EIP-3: Removing sequencing from expansion rewards

Proposals
#1

Background

In the current incentive design, expansion rewards are distributed in sequence. Coupon holders get all the rewards first. ESD holders and LP providers get their rewards, only after coupons are redeemed or expired.

There’re 4 undesired side effects because of this sequence.

  1. During coupon redemption phase, ESD holders have no incentive to cooperate with coupon holders to keep the peg up.
  2. ESD holders have incentive to book short-term profits above $1, knowing coupon holders will keep bringing the price up.
  3. Coupon expiration is good for ESD holders, because they get to own a bigger share of the pie after the coupons expire.
  4. Liquidity dries up since LPs have no incentive during coupon redemption phase, in spite of ESD trading above $1.

This makes coupon holders and ESD holders work against each other, even though both share a bullish view on ESD.

Proposed Solution

The objective of the proposal is to make coupon holders and ESD holders cooperate and work in alignment with the protocol, with minimal changes.

  1. Remove the sequence in sharing the rewards.

This will avoid competitiveness between ESD holders and coupon holders. Expansion rewards can be offered in 80:20 ratio, between coupon holders and ESD holders, distributed in parallel. ESD holders include both DAO and LP. Liquidity will remain high at all times above $1, even during coupon redemption phase.

  1. Increase the expiration time to 180 epochs.

This will make ESD holders more cooperative to reach expansion faster during coupon redemption phase. The alternative to keep the price lower till coupon expiration is no longer rational.

4 Likes
#2

Thanks for putting this together! Some thoughts:

  • I like the idea of diverting some expansion during coupon redemption periods to the bonded DAO. This incentivizes voting during contractionary periods, and adds additional incentive to hold ESD.

  • I think a 50/50 split is more appropriate if the coupon expiry window is doubling.

  • The proposal doesn’t do this, but I want to emphasize that rewarding LPs at this stage would be a bad idea.

  • We need to properly handle the edge cases in the code, particularly around when all coupons become redeemable.

1 Like
#3

My concern with sharing 50:50 rewards between coupon holders and ESD holders is that, buying coupons are no longer attractive, given the expiration risk.

Coupon rewards redeem into ESDs and burn the coupons, while bonded ESD rewards compound. 50:50 is heavily skewed towards ESD holders, making coupons irrelevant.

#4

The marginal expiration risk added by 50/50 split is mitigated by doubling the expiration period.

You are advocating for 80/20 and double the expiration period. That is an effective lengthening of the current expiry period and an entirely different discussion IMO

#5

I’m not too particular about doubling the expiration time, if not for the current scenario we’re in. We need time to code, test and pass a proposal.

We can roll back the expiration length to 90 epochs subsequently.

#6

I’d advocate for an 80/20 split between coupon redemptions and LPs. When there are no coupons, the DAO gets 80%; when there are coupons, coupon holders get the 80% instead.

Given the added incentives here, I don’t think it’s necessary to extend the redemption window. If we do, maybe extend 20%, but I think the incentives for added liquidity above $1 would balance this out anyway.

#7

In my opinion, LPs don’t play any significantly better role during coupon redemption phase, comparing normal expansion phase.

Maintaining the status quo of rewards between DAO and LP at all times will keep the price stable.

If we incentivise LPs more during coupon redemption, DAO holders will self half to get into LPs, which will create additional sell pressure. We should avoid this especially after coming out of contraction.

2 Likes
#8

Curious your rationale here @delitzer. This is a different situation than expansionary periods, so I don’t think the 80/20 split is a useful prior.

I am also not sure it’s a good idea to incentivize any liquidity when the protocol is not in a purely expansionary stage.

1 Like
#9

@willprice It is an expansionary period: price is above $1 and supply is growing. Incentives to LPs increase liquidity and make it less likely that the price gets pushed back under $1, in particular by coupon redeemers who may be looking to take profit.

I don’t think paying out DAO holders during this period is particularly helpful. I don’t think sharing rewards with DAO holders addresses current situation #2 (“ESD holders have incentive to book short-term profits above $1, knowing coupon holders will keep bringing the price up”). In fact, paying the DAO holders likely makes this worse.

2 Likes
#10

Incentivising liquidity more than status quo, during coupon redemption phase will encourage profit booking by coupon redeemers. Because it becomes the best time with lowest slippage.

If we incentivise only LP and not DAO, bonded ESDs in DAO will sell half to provide liquidity, which will further increase sell pressure.

1 Like
#11

These are solid points. So you think maintain the same 80/20 for DAO/LPs during this period, but the balance for redemptions vs DAO/LPs is still TBD, correct? Seems like at least 50% should go to redemptions, so low end is 50 redemptions, 40 DAO, 10 LPs. Your proposal of 80/20 would then be 80 redemptions, 16 DAO, 4 LPs, likely the other end of the range. I’d probably lean towards a heavier weight on redemptions to start, as you proposed.

#12

Yes. My proposal is to maintain status quo between DAO and LP. Which is currently 80:20. This will keep the price stable.

Ratio between coupon holders and ESD holders is TBD. I think it has to be above 50% for coupon holders.

#13

After thinking about this more and reflecting on @delitzer’s points , I’ve changed my mind:

  • DAO should receive no additional rewards. There is too much passive ESD bonded, and the marginal change in behavior that incentives would create is minimal.
  • Bonded LPs should receive a fraction (20%) of expansions during coupon redemption periods. Liquidity is a service, and LPs solely bear the risk of providing it.
  • The LP unstaging period should be lengthened from one epoch to at least five epochs. This will prevent a liquidity drought every time the TWAP drops below a dollar, and direct rewards to LPs who believe in the long-term success prospects of ESD.

In summary:
:x:additional rewards to DAO
:white_check_mark: 80/20 split between coupon redemption and LPs
:white_check_mark: Increase coupon expiry period from 90 to 120 epochs to compensate for reduced redemptions per positive epoch.
:white_check_mark: Require LPs to wait 5 epochs in staging (fluid) before being eligible to withdraw

3 Likes
#14

I like what I am reading here.
I understand the argument to reward bonded ESD to get users to vote and hold during redemption, however the rewarding of LP seems very important to me.

Properly incentivizing liquidity (positively or negatively) allows for easier or more difficult price manipulation:

  • In debt generation, low liquidity helps users going back to 1$ while de-incentivizing selling tokens because of high slippage
  • In coupon redemption territory, we reach the stability point we would ideally like to keep and we are clearing debt. This point being critical for the system, increasing LP reward at this stage makes price manipulation more difficult and LP providers should be rewarded for this service to the protocol. Some ESD holders might want to sell their incentive is to go past coupon redemption to expansion for nice reward, which should limit profit taking
  • In expansion, steady growing LP helps with stability to avoid price volatility and should be rewarded as well

TL,DR: no reward for bonded ESD during coupon redemption, higher rewards for LP (70 for coupons / 30 LP?), increase coupon expiry (+20/30% compared to now?), come back to 80/20 for bonded ESD / LP in expansion post clearing

#15

I agree with much of this, except I would increase the LP incentive during coupon redemption period in exchange for increasing staging time before withdrawing.

#16

If we incentivise LPs alone, we’re incentivising selling half of bonded ESDs to get into LP.

This creates additional sell pressure, which we should avoid especially when coming out of contraction.

Let’s maintain the status quo between DAO and LPs during coupon redemption phase.

2 Likes
#17

First, thanks to @freakyarch for beginning to formalize this conversation which originated in Discord!

There are four categories of stakeholders, which collectively participate in the process of maintaining ESD’s Target Price:

  • Coupon purchasers - reduce supply below $1, increase supply above $1
  • Liquidity providers - take on the risk of ESD declining in price, to facilitate orderly trading
  • DAO voters - locked during votes, otherwise generally liquid to sell above $1
  • Ecosystem float (ESD on exchanges, in DeFi, etc) - fully liquid to sell above $1

The suggestions by @freakyarch, @delitzer, and @willprice would re-orient some of the risks / rewards between these groups.

1. I firmly agree that the allocation of expansion (when above TWAP $1) doesn’t need to entirely prioritize Coupon redemption – it’s possible that other recipients / use-cases would create a more balanced ecosystem.
2. Any reduction in the expansion being allocated to Coupon redemption should be offset by increasing the redemption period; I agree with many of the suggestions; if the expansion allocation is reduced by 20%, the redemption period should be increased by at least 20%, otherwise Coupon Holders bear additional risk through this proposal; 120 epochs is arbitrary, but “feels” correct.
3. If expansion is re-allocated, I don’t understand why bonded DAO tokens should take “preference” over other stakeholders (including Coupon holders) – they take on very little risk for the system.
4. Liquidity is a virtue, and the ecosystem–across all stakeholders–is better off when there is more liquidity for ESD. Liquidity providers also bear the risk of a decrease in price. Incentivizing them to participate, and remain committed is something that the community should encourage–increasing the bond/unbond period for LPs to take advantage of expansion incentives makes a lot of sense.
5. From an implementation perspective, an 80/20 split between Coupons Redemption and Liquidity Providers has an elegant symmetry to the 80/20 split between DAO and Liquidity Providers, when Coupons have been fully redeemed.

In short, I agree with the suggestions above. It would be nice see them packaged into a new implementation.

Risks
As @freakyarch mentioned, this can and will change marginal behavior at ~TWAP $1, namely that users might be strongly incentivized to participate as Liquidity Providers. The math on short-term returns is very strong.

Users without USDC might be encouraged to sell ESD to participate (but, I hope nobody is all-in ESD–please be responsible), which would be self-defeating, and prevent TWAP $1 in the first place–limiting this risk IMO. The opposite is more likely to be true; users purchasing ESD around TWAP ~$1.

4 Likes
#18

There are some implications to lengthening the wait for LPs to withdraw that needs to be considered:

  • We are essentially locking LPs out from contributing to returning to peg by selling for 5 epochs once coupons are redeemed and we are back to expansion.
  • LPs that do not have or have run out of USDC to compound their rewards, will be locked from removing their rewards for 5 epochs, if they have to stay in staged for 5 epochs. Is it possible implementation wise to allow withdrawal if state is frozen, so they can unbond & bond LPs same epoch and claim rewards next epoch?
  • when we go back below 1, we will be up against significant liquidity (~25% of MC) to push price back up above 1 for at least 5 epochs.

I would also like the participants to re-evaluate the coupon expiry time, considering eq. said that 90 epochs was chosen arbitrarily and we have since reduced epoch time from 24hr to 8hr, I am not sure what the right number is, but I think it is worth pondering over. For this proposal though I agree with proportional increase to epoch time based on how much the rewards are cut by.

1 Like
#19

We are essentially locking LPs out from contributing to returning to peg by selling for 5 epochs once coupons are redeemed and we are back to expansion.

this is a feature, IMO. The alternative is creating additional cycles around liquidity provision and removal (similar to what we saw at the end of the last expansion)

LPs that do not have or have run out of USDC to compound their rewards, will be locked from removing their rewards for 5 epochs, if they have to stay in staged for 5 epochs. Should be allow withdrawal if state is frozen, so they can unbond & bond LPs same epoch and claim next epoch?

Yes. I see no reason for the transition from rewarded to claimable states to change at all.

when we go back below 1, we will be up against significant liquidity (~25% of MC) to push price back up above 1 for at least 5 epochs.

ESD is attempting to be a stablecoin and medium of exchange. Such an asset needs liquidity even during deviations from the peg - to the upside or the downside. It is simply not sustainable to have liquidity evaporate the moment the TWAP drops below a dollar.

I would also like the participants to re-evaluate the coupon expiry time

agree that proportional changes are fine here but that this should be considered separately.

1 Like
#20

Blockquote
It is simply not sustainable to have liquidity evaporate the moment the TWAP drops below a dollar.

At the same time it makes it easier to push back the price to 1$

Blockquote
LPs that do not have or have run out of USDC to compound their rewards, will be locked from removing their rewards for 5 epochs, if they have to stay in staged for 5 epochs

Can we imagine increasing the length of LP lock only during coupon redemption (TWAP > 1$ & coupons to redeem while bucket not enough to allow all coupons to redeem)?