Back to Discussions
Recommended Adjustment to Vault Reward Schedule on Kintsugi and Interlay
16 Comments

Tl;dr.

  • KINT and INTR Vault block rewards are currently too high, considering market conditions and in comparison to competitors / projects with similar risk profile.
  • This proposal suggests to adjust Vault block rewards to a more sustainable level, updating the KSM / DOT collateral requirements to competitive and more capital-efficient rates, and introducing regular review cycles to assess ROI for protocol and Vault operators.
  • The saved funding will remain with the network treasury, to be used as governance decides, e.g.. to bootstrap kBTC/iBTC growth as market conditions improve and help build liquidity in the native AMM and lending protocols set to launch on Kintsugi and Interlay in early 2023 (more details in upcoming Interlay 2.0 whitepaper - read 2.0 roadmap announcement).

Introduction: Vaults on Kintsugi/Interlay

Vaults are the heart of the Interlay & Kintsugi networks: they secure BTC while kBTC/iBTC is being used in DeFi. To do so, they deposit collateral in different assets, insuring users against exchange rate fluctuations and loss of BTC. As such, the costs of running a Vault are twofold:

  1. Infrastructure costs and maintenance labor for running the Vault software
  2. Capital (opportunity) costs for putting up collateral

During the bootstrap phase (while the Interlay and Kintsugi protocols do not generate sufficient bridge revenue to cover all operational costs), the treasury pays block rewards to Vaults in the native KINT (on Kintsugi) and INTR (on Interlay) tokens. 30% of the treasury was proposed to be used for this purpose over the first 4 years of network operation, 12% in the 1st, 9% in the 2nd, 6% in the 3rd and 3% in the 4th year. Block rewards are split among Vaults proportional to how much BTC has been minted with them (for an individual Vault: vault_BTC / total_BTC_minted).

Problem

The current KINT and INTR reward emission is unsustainable:

  • ~80% APR on Kintsugi
  • ~65% APR on Interlay

as opposed to what can be observed on the market from competitors/node operations comparable to the Kintsugi / Interlay risk profile: between 12% and 22.8% [3].

This creates a negative feedback loop: there is too much emission, disproportional to the adoption curve, and Vaults are incentivized to sell KINT/INTR in anticipation of being able to repurchase it cheaper at a later point in time. This in turn creates a downwards spiral of both token prices that slowly but steadily burns treasury funding and reduces returns for Vaults (despite the high emission).

Goal

The goal of this proposal is to adjust Vault block reward rates to be more sustainable while keeping incentives for operators high enough - and more reliable.

Solution

1) Reduce collateral thresholds.

The latest collateral risk models [1] suggest the following collateral thresholds for KSM (on Kintsugi) and DOT (on Interlay):

KSM (on Kintsugi)

  • liquidation: 130%
  • premium-redeem: 145%
  • secure (min. when minting): 160%

DOT (on Interlay)

  • liquidation: 125%
  • premium-redeem: 140%
  • secure (min. when minting): 155%

This change reduces the capital requirements of Vaults by ~40%.

For more information on the meaning of individual thresholds, see [2]

2) Adjust Vault block rewards to competition.

The proposal is to adjust the Vault rewards such that:

  • APR on Kintsugi is ~30% assuming the amount of collateral locked at the time of this proposal. This proposal updates the block reward from 0.456621004566 KINT to 0.102803978513969 KINT.
  • APR on Interlay is ~25% assuming the amount of collateral locked at the time of this proposal. This proposal updates the block reward from 45.6621004566 INTR to 9.866085649312430 INTR.

3) Regular Review & Adjustment of Vault Economics.

Rather than having a static model, we propose that Vault economics be reviewed on a regular basis, allowing to react and adjust for external/macroeconomic events.

The community, including Vault operators, can make recommendations/requests to adjust the Vault rewards schedule via the Governance portal (guides provided).

The process to adjust rewards will be greatly simplified over the next months with upcoming improvements to governance (more details are incoming in the Interlay 2.0 whitepaper).

Expected Result

The optimization of collateral thresholds combined with the adjusted reward schedule is expected to reduce emissions as follows:

  • KINT emission by ~78%
  • INTR emission by ~79%

Why should community vote in favor of this proposal?

This proposal makes a big step to improving the economic stability and outlook of both Kintsugi and Interlay.

The current emission schedule was conceived during a growing crypto market, in anticipation of reaching significant iBTC targets by EOY 2022. However, due to recent macro-economic events in crypto (Luna, 3AC, Celsius and now FTX), as well as the two hacks (Nomad impacting Moonbeam and aUSD on Acala) affecting Polkadot, the growth of DeFi on Dotsama has not been as fast as hoped. As a result, kBTC and iBTC utilization rates are low, with most BTC sitting on Kintsugi/Interlay.

The high initial emission was also a measure to reward early technical risk and to cover the high collateral requirements imposed at launch (260% over-collateralization for KSM/DOT Vaults). The technical risk of the bridge can be considered significantly reduced by now. The bridge has been operating successfully since March 2022 on Kintsugi and since August 2022 on Interlay, withstanding the recent market turbulences. The removal of the aggressive BTC theft reporting [4] has also greatly reduced the risk of wrongfully slashing Vaults. The proposed reduction of collateral thresholds to competitive rates also constitutes a big step towards Vault capital efficiency.

References

[1] Interlay Vault collateral risk modelling framework. https://github.com/interlay/collateralization-analysis

[2] Vault collateral thresholds. https://docs.interlay.io/#/vault/overview?id=collateral

[3] APR comparison to similar projects. https://api.ipfsbrowser.com/ipfs/get.php?hash=QmPhP8utFLhharELDeN6Njbc2234DGxnc4A73wxf2JHwPJ

[4] Kintsugi proposal discussing removing theft-reporting. https://kintsugi.subsquare.io/democracy/referendum/52

--
EDIT 23 Nov 2022 15:34 UTC: added link to Interlay 2.0 roadmap announcement

Edited
Reply
Up
Share
Comments

The “secure” thresholds are uncomfortably close to the premium redeem and liquidation thresholds.
So far the premium redeem threshold of 200% has been the important threshold that vaults pay attention to, to avoid an extra 10% fee on redeems. It has been triggered many times by vaults whereas I'm not aware of the liquidation threshold being hit. Against the 200% premium threshold, we’ve usually seen vaults prefer to stay in the 230-250% range, with anything below about 215% requiring immediate action. Therefore a good ratio between initial “secure” threshold and the premium redeem threshold would be around 240%/200% or 1.2. However the proposed new thresholds for KSM are much lower at 160%/145% = 1.1. Therefore I would recommend shrinking the gap between liquidation and premium redeem and increasing the gap between premium redeem and secure thresholds. Specifically, for KSM I’d suggest reducing premium redeem to 140% and increase initial secure threshold to 165%. Similar change for DOT (135%, 160%).

I think it's a good idea to raise the vKINT/vINTR staking reward rate above the vault rewards rate, which may encourage more vault token staking instead of sell pressure. I think the vault rewards rates are slashed too far, though, but I’m biased as a vault operator. At higher rates I’ve been able to pool community resources to run a vault, but if vault rewards reduce to only about 10% higher than staking rewards I don’t think there’s enough incentive for people to pool together funds in a community vault.

I really look forward to LKSM/LDOT vaults and injecting treasury funds into the new DEX to supply “Treasury Owned Liquidity” as another way to stabilize the token price.

Reply
Up

Therefore I would recommend shrinking the gap between liquidation and premium redeem and increasing the gap between premium redeem and secure thresholds.

IMO the ability for vaults to set higher custom thresholds addresses this concern. So any vault that prefers higher secure thresholds, the vault should set a higher threshold. Making the gap between liquidation and premium redeem smaller, increases likelihood of liquidations which is not ideal.

I think it's a good idea to raise the vKINT/vINTR staking reward rate above the vault rewards rate, which may encourage more vault token staking instead of sell pressure.

That doesn't solve the actual problem: emissions are too high in this bear market. It just bleeds out the treasury without adding much value.

At higher rates I’ve been able to pool community resources to run a vault, but if vault rewards reduce to only about 10% higher than staking rewards I don’t think there’s enough incentive for people to pool together funds in a community vault.

Three's always the option to adjust rewards later on.

I really look forward to LKSM/LDOT vaults and injecting treasury funds into the new DEX to supply “Treasury Owned Liquidity” as another way to stabilize the token price.

LKSM vaults are live on Kintsugi :)

Edited
Reply
Up

Technical Proposals

The technical proposal overwrites the storage items of the vaultAnnuity.rewardPerBlock amount. This approach is chosen as the scheduled rewards are batches of [tokenTransfer, update_reward_per_block]. However, we can't use update_reward_per_block since it will try to distribute the total rewards over 1 year.

Kintsugi

Current value: 456,621,004,566 planck
New value: 102,803,978,514 planck
New value in little-endian hex: 0x123998ef170000000000000000000000 (use https://www.save-editor.com/tools/wse_hex.html to verify)

Extrinsic

https://polkadot.js.org/apps/?rpc=wss%3A%2F%2Fapi-kusama.interlay.io%2Fparachain#/extrinsics/decode/0x000504803c20031a1af83128241fb740caa6b146eb816a34663a266db87ec8fa747c29bb40123998ef170000000000000000000000

Proposal

https://polkadot.js.org/apps/?rpc=wss%3A%2F%2Fapi-kusama.interlay.io%2Fparachain#/extrinsics/decode/0x020208460ad4000504803c20031a1af83128241fb740caa6b146eb816a34663a266db87ec8fa747c29bb40123998ef170000000000000000000000460015995c75360ecd53dd03149a7f93932b92f1b71d0057507134fc0f931050c57a0b005039278c04

Interlay

Current value: 456,621,004,566 planck
New value: 98,660,856,493 planck
New value in little-endian hex: 0xad32a5f8160000000000000000000000 (use https://www.save-editor.com/tools/wse_hex.html to verify)

Extrinsic

https://polkadot.js.org/apps/?rpc=wss%3A%2F%2Fapi.interlay.io%2Fparachain#/extrinsics/decode/0x000504803c20031a1af83128241fb740caa6b146eb816a34663a266db87ec8fa747c29bb40ad32a5f8160000000000000000000000

Proposal

https://polkadot.js.org/apps/?rpc=wss%3A%2F%2Fapi.interlay.io%2Fparachain#/extrinsics/decode/0x030208460ad4000504803c20031a1af83128241fb740caa6b146eb816a34663a266db87ec8fa747c29bb40ad32a5f81600000000000000000000004600b51e9a6cc54ceb06bd43a7fe49f07f86e7ebdd6c669004d68a75f710532667970b00a89c134602

Edited
Reply
Up
The current KINT and INTR reward emission is unsustainable:

~80% APR on Kintsugi
~65% APR on Interlay

I don't think this is accurate since most vaults are self minted.So the actual APR is much lower.When there is no demand for KBTC/IBTC, BTC is also the cost that the vault operator needs to pay, and needs to bear risks.

The goal of this proposal is to adjust Vault block reward rates to be more sustainable while keeping incentives for operators high enough - and more reliable.

The goal appears to be to reduce the block reward. But why reduce it? We saw:

This creates a negative feedback loop: there is too much emission, disproportional to the adoption curve, and Vaults are incentivized to sell KINT/INTR in anticipation of being able to repurchase it cheaper at a later point in time. This in turn creates a downwards spiral of both token prices that slowly but steadily burns treasury funding and reduces returns for Vaults (despite the high emission).

Simply put, there is no actual demand for INTR/KINT, so the vault operator chose to sell the rewarded INTR/KINT, which caused the price to drop.

Then let's consider:

  1. Why did the vault operator sell INTR/KINT? Is it because there are too many rewards?
  2. Why did the price drop?

My considerations are:
The total amount of INTR/KINT is unlimited, which means that if you put INTR/KINT in your hand, it will depreciate every moment. Therefore, if the reward is reduced, the depreciation is only reduced, but the depreciation is still happening. In this way, no one will choose to hold INTR/KINT.
Therefore, no matter what the block reward rate is, the price will keep falling because there is no demand for INTR/KINT.

I have the following suggestion

  1. First set the total amount of INTR/KINT, and then reduce the block reward.
  2. Provide a function similar to "Layer 2", users can safely provide BTC to share the rewards of the vault operator. The vault operator can set the commission by itself.
Reply
Up 1

" incentivized to sell KINT/INTR in anticipation of being able to repurchase it cheaper at a later point in time"
That was funny! What evidence do you have for the second part of your assertion?

I agree completely with the analysis of JZCg in terms of lack of utility for the token.

It seems to me that the vast majority of the Kintsugi users are vault operators, in other words KINT farmers. I predict that many will leave if you decrease the vault rewards. Personally, I have scaled back my activity to about 1/10th mainly due to losing my KSM short hedge on FTX. But even without the FTX debacle, I was preparing to leave as soon as the APR no longer justified the risks.

For me, there is insufficient work done on analytical tools and strengthening of the UI accuracy:

  • dashboards only show 4 days of history and don't use accurate prices
  • APRs don't take into account real block times
  • Hard coded 24-hr logic when the chain uses 48hr for premium redeems

These deficiencies are manifestation of the risks of putting one's money into a degen yield farm. Maybe Kintsugi is still the most lucrative one out there, but I doubt it's 4x over par.

Fixing emission rates without having a complete rethink of the whole economic model seems to me a very risky approach.

Reply
Up 1

The proposal seeks to reduce the rewards on Kintsugi to ~30% APR.

Some thoughts that I have regarding this change are:

  • The APR is only met with 100% collateral usage. In the real world, you're facing redeems and issues and you'll never reach constant collateral usage and thus constant rewards
  • Vault operators carry the risk of losing collateral through premium redeems or liquidations
  • Vault operators have the occurring infrastructure and operational costs to keep the vault running and maintained

Given the unstable rewards and involved costs it will become more or less similar effective to just staking the collateral which is why I would then re-evaluate vault operations.

Reply
Up 1

@wd9Y...BDKK

Agree.
We need more INTR/KINT application scenarios.

Reply
Up 1

@wdBL...JZCg
I don't think this is accurate since most vaults are self minted.So the actual APR is much lower.When there is no demand for KBTC/IBTC, BTC is also the cost that the vault operator needs to pay, and needs to bear risks.

If there is no use for kBTC/iBTC, the vault operator should not self-mint it in the first place. Vaults should only self-mint if they want to use kBTC in DeFi, that's the whole purpose of Interlay. Otherwise you could just hold your BTC in your cold wallet. So this argument that self-minting reduces the APR does not sound very reasonable to me. On Interlay we currently have 89 iBTC minted. Only 22 of those are LPed into pool. Since there are no other use cases as of now, the rest most certainly remains under self-custody by vaults who self-minted and only farm rewards but do not contribute to the growth of the ecosystem at all.

Another thing to consider is, that the ecosystem also has to provide incentives for liquidity bootstrapping for the AMM and lending protocol launch. This will help to increase volume of the bridge and in turn also the rewards. I think in the end, vault operators will be more interested in earning BTC fees than INTR rewards, so I would considers this reduction in the light of the overall strategy for Interlay and Kintsugi.

I agree with the need for more INTR/KINT utility though and this is actively being worked on. This proposal is only the first step into a completely reworked tokenomics.

Reply
Up

@wdBr...Pm46

Given the unstable rewards and involved costs it will become more or less similar effective to just staking the collateral which is why I would then re-evaluate vault operations.

On Kintsugi LKSM is already supported as a collateral and there are plans to extend this offer to more staking derivatives as well as to allow the same on Interlay. This means that you can actually get the vault rewards on top of your staking rewards, so these two options are not competing with each other.

Reply
Up

@wdB3...ay33
"the ecosystem also has to provide incentives for liquidity bootstrapping for the AMM and lending protocol launch"

Instead, Kint/Intr treasuries should deposit KINT/INTR and submit a proposal to KSM/DOT treasury to deposit an equivalent amount of KSM/DOT. Therefore the treasury would provide liquidity for the KINT/KSM and INTR/DOT pairs and you wouldn't need to provide liquidity incentives for mercenary capital for those pairs. ("Treasury Owned Liquidity") (kBTC/iBTC pairs would be different since treasury doesn't hold any kBTC/iBTC, but it could slowly be bought over time. For those pairs, LP incentives would be needed at first.)

Reply
Up

@wdCm...tn4a

I totally agree that treasury owned liquidity should be prioritized over attracting LPs via incentives. However, this only works for KINT/INTR at the moment. Given that the lending protocol will likely whitelist other assets as collateral, there should be enough liquidity on-chain to create profitable liquidation opportunities for distressed debt and that will requires initial incentives.

Also, your post implies that LPs are bad, doing it only for the money, whereas vaults are the good guys who deserve the incentives. Given the comments above, it seems that many vaults behave very much like the 'mercenary capital' you described. Once APY becomes to low they will move on and farm elsewhere. The ecosystem needs vaults and LPs that also provide capital at realistic APYs and those can only be sustained by bringing real volume to the ecosystem, which can only be achieved by bringing more use cases for kBTC & iBTC.

Reply
Up

Thanks for starting an active discussion!

Please check out the Interlay 2.0 roadmap announcement that addresses quite a few points raised so far: https://medium.com/interlay/interlay-2-0-roadmap-c6387d9d1abb

Edited
Reply
Up

FYI, the mentioned proposal for switching to a capacity-based Vault block reward distribution model has been summarized in this discussion post:

https://interlay.subsquare.io/post/13

Reply
Up

Please also consider point (3) - the recommendation here is to stabilize Vault revenue and review it regularly, scaling as the network gains adoption and DAO revenues start to pick up.

Reply
Up 1

Hi,

I'm sorry there is something I don't get in your calculations.
If you :

  • Divide vault block reward by 4.5
  • Reduce collateral threshold to 160%
  • Move to a formula where reward = (vault capacity / total vault capacity) * block reward
    I don't see how this isn't just equivalent to dividing current APY by 4.5 directly

Indeed, reducing collateral threshold has no impact on APY with the new formula.
With the same amount of collateral provided total and collateral provided for a given vault, the given vault will just have its current APY divided by 4.5, meaning much less than the announced 25%.
The only thing changing is that the bridge will have more capacity and that's it (oh, and the slashing or liquidation risk is increasing)

Let's say capacity-based Vault block reward distribution model is implemented :

Reward.Now = Reward w/ secure threshold at 260 = (vault capacity / total capacity) * current block reward
Reward.Then = Reward w/ secure threshold at 160 = (vault capacity / total capacity) * new block reward

We take the hypothesis this does not change : (vault capacity / total capacity) so we get
Reward.Now / Reward.Then = current block reward / new block reward
= 4.5 * new block reward / new bloc reward
= 4.5

==> Changing vault block reward by a ratio of 4.5 directly changes APY by 4.5, and secure threshold level has no form of impact

Edited
Reply
Up

And in practice it's even worse, as there is currently ~10% available capacity in the bridge, implementing the capacity-based Vault block reward distribution model already reduces from ~10% current APY.

So see, for a vault at secure threshold:
Current APY ~65%
Apply capacity-based reward
New APY ~59%
Apply new bloc reward (/4.5)
New APY ~13%

Reply
Up