[G-UNI DAI / USDC 0.01% tier] Onboard G-UNI for UniV3's new 0.01% fee tier

Summary:

Onboard a G-UNI DAI / USDC 0.01% fee tier as Maker Collateral type to stay competitive as Uniswap V3 LP on this token pair. Recently Uniswap onboarded the 0.01% fee tier specifically targeted at stablecoin pairs and increasing stablecoin swap traffic to Uniswap. Volume is quickly migrating away from the 0.05% tier to the 0.01% tier for DAI/USDC – on this pair there is very little slippage for most trades across both of these tiers, so fee amount is what mostly effects the execution price. Over the last 48hrs the 0.01% fee tier has already slipped the volume of the 0.05% tier (even with 20x less liquidity) and over last 24 hrs is earning more in fees (even though fees are 5x less). Long term it seems extremely likely that most stablecoin volume will be on the 0.01% fee tier, so Maker’s leveraged LPing should definitely be moving there in my view.

G-UNI DAI/USDC 0.01% tier deployed here

Proposed tick range:

The 0.01% fee tier has a tick spacing of 1 tick providing maximal flexibility of range selection. So what fixed long term range should we set for the G-UNI Pool? Tick -276324 is roughly the center point with a price of 1.00000264384 DAI = 1 USDC. The trade-off is we want to be as concentrated as possible (to earn maximal fees while in range) while never/infrequently going out of range (to not lose out on fees altogether). My proposal would be a (-276322, -27626) 4 tick range which represents a price spread of 0.9998 - 1.0002

Everything else should remain the same for this new G-UNI pool as with the original 0.05% fee tier pool (no changes to implementation etc.).

Onboard G-UNI DAI/USDC 0.01% pool ?

  • Yes
  • No
  • Abstain

0 voters

Next Steps
This poll will end on 2021-11-25T00:00:00Z. If successful, an onboarding poll will be created which will be put on-chain the following Monday. -@LongForWisdom

5 Likes

I am actually not sure if we are doing anybody a favor if we onboard this - given the amount of gas one needs to spend for opening a vault. Does anybody have run some numbers on the expected fees or on how much you need to deposit to have this profitable without committing to an unreasonable time?

3 Likes

Made edits to fix some stuff, apologies for the inconvenience, please vote again.

I’m no expert, but I recently read Rekt - Uniswap V3 LP - REKT

3 Likes

It will be similar to the 5bps pool, but less fees. Volume seems to be much higher though, so perhaps we can get similar TVL. It’s worthwhile to add this as I agree most volume will run through this pool especially as the TVL keeps growing.

5 Likes

I mean technically users can minimize gas using tools like InstadApp to flashloan large amount

The 0.05% fee tier for G-UNI was highly profitable before the introduction of the 0.01% fee tier. It averaged a 4-5% return with a 20 tick range (and with 20x leverage it was around a 100% return on principal). The fees are reduced by 5x on the new fee tier, but also the concentration is increased by 5x with a 4 tick range. Also, volume should increase as this pool has much more competitive rates and should attract more stablecoin volume to Uniswap. Even if APR is less than half of what it was on the 5bps fee tier, with 50x leverage it should still be a decent return.

4 Likes

I was reading around on Gelatio network regarding their methods etc. I have not used their system simply because I have my own v3 modelling and approaches.

My question here relates to the spread on the G3UNI we are onboarding. What happens if the DAI USDC price exits the Gelatio spread (say to the downside). Do we end up potentially with G3UNIV3USDC-DAI vault with collateral under water price wise? I guess with LR of 102 we should be ok. I am just concerned about the situation this could put vault holders in that leverage.

2 Likes

The 0.01% fee pool has already taken over most swap volume within a week of launch, and with only $8 million in liquidity versus $250 million for the 0.05% fee pool.

It’s probably inevitable that all swap volume moves to the low fee pool, which would eliminate demand for using our existing vault (users would face a negative return). So I think the sooner we can focus on the lower fee tier pools, the better.

4 Likes

Hi Frank here from Oasis.app :wave:

While I agree with the notion that the 5bps pool will probably not reach much volume anymore, I highly doubt if the 1bps pool will gather enough fees for the integration with maker to make sense.

I’ve made a spreadsheet with some basic calculations, made here:

For a user of this 1 bps pool, through GUNI & Maker with the assumptions in the sheet, it would take quite a high initial deposit amount, to have a relatively quick break-even point of the fixed costs (gas costs for the user).

Assuming:

500M total pool size
10k fees per day on average (2x the current amount)

Then the non leveraged yearly APR is 0.73%. With a stability fee of 0.5%, the net non leveraged yearly APR is 0.23%

With a 50x multiple, the net yearly APR excluding gas costs is 11.50%

Including gas costs this means with 10k initial deposit: 963 days till break-even
with 100k initial deposit: 96 days.

I would like to be wrong, but I think this race to the bottom is not benefiting anyone in the long run, also not Uniswap users.

8 Likes

11.50% APR is still acceptable for many people.

also, your assumption of 500M total pool size is fair, but in practice today even if you assume ALL current TVL in the 5bps pool moves to the 1bps pool, we still only get to ~$300M total pool size which gets us closer to like ~30%-40% APR, which is definitely acceptable for many people.

especially if we allow some flexibility in ranges, this could work well (eg; support a tighter-range and wider-range GUNI pool).

one step even better, skip GUNI and allow any UNIV3 NFTs that have their ticks within a certain range bounds directly, giving people full individualized flexibility on their range.

1 Like

Thanks Frank for the thorough report. It seems the major two factors are:

  1. The amount of volume we might expect given a migration of most of the liquidity to the new 1bps tier
  2. The size of the stability fee

I am curious about (1) in that I think it is possible we really will see a significant bump in volume there- though this is pure conjecture. Just noting that this highly competitive fee tier may attract new volume to uniswap stablecoin pairs.

(2) is not up to me to decide though I suppose it is plausible that the Stability Fee is again lowered at some point in the future to make the margins better? (it just recently lowered from 1% to 0.5%)

Given this, and also given that for larger holders or participants opening the position for very long time frames 11.5% APR might still be worthwhile, I still see the benefit in migrating the G-UNI position sooner rather than later.

Also note that if the 1bps tier really does take the lion’s share of the volume, then the current GUNIV3DAIUSDC1-A on the 5bps fee tier will probably have even worse returns than your projections for 1bps (The 5bps pool could well up with less than 1/5th of the volume of the 1bps pool which would mean less total fees captured by staying put rather than migrating).

not sure what the benefit of individualized ranges would be. G-UNI can theoretically be set as tightly as possible, plus it adds the benefit of auto compounding the fees for you, which for these type of stable pools is especially attractive. Definitely worth the small gas overhead of entering the pool.

1 Like

Lets add the fact that PSM tin/tout=0 will basically funnel large transactions through the PSM.

Put simply the race to the bottom here with Maker putting large scale fees effectively to 0 is going to squeeze trading liquidity to the margins and fee compression will effectively squeeze out liquidity.

The numbers being posited today:

are not going to be the same tomorrow.

Add another fact. I read that Gelatio fees are 0% which are due to expire end of 2021.

Thanks a lot for the calcs Frank, your sheet is very helpful! Nicely done :white_check_mark:

However, I would say the assumptions here are debatable. Firstly, I don’t think that having $500M in liquidity for this pool is realistic given the current volume and only assuming a 2x in fees over the long run. At some point we will find an equilibrium, probably around the 200-300M mark where it does not make much sense to add more liquidity because other options for stable lending are more attractive and less gas intensive. By that time, the DAI / USDC pool will consist entirely by G-UNI MakerDAO LPs, all others would have already left the pool.

Secondly, 10k initial investment for earning $ on stables on Ethereum is certainly also fairly low these days given current gas prices. For these type of principals, layer 2s like Arbitrum will most likely be more attractive to earn a passive yield. I see where you are coming from, wanting a wider range of smaller users to participate as well, however this I think is probably not possible on Ethereum in the short to medium term, especially since I assume gas prices won’t go down any time soon.

If we change these assumptions slightly, e.g. $50k principal and $200M TVL, then you can get to a Net yearly APR excluding gas of 66.25% and 33.44 days to break even (incl proxy), which imo sounds reasonable (this assumes fees will not exceed $10k, which I think they probably will over the next weeks).

Also, I think it might be worth discussion lowering the stability fee to somewhere around 0.25% to make this collateral type more attractive. This would offer a net APR of 78.75% with 28 days to break even.

3 Likes

Add another fact. I read that Gelatio fees are 0% which are due to expire end of 2021.

This is not really relevant for G-UNI, this is for Gelato’s general keeper as a service product. The G-UNI pool in discussion simply take a 1% cut of all earned fees, which in the grand schemes does not affect the APR too much.

2 Likes

On a cursory read it wasn’t clear to me even what fees would be other than they wouldn’t be 0% rather quickly. So if the fees are 1% cut of all fees earned then agree it won’t affect final APR much.

It will be interesting to watch how this plays out.

Thanks for the clarification!

2 Likes

I think we should use a D3M instead of a regular vault. But if we do use a regular vault then it would probably make sense to reduce the stability fee

2 Likes

Having thought about it more, I think Maker should do the following:

  1. Figure out pricing for this 0.01% pool that would work from a Maker cost point of view.
  2. Figure out what the remaining APY would be for the user on that pool, considering this pricing + some expected fee volumes, without using a multiple of that APY using Maker.
  3. Figure out what multiple + what max TVL makes sense for that APY.
  4. Figure out what the estimated break-even point would be for users of that pool, based on minimum vault debt criteria.
  5. If all of the above make sense for Maker & Vault users, then onboard this collateral with those parameters, AND off-board the 0,05% fee pool at the same time.
2 Likes

I will add that because this is a stable-stable pair we can reduce the oracle costs by a lot since the price barely changes. Current plans is to start updating once per day which will give us a 24x reduction in costs. We can treat the oracle costs as basically negligible for this reason.

3 Likes