A New Mental Model for Defi Treasuries

Author

The Defi bull market, started by COMP liquidity mining in Summer 2020, has turned many Defi protocols into rapidly growing revenue monsters. You would assume that this puts them in a comfortable financial position, and a superficial look at DAO treasuries seems to confirm that. For example, OpenOrgs.info suggests that the top Defi protocols are sitting on many hundreds of millions or – in Uniswap’s case – even billions of dollars.

However, almost all of this supposed treasury value comes from the projects’ native tokens such as UNI, COMP, and LDO, as the following charts show:

While we agree that native tokens inside a project’s treasury may be financial resources, counting them as assets on their balance sheet does much more harm than good and is often used as an excuse for poor treasury management

To elucidate that point, allow us to take a quick detour into traditional accounting.

Native tokens are not assets

While Defi tokens are not considered equity in the legal sense, we can still learn from how traditional companies account for their shares. Simply speaking, the float (all shares available for public trading) and restricted shares (employee shares that are currently vesting) together make up a company’s outstanding shares.

These outstanding shares are a subset of authorized shares – a self-imposed soft cap on total issuance. Crucially, shares that have been authorized but not issued are not counted to the company’s balance sheet. And how could they? Counting unissued shares would allow a company to arbitrarily inflate their assets just by authorizing more shares without ever selling them.

We hope you see the connection to native tokens in DAO treasuries: these are the crypto-equivalent of authorized but unissued shares. They are not assets of the protocol but merely report how many tokens the DAO could “legally” issue and sell to the market.

Whether a DAO authorizes a small or very large number of tokens into its treasury is hence meaningless: it says nothing about its real purchasing power. To illustrate that point, imagine that Uniswap tried to sell as little as 2% of its treasury. When executing this trade via 1inch, which routes the order to many on- and off-chain markets, the price impact on UNI would be almost 80%.

Real Defi treasuries

Ignoring authorized but unissued shares allows us to get a different, much more accurate picture of Defi treasuries. For this exercise, we broke the non-native down further into the three categories of (1) stablecoins, (2) blue chip crypto assets, and (3) other non-stable crypto assets. Using this new categorization, Uniswap has ~0 assets in its treasury and only Lido and Maker have >$50m. 

But why are treasuries of this size problematic?

First, we have seen that it’s not enough to issue new shares, you also have to sell them on the market. This causes price impact, which quickly becomes a constraint for larger sales. But further, the price the market is paying for your native token is not guaranteed but highly volatile.

Second, that price depends on the overall market conditions. The crypto market has gone through several speculative cycles in which tokens can reach euphoric valuations but can also collapse 90%+ and stay there for a long time.

Third, times when Defi projects urgently need liquidity can correlate with project-specific risk: for example, when a project experiences a large insolvency event due to a bug or hack and wants to make users whole, the token price tends to be depressed as well – especially if holders expect a dilution event. 

Case study: Black Thursday exposes MakerDAO’s treasury

The risk of holding insufficient treasury reserves is not just theoretical, as MakerDAO experienced firsthand during the market crash on March 12, 2020 (commonly referred to as “Black Thursday”). A lack of liquid assets put the MakerDAO credit system at risk of collapse, and although the crisis was eventually defused it has led to a significant erosion of token holder value. Let’s see how it played out: 

From MakerDAO’s launch in 2018 through March of 2020, the DAO had used net earnings to buy back and burn MKR tokens (returning capital to token holders), with a total of 14,600 MKR burned at a cost of over 7 million DAI. MKR token price through this period averaged around $500. 

Then Black Thursday came around, and due to sharp price declines and a congested Ethereum network, Maker failed to liquidate under-water positions in time, causing 6 million in losses to the protocol. After deducting the 500,000 DAI in MakerDAO’s treasury at the time, it had to cover the remaining 5.5 million in losses through auctioning MKR tokens on the market. Maker ended up selling a total of 20,600 MKR at an average price of roughly $275.

It took until December 2020 for Maker’s accumulated earnings to reduce the token supply back to the original supply of 1 million MKR through buybacks, at a total cost of over 3 million DAI (again with an average MKR price of roughly $500). 

Image: The Makerburn site shows significant token dilution caused by the Black Thursday crash.

To summarize the financial impact, $6 million in credit losses from Black Thursday erased $10 million in earnings accumulated over 3 years. The $4 million in additional losses could have been avoided if Maker held more treasury reserves in stable assets like DAI, as they could have used these funds to cover insolvent loans without needing to sell MKR at depressed prices. Or put another way, Maker could have seen up to $4 million in additional value accrual by holding a larger treasury.

While it’s difficult to evaluate treasury needs in advance, the 500,000 DAI Maker held as of Black Thursday was almost certainly too little. It represented only a 0.35% capital buffer for the protocol’s 140 million in outstanding loans, where most traditional financial institutions hold at least 3-4% in risk capital. And this is before taking into account operating expenses and salaries, which could cause further forced selling during market downturns if they aren’t covered by non-native treasury assets.

Understanding buybacks and dividends

That many Defi projects naively look at their token as a treasury asset and might have to sell it at the worst possible time is the result of a missing framework for how to do it better. While there are many ways to run a protocol, practitioners might benefit from the following guidelines.

Rule 1: The goal of a DAO is to maximize long-term token holder value.

Rule 2: When put into action, rule 1 suggests that each dollar a protocol owns or receives as revenue should be allocated to its most profitable use, discounted to the present day. Options typically include saving the money in the treasury, reinvesting it into growth or new products, or paying it out to token holders via token buybacks or dividends.

Only if the money has a higher return for token holders outside the protocol (after taxes), it is correct to pay the money out instead of saving or reinvesting it. In practice, we see many Defi protocols pay out funds that could be used for growth or saved in the treasury for future expenses. According to our framework, that is a big mistake. In the case of Maker, we have seen how it sold cash for tokens but then had to buyback that same cash with tokens at a much higher cost-of-capital.

In general, we recommend divorcing the idea that paying dividends or buying back tokens is somehow “rewarding” token holders while internal reinvestment is not. The most rewarding decision for token holders is the one that maximizes the return on each dollar, whether internally or externally.

Rule 3: When adhering to the above rules, a DAO becomes an acyclical trader of its own token. If the DAO sees its token as overvalued and internal reinvest has a good return, it should sell tokens for cash and reinvest that cash into the protocol. This is almost certainly the case in all bull markets. When the DAO sees the price of its token below the fair value, and it has excess cash without a high internal return, then it can buy back the token. This is almost certainly the case in all bear markets.

Toward better treasury management

Finally, we want to share our views on how DAOs should manage their treasuries. We came up with the following rules:

Rule 4: DAOs should discount native tokens from their treasury immediately – they are the crypto equivalent to authorized but unissued stock.

Rule 5: DAO treasuries need to survive the next bear market. It might not happen next week or next month, maybe not even next year. But in a market as driven by speculation as crypto it will happen. Build a treasury that will last you 2-4 years even if the entire market collapses by 90% and stays there for some time. 

We recommend 2-4 years specifically because you want enough to survive even the longest crypto winter by known standards, but not so much that you get rich and lazy, or too distracted by running your protocol like a hedge fund. 

Considering the known operating expenses of major DAOs with large dev teams and liquidity mining programs, very few if none satisfy that condition today. That means, most or all of them should use the bull market to sell tokens and build real treasuries with stable assets that will not only make them survive the coming bear market but hopefully put them ahead of their competitors.

Rule 6: DAO treasuries should understand their application-specific liabilities and hedge them. For example, a lending market might plan that some % of loan positions will fail every year. While they don’t say so explicitly, it is implicitly understood that the lending market underwrites that risk. So the underwriting becomes a regular cost on their balance sheet and can be hedged accordingly. Meanwhile, a leaner protocol like Uniswap may underwrite no additional risk and hence could do fine with a much smaller treasury.

Article by Hasu and monetsupply

Acknowledgments: Larry Sukernik, Georgios Konstantopoulos, Dan Robinson, Tarun Chitra, Ali Atiia

Share this post

Share on twitter
Twitter
Share on facebook
Facebook
Share on linkedin
LinkedIn
Share on reddit
Reddit
Share on email
Email

10 Replies to “A New Mental Model for Defi Treasuries”

  1. IMO a lot of the points that support ignoring native token in treasury evaluation can be solved with lending markets.

    If you can mint synths or provide native tokens as collateral (without the lending out your token) as a means to finance sudden/major costs, there won’t be token dilution in the market. Native tokens then are a means to finance these sudden events, and are really valuable assets.

    As for a TradFi example, if the CEO of a company can’t feasibly sell his stocks without causing market turbulence, but can still borrow money against his stocks, are his stocks worthless, or is it worth the amount of loans it can collateralize?

    1. To use this example though, if a company say a bank gets hacked and loses all its deposits, it’s likely that the amount one can borrow against stocks in this company is significantly smaller if even possible.

      1. With permissionless lending pools like Fuse, there will always be people who are willing to take the risk. Though not on a permissionless pool, Alpha Homora still managed to get a loan from Cream with ALPHA as collateral after its hack earlier this year.

        The interest rate from permissionless pools might skyrocket though, which would hamper the native token’s utility, but I’d argue its value is still far from zero.

  2. If the argument is that DAO treasuries need to be set funds to use during bear markets then it makes no sense to consider BTC and ETH ‘blue chip’ assets. To use the Maker example, if they had held their treasury in ETH and then sold it during March 2020 for sub $100 to cover the shortfall it would have been arguably worse than selling MKR for $275.

    The barbell approach of holding stables + native token is probably a better approach. Stables could be used to provide a yield and also be sold without impairment during a bear market.

    A higher token prices is positive for sentiment/morale/visibility so holding and/or buying tokens can have positive reflexivity so be preferable than liquidating for BTC/ETH.

  3. I think there are several other treasuries with big non-native treasuries you haven’t talked about. Besides all the Olympus DAO forks which have big amounts in treasury, there is also Keeper DAO which has a $100 million treasury in non-native assets.

  4. This articel is f*cking awesome.
    In my protocol we were using some of this rules but being able to see it here its also a nice confirmation .

    Can I translate it to Spanish and publish it?

  5. I want you to join BlackBull Partners. We’ve been in operation for 10 years and work with over 20,000 websites across the industry including many of your competitors.

    Why join BlackBull Partners?

    – 20k+ active partners
    – CPA up to USD$1,000 per client
    – 15 support languages in 180+ countries
    – 2,000+ multilingual marketing tools
    – Customisable live S2S tracking system

    Who are BlackBull Markets?

    – 1:500 leverage
    – no commission trading
    – spreads from 0.00 pips
    – Award-winning trade execution speeds
    – $0 minimum deposit to open an account
    – New Zealand owned and operated business
    – 24/7 dedicated client support
    – Multi-region regulated licences
    – Platforms include TradingView, MT4, MT5, WebTrader, BlackBull CopyTrader, BlackBull Shares, BlackBull Trade and more.
    – Trade 26,000+ instruments (forex, shares, crypto, indices, gold, silver, and oil)

    I’d be happy to discuss other options with you as well. You can register at http://partners.blackbull.com or get in touch if you have any questions.

    Kind regards,

    Benjamin Boulter
    BlackBull Partners

Leave a Reply

Further reading