RAI Incentive Revamp

General Discussion
1

I’d like to give a quick overview of my personal thoughts around revamping FLX incentives for RAI in the next weeks.

Current Incentives

Currently we’re giving out 474 FLX per day in various incentives programs such as Uniswap v2 RAI/ETH, Aave RAI borrowers and Uniswap v3 RAI/DAI.

There are a couple of issues with the current distribution:

  • It gives out too many incentives for the wrong things. For example, the RAI/ETH pool is mostly unused at the moment besides liquidations and arbing
  • Aave liquidity ballooned with the help of stkAAVE and Idle rewards as well as due to partners such as Fei that are now lending RAI. This means we have Aave liquidity that’s sticky even without heavy FLX incentives and current incentives may give out too much

Solution

In the next month (and once the audit for it is done) we’d like to onboard RAI into Curve. Assuming the Curve community accepts RAI, we would like to do the following changes to our incentives:

  • Decrease RAI/ETH incentives from 120 FLX per day to 0 over 3 to 4 weeks
  • Allocate between 55-70 FLX per day to incentivize the Curve community to hopefully allocate a gauge for RAI
  • Potentially allocate 10-15 more FLX for RAI/DAI Uniswap v3 or RAI/ETH Uniswap v3

We hope that these changes can help us increase liquidity and the supply of RAI while decreasing the amount of FLX spent on lesser used pools. In addition, we hope to decrease the amount of cash we throw to unproductive uses such as Uniswap v2.

Note that one major issue with giving up on Uniswap v2 is that all our liquidity sources would suddenly become concentrated. This can be really dangerous during a market shock so we may want to allocate some FLX to RAI/ETH Uniswap v3. In addition to this we should upgrade the keeper infrastructure so it uses flashloans from Aave and RAI/ETH Uni v3 to make sure there are plenty of liquidity sources in case of liquidations.

Lastly, we’re considering a decrease in Aave borrow incentives from 40 FLX to 20 FLX in the next few days as well as start to decrease RAI/ETH Uniswap v2 incentives soon.

Tokenomics

This incentive revamp will soon be accompanied by a new tokenomics plan for FLX which would hopefully make the incentive distribution more manageable and sustainable while potentially moving away from buyback and burn.

2 Likes
2

As for Uni v3 RAI/ETH incentive, should we require upper limit on concentration factor?
and what do you think of full range option?

1 Like
3

I think we should indeed limit how much people degen and concentrate.

Full range is a bit suboptimal, was thinking of something like 65% away from current market price left and right. Can do less of course but not more. This should give a 2.5x liquidity boost vs full range.

1 Like
4

I agree with the need to reallocate to Curve and adjust incentives.
I disagree with the proposal to reduce net rewards below the current 474FLX/day. This should be maintained.
I disagree with the language used here. It’s insulting that the team feels they are “throwing away” tokens by giving them the community participating in the protocol.
I would maintain the Aave incentives as they are already modest. If they are reduced suggest 5 or 10% reduction is more appropriate than a 50% chop.

5

I also wouldn’t use the term “throw away”. Some FLX farmers are long-terms believer in the projects with no intends to sell. It’s not the case of everyone of course, but we shouldn’t generalize.

I don’t think there is any urgency in reducing the amount of tokens being distributed. We should mostly focus on better distributing those tokens, put them where they are the most efficient.

Curve looks promising, let’s see how that goes. I am currently researching the strategy for Curve. It seems that we would need a bigger bribe push at the beginning. Curve gauges votes are automatically renewed on a weekly basis, which means that votes tends to stay in the last gauge that offered a bribe and this tends to be quite sticky. You need to offer a large bribe at the beginning to make to the switch worth it.

6

Sorry for that, giving to community is important and it’s also important to use resources efficiently which we’re not really doing with Uniswap v2.

1 Like
7

Curve integration is very appealing - but I’d personally be hesitant about putting RAI into a pool with USDT.

Is there a way we could collaborate with the Curve community to set up a pool with USDC-DAI-RAI?

8

At the moment not really and I doubt the Curve community would allocate as many incentives to a pool like that vs RAI Metapool.

We’re in a position where we don’t even add stETH as collateral in RAI so we preserve purity. We really have to find something else to grow and that something is probably Curve. People want RAI to grow and succeed and the only places where we can take more risk are integrations.

9

I don’t know enough about how the Curve bribes work. Can someone please explain? 55-70 seems like a lot.

I think any curve pool with RAI in it will see considerably more volume than a comparably-sized pegged coin pool just because RAI is more volatile which creates more arbitrage opportunities.

Also, I like the aave rewards. I like the idea of keepers using Aave for liquidity source. Please don’t axe them 50% all at once.

ugh. I agree with the sentiment 100%. Especially because funds put in curve tend to stay there for a long time.

I think we can get away with not having USDT in the curve pool.

At first, this seems preposterous:

image
image1198×338 37.4 KB

USDT has more uniswap volume than DAI! We’d potentially be missing out on 20% of stablecoin volume/liquidity by omitting USDT.

Then I take a closer look at what USDT is actually being traded against

image
image1199×618 64.9 KB

It’s 1: ETH (0.3%). We will have our own ETH/RAI liquidity so it seems unlikely trades will get routed RAI–>USDT–>ETH very often.
2. USDC. We don’t need this liquidity either. Nothing is going to route RAI–>USDT–>USDC.
…
DYDX is the first interesting one. We may miss out on some DYDX volume if we omit USDT.
AVINOC… No idea wtf this is.
On page 2 the volume becomes negligible ($200k/week for the number 11 pair).

So basically if we omit USDT, the main expense is it probably becomes less efficient to trade DYDX. That really seems like DYDX’s problem for relying on USDT.

Obviously the way the RAI/DAI incentives were set up strongly incentivized capital concentration.
That’s because if your RAI/DAI v3 position goes out of range it’s really not that big a deal. Just pay ~0.05% to rebalance it.

On the other hand, RAI/ETH v3 is not a farm. If you treat it like a farm you will be punished by IL. I think people will learn this pretty quickly (within a few days of starting) so I’m not concerned about over-concentrated liquidity.

Here’s a screenshot of flipside’s v3 position calculator for WETH/USDT, which is one of the pools we want WETH/RAI to replace:

image
image1581×753 218 KB

It’s seeing ~66% APY organically with a range position that is ~10% above and below the ETH price. (almost 100% on the USDC/wETH 0.05% pair).

Unfortunately flipside doesn’t have RAI/ETH on there but I’m guesstimating the same position would see ~30% APY on RAI/ETH v3 if RAI/ETH v3 had enough liquidity to make it worthwhile to route through there. (based on comparing v2 position fees)

One thing where RAI can’t beat USDT (yet) is CEX liquidity and CEX<–>Uniswap arbitrage. I dont know what it takes to get RAI listed on all the Asian futures exchanges. I think RAI has good product-market fit for Chinese users – it’s just a matter of marketing it to them.

So, anyways, I think if you can subsidize reasonably managed RAI/ETH liquidity positions by like 50%/yr then you can potentially win over some WETH/USDT positions. That would require quite a bit of FLX subsidy.

Magical things happen when you get AMM liquidity in multiple different uncorrelated pools. Polly Finance is awesome for this but unfortunately it’s on polygon so not generating much volume for the RAI/ETH pair.

Anyways, here’s what I’d do:

  • Incentivize RAI/ETH v3 without mint to attract LPs, not farmers (or teach our farmers how to LP).
  • Figure out how to market to China
  • Get new projects to incentivize RAI liquidity over stablecoins (after we have sufficient RAI/ETH liquidity)
10

I agree with some of the sentiment and ideas already echoed here. I would add one observation: incentives likely are responsible for the minting of well over two thirds of RAI in existence. There is $60M RAI out there and $16M in Uni2 plus the $30M in Uni3 RAI/DAI, a good chunk of it which was bought out by DAOs. Without the community and without incentives, there is no RAI, there is no TVL, there is no FLX.

The problem is dual - we need both more minters and more users. The -0.5% is nice for minters, but not enough to generate a self sustaining supply wheel. (Note DAI relies mostly on USDC minters to increase supply dramatically when needed eg immediate profit).

The other problem is organic use for RAI - agree with Dr Ghosto that we need more new projects to incentivize RAI. Is there a way to reward projects for choosing RAI?

The best of both worlds is when minters are users - if we can get other rewards for RAI, that could drive TVL which drives FLX - then more RAI mint, more RAI liquidity etc - that would be key for driving the system upwards. DAOs are nice but they don’t mint and modest RAI price changes don’t drive mint either.

Ultimately the holy grail (I think) would be to figure out a way for the protocol to make money on the use of RAI itself (I thought Dr Stone mentioned having some ideas here) and then potentially further subsidize negative interest rates to drive mint.

11

Curve incentves are super competitive. Anything below $100K per week looks like peanuts, $150-$250K is kinda sweet spot and there are outliers like MIM that give millions per week but seems like they plan to decrease incentives.

We can cut Aave over a couple of weeks but we really need to do it and transfer to Curve and RAI/ETH Uni v3.

Assuming we keep RAI/ETH V3 incentives small I’m ok with LP only.

Why do you think RAI is a good fit for Asia based users? I think they mostly care about liquidity and USDT is good enough in that regard.

12

I think it’s a slight exaggeration to say there’s no RAI without incentives. There’s more than $23M that was bought for the long run and we have tens of millions to fulfill from other protocols wanting RAI. We just have to manage to get more liquidity and we can more easily make RAI sticky.

I’m not completely sure I agree with more users. RAI has been pretty B2B and this is where it found its niche. One group buying $10M RAI to hold for years makes up hundreds or even thousands of traders.

Organic demand for RAI currently comes from DAOs and protocols which is exactly what we want. Personally I never expect RAI to become a high-frequency trading tool but rather a lower velocity reserve asset.

On the topic of other rewards for RAI, Curve is the game. We didn’t do Curve until now which is actually great because it forced us to reach out to others and see that they really want RAI.

We thought more about protocol making fees from RAI but it doesn’t really work without excessive governance.

13

I agree with keeping the incentives small at first. I think RAI/ETH v3 incentives might actually indirectly subsidize Aave as well.

Hypothesis: If you add RAI/ETH v3 incentives (without mint) then it will increase RAI demand and RAI price. (And I think you said you were cranking up the controller strength 50-100% soon?)

This means RAI shorts become profitable again and will increase RAI borrow demand and thereby RAI lending rates.

If the market acts somewhat rationally that’s what I think will happen, anyways.

Why do you think RAI is a good fit for Asia based users? I think they mostly care about liquidity and USDT is good enough in that regard.

I don’t speak any Asian languages so all my information comes 3rd hand but I’ve heard that China/HK are starting to go after money transmitters (who are mostly handling USDT). Example of a big one. Again, take my opinion on this matter with grain of salt.

14

Hehe of course it’s a rhetoric exaggeration from the King of Rock n Roll! :smiley:

However, I don’t think the B2B can be 100% of the story for RAI. Let’s just assume a DAO wants to hold $10M. Then someone else needs to provide the service of locking up either $30M+ of ETH and monitoring passively or lock up $20M of ETH and monitor actively. Either way, that participant isn’t free. I don’t think the 0.5% fee on 1/3 of ETH sufficiently cuts it. Here is where the protocol needs the community - the mint and manage.

The things that make RAI compelling as a reserve asset (backed by ETH exclusively) also make it compete for the most eligible collateral asset. That ETH can be used anywhere, so the cost will be competitive. At the very minimum it could be staked with lower hassle and generate 6-7% with minimal risk. Or it can be levered more against DAI/USDC on Aave etc. MIM is on fire because it uses the most degen collateral everyone has but can’t plug into the “safe” places.

15

Yup we saw the controller can be slightly stronger.

Interesting news on USDT, will check :ok_hand:

16

MIM is on fire because they focus a lot on Curve but I agree we got lots of competition to get ETH.

This is why we also wanna focus on Curve for the foreseeable future so there’s a compelling reason to mint and LP + use the Curve LP token in a saviour.

17

I really like the idea of using curve tokens as saviors. That’s much more attractive than the current savior model as V2 pools simply aren’t that efficient at this point. Additionally, I’d imagine a lot of users would like saviors but don’t necessarily want levered ETH exposure.

Quick idea, what if the reflexed DAO sold FLX bonds for RAI Curve LP tokens? I don’t think you’d need to do it through Olympus pro since Reflexer is already a pretty large community, so you wouldn’t have to give their treasury the haircut. It would also be a really attractive way of creating long-term liquidity without having to continuously issue FLX. I’d imagine could be done in addition to saviors.

The biggest downside is probably regulatory risk as bonds can look like an ICO, although if the contract is written with a gitcoin grant and the bonds are approved/initiated with a governance proposal it’s probably relatively safe?

19

Heh thanks, also think Curve LP will be really attractive. Can’t wait to try it out :rocket:

Bonds do have their own problems that no one until now seemed to have mentioned. We don’t plan to use them and instead I’ll put some details around long term incentives to provide RAI liquidity in an upcoming post.