the central bank is buying Treasuries from the banks to do that.

As we can see, the balance sheet of the central bank is expanded and that is creating new money in the monetary base. Yet, it doesn’t change anything for the money supply or the rest of the economy. If banks don’t do anything, nothing happens.

Sounds strange? Well, look no further than what happened with Quantitative Easing in 2008. The monetary base increased sharply (x2), but M1 (our money supply) didn’t change as much.

Conclusion

As we have seen, it is commercial banks that are creating money. Reality is obviously more complex as the government can be more active than in our short presentation. Nevertheless, a balance sheet view is always useful to understand what is really happening.

The direct conclusion is that if stablecoins are seen as money (usable for payments) AND the stablecoin issuers aren't backed only by money (money base or money supply) THEN there is monetary creation.

Bagehot, Walter. Lombard Street : a description of the money market.

Link to canonical version

One particular interest I have is to make stablecoins fungible. I’ve been advocating for solutions, got fees removed for DAI (published the story of the DAI peg), and improved the peg of FEI. This article makes the case of the importance of DeFi stablecoin fungibility and provides initial thoughts on how to implement a decentralized clearinghouse, that is called ClearingDAO.

Why it matters

At its core, $1 should equal $1 no matter its form. It’s the basics of the name pegged stablecoin. To be money-like it should be accepted at face value with no question asked. It’s something that existed even 18th century Scotland or in some parts of the US (the Suffolk System) a bit later. If some private actors were able to solve this problem centuries ago, how could we not today?

It is already a reality for CeFi as FTX already embraces fungibility between USDC, TUSD, GUSD, USDP, and BUSD. All those are just ways to send US Dollars (notice USDT is excluded, as well as all DeFi stablecoins).

In their recent RealVision video, Raoul Pal and Darius Sit from QCP Capital discuss how banks in Asia are looking to get fungible stablecoins.

Therefore, the trend is clear, stablecoin fungibility will be a key feature in CeFi. DeFi needs to catch up to stay relevant.

Moreover, stablecoin fungibility is also key to unlock some DeFi usage. It is of prime importance for Real-World Assets but also

Some thought on a decentralized clearinghouse

Where we are now

My previous thoughts were based on a hierarchy of money structure managed by bilateral or unilateral arrangements. While such design work, there are two pitfalls:

• It is not unified and creates a huge mess on the UX side. 1inch can route well through PSM from USDC to USDP, but doesn’t seem to work on the reverse. It doesn’t take into account the DAI-FEI PSM either (which changed recently).

• It is also quite an economic challenge and takes unneeded space on the balance sheet of stablecoin issuers. To keep liquidity, Fei is lending capital to MakerDAO at a 0% interest rate (i.e. holding DAI) and MakerDAO has no incentive to hold FEI liquidity. The core issue is the lack of interest rates.

The historical solution to this problem is to provide for clearing.

Having a protocol for DeFi stablecoins clearing would be a great advantage for DeFi stablecoins. Customers would be able to hold whatever stablecoin they prefer (from a risk/reward perspective, for a brand reason, for ESG concerns, …) and be confident to be able to use it at par. Such a ClearingDAO could have many shapes and properties but would solve temporary imbalances between stablecoins. We will consider that the interface would a simple swap interface.

To see how ClearingDAO would work, it would need to solve internal and external imbalances.

Internal imbalances

Internal imbalance occurs when one protocol is issuing more stablecoin than there is demand for it while another protocol is having too much demand and not enough supply. A good example is no other than MakerDAO which needs to rely on $5B of USDC to bridge the gap due to high demand (and comparative low supply). On the other end, Fei had to reduce asset exposure quickly during the last crash as their DAI reserves got emptied and the peg was at risk.

We will assume that each DeFi stablecoin can register itself to the ClearingDAO and provide its own stablecoins in significant volume (it doesn’t have to be limited to stablecoin issuers but it makes mainly sense for them). For instance, MakerDAO could deposit 100M DAI. This would not have any impact on DAI as at this stage no one can do anything. MakerDAO is the only actor able to deposit 100M and get nothing because the cost of capital of DAIs that are not in circulation is 0%.

Participants can also add other stablecoins than their own. For instance, MakerDAO could add 50M FEI that it owns. From there, we can allow people to swap DAI to FEI (up to 50M) and FEI to DAI (up to 100M).

This is not different from current PSMs. If we add another stablecoin issuer we can see how inefficient we become. In the representation below, MakerDAO owns 50M of FEI (i.e. extending a 50M free loan to Fei Protocol) and Fei Protocol is extending a 75M free to MakerDAO. We could be more capital efficient.

The intuition would be to net it with a 25M loan from Fei Protocol to MakerDAO. It could already be done trustlessly by MakerDAO getting back the FEI it owns from the contract, converting them to FEI, then adding the resulting DAI (or anyone with a flashloan).

So far we still haven’t addressed the motivation of entering in such a protocol and how to limit exposures in a multi-stablecoin system.

The solution is to provide credit facility arrangements between each actor. Fei Protocol could extend a credit facility to MakerDAO (holding of DAI) up to 100M at 0% and a credit facility to FRAX for 50M. A 0% credit facility would be unilateral as there is no reason to avoid it for the other party. This would solve the credit exposure issue.

We can also add bilateral agreements with a non 0% interest rate and/or a non 100% collateral ratio. MakerDAO might agree to extend a loan facility to Fei protocol at 1% with a 200% collateral ratio and a 100M limit. We could imagine plenty of variations in the loan structure, but the important part is that both parties should agree to it.

It is also important to note that the 200% collateral ratio is not a limiting factor. Stablecoin issuers can mint as much of their own stablecoin as long as it is only used as collateral and not put in circulation. It is only limited by the risk appetite of stablecoin holders.

ClearingDAO should perform swaps in a way that limits debt exposures. While I haven’t explored much, it is probably a difficult problem with more than two actors and heterogeneous bilateral agreements. Some simplification might be required.

The key point of the protocol would be the ability to perform adverse clearing. Any lender can stop an agreement at any time (most likely with a notice period) and perform a liquidation of the foreign stablecoin loans (though the sale of the collateral it owns directly or as collateral).

Now that internal imbalances are taken care of, let’s see how we can deal with external imbalances.

External imbalances

External imbalances occur when, on aggregate, the overall demand for DeFi stablecoin fluctuates faster than supply. CeFi stablecoins tend to be backed by real assets like Treasury Bills only. It is not possible to mint them without a cost of capital. Even beyond this issue, it is also unlikely to get them as ClearingDAO participants.

Therefore, the simpler solution is to allow people to trade from such CeFi stablecoins and allow each ClearingDAO participant exposure to some CeFi stablecoin (i.e. 0% loan to some external stablecoin issuers). This is how MakerDAO has already solved the problem with the USDC PSM. Nevertheless, the issue with a 0% loan is that no one is incentivized to repay it. As it is backed only at 100%, any clearing would be at a cost (as marginal as it is to compensate market makers). This is actually an incentive to redirect some CeFi stablecoin holdings to DeFi stablecoin holding even with a low-interest rate.

Sadly, it would not be possible to work the other way around, i.e. having CeFi stablecoin providing loans to the DeFi stablecoin. We can only expect that, should a market opportunity presents itself, some actors will find a way to make a profit. Nevertheless, the current DeFi stablecoins issuers are still holding mainly short-term assets and should be able to shrink their balance sheet quickly if needed. So it might not be a huge problem for now.

Conclusion

As we have seen, the creation of ClearingDAO would be kind of recreating an interbank market that would lead to a benchmark rate. In a sense, the ClearingDAO would be a stablecoin swap protocol (always at par) on the front-end and a lending protocol in the backend (between stablecoin issuers).

While ClearingDAO solves an issue when an imbalance occurs, it should not only be seen as a lender of last resort solution. Imbalance can be a business decision.

Therefore, this DeFi primitive would increase DeFi stablecoin value proposition and unlock new opportunities.

Those are only initial thoughts to open a discussion and no real project is planned.

Summary

• The price stability of DAI was expected to be managed by rate settings and speculative arbitrage
• This didn’t work in the aftermath of Black Thursday when DAI liquidity was needed to close borrowing positions, nor during DeFi Summer
• A risk-free arbitrage was introduced with a low Collateral Ratio USDC-A and finally the PSM
• The balance sheet of MakerDAO can now be expanded by DAI demand and no longer only by DAI supply
• Fees (tin) on the Price Stability Module (PSM) were removed on November 10th 2021. Now 1 USDC = 1 DAI = 1 USDP
• No significant change in USDC exposure was attributable to this change
• Decentralized trading volumes of the DAI/USDC pair increased as the PSM volume was additive (but might be due to the Uniswap v3 1bps pool)
• Impact for DeFi users is not significant as only a minority of traders use the PSM. Uniswap V3 1bps tranche was a more significant improvement. This might be due to high gas costs of the PSM and user preference for Uniswap.
• Crypto-lending is highly volatile which impact the balance sheet structure
• Real-World Assets could reduce balance sheet structure volatility

Link to the canonical version of The history of a DAI at par value.

Introduction

MakerDAO introduced DAI with a target price of 1 US Dollar. As we will see this quest for the peg took years. You can see below the evolution of the price of DAI against USDC since the end of 2019.

We will see how MakerDAO first relied on interest rate setting to balance the supply and the demand of DAI, then introduced a USDC-based vault to provide an arbitrage (speculative at first then risk-free). Finally, the Price Stability Module (PSM) was the biggest innovation of MakerDAO at the end of 2020. It solved the peg issue that MakerDAO was facing since DeFi Summer and replaced the stablecoins vaults that were coming to a limit.

We have explored previously that stablecoin issuers should fit two main constraints: solvency and liquidity and this paper is about the liquidity part. It was also discussed at length in Stablecoin design — Liquidity. We have later discussed a solution to fix the par value of stablecoins.

Some might challenge the need for liquidity as DAI is not contractually redeemable for anything until an Emergency Shutdown. One historical example is the Bank of Amsterdam that moved to a fiat standard in 1683 but used open market operations to keep the price of the bank guilder within a range (see How Amsterdam Got Fiat Money).

This paper starts with a historical analysis of the DAI peg by analyzing changes to address the problems at hand. We focus particularly on the balance sheet impacts of the implemented solutions. Then, we discuss some ideas before concluding.

Historical analysis of the peg

Early days

The early days of DAI (MCD version, but the SAI version was quite similar) were quite bumpy. DAI traded usually below the peg and that was solved by raising the interest rate (Stability Fees) on the Eth Vault (ETH-A SF). There was too much DAI supply and not enough DAI demand. You can see that each ETH-A SF raise was impacting the DAI price (it is unclear what caused the spike at the end of February). It should be noted that, at the time, the DAI Saving Rate, the DSR, (the ability for DAI holders to get an interest) was moved in line with ETH-A SF. They both achieve the same goal (increasing the price of DAI) but while increasing ETH-A is decreasing DAI supply (borrow), increasing the DSR increase the DAI demand (to get the yield).

This shows that dealing with ETH-A SF and the DSR were kind of effective to keep the DAI price close to the peg. Effectiveness should nevertheless be measured by the y-axis scale.

Black Thursday

On March 12th 2020, Black Thursday unfolded. Crypto prices plummeted (ETH fell by 43%) and it was the worst day for MakerDAO which led to $4.5M of underfunding and loss of collateral from borrowers due to a zero-bid issue. You can read a comprehensive Glassnode account here.

What matters for this analysis is that the price of DAI jumped and stayed at an elevated level. This would be strange for any TradFi practitioner. If DAI is a liability of MakerDAO, how could its price increase when the whole protocol becomes insolvent? The solution to this riddle is that borrowers need those DAI to close their vault and get their collateral back. When everyone wanted to close their vaults, DAI was a scarce asset as the debt could go only be paid in DAI.

During those events, USDC-A and USDC-B vaults were added to provide a backstop and limit the price of DAI (the limit being $1.25 at the time).

DeFi Summer

DeFi Summer is assumed to have started with Compound providing liquidity farming. On June 15th, 2020, Compound started rewarding its users. On July 2nd 2020, Compound changed the rewards and it impacted the price of DAI as was forecasted by Cyrus.

Quite quickly, the world was amazed that there was now more DAI on Compound than issued by MakerDAO (due to recursive farming).

The balance sheet of MakerDAO expanded to provide those DAI but not as fast as needed. As you can see below, some arbitrageurs were using the USDC-A vault when the price of DAI was volatile. Nevertheless, it was a speculative arbitrage and not a risk-free arbitrage, therefore the market anchored an ever-increasing “base” value of DAI.

Lowering the stablecoin vault liquidation ratio

On September, 15th 2020, the collateral ratio for USDC-A was decreased from 110% to 103% (it was 125% initially a few months ago). Quickly after, on September, 19th, liquidation ratios for USDC-A, GUSD-A, PAX-A and TUSD-A were decreased to 101%. Those actions created a risk-free arbitrage as the DAI price was above 1.03 before the first change and above 1.01 before the second change.

This action immediately provided a median transaction price cap at 1.01 (the bands being the 30bps of Uniswap fees). Nevertheless, with a Stability Fee of those vaults (4%) and no liquidation, the solution was not sustainable. As of writing, there is still $335M locked in those vaults. With a 1% Stability Fee, some vaults are becoming undercollateralized (example). A Stability Fee is present to render all those vaults liquidable in the future, but it generates some undue revenues (lending revenues that will be matched by a liquidation loss at some point). It is not impactful as long as it gets managed within a decent timeline (i.e. within 12 months).

On the MakerDAO balance sheet, we saw a huge inflow of USDC-backed loans resulting in the doubling of the balance sheet. The effect started a bit before each executive implementation as people front-runned the opportunity.

Introducing the PSM

On December 24th, 2020, the Price Stability Module (PSM or MIP29) was introduced. The PSM allows anyone to deposit USDC to get DAI and the opposite (which is limited by the amount of USDC in the PSM). The discussion on the need for a PSM started in early July. The PSM is controlled by two main parameters tin and tout corresponding to the fees that are taken for allowing the swap.

The tin parameter started at 1% (same effect as USDC-A with a 101% LR) and was reduced linearly to 0.1% over 7 days. The Debt Ceiling (DC) was limited to 3M, then 30M during which the facility it was limiting the usage. It is only after January 27th that the PSM DC was no longer a limiting factor.

As you can see below, moving the risk-free arbitrage from 1% to 0.1% didn’t have any impact on the stablecoin exposure. Some USDC-A exposure was migrated to PSM-USDC-A, but overall the stablecoin exposure was quite stable.

The PSM still improved the pricing of DAI that was closer to $1. Nevertheless, the trend started before the introduction of the PSM (indicated by the fact that the median price was quite below the risk-free arbitrage of $1.01).

Hardening the peg: 1 DAI = 1 USDC = 1 USDP

On November, 10th 2021, the fees (tin) on the USDC and USDP PSM were removed after a signal request. The intent was to move toward a space where all stablecoin are tradable at par by anchoring to a risk-free arbitrage.

One of the concerns was that such a move would attract a significant amount of USDC. As for the move from 1% to 0.1% in the PSM, it wasn’t the case and no significant impact could be found in the balance sheet evolution.

.

The price of DAI wasn’t impacted directly either. It has been around par since October 17th.

This later date is explained by borrower behavior. As you can see below there is a clear correlation between the increase in Nexo loan exposure at MakerDAO (see MakerDAO Borrower Concentration Analysis) and the amount of DAI on Curve. This makes sense as Nexo’s behavior is to sell the DAI it borrows. While correlation is not causation, it seems to be strong enough evidence. This also explains why the PSM was losing USDC before the fees change.

Therefore, the main impact we can attribute to the suppression of fees is an increase in PSM volume. Since fees are removed, there are inflows and outflows every day and volumes are far above history.

The impact in trading volume can be seen on DAI/USDC pair trading per decentralized exchanges. Curve was the main place before the PSM but this immediately moved to the PSM after fees were removed. Quickly after, Uniswap v3 added the DAI/USDC 1bps pair and took some market shares as well.

It is also interesting to note that the PSM volume seems to be an addition to the preexisting volume (PSM without fees was introduced week 45 and the DAI/USDC 1bps Uniswap pools got significant liquidity on week 46.

Finally, the chart below shows the daily spread over swaps done on Dex. One can see that 50% of the transactions are done within the Uniswap fees spread (20bps until November, 17 when the 1bps pool got $7M of liquidity creating a spread of 2bps). While removing the fees improved the number of transactions through the PSM, it’s still below 10% of the transactions (but with a higher amount per transaction compared to other venues).

Discussion

After showing the history that brings us to a point where DAI is now traded at its par value, we will now discuss some learning and ideas that could inform MakerDAO community decisions in the future.

Rates and speculative arbitrage are not enough

As we have seen, initially, rate settings (both on the lending side through Stability Fees and on the DAI Saving Rate) were the only tools to guide the price of DAI toward the peg. Before Black Thursday, the system worked quite well and it could be argued that with enough liquidity available (as we have today with billions on Curve), MakerDAO would have been able to achieve a strong peg.

This stopped working even with Stability Fees at 0% and DAI Saving Rate at 0%. The DAI demand was above DAI supply leading to an adjustment upward in price. First, it was due to the need of DAI to close borrowing positions (which the MakerDAO whitepaper interestingly labeled as a “benefit”). Then DeFi Summer increased the DAI demand for liquidity farming. We are now at a stage where stablecoins are a more useful use case for some people than crypto-speculation.

The USDC-A vault facility was introduced to provide tools for speculative arbitrage (removing the fluctuation risk of the underlying asset). Sadly, there weren’t enough arbitrageurs to fix the peg. Moreover, as the underlying problem was not solved (DAI demand > DAI supply), the arbitrager stopped taking risks.

The problem was fixed only after a risk-free arbitrage was introduced and the balance sheet was allowed to expand due to DAI demand.

Ensuring the peg is an impactful balance sheet decision

Currently, MakerDAO is providing a tight peg with regulated stablecoin, the main one being USDC. As we can see in the chart below, fluctuation of stablecoin reserves as a ratio of the balance sheet size is quite important with liquidity reserves taking up to ⅔ of the balance sheet. There were two 30%+ fluctuations in 2021. The liquidity part of the balance sheet is important and can fluctuate wildly.

Crypto-lending require a high level of liquidity reserves

The chart below illustrates that fluctuation in terms of stablecoin/balance sheet was mainly driven by DAI supply (users taking loans) than DAI demand. While DAI demand was almost a straight line in 2021, DAI supply fluctuated significantly and was correlated with crypto-market sentiment.

Therefore, the type of assets (other than liquidity reserves) in the balance sheet has an impact on the liquidity reserves needed. As we have seen earlier, the liquidity reserves were quite reduced when Nexo started to aggressively mint. How important is it for our borrower to be able to mint additional DAI at scale without hurting the peg? The answer to this question is important to understand the minimum liquidity buffer to keep.

Obviously, the historical “low” volatility of DAI demand shouldn’t be taken for granted either as regulation is unpredictable and stablecoin competition is growing.

Impact of RWA and money market lending

Started only this year and still quite small, Real-World Asset lending (RWA) seems so far unaffected by crypto-volatility. Adding RWA to the asset mix will probably stabilize the need for liquidity. It should be noted that RWA will more likely be less liquid, therefore bringing their own liquidity considerations

Even more recent was the money market initiative (Direct Deposit DAI Module, D3M or MIP50). This module adds some DAI on Aave when the DAI borrow rate there is above a given threshold. As Aave is crypto-backed lending, it shows the same volatility of borrowing demand as MakerDAO. Therefore, rates are low when speculative demand is low.

It can nevertheless act as a buffer in case of strong borrowing demand. Indeed, while Maker is engaged to keep interest rate fluctuations low on its platform, it didn’t make such a promise for its lending on Aave. In case of liquidity risk, MakerDAO could stop its Aave DAI lending without much commercial consequence. In such a framework, it could make sense to provide a lower borrow rate on Aave but be upfront that Aave borrowers don’t have the same rate protection that MakerDAO borrowers enjoy.

Conclusion

As we have seen through this paper, achieving the par value of DAI (1 DAI = $1) took years in the making. It was designed to work with rate setting as the only policy tool. It quickly ran into the problem of not having enough borrowing demand to fulfill the need of DAI. Speculative arbitrage wasn’t enough as the credibility of MakerDAO to solve the peg issue decreased.

While the balance sheet expansion of MakerDAO was only possible due to borrowing needs (DAI supply), the PSM enabled the DAI demand side to alter the balance sheet as well. The PSM, therefore, represents the imbalance between DAI supply and DAI demand. The adjustment variable is no longer the price.

Going forward, we recognize that the current balance sheet structure of MakerDAO is fairly volatile as it is based mainly on crypto-speculation. The RWA and money markets initiatives are great steps towards reducing the volatility due to crypto-markets.