Upcoming COMP farming change could impact the Dai peg

I have been watching the evolution of COMP farming extremely closely over the last two weeks, and I’m concerned that an upcoming governance change to their protocol could significantly affect the Maker ecosystem. On Friday the balance of power among their collateral assets is going to change dramatically

Up until now, Dai has been left out of the COMP farming craze. This is because COMP is currently rewarded to farmers paying/earning the most amount of interest. Because the BAT rate maxes out very high (~37%), BAT farming has been exceedingly popular. By comparison, at 100% utilization rate, Dai only pays 15% interest.

Under the new rules, COMP will be allocated to every dollar borrowed, as opposed to interest paid. This means that farmers will now try to borrow the most amount of money, and will naturally try to pay the least amount of interest. My expectation is that the two most popular farming assets will be USDC and Dai due to the shapes of their (attractive) interest rate curves. While USDC is easier and safer to farm (fixed oracle price of 1, larger supply, larger potential supply, greater liquidity), Dai will earn more COMP up to a 98% utilization rate. Above 98%, USDC does technically become slightly more attractive but I think that is somewhat irrelevant.

DAI Interest rate curve

USDC Interest rate curve

There are more or less two types of COMP farmers: natural Dai holders and natural ETH holders.

  1. Natural Dai holders currently hold roughly 130 million Dai (total Dai supply). The optimal tactic for these people is to deposit Dai into Compound, borrow Dai, re-deposit, etc. This recursive loop can gain up to 4x leverage. In other words, 1 million Dai can create 4 million cDai (earning plenty of COMP). Additionally, there is only a very tiny liquidity / liquidation risk here. As long as you cover your interest rate payments, you bear basically no price risk.
  2. Many farmers are natural ETH holders. They will first deposit ETH, borrow Dai, and then re-supply and go through the loop. These farmers carry liquidation risk (if ETH crashes, or if Dai spikes up). If they are also re-suppliers (what I call ‘double-dippers’) then they also face liquidity risk (cannot withdraw Dai if the utilization rate stays at 100%). Single-dippers (merely depositing ETH and borrowing Dai, without re-supplying), can always pay back their loan since they have the Dai sitting in their wallet. Farming is still extremely profitable for single-dippers.

Due to the high APY for COMP farming (circa ~70-100% APY), there is a chance (likelihood, even) that we see an unprecedented demand for Dai. Much of the natural supply for Dai could also be locked up in COMP farming, thinning out sell-side orderbooks. It is almost certainly more profitable to farm COMP than to sell Dai for a small premium. This could break the peg sharply. Even the creation of new supply (generated through Maker) might go to COMP farming rather than selling. All of the debt ceilings might instantly be maxed out (except maybe USDC-B).

Does the community generally agree with the above analysis? I’d really like to get some other perspectives. My first instinct is that the typical response to such an event would be the threat of Emergency Shutdown. Only if Dai holders are at risk of taking an instant loss (if they are purchasing Dai significantly above $1) might they think twice about undertaking such an activity. But I think we should explore other options such as special risk parameters for ETH or USDC (lower LR, higher DC), or maybe even reach out to and participate in the Compound governance process. It would be ideal if we could have more than the 4 remaining days to plan for what’s about to come.

Thank you to Will Price for walking me through these scenarios.


You can’t borrow the same asset you deposit or use as collateral on Compound

I am curious to know why DAI rate curve is so flat compared to other stable coins in the first place. But even if there is a good reason for that, all rate curves including other parameters should be studied upfront when deciding for a new COMP incentivization plan. It feels to me Compound is rushing to fix one thing and making things worse elsewhere. Fundamentally COMP incentivization plan has a problem because Compound enables recursive leverage building with the same asset.


Yes, you certainly can as long as its an approved collateral asset. To my knowledge only W-BTC and USDT are ineligible.


No you can’t, you should go to compound interface and try it out. You can’t borrow the same asset you are using as collateral

Won’t that just incentivize arbitragers to mint more DAI with the available collaterals and sell for an instant 1-10% profit depending on how much the peg slips? Eventually this would attract so much attention to MakerDAO that everyone would rush to lock up more collateral to mint DAI. I think this is a different scenario than Black Thursday in that minters will actually come this time as opposed to everyone trying to up their CR. This isn’t a fear induced situation like in March so I’m not as concerned, although it’s a great point to bring up and definitely should be talked about. What would your proposal be to mitigate this from happening?


I’m concerned that people will short DAI by borrowing on Compound (and earning positive APY) rather than coming to Maker.

1 Like

You can their interface doesn’t allow it but you can do some console hacking to get it to work, or just interact directly with the smart contracts.

@cyrus I’m very concerned by the points that you raised. What incentive is there for market-makers to mint Dai and sell at a premium if the Compound yield has a significantly higher expected return.


It’s possible to do it not through compound UI but you can write a smart contract that does it and lot of people are using the 4x loop to maximize comp earnings as he mentioned above

1 Like

They need to sell the DAI they borrow in order to short it…

I agree with @Aaron_Bartsch. An instant 5-10% profit opportunity (if DAI is at $1.05-1.10) is much more attractive than farming COMP (would need to farm ~15-30 days for that kind of return). It’s just important that the Debt Ceilings are high enough, especially for ETH.


I’m not so sure. In order to have confidence in that trade, you have to have an expected time horizon for Dai to return to the peg. If you can make 10% guaranteed over the course of a month, wouldn’t you rather have that than 10% instantly and hoping the peg returns within a month? I think the upper bound on the peg here is significantly higher than $1.05-$1.10.


as part of the new patch “The function has also been updated to require an external account (not a smart contract), reducing the risk of griefing” so maximising COMP reward using leveraged borrowing won’t be an option any longer since smart contract address won’t be considered to receive a reward


Something like this?
Mint Dai > sell for USDC (lock in profit) > farm COMP with USDC

1 Like

People can borrow DAI on compound and supply USDC on compound to yield farm too. You’re earning yield COMP farming and you can make bonus on the DAI short trade.


Would it make sense to provide a high leverage ETH-B ramp with a higher Risk Premium to take on these farming users without clobbering the ETH-A DC? Presumably these farmers want as much Dai as possible so they may prefer ETH-B at the higher rate.

1 Like

I’m not sure if ETH-B is the answer. That changes the problem from a slightly risky peg, to actual system risk. Can the current liquidation system handle ETH-B?

Anyways I do agree that there should be some debt ceiling increases in place to absorb some of the increased DAI demand. Though as @monet-supply said, they might just prefer to short DAI on compound.

1 Like

True, I see your point… If there is no confidence that the peg returns within a reasonable time frame, I’d much prefer the COMP return. Especially since my capital remains liquid (unless utilization is at 100%…) and I can move to better opportunities if they open up .

Btw. there is a new DAI interest model in the works - the formal proposal should come tomorrow. Here is the discussion: https://compound.comradery.io/post/1783
All the proposed changes should help, especially setting the DAI APY at 80% (instead of 90%) to 4%.

Also, I expect a lot of farmers to just supply and borrow ETH - to remove the liquidation risk. The interest spread between borrowing/supplying is between 2-4%, so saving a percentage point on interest payment by borrowing DAI is probably not worth the additional risk. Even more importantly, I can borrow close to 75% by staying in ETH, whereas with DAI I’d probably stay well below 60%
Edit: Well not “a lot of farmers” since you can’t borrow the asset you’re supplying through the UI… but probably the bigger farmers.


So one of the main fears is people buying up a lot of DAI to earn the high rate. Though it’s questionable how much stablecoin holders would bid up DAI to COMP farm.

Yes with 1m DAI you can you can borrow and lend another 3m DAI.

But with 1m USDC you can also borrow and lend 3m DAI. And if you have 1.5 m USDC collateral, you only get liquidated if the peg goes to $1.5. (Basically you just tone down the leverage a bit and still don’t have liquidation risk)


Yeah good point @vires-in-numeris.

It may make less sense to COMP mine with DAI if you have ETH collateral.

Can use $1m ETH to borrow $3m ETH.

Or you could use it to borrow 1.5m DAI and lend 1.5m DAI. (At $1.0 peg, liquidation if ETH drops -33.33%, At $1.05 peg, if ETH drops -30%)

I’d sleep better at night if I didn’t need to worry about ETH dropping 30%.

1 Like

(This is my personal opinion and I may be completely wrong!). The safe way to leverage yield farming is never with DAI, because DAI can spike up and your leverage position at compound could get liquidated, currently their oracle has a 1.05 DAI/USD top limit. This could be changed and leverage positions would get liquidated. The correct way in my opinion, is possibly take ETH/WBTC/BAT mint DAI --> sell for USDC which is always $1, lever up at compound to earn COMP. This might be a giant gift for MKR token holders because they will need to raise interest rates to match this new demand for minting DAI. Defisaver and Instadapp who are helping with this leveraged positions, need to help unwind the whole trade if ETH/WBTC/BAT are falling. So again if ETH/WBTC/BAT fall in price…, there is a DAI demand problem with closing out vaults, so USDC vaults debt ceiling might need to be raised to handle DAI spike in price when volatile crypto falls… Lots of variables at play, would love to hear more analysis from others, but overall I think this might be a positive event for Maker.