CIP-37: Increasing liquidity for COW token via a programmatic order and an FM-AMM

[AndreaC]

Hi @Yakitori

I personally agree with you that some of the liquidity already deployed could be better used. I’m just saying that it is a separate discussion: unless you believe that we can make the price impact of a 100k trade disappear simply by repurposing existing liquidity, then we need to deploy more funds (and a better mechanism), which is what this CIP is about. But I also hope that, in a separate discussion, the DAO will rethink how the existing liquidity is used.

Selling would dilute holders at a low price and without a strong buyback inflow it would weight on the price, that is already very volatile.

On this, I disagree. First, I don’t understand why you say at a low price: we would sell at market price, which is currently quite high by historical standards, and we will have a floor below which we would not sell. In addition, we have a number of initiatives that, collectively, will reduce the COW outflow by more than 5M. So, overall, during the period the program is running, the DAO will be spending less COW than historically. I also disagree with the dilution argument. Successive fundraising is a basic fact of startup investing, which implies that old investors are diluted. But if the funds raised are used to create additional value, then this is also beneficial to the old investors.

[AndreaC]

To be clear, we propose to start selling now and continue doing so for a period of 9 months to a year. So, liquidity accumulates over time. We also propose to sell 5M COW, not 10M

If I understand your requirements, you want 10X in liquidity (so that whales and funds can invest), right now and without deploying additional resources (else we may depress the COW price). We don’t know of any mechanism that can achieve it.

[cryptocondom]

Would suggest paired liquidity with ETH from the GNO to ETH vault to supplement the 9-12months of COW liquidity tested via cow/eth COWamm.

If your dont know any mechanism that can achieve 10x liquidity…allow me to propose a cex listing with POL to increase liquidity.

[luude]

I also agree with the concerns raised in this thread above and the notion to use the liquidity from the GNO:ETH pool to be redirected to COW:ETH.

There are also more active strategies that can be conducted using far superior tech to a balancer v2 pool which would allow for much more liquidity utilisation and thus facilitating tighter spreads for larger actors while keeping the initial outlay of liquidity lower.

If for example CowDAO conducted a more hands on active approach to liquidity management using a full range Balancer/Uni v2 style + a CLMM concentrated liquidity strategy we could see marked improvements for liquidity depth using a fraction of the size we would have to if just seeding a balancer v2 pool.

[thegeneticjackhammer]

I agree with cryptocondom on this.

Liquidity is needed now and I think it’s a matter of urgency. COW DAO is one of the best projects in this space but it’s crazy how much the price can shift on small buys/sells.

[AndreaC]

Well, cex listing is unfortunately not up to us

Anyway, it is an avenue we are pursuing, but it has many of the problems we discussed: the project being listed needs to provide the initial liquidity, which would mean selling COW for whatever other token COW is traded against, which may depress its price, …

Would suggest paired liquidity with ETH from the GNO to ETH vault to supplement the 9-12months of COW liquidity

As you mentioned, FMAmm is for the moment a theoretical idea in a research paper. We want to make it a product and test it by slowly increasing its TVL. If it works and has all the benefits that the theory predicts, then I agree with you that we should supplement its liquidity with additional funds. But I think it is premature to commit to doing so right now.

[garrettz]

starting slow makes sense, but this is not an argument for selling cow over using eth currently earning just 3% in the treasury.

[cryptocondom]

Fortunately it is up to the protocol. You never know wen a pesky cex listing(s) could hit to increase liquidity. In the mean time, would suggest we modify this proposal to silo the two things being proposed:

1. creation and deployment of the CoWAMM as a proof of concept
2. addition of liquidity via re-allocation of eth from the GNO-ETH to COW-ETH pair.

If you would like to addend your current proposal to say, 1m COW allocation to test it in production, I would be more than happy to support this proposal with more than 6m+ cow votes. I am willing to write a second proposal for the re-allocation of COW liquidity from GNO-ETH to address our current liquidity shortcomings. Thoughts?

[Yakitori]

@AndreaC

I’m just saying that it is a separate discussion: unless you believe that we can make the price impact of a 100k trade disappear simply by repurposing existing liquidity, then we need to deploy more funds (and a better mechanism), which is what this CIP is about.

You make actually an interesting point here - what is the liquidity needed ? I don’t know about FM-AMM, so I can only use as reference the current Balancer pool.

If we use the classic xy = k formula, if we swap 40 ETH only in the current COW/WETH pool (which has 216 ETH for 1,616,000 COW), we get a slippage of 40%. There’s the GNO pool, but let’s assume we remove it to ease the computations.

Multiplying the liquidity by a factor of 10 would decrease this slippage to 3.7% or in other words we’d need to have 5.4M$ of ETH in the pool to get a reasonable slippage, which is around one fifth of the MC (per CMC).

I’m curious about how the FM-AMM would help in this case.

@AndreaC

First, I don’t understand why you say at a low price: we would sell at market price, which is currently quite high by historical standards, and we will have a floor below which we would not sell.

COW launched at the start of the bear market and had 0 value accrual to its token for more than a year. This is isn’t a good reference. If you look at the market cap of the project, it’s actually pretty low, at around $30M, vs 0x ($282M) or 1inch ($516M).

I think that it’s another good example of the general and a thousand times tested rule “DAOs should not trade” - aside of course from portfolio rebalancing.

@AndreaC

In addition, we have a number of initiatives that, collectively, will reduce the COW outflow by more than 5M. So, overall, during the period the program is running, the DAO will be spending less COW than historically.

Let’s then collect the ETH from the solvers and add it to the liquidity instead of selling. You’re suggesting we’d market sell COW while we market buy COW from solver revenue. That doesn’t make much sense.

@AndreaC

Successive fundraising is a basic fact of startup investing, which implies that old investors are diluted. But if the funds raised are used to create additional value, then this is also beneficial to the old investors.

I fail to understand the underlying value created by additional liquidity. We’re all very happy if we get new members of the DAO, but this doesn’t add underlying value to the project.

[tanglin]

Completely agree here with Yakitori!

First of all let’s be clear, 1.5M at current prices for the product COW is, is nothing. Nevertheless, adding selling pressure to a per-se inflationary token for the time being is suboptimal, so why not first addressing other potential alternatives, and then recurring to the idea of selling COW if its necessary at all after the refinement process.

Side note: Wouldn’t it possible to do it OTC and simultaneously gather these ‘whales and funds’ while increase the liquidity? Quite sure there are a bunch of good, high quality players looking to enter COW. I added +1M COW in the past months and was quite hard to do so without getting eaten by price impact and volatility, so pretty sure some would look at it with good eyes.

Now back to refinement, as mentioned before, resources could be better allocated, starting with the GNO/ETH & GNO/COW pool.

• Take the GNO of the current GNO/ETH pool and proceed with liquidation. Being 10 times the market cap while being listed on the most popular CEXs means that they are better prepared for any additional source of sell pressure. Furthermore its 370,000 USD we are talking, not 1.5M. Also Andrea made a good point when it comes to diversification of portfolio, but that also applied to GNO which currently amasses a huge portion of the portfolio.

• Take half of the GNO and COW from the GNO/COW pool. And start generating staking revenue instead of the current losses in which the position is currently incurring.

• Take those $740,000 USD worth of ETH and pair it with the now freshly available $COW, that would start generating revenue today without increasing the supply in circulation.

If after increasing the liquidity by $750,000 USD + increasing the efficiency of the pool by eliminating a unnecesary trades + using a more efficien AMM model -yet to be proven- there is a need for an additional increase, I would be more than happy to vote in favor of an additional allocation of 2.5M COW to be sold over the course of 9-12 months, which would even (at current prices) the liquidity added, resulting in a net positive impact in the market due to the increase efficiency.

With current inflation and treasury net results, we are at a big loss, therefore using a part of the treasury to increase liquidity + effiency + test a promising COW product sounds like a rationale move, otherwise why is that we have a treasury in the first place. I understand the need for back-up, but also have experienced big treasuries sitting idle while projects die off many times, in a rapidly evolving and highly volatile market taking measured risks is a must.

Btw the current marketcap to liquidity ratio is <1/10 — quite healty for any token—, this proposal would set it to ~1/5 or so which is ideal. As @Yakitori mentioned is more a problem of low marketcap and effiency rather liquidity, but we can attack both verticals at the same time.

[adjum]

@AndreaC

First, I don’t understand why you say at a low price: we would sell at market price, which is currently quite high by historical standards, and we will have a floor below which we would not sell.

Perhaps I can take a stab at this. Imagine the algorithm is of the form, “Sell 20,000 CoW at market rate at noon each day provided price is above $X.” I foresee multiple challenges:

1. When market prices are at $X + epsilon, the market is incentivized not to purchase $COW, knowing that 20,000 CoW is about to be market sold, thus making prices cheaper by delaying one’s purchase. This would be especially true as we get ever closer to noon, distorting market dynamics in ways I believe CoW is eager to avoid.

2. When market prices are at $X - epsilon, market participants with short timeframes are incentivized to sell, as there will be a 20,000 COW sale forthcoming, thus preventing price from appreciating further for that short timeframe holder.

Therefore, the current proposal not only creates selling pressure as we approach $x, but also reduces buying pressure as we pass $x, thus resulting in an $x threshold creating a price ceiling in-effect, even if not encoded nor intended to serve as such.

I can appreciate that one may argue that neither of the above would be a material concern if the 20,000 COW sale is immaterial. However, for that I have two thoughts:

1. Total daily trading volume is currently ~1.3M COW, making a 20k sale ~1.3% of total daily volume. That feels material to me.
2. If one were to reduce the daily sale to a level where the price impact is immaterial, I suspect it would undermine the pace at which one is trying to seed the AMM, and thus the intentions of the proposal.

I fear the null set may be the only outcome to the conundrum above (i.e., one can’t both make material progress on building up the AMM via constant COW sales at a healthy volume / pace && not have a material impact on market dynamics); therefore, I believe an alternative should be pursued.

A number of other people within this thread posted a variety of interesting ideas that I believe have merit, and so should be pursued, as alternative means for increasing liquidity without having to repeatedly market sell CoW at what I believe to be low prices.

However, in the event that someone obligated us to pursue this option, at the very least I would suggest making the selling threshold dynamic, and monotonically increasing over time at a rate to which the community agrees. In so doing, while the market sales would continue to distort market dynamics and also sell CoW tokens at a price that I consider far too low (both of which I would like to avoid), at least the in-effect ceiling being created by the program would be ever-rising, where perhaps there is a growth rate that would enable everyone to be reasonably comfortable (knowing that sales would only occur regularly if prices were also regularly rising throughout).

Thank you.

[octoshi]

Hi, I’m very excited to see this proposal, it reminds me how smart the team is and why COW is my biggest altcoin position.

I agree with CC here, liquidity is very important and is a key step in the future of COW, but there are better ways to do it than selling COW, especially when we are still very undervalued.

Sure you can think of smarter solutions than me, but we have a lot of money in the treasury, as well as low volume pairs that could be recycled into liquidity for COW.

Additionally, liquidity could be added little by little, as we reach certain price targets.

Thank you and let’s keep working smart!

MOOOOO

[koeppelmann]

Hey everyone - Martin from Gnosis here.
We are quite excited about CoW-AMM - we think it they could really bring back the promise that AMMs once had: that none-professionals can act as market makers.

---

Regarding set up strategies and liquidity: As Cowswap was initially developed within Gnosis - as part of the spin-out process - both GNO holders and GnosisDAO have been given COW tokens. We are now discussing how GnosisDAO can make better use of those tokens (and a bunch of other projects Gnosis holds token off) but also create long-term incentive alignment between those projects.

The proposal is to create larger AMM-pools (and ideally we can use COW-AMM) between GNO-tokens and projects token. I made a full post here.

My concrete proposal in this case is that both DAOs together create a COW/GNO AMM pool on Gnosis Chain. Gnosis would contribute COW tokens and COW can contribute its GNO tokens. While certainly traders want liquidity on Ethereum - the performance for the pool will be significantly better on CoW as on Ethereum gas costs heavily eat into LP profits. And of course - we (Gnosis) have way more interest in liquidity on Gnosis.

[AndreaC]

@Yakitori

I fail to understand the underlying value created by additional liquidity. We’re all very happy if we get new members of the DAO, but this doesn’t add underlying value to the project.

This is maybe where the problem is: a CIP is a technical proposal centered around what we want to do, and does not necessarily discuss why.

So, from a strategic perspective, the non-exhaustive list of the benefits of the proposal is:

1. expanding from the DEX aggregator market to the DEX market, which is orders of magnitude larger in terms of volumes, market cap, funds raised from VC, fees generated, and overall importance in the DEFI stack. We do so with FM-AMM, a product that, at least in theory, is far superior to the current standard (Uniswap v3). This product cannot be easily copied because it “works” only if trades are first batched. How about capturing parts of the biggest market in DeFI (that of DEXs)?
2. eliminating arbitrage profits and sandwich attacks eliminates 95% of MEV. Imagine what Ethereum would look like if almost all profits from MEV disappeared: no more builders, MEV bots, BPS. Again, we are bringing up a significant challenge to some of the biggest companies in Crypto.
3. a brand new mechanism for token distribution and liquidity provision (how many startups raise millions just to do that?!?)
4. more liquidity to the COW market,

To be honest, in comparison to these benefits, discussing whether we should sell 5M COW, or 4M, or 3M but also redeploy some other funds feels like a waste of time. But since we are at it:

• 5M feels like the right ballpark because it is real money and allows us to test the mechanism at a decent scale. At the same time, it is less than what we plan to save from other initiatives. So on net we will be selling fewer COWs than last year.
• We don’t want to tie these other initiatives directly with the deployment of FM-AMM. We don’t want to slow down FM-AMM because something unexpected happened with implementing these other initiatives. To move fast, we need to move in parallel.
• We will happily consider deploying other existing funds to FM-AMM once it is battle-tested. We don’t want to do it now because:

1. Shuffling funds already deployed to support the liquidity of the COW token doesn’t change much in terms of total liquidity. Also, we may run into situations in which we need to pause FM-AMM to change/upgrade some parts of the infrastructure. If FM-AMM contains funds previously deployed for liquidity purposes elsewhere, we may end up with less liquidity than now.
2. Using other funds that do not involve providing liquidity for COW opens a whole discussion about the DAOs optimal treasury management strategy. Again, we are happy to have this discussion in parallel but don’t want to wait for its concludion before moving on with FM-AMM.

CoWSwap has the chance to be the first to deploy a “MEV-capturing” AMM, something that has been discussed at length and would be a game changer in the DEFI space. We want to move fast without getting bogged down into unrelated discussions, such as whether this pool or this other pool has a higher API for the treasury (these discussions are important but can run in parallel).

cc: @cryptocondom

[Chris]

I have now read the entire discussion regarding the proposal. COW is undoubtedly one of the best protocols out there from a fundamental perspective. However, if there’s one thing I’ve learned in this space, it’s that price is a far more effective marketing tool than fundamentals, and good marketing is crucial for any project. In my opinion, any selling pressure on COW would be a mistake. There are enough interesting alternatives discussed in this forum regarding liquidity. Anything that hinders fair price discovery for COW would be a mistake when such a gem of a project is already very undervalued by the market. Andrea, you mentioned that “historically the price of COW is very high,” but history doesn’t matter in this case because it’s still a small project with a 60 million market cap, which is (and I believe everyone here agrees with that) still a significant undervaluation of this project.

[AndreaC]

@Chris , your logic is untenable: it would prevent the DAO from doing any fundraising! But Cowswap needs additional resources to reach its maximum value and not be eaten by competitors. Liquidity for FM AMM is just an example of this broader point.

[tanglin]

Completely understand (and deeply agree) with your describred implications and benefits of this CIP, and the need for acting fast! But selling 5M COW over 9-12 months doesn’t sound fast to me, taking into accout there are multiple sources of available liquidity. Why not use them instead to boost the DEX side (e.g, just by using the existing ETH in the GNO/COW pool we would be saving 2-3 months of time).

When it comes to selling the 5M COW, why not doing so OTC with some vesting period over 9 months (or maybe even without). We all agree on the fact that liquidity is an issue, but thats mainly due to:

• Current asset distribution inneficiency
• Low exposure from the DAO to the COW token, if we believe that COW is currently undervalue, which I think we all agree on, exposure should be increased
• Current low marketcap

So again, by focusing on porfolio rebalance and then evaluate the possibility of selling COW we would be takling the core issues orders of times faster (anywhere from 2-3 times to 9), in a more efficient way, and with way less impact on holders than with the current propositon.

I would be more than happy to add extra 1M COW to my holdings if there was a OTC or community raise to finance the new DEX model, but only after the treasury is properly balanced, otherwise I would be financing treasury inneficiencies

[Anna]

The scope of this CIP draft comprises a broad spectrum of advantages, a few already highlighted in detail by Andrea:

1. expanding from the DEX aggregator market to the DEX market, which is orders of magnitude larger in terms of volumes, market cap, funds raised from VC, fees generated, and overall importance in the DEFI stack. We do so with FM-AMM, a product that, at least in theory, is far superior to the current standard (Uniswap v3). This product cannot be easily copied because it “works” only if trades are first batched. How about capturing parts of the biggest market in DeFI (that of DEXs)?
2. eliminating arbitrage profits and sandwich attacks eliminates 95% of MEV. Imagine what Ethereum would look like if almost all profits from MEV disappeared: no more builders, MEV bots, PBS. Again, we are bringing up a significant challenge to some of the biggest companies in Crypto.

3. Showcasing a new tool for DAOs for automated liquidity provision - which is aligned with the overall strategic focus of CoW Protocol’s targeting of DAOs. Showcasing a use case for FM-AMMs for high liquidity holders is essential to bootstrap liquidity - which allows to move FM-AMMs from a great-in-theory project to a project with actual liquidity and adoption. Having a live implementation of the model is crucial for adoption by externals.

4. While CoW DAO is on track to cover current operational costs from it’s different fee models (CIP-34 to experiment with fee models passed two days ago; additionally a revenue model for MEV Blocker is being discussed) within the next 1-2 years, it would be beneficial for the DAO to collect further means of funding in order to scale it’s operations to guarantee being competitive in a fast-paced, highly competitive market. For this, a fund raising round could be envisioned for Q3 this year. Proper fund raising rounds have advantages: they allow for vested token releases, however, they also take time. In the meantime, while the token price is at double the valuation of the series-A round, selling some portion of the token can be used as safety net - nobody knows how the market develops.

5. Increasing liquidity of COW, whilst only making it accessible to CoW Protocol, ensures that COW related trading volume is fully captured by CoW Protocol, rather than competitor projects, which is still occasionally the case.

6. Increasing liquidity for COW.

It seems there’s consensus in this Forum on rolling out CoW FM-AMMs, the discussions largely evolves around the source of the liquidity:

Regarding the discussion to leverage GNO:

The idea to leverage CoW DAO’s existing GNO to bootstrap funding of the FM-AMM is a very short sighted idea in my opinion:

• GnosisDAO holds a significantly larger position in COW than CoWDAO holds in GNO
• GnosisDAO is contributing to COW liquidity
• GnosisDAO is continuously supporting the CoW DAO ecosystem, across all its product offerings, and is concretely considering bootstrapping the CoW FM-AMMs on GnosisChain by deploying liquidity on a broad spectrum of pools (more info here: https://forum.gnosis.io/t/token-holdings-of-gnosisdao/7985)

There’s a mutual understanding of alignment and partnership between CoW DAO and GnosisDAO that outweighs any short term gains from selling GNO.

Regarding the concerns of providing an unintended ceiling:
There’s increasing interest by investors to buy into COW, which large investors currently avoid doing due to missing liquidity. The assumption is that there’s more demand in COW than the size of the next fund raising round will comprise. Some interested parties are liquid funds and wouldn’t adhere to the standards of a proper fund raising round (i.e. long term vesting schedule). Passing on their investment due to missing liquidity of COW, is a lost opportunity.

Regarding the alternative to focus on CEX listing:
There are increasing efforts to get COW listed on top CEXes, however, it is too early to tell the outcome of respective efforts and the likelihood of getting listed correlates highly with the progress of CoW DAO’s roadmap and ability to deliver on innovative products.

While I agree with Andrea, that making small adjustments to the sell amount and the time period won’t have significant impact, I would consider following adjustments to address and reduce some of the community’s concerns:

• Given the current market price of COW, selling 5M COW with an implemented floor price that averages the sell price to at least 30 cents, would provide at least 1.5M USD in funding. Aiming at 1M USD and thus a 2M USD funding of the pool should be sufficient for the experiment, hence reducing the total amount of COWs sold to 3.5M and the total amount of COW used for this experiment to 7M should suffice.
• The optimistic time period of the current proposal covers 9 month in case the daily value of COW is continuously above a certain threshold. Preferably the experimentation period of this proposal has already ended by the time a new fund raising round is concluded, to display clear results and avoid any further concerns about token emissions related to this proposal. Reducing the experiment to a total of 7M COW would result in an - optimistic - 6 month experimentation period and thus align with a strategy to fund raise in Q3.

[Yakitori]

@AndreaC

This is maybe where the problem is: a CIP is a technical proposal centered around what we want to do, and does not necessarily discuss why.

@AndreaC

To be honest, in comparison to these benefits, discussing whether we should sell 5M COW, or 4M, or 3M but also redeploy some other funds feels like a waste of time.

The basis of fair and good governance is that you explain why you want to pass something and take into account the feedback if you want people to vote for it. Decentralisation is not insiders pushing covert agendas on voters that are required to be polite enough to vote it or leave.

In our case, those COW tokens are common property, so we’re fully entitled to understand how you came up with this amount and discuss if this is the right one. Governance prevents us from taking bad decisions, it’s not a “waste of time”.

Regarding the FM-AMM:

I think your proposal isn’t the right one if you want to deploy liquidity fast into the FM-AMM.

• It’s going to take 6 months to a year before we get “real money” available.
• You emphasize that we need to act fast.

Those two premices are contradictory. We could add the following:

• We’re going to buyback and burn an estimated 4M COWs in the next year.
• You want to sell 5M COWs in the next year.

Again, this is contradictory. Why not just keep the ETH?

And last:

• We shouldn’t open the subject about treasury management now
• We should sell assets from the treasury.

Also, if the FM-AMM is a great innovation (and I believe you on this!), I think it would make sense to use some funds from the treasury in it, which would be a great example of skin in the game. Especially if we want other DAOs to use it for their own liquidity uses.

Even if we have discussions about it, it will probably be much faster than your proposition, that would yield only $166k/month for the FM-AMM.

We’re also going to get fees from swaps. From my calculations, it would yield around 300 to 400k$ per month, depending on the fee collection method. This could be used as well.

@AndreaC

it would prevent the DAO from doing any fundraising!

COW doesn’t need fundraising as long as it has a treasury of 10M$. The Karpatkey report mentions that the DAO own 6.3M $COW, so the current proposal of selling 5/6th of it would clearly hamper any future fundraising efforts.

[Yakitori]

Thanks a lot for the context!

Just for the reference - COW DAO raised 23M$ in 2022 - is the Karpatkey report a truthful image of what is left, or are there more assets available?

Overall, I’d like to hear your thoughts about this:

There are currently 1k ETH in the treasury (again Karpatkey’s numbers). Why not taking half of it to bootstrap the FM-AMM now?

• We’d be able to deploy fast, and earn the fees associated.
• Prospective investors would be buying COW on the market and the DAO would earn the ETH.
• It would show that we eat our own dog food and trust our devs (the bulls that lead the cows! )

Just as a sidenote, COW price has been doing +50%/-30% swings in the last month. I don’t think that using the current one ($0.3) as a reference for action is right, as it could go back to goblin town fast given the wild volatility.