Translation: Hasu: A New Conceptual Model for DeFi Vaults
The author of Uncommon Core, Hasu, wrote an article "A New Mental Model for Defi Treasuries" discussing a new mental model for DeFi treasuries. Here is the translation:
The DeFi bull market, kicked off in the summer of 2020 with COMP liquidity mining, has turned many DeFi protocols into rapidly growing yield monsters. You might think this puts them in a comfortable financial position, and it seems to confirm this by looking at the surface of DAO treasuries. For example, OpenOrgs.info shows that top DeFi protocols hold hundreds of millions of dollars, with Uniswap alone holding even billions.
However, almost all of these assumed treasury valuations come from the native tokens of the projects, such as UNI, COMP, and LDO, as shown in the image below:
Ad - Scroll down for more
Treasury of major DeFi protocols: Amount, blue represents native tokens
Treasury of major DeFi protocols: Proportion, blue represents native tokens
While we agree that the native tokens in project treasuries may be financial resources, considering them as assets on the balance sheet is more harmful than beneficial, and it is often used as an excuse for poor fund management.
To clarify this, let's quickly skip traditional accounting.
Table of Contents
Native Tokens Are Not Assets
While DeFi tokens are not considered equities in a legal sense, we can still learn from how traditional companies account for their shares. In essence, freely tradable shares, which are all shares available for public trading, and restricted shares, which are currently locked employee shares, together make up a company's outstanding shares.
These outstanding shares are authorized shares, part of the self-imposed minimum issuance set by the company. Importantly, authorized but unissued shares do not appear on a company's balance sheet. Why is that? Including calculations of unissued shares would allow companies to inflate their assets arbitrarily by authorizing more stocks without selling them.
We hope you can see how this relates to native tokens in a DAO treasury: these are authorized but unissued stock-like crypto equivalents. They are not assets of the protocol, yet DAOs rarely disclose how "legally" many tokens they have issued and sold to the market.
Therefore, whether a DAO authorizes a small or large number of tokens into its treasury is meaningless: it does not explain their actual purchasing power. To illustrate this point, imagine Uniswap trying to sell as little as 2% of its treasury tokens. When this trade is executed through 1inch, it directs the order to many on-chain and off-chain markets, affecting the price of UNI by close to 80%.
Real DeFi Treasuries
By not considering authorized but unissued shares, we can have a more distinct and accurate understanding of DeFi treasuries. In this exercise, we further categorize non-native assets into three types: 1 stablecoins, 2 blue-chip crypto assets, and 3 other unstable crypto assets. Using this new classification, Uniswap's assets are approximately 0, with only Lido and Maker holding over $50 million in assets.
Treasury compositions without native tokens: yellow for unstable coins, red for blue-chip crypto assets, blue for stablecoins
But why would treasuries of this scale become problematic?
Firstly, we have seen that issuing new shares is not enough; you also have to sell them in the market. This can lead to price impacts and become a constraint during massive sell-offs. Moreover, the price paid by the market for your native tokens is not guaranteed and highly volatile.
Secondly, this price is dependent on the overall market condition. The crypto market has seen several speculative cycles where tokens can reach pleasing valuations but can also crash over 90% and remain in that state for a long time.
Thirdly, sometimes when a DeFi project urgently needs liquidity, it may be associated with project-specific risks. For example, when a project faces a major bankruptcy event due to a bug or hack and aims to compensate users, token prices typically suffer, especially when token holders anticipate chip dilution.
Case Study: Black Thursday Exposed MakerDAO's Treasury
The risk of holding insufficient reserves in the treasury is not just theoretical, as MakerDAO experienced during the market crash on March 12, 2020, commonly referred to as "Black Thursday." The lack of liquid assets put MakerDAO's credit system at risk of collapse, and although the crisis was eventually mitigated, it led to a significant decrease in token holder value. Let's see how it unfolded:
From its launch in 2018 until March 2020, DAO had been using net profits to repurchase and burn MKR tokens, returning funds to token holders, totaling 14,600 MKR burned at a cost exceeding 7 million DAI. During this time, the average price of MKR tokens was around $500.
Then came Black Thursday, where due to a significant price drop and Ethereum congestion, Maker failed to liquidate underwater positions in time, resulting in a $6 million loss to the protocol. After deducting 500,000 DAI from MakerDAO's treasury at that time, it had to auction off MKR tokens in the market to cover the remaining $5.5 million loss. Maker ultimately sold a total of 20,600 MKR at an average price of around $275.
It wasn't until December 2020 that Maker's accumulated profits reduced the token supply to the original 1 million MKR supply, with a total cost exceeding 3 million DAI and an average MKR price of around $500.
The Makerburn website shows a significant dilution in tokens due to Black Thursday crash
Summing up the financial impact, the $6 million credit loss on Black Thursday wiped out $10 million in accumulated profits over three years. If Maker had held more reserves in stable assets like DAI, they could have avoided an additional $4 million loss, as they could have used those funds to pay off the debt without selling MKR at a lower price. In other words, Maker could have gained up to $4 million in additional value by holding larger reserves in the treasury.
While it's challenging to assess funding needs in advance, by Black Thursday, holding 500,000 DAI was almost certainly too little for Maker. It represented a protocol with $140 million in outstanding loans but only a 0.35% capital buffer, whereas most traditional financial institutions hold at least 3-4% of risk capital. This doesn't even account for operational costs and personnel expenses, which if not covered by non-native treasury assets, could lead to further forced selling during market downturns.
Understanding Buybacks and Dividends
Many DeFi projects naively treat their tokens as treasury assets and may have to sell them at the worst times due to a lack of a better framework. While there are many ways a protocol can operate, operators can benefit from the following guidelines.
Rule 1: The DAO's goal is to maximize long-term value for token holders.
Rule 2: To act on Rule 1, it is suggested that every dollar a protocol owns or generates as income should be allocated to its most profitable use, discounted to the present. Options typically include storing funds in the treasury, reinvesting them for growth or new products, or distributing them back to token holders through buybacks or dividends.
Spending this money only when it has a higher return to outside token holders after tax than saving or reinvesting it is the correct choice. In practice, we see many DeFi protocols spending funds available for growth or saving for future expenses on token holder payouts. According to our framework, this is a significant mistake. As seen with Maker, we observed how it liquidated tokens only to later buy them back at a higher cost.
In general, we recommend that forgoing dividend payments or token buybacks is a form of "reward" to token holders to some extent, whereas internal reinvestment is not. The most valuable decision for token holders is to maximize the return on every dollar, whether internally or externally.
Rule 3: When in compliance with the above rules, the DAO will become an infrequent trader of its own tokens. If the DAO believes its tokens are overvalued and internal reinvestment has a good return, it should sell tokens for cash and reinvest that cash back into the protocol, as is the case in almost all bull markets; when the DAO sees its tokens' price below fair value and it has excess cash with no high internal return, then it can buy back tokens, as is common in most bear markets.
Achieving Better Treasury Management
Finally, we'd like to share our views on how DAOs should manage their funds. We propose the following rules:
Rule 4: DAOs should immediately discount their native tokens in the treasury as they are authorized but unissued stock-like crypto equivalents.
Rule 5: DAO treasuries should be able to survive the next bear market. It may not happen next week or next month, or even next year. But in a speculative-driven market like cryptocurrencies, it will happen. Building a treasury that can last 2-4 years, even if the entire market collapses by 90%, can sustain for some time.
We specifically recommend 2 to 4 years because you need ample time to survive in a long crypto winter with known standards but not enough to make you rich and lazy or to run your protocol distracted like a hedge fund.
Considering the operational costs that major DAOs with large development teams and liquidity mining would have, currently few meet this criterion. This means most or all of them should take advantage of bull markets to sell tokens and build a truly stable asset treasury, not only to survive in future bear markets but also to potentially outperform competitors.
Rule 6: DAO treasuries should understand specific debts for their applications and hedge against them. For instance, the lending market may expect a certain percentage of loan positions to fail each year. While it may not be explicitly stated, people generally understand the risks the lending market bears. Therefore, underwriting becomes a regular cost on the balance sheet, which can be accordingly hedged. Meanwhile, leaner protocols like Uniswap may not bear additional risks, so it doesn't matter much even with a smaller treasury.