Use Treasury Funds to buy Convex ($CVX) rather than buying CRV for Bribes


Convex Finance is the synergistic flywheel of Curve Finance. It is composed of anonymous CRV developers, at least several of which are closely aligned with CRV, including Charlie Watkins, the chief dev at CRV.

CREAM is presently contributing to Andre Cronje’s bribe project on Curve. Essentially, this is resulting in a biweekly payment to any veCRV wallet addresses who vote for the CREAM v2 gauges.

The bribe protocol is currently able to be followed here:

Only 1,309,219.18 veCRV (or 0.49% of all veCRV at 267,394,811.02 total) participated this week in the Bribe protocol. Soon, the number of veCRV eligible to participate in Bribes will increase dramatically on 9/16/21 when Convex initiated voting rights for locked CVX.

Only locked CVX can vote for a gauge pool and carries a ratio of voting rights commiserate with the amount of CRV locked into the convex platform.

i.e. If 30% of the 21million circulating CVX are locked, then these 7 million CVX would control all presently locked cvxCRV (90.5m). As such, 1 CVX would carry the voting power of 12.9 veCRV.

Convex Finance controls 33% Curve Finance governance by virtue of perpetually locking 90.56 million veCRV as cvxCRV into its platform. Because of this, 1 CVX controls 4.6 veCRV at present. TVL on CVX is 7.7Billion, contributing to 54% of the TVL on Curve. Recently, CRV has announced its intention to expand to AVAX, Sol, Arbitrum, Celo. CRV is already on Polygon, Fantom, Ethereum and xDai. Essentially, our participation in gauge voting has supported our belief that CRV is the backbone of defi and likely to continue to outperform other, similar protocols.

By virtue of it’s flywheel, Convex finance is making a mean of 1million per day in protocol fees since its inception. Vote locked CVX will be eligible to vote from Sept 15th 2021 and onward. APY on vote locked CVX is expected to be 20%+. This would provide the treasury with a source of income from its investment as well as a guaranteed biweekly vote for the CREAM v2 related gauges.


I propose that the CREAM treasury allocate funds to CVX purchase and vote lock rather than paying for the Bribe Protocol. It is important to note that this proposal is NOT suggesting CREAM build on the Convex flywheel. That comes with complex tokenomics issues that are not applicable to this proposal.


This proposal is strictly to suggest that the treasury allocate funds to buy CRV voting rights thru the purchase of CVX. This would benefit CREAM as follows:

  • Perpetual, weekly GUARANTEED voting rights for the CREAM v2 gauges by an expected ration of 1 CVX controls the vote of approximately 12.9CRV.
  • Buy the purchase of CVX, CREAM would be able to diversify its treasury and receive income from yield generating, vote locked CVX on the order of 20%+


Risks associated with this proposal include the fact that CVX may not lock as much CRV in the future as predicted, essentially diluting future voting power for token holders. Based on their flywheel Tokenomics with high APY in CRV/CVX and 3CRV rewards, this is unlikely but a possibility.

Another risk is that this is a long-term commitment to CRV gauge voting and if CRV ever voted to stop gauge voting, the value of CVX to the platform would be diminished. I think this is unlikely since CRV’s plan all along has been to provide permissionless, v2 gauges with veCRV based governance…but its still a risk.


I am very much in favor of this strategy. We have fairly substantial reserves we can use to buy CVX without even selling CREAM tokens directly. There are plenty of advantages to this proposal on top of buying cheap CRV votes we can get a high growth asset on our balance sheet, earn yield from staking CVX, and build a relationship with one of the biggest holders of crvIB pool tokens and the top asset management platform (sry YFI).

Disclosure: I bought CVX last week specifically because of @afirebrand 's proposal and i am a big believer in CRV + CREAM and want to make sure we implement this strategy, even if it’s not from directly from CREAM treasury (which we should do too)

Reposting my response to cream team from discord here:

We don’t have to use/sell CREAM we can buy with existing reserves which are in the millions of dollars i believe

  • How does this investment translate into the protocol fee → Ironbank is our biggest market and 2nd biggest revenue generator. Using CVX to boost crvIB pool will increase liquidity and TVL and therefore fees to CREAM stakers
  • For every $1 spent in CREAM, how much incremental fee will be generated? → right now locked CVX earns 16% APY [1] so at a minimum it earns $0.16 per $1 spent (until we get diluted by other DAOs). Then you have to account for additional fees from boosting crvIB pool, which increases TVL and protocol usage.
  • We already spent $100k to bribe CRV pool [2], I didn’t see any analysis on expected results nor a DAO vote before going forward. Why the double standard when he’s given more public/thoughtful discussion with community than cream team themselves? Bribing is purely a loss while buying CVX earns yield and can have price appreciation, its a no brainer if we are experimenting with new incentive mechanisms.
  • For the same spend efficiency of liquidity incentives, are there any other better alternatives? → Related to above, i haven’t seen data on effectiveness of CREAM’s CRV bribing, however there is data on the effectiveness on buying CVX. Right now 1 CVX = 14 CRV of voting power [3] so ~$1M of CVX has same voting power as all CRV currently voting on crvIB [4] which is valued at ~$35m, a 35x more effective expenditure than buying raw CRV and locking it up.
  • Can we quantify the impact of risk mentioned in the proposal? What happens if CVX price takes a nose-dive? → it doesnt matter. We still control the underlying CRV that votes to boost our crvIB pool, CVX price doesnt matter.
  • Is the reward worth the risk taken here? → absolutely. We just yeeted $100k into the air, why not spend $100k buying yield bearing assets that sit on our balance sheet?

[3] Dune Analytics

To address a few of the points that @Eason brought up:

-A lot of protocols are doing this because CVX is a yield baring asset which has governance control of CREAM in order to incentivize the gauges of their native tokens. In this case, these protocols, like CREAM itself, have participated in the bribes.crv, and noted a commiserate increase in gauge APY secondary to bribe participation. For example, on 9/4/21, $SPELL increased their weekly bribe to $400k secondary to the fact that they saw such a marked increase in LP return based on their incentivized gauge weight

-I would suggest we use the same amount of CREAM on a CVX investment that was previously used unilaterally for bribes. Only this time, the community is choosing to participate in this investment and supports the allocation of funds in anticipation of future treasury revenue from this yield baring asset as well as from token appreciation itself

-Is the risk worth the reward in this case? Well, I would argue yes since the protocol has already spent 1million on this exact same cause with nothing to show for it other than a few weeks gauge incentivization. By spending the same amount of money to buy our own gauge votes, we can incentivize gauges that prioritize CREAM LP pools in perpeteum.

Hey Kiba, afirebrand,

Thank you for the proposal and response! One thing about the proposal that’s not clear to me is: How much CVX do you propose to purchase and what is the lock-up duration for it?

Essentially, this proposal is equivalent to committing a supply-only liquidity incentive for a certain period of time. The key(or a fairly easily overlooked thing) I want to point out here is that Liquidity & TVL aren’t equal to fees. Fees come from borrows. By locking into CVX long-term meaning that your assumption is the borrow will always go up proportionally with the TVL growth. In general, for a double-sided marketplace, locking resources into 1 side of it is a quite risky thing to do, especially in such a dynamic environment. The lack of flexibility here will potentially make us fall into situations where borrow isn’t growing while we continue to spend incentives on supply. We will also get stuck when the spend efficiency drops, i.e. there’s nothing we can do when the marginal cost of CRV incentive gets too high as our resources have already bought the vote.

In terms of CVX APY, the opportunity cost should also be taken into consideration if the idea is to use treasury to buy it. Currently, the asset in treasury is already earning yields so the real additional yield from CVX APY would be the difference in yields plus the variance in CVX values(which would bring additional risk to the balance sheet value). If the goal is to have yield-bearing assets on our balance sheet then I believe there are other options with better risk-adjusted returns.

Again, I want to thank both of you to initiate the discussion here. Please do understand that I am not trying to get people to vote against the proposal. I am simply just raising questions/concerns and requesting more data to ensure that we have all angles covered before making decisions.

How much do you propose to invest and for how long → what is the expected impact in terms of both supply & borrow → how does that compare to having an equal amount of liquid asset at hand → what’s the loss/risk when the marginal cost gets too high.

Appreciate the response. I would propose the treasury purchase up to $500k of $CVX and lock the funds for the duration of 4 months. This is the minimum mandatory locking period for vote locked CVX and would result in a minimal impact on treasury liquidity at hand.

At the present valuation, $500k of $CVX (presently trading at $12.40) would purchase CREAM the right to approximately 40,000 CVX. This is equivalent to the voting rights of 520,000 veCRV as noted by the present ratio of 1 vote locked CVX controlling 13 veCRV votes.
Reference: Dune Analytics

Although your points on liquidity and TVL are well noted, I would note that CREAM’s own TVL is significantly less than Convex’s 8billion TVL and Curve’s 12.5billion TVL. I think most individuals in DeFI are very well aware that TVL can be liquid and that this number is necessarily correlated to fees. That said, the growth of Convex and Curve in TVL is undeniable. In the space of 4 months (121 days as of today), the Convex protocol has:

  1. Locked 8B in TVL
  2. Achieved protocol revenue of 130million +
  3. Locked 100m CRV into the Convex platform as cvxCRV

I think we can all agree that short and midterm growth of the Curve and Convex ecosystems is likely to be positive as multiple near-term catalysts exist, including:

  1. Launch of Curve on Arbitrum
  2. Launch of Curve on Fantom
  3. Launch of vote incentives/bribes for vote locked CVX on 9/16/21
  4. Curve v2 Pools launch

By purchase of CVX, we gain the opportunity to obtain an appreciating asset with exposure to both CRV and CVX tokens thru protocol revenue rewards paid at 13% APY (paid as cvxCRV) which can then be re-staked for 69% APY. As previously noted, the point of purchasing CVX is NOT strictly for yield baring income. Yes, there are definitely better assets with a higher APY. However, there are NO other assets with as much power to control the Curve gauge weight as much as the vote locked Convex governance token has. This is the reason why MIM and Alchemix have both passed proposals to purchase CVX for their own balance sheets. This is also why Yearn has stopped dumping their yield harvested CVX and why Synthetix, FRAX, and others are also considering proposals to do the exact same thing…just like we are.

I’d leave you with a final thought to consider. Adam Cochrane, a well known crypto influencer and venture capitalist noted Monday that Convex Finance is a “Kingmaker” because of its vote locked control over Curve gauge weights.

As such, if we believe that promotion and incentivization of our Iron Bank LP pools is important for CREAM growth, then I think that CVX + CREAM is a natural progression of CRV bribes + CREAM.

Thanks for the response @eason. It would be very helpful if the CREAM team could provide the report/analysis they used for bribing to further this conversation. From the community’s perspective we have no idea what went behind that initial decision nor the results of the bribing program which should already be available since it only lasted one week and happened a month ago. All we have is the information available to us, none of which is from the cream team. Transparency helps us work better together :slight_smile:

Speaking for myself, I’m still evaluating if it is a good idea so haven’t proposed specific numbers yet.

Agreed. It would be great to hear how this factored into the CREAM teams decision when they did bribing because the same argument could be made against that.

I sent a proposal directly to CREAM team to start doing DAO to DAO loans a week or two before the Pleasr deal and proposed using IronBank for a DAO loan at Index Coop 7 months ago. This is a huge area for demand, we could be issuing millions of dollars of loans to high quality DAOs driving up utilization. PoolTogether and IndexCoop are two notable DAOs that are ready to pull the trigger once they have a finalized offer they can bring to their DAOs for voting.

I didn’t get much of a response from the CREAM team so we’re developing this on other lending protocols. A proposal for a dedicated Fuse pool is going up soon and im working on getting an Aave grant to build a market for DAO loans. We already have two specialized DAOs to contract out credit rating analysis and accounting reports on loans. I would love to do it through IronBank but I don’t feel we’ve been given the opportunity to do so. Huge opportunity to grow borrow side passing right by CREAM.

There are plenty of other ways to increase usage on CREAM like integrating with wallets like Zapper, Zerion, InstaDapp, Furucombo, etc. That takes a lot of dev work and we have a massive treasury to fund that through bounties/grants if we want to. Getting direct integrations from other DeFi protocols to build on top is another way, IronBank also provides a great opportunity for that with p2p undercollateralized lending.

Basically my response is “do all the things” and I’m here to help you guys if you ever want it <3

I mentioned yield but as @afirebrand mentioned, this as an investment in protocol growth not a treasury diversification play. We were asked what the expected ROI is and that’s unknowable until we do it, same as with the bribing scheme. The main difference is we have the ability to recover funds if our CVX experiment fails and we have some baseline returns from farming where as we are guaranteed complete loss of funds from bribing. Risk/reward profile for bribing, from the info available to me, is definitely worse than buying CVX since you can’t even guarantee you get any CRV votes for the money you spend. If you can justify the bribing scheme to yourselves I don’t see how CVX is any different or worse.

We do have $100m of CREAM in the treasury so even a substantial investment would be a fraction of our treasury value,. Market selling is an issue unless we find an OTC buyer. Preferably we borrow against our CREAM and take a loan on IB to fund this but that reduces overall yield and has some risk of liquidation. IronBank has 5,500 CREAM left in the collateral cap which is ~$1m and stables have 5% APY borrow rate which seems like a pretty reasonable bet to make given the potential. Basically 0 change of getting liquidated since we could post $20m in collateral making liquidation price $16(?) and the CRV APY alone would cover interest payments with CVX APY being profit. Then factor in our actual CREAM protocol growth/fees from using CVX to vote for crvIB incentives and the risk/reward is dramatically in our favor.

Side Comment

At the moment I think we’ve lost the lending protocol game on mainnet and all L2s. We have less capital, less integrations, less reputation, less community, and less liquidity than the majority of our competitors. What that does give us is more potential to experiment and push the boundaries. We HAVE to innovate and be first movers in order to stay competitive whether thats at the product level with IronBank undercollateralized loans or the economic level with schemes like bribing/CVX. Are there risks to doing this CVX deal? Definitely. But there is a much bigger risk if we don’t do anything and lose the edge we do have. Thats a big reason why I proposed the Listing Committee, our v1 edge was long tail assets and we were falling behind on that. In my opinion, IronBank has the most potential and pushing on that as hard as possible is the best way forward for CREAM. CVX is a very reasonable bet to ensure our best product can succeed in a hyper competitive market.

Thanks! Can you update this into your proposal post? Having the numbers is really helpful. I think then here I can run some math to help the community project/estimate the additional fee that would be generated from the incremental crvIB pool over the 4 month period:

  • Currently, the borrow APY of Iron Bank DAI/USDC/USDT are around 5%. Reserve factors are 10% for all of them.
  • Assuming that the borrow volume will increase at the same rate as growth rate of the supply(as the result of additional reward because now we get to vote for it)
  • In order to generate an additional $500K protocol fees over the 4 month period, we will then need
    `$500K / 10% Reserve Factor / 5% Borrow APY * (12 months / 4 months)
    = average $300M increase in supply for the whole 4 month period to just break even.

This means that with the $500K purchase of CVX, the votes we get will need to boost the crvIB pool growth by ~80% AND the borrow volume will need to grow by 80%(or roughly another $120M) just to not have a negative return on this investment. So my remaining question will be:

  • How much reward APY boost do you think these additional votes will get us?
  • Does that look like it can increase the volume by 80% for the whole period of time?

Update - should've also taken CVX price into consideration here. The net cost should be the (CVX value after 4 months lock up ends) - $500K. I think then this would've made our expectation on the return more reasonable. We can see the $300M number above as the maximum TVL increase required to break even.

Meanwhile, I do agree that we need to innovate, experiment and push the boundaries. However, from the information presented here, this doesn’t feel like an innovation per se as other protocols are already doing this. Also, I think that the fact you mentioned that we are a protocol with less capital is exactly the reason why we should be extremely cautious about going into a play that favors parties with more capital.

Thanks, the forum does not permit me to edit the original proposal or I would add the proposed amount.

As I’m sure you are aware, there is no way to guarantee that a consistent APY boost can be achieved for the whole 4 months of locking because the amount of tokens locked is variable from day to day. At present, 8 million CVX, or 32.5% of all circulating CVX tokens are locked for the first voting epoch. If more tokens are locked, then the amount of APY achievable would indeed decrease.

Yearn has consistently been using their yveCRV votes to boost the IB LP pools. As such, any CREAM APY that is generated by CVX voting would only serve to augment the yveCRV in a synergistic capacity. I’m not sure why you think that the entire $500k needs to be ROI’ed within the duration of the CVX 4 months lock period. Locking can and would be renewed to continue to generate yield and voting rights for the protocol. With respect, that’s ridiculous and absurd.

This is a long term investment in CREAM’s future. If we did not believe in the CRV + CREAM story, or the CVX + CREAM synergy…then why did we participate (unilaterally I might add) in the first place??

Got it. I see your point now - I guess this is more of a philosophical difference between you and I about investment. For me, it’s much more ridiculous and absurd not to think clearly about potential return(or at least how to measure it) for any significant amount of investment. Even if we don’t want to drill down to fees generated, we should at least have some expectations around:

  • How much reward APY boost do you think these additional votes will get us?
  • How much TVL boost do you think it can generate?

That way, we will know how to asses the result and decide whether to continue the investment(assuming that the proposal is passed for the $500K / 4-month period) after the first epoch ends.

Absolutely agree that return on investment MUST be considered for the good of the protocol. That said, if you find ANY investment that can pay 500k back in 4 months, please feel free to let me know because i need to switch over my 401k! Seriously though, 13% APY in interest baring CRV yield will be able to pay back the loan to ourselves in a rapid manner.

500k at 13% APY will more than cover the 5% cost of our loan terms even if we just farmed and dumped the CRV we harvested (which I would not recommend as this too is an appreciating asset which could then be restaked for 70% APY to net more CVX and CRV).

Furthermore, the voting right we would achieve for owning CVX would permit us to achieve an approximately 2.5% increase in CRV for IB pools ontop of any additional boost that yveCRV voting would give us thanks to our control of $ Bought / CVX price * CRV per CVX (see @Kiba’s most excellent calculations from the CURVE discord governance posts).

Lastly, I’m not proposing that we buy $500k/4month period. I’m proposing a 1 time $500k buy of CVX. As you previously noted, we have to mitigate risk and a one time buy would allow us to assess the investment over a defined period of time and then consider if future buys would be viable and profitable to the protocol.