Nouns fork is perpetually vulnerable to a dishonest minority

tldr; The DAO fork, while protecting against a dishonest majority, may have opened up an attack vector for a dishonest minority to repeatedly drain the Nouns treasury.

Nouns is perhaps one of the most active and fully decentralized DAOs out there. It is continuously at the forefront of funding public goods, art, and governance experiments - the upcoming “DAO fork” is another interesting one to observe.

The fork is introducing a net new mechanisms that did not exist in traditional corporate governance: with only 20% ownership of a company, you cannot simply demand 20% of the company’s assets. As a result, in a traditional structure → it’s 51% or no cookies.

In the status quo, the honest minority can easily be screwed. We see this happen to founders, executives, honest minorities ‘pushed out’ over and over again in corporate history.

Hence the introduction of the DAO fork — a minority protection mechanism.

However, I suspect with this, it may also introduce a new attack vector, one of the dishonest minority of arbitrageurs. Not a one-time attack, but a repeated dishonest minority attack. Simply put, imagine if you no longer need 51% to mount a corporate raid, only 20%. How much more profitable would Carl Icahn be?

History is being made as we’re witnessing a minority raid unfold on a $50M treasury.

What’s happening?

The last 72 hours has been fascinating, as cards are finally starting to be played. Both sides are starting to realize that the probability and impact of a DAO fork is very much happening. Let me explain how both sides are playing their cards.


  • [Prop 311] Fork appears on core dev roadmap

  • [Prop 324] Arbers propose to halt the auction to protect their gains

  • [Prop 330] Honest majority propose a poison pill (perhaps accidental) to reduce the total amount in the treasury such that the arb is no longer attractive. Re-proposed here: [Prop 336] — first one was cancelled

  • [Prop 332] Arbers propose a protective wrap — a clever maneuver where you reduce the amount of liquidity in the treasury and stash it away so it can’t be poison pilled. (This would be similar to as if a corporation buying up bonds so that management stops wasting money on Facebook Ads).

What is really starting to surface here, and Nouners are starting to realize, is that the new system — while does a good job of of protecting against the dishonest majority — is now vulnerable to a dishonest minority.

Who are the minority arbitrageurs?

First, it’s important to delineate what is happening here. Arbitrageurs are not “spend conservatives” trying to conserve capital and seek higher ROI projects. What Nouns is currently facing against is a classic corporate raid.

But unlike the classic corporate raid of a 51% attack, this is a minority raid — something unseen before in corporate web2.

Voting history graph of a likely arbitrager's behavior
Voting history graph of a likely arbitrager's behavior

Consider the above voters, who bought Nouns recently at below book value and have voted no on every prop except the v3 audit, core dev funding, and vote refund – props that either directly enable their arb or save them gas money.

To brand them ‘honest minority’ is an insult to those who wants to build the ecosystem but just with a different operating model. These are folks with a short-term agenda of robbing the DAO. These voters are the Carl Icahn of web3. Let’s not kid ourselves. This is a corporate raid.

If you click around some of these profiles, you’ll tally up anywhere from: 120-150+ votes. This along with the assumption of a 10% swing of opportunistic individual voters (which you can estimate to be around the number in the Governance Pool), you’ll have ~22% of current Nouns supply. Enough to trigger the DAO fork.

The long term vulnerability

The fork is still a step in the right direction for Nouns to achieve full decentralization with minority protection. However, it currently leaves a gap against the dishonest minority, a vulnerability that will likely be continuously exploited until it’s closed with an update to the mechanism.

In the event of a DAO fork, there’s a few possible scenarios:

Low probability, Scenario A: honest majority, honest minority

  • This is what the best case scenario of the intended DAO fork mechanism, an honest minority trying to exit with good intentions.

High probability, Scenario B: honest majority, dishonest minority

  • Cost to the honest majority: -20% of treasury and book value

  • Dishonest minority: takes a 20-30% return over 3-6 months → 70-100% IRR. This is juicy enough returns to set a precedent to raise capital on or rally troops once again for a second corporate raid.

Low probability, Scenario C: dishonest majority, honest minority

  • This scenario is a classic 51% attack or hostile takeover, it will exist with or without the DAO fork mechanism.

Why is B far more likely? Because by definition a dishonest minority would form before it becomes a dishonest majority, therefore Scenario B will trigger first, every time. In a steady cadence, the dishonest minority will resurface on a predictable cadence, and will periodically drain the treasury.

Similarly, when it comes to coordinating an exit. While the honest minority might side with the dishonest minority temporarily, the dishonest minority will only raid when they’re the book value is sufficiently greater than the auction price. Making Scenario B a much more likely outcome than A.

Post-DAO split can technically be higher based on market value per Noun post split.
Post-DAO split can technically be higher based on market value per Noun post split.

Why does this happen post DAO fork?

  1. The treasury will be drained by ~20-25%.

  2. Due to the last wave of defeated props, there will likely be a spending uptick post-fork, as a lot of builders and projects will put up their props again.

    • Nouns traditionally spends ~800 ETH per month on public goods, this has been cut in half in the last 2 months. We will likely see a resurgence of props right after DAO fork causing downward pressure on the book value.

    • Post fork, with the additional supply in Nouns’ treasury, it is also very likely that Nouns will start paying projects out more in Nouns rather than pure ETH. This will increase supply and add sell pressure in the market, further dropping auction price.

  3. This extra expense will likely drive down auction value of Nouns. In fact, it’s very possible that the raiders would want to launch a campaign to drive down auction value as much as possible at Nouns.

    • At a minimum, the Nouns auction will lose the buy floor of the arbitrageurs. While this is probably a good thing, it will create, once again, a gap between book and auction price.

    • Just like before, a risk premium (uncertainty in future prices) + illiquidity premium will form up. The illiquidity premium is lower than pre-fork, but should roughly be thought of as cost of coordination to assemble 20% of Nouns supply

  4. This gap of Book value > Auction value will once again open up again until the cost of buy up 20% of Nouns is worth the arb, then it begins again.

  5. This will rinse and repeat until the Nouns treasury is too small to juice out a profit from the arb.

Duration of each cycle is likely 6-12mo
Duration of each cycle is likely 6-12mo

While seemingly small at first, the dishonest minority may be more harmful than the dishonest majority in the long run because the barrier of entry to becoming the dishonest minority is so much lower. The goal here is to patch the incentives and system such that the DAO is protected against the dishonest minority treasury draining loop.

This is not the end

As with all things financial, there is no such thing as “problem’s gone”. Problems only disappear when the last dollar is squeezed from an opportunity (the arb). Make no mistake that this sort of arbitrage, once it happens once — will happen again and again. There are definitely ways to defend against a corporate raid (one example might be locking excess ETH into a timelock to reduce the arb), and it’s also important to acknowledge the costs associated with each solution. However, the real question is: is this a cost worth bearing? I would argue ‘yes’, at most you can let this kind of raid happen once — but you must not let this type of raid be a repeat incident. Or else, there will be no treasury left to govern.

Subscribe to cfeng.eth
Receive the latest updates directly to your inbox.
Mint this entry as an NFT to add it to your collection.
This entry has been permanently stored onchain and signed by its creator.