Add an admin fee to be partially burned & distributed to BNT holders

OLD GOV (LEGACY)
1

Hey Bancor DAO! I’d be happy to discuss an introduction of an “Admin fee” to increase BNT’s utility and value proposition for holders & stakers alike: incentivizing BNT holders to stake their tokens, create an additional deflationary mechanism and to fund the community’s DAO developments.

Summary:

Introducing some kind of admin fee (e.g., 0.02-0.05%) taken from liquidity providers who provide single-sided, non-BNT liquidity in the system (i.e., “TKN” liquidity providers).

The purpose of such a fee is to add new incentives to hold & stake BNT, create an additional deflationary mechanism via burning, and incentivize long-term protocol participation of LPs. This will also give BNT holders and BNT liquidity providers an additional source of revenue, on top of and regardless of the pools they provide liquidity to. Furthermore, part of these fees can be used to create a treasury fund for governance-driven grants such as technical development grants, third-party integrations, community marketing and other project collaborations.

Abstract:

Let’s say Alice has 1000 DAI, and wants to buy ETH via Bancor. BNT/ETH pool’s swap fee is 0.1%, and BNT/DAI pool’s swap fee is 0.1%. As two hops are required, she pays 0.2% to buy ETH through DAI, which equals $2. With the existing implementation, these two dollars go directly to the liquidity providers of these pools. If we add the admin fee, 10% of the $2 (or $0.20) could be sent to a smart contract that market buys BNT and distributes the proceeds to all BNT holders. Users who are staking BNT in a protected pool (i.e., holders of vBNT) may also receive a greater share of the fees based on how long they’ve been holding vBNT.

Note: as the fees accrued are in various ERC20 tokens, a “buyback” mechanism must be in place that automatically converts the accrued TKN to BNT by market buying BNT, which creates a buying pressure on the secondary market (and generates fees again).

Once the TKN side is converted to BNT, all tokens are sent to the main admin fee smart contract. From there, a percentage is distributed to all BNT holders, a percentage is distributed to BNT stakers based on how long their BNT has been staked in the protection contract, and a percentage is sent to a governance-controlled treasury contract and used for further community development. (FYI: the Curve DAO had a proposal for a similar mechanism, the details of which can be viewed here

Every BNT token holder who provides liquidity in the protocol receives vBNT in return, which is used to vote on-chain for proposals. Should we introduce a source of revenue for vBNT holders, a stickiness factor is essential, to promote LPs keeping their BNT staked in protected pools. The longer you stake your BNT in the protection contract, the bigger the boost you get and therefore more rewards as opposed to those who “lock” for a shorter time frame (this does not mean we must enforce a lockup period).

Why should I want this?

As a trader: This proposal does not increase the fees traders pay for executing orders: it still depends on the pool’s fee (e.g., 0.1% fee for ETH/BNT). By increasing the incentive to stake BNT in protected pools, the admin fee can lead to more liquidity entering the system, bringing down the cost of swaps.

As a vBNT holder (a user staking BNT in a protected pool): Previously BNT LPs received swap fees only for the pool they supplied liquidity to, and only to the BNT side unless they provided dual-sided liquidity. If implemented, the admin fees would provide an additional source of revenue from ALL the pools. The underlying asset of vBNT which is BNT also gets another value proposition, which should make it more attractive to hold & stake.

As a TKN Liquidity Provider: In the short term, as an LP providing TKN-only liquidity, the admin fee would slightly decrease APR from swap fees. However in the long term this proposal’s goal is to increase liquidity depth by increasing the incentive to hold and stake BNT, which will ultimately allow for more single-sided TKN deposits, generating more volume which translates into higher swap fees & APR.

BNT holders: The protocol will swap the TKN accrued (through bancor.network) to BNT, which creates automatic and consistent BNT buying pressure. Additionally, %X of the BNTs would be burnt, while %X could be re-distributed specifically to vBNT holders

Would appreciate any feedback and input. As you see I left the percentages part as a variable, as I’d like to see what numbers the community members think might be a good fit.

4 Likes
2

Supported by me, anything that creates rarity is a bonus.

2 Likes
3

I’m definitely in favor of this cashflow going to vBNT holders as opposed to BNT holders. The way that the protocol is designed, the success of the project uniquely relies on as many BNT being staked as possible and the price/marketcap of BNT being allowed to appreciate. Any BNT that is held on exchanges is vehemently holding back the liquidity blackhole from taking effect. With BNT moving off exchanges to take advantage of this there will both be more TVL captured (due to pools needing half BNT) and less BNT on exchanges to act as downward weight on price.

2 Likes
4

Do we know the impact on vBNT APY assuming an admin fee of 0.01% on all trades?

1 Like
5

This is very strong. To my mind, this achieves a similar effect to token burning, without burning anything. The wording in this proposal is “BNT stakers” and not “vBNT stakers” - the difference is small but important. I support the wording as it appears here; I want people who are disinterested in governance to have the option to not be involved if they choose. Therefore staking BNT, and not vBNT, is sufficient to qualify for rewards. May I ask you to comment?

6

May I suggest adding some verbage here: The admin fee can be treated as a minimal contribution to maintaining the health of the protocol, and ensuring the sustainability of impermanent loss protection for everyone.

7

Interested to discuss the distribution and burn percentages with you. How would you feel about a 25% burn and 75% distribution?

8

Hi everyone!

I agree 100% with this proposal. After discussing with a few members of the community, I understand now that buyback-and-burn is the way to go as stated above. Below are some of the points I expressed in the past about buyback-and-burn instead of buyback-and-distribute or make. I also added the response from Michal regarding why buyback-and-burn was the way to go.

JBMasterCrypto:

  • I think the reason sushi and curve are distributing tokens is because buying in the market and burning has not been proven to be effective for capital tokens or tokens that have governance. Also, it seems that the market does not like the burning mechanisms as it does not give you a tangible way to measure your earnings. I recommend reading the following article:

“The key point is that removing capital units from circulation works by improving participation ratios for outstanding stakeholders, not inherently because the asset is made “scarcer”, and not necessarily because the total value of the equity increased. Reducing the number of shares may boost the price, but it doesn’t change the overall value of the system. It’s the same for capital tokens in cryptonetworks: buybacks do have a positive influence on price, but burning doesn’t create new value, it only redistributes current value among a smaller group of people. There’s a reason why companies that do a lot of buybacks are associated with low growth.”

  • I think we are already competitive in other areas and we are providing value to BNT and TKN stakers by improving their LPing experience with IL protection and single side liquidity provision.

  • I think by buying in the market BNT with TKN fee sharing we would already achieve network inflation equilibrium since people re lock their BNT to compound their earnings. What ultimately provides value to the network is the fact that the fees accumulated are used to buy BNT in the market. That BNT is then re locked in pools lowering the inflation rate.

Michal:

"Thanks for your answer, appreciate.

I am familiar with this article, and I would 100% agree with your point, if we do not mint tokens to cover IL.

Current Bancor LM program (and every other on the market) is limited and eventually will end capping the total circulating supply.

IL insurance is not capped and tokens issuance grows linearly with the utilization of the network, therefore we need a burning mechanism with the same principles that grows linearly together with the network utilization.

Also it fits mathematically, TKN side is charged perpetually for the insurance where BNT side is assured that they are not diluted.

I am more than happy to reconsider distribution of portion of the confiscated TKNs towards vBNT, after concerns about BNT being overly diluted are calmed down and we see the inflation on the acceptable level."

1 Like
9

Fantastic post Avi. Great to see this level of detail & depth into further utility & adding value to the BNT token & incentivizing participation for on-chain proposals. I agree with a BNT + vBNT + treasury allocation, however you mentioned that [quote=“Avi, post:1, topic:245”]
From there, a percentage is distributed to all BNT holders , a percentage is distributed to BNT stakers based on how long their BNT has been staked in the protection contract, and a percentage is sent to a governance-controlled treasury contract and used for further community development.
[/quote].

Perhaps rather than admin fee = length of vBNT holding/ BNT staking, we have a vBNT duration cutoff. What you propose currently might discourage new participation in the proposals, etc. as they will always be lagging behind. I understand that early suers should be rewarded, and I don’t wish to take away from that, but I think a cutoff point of vBNT is better rather than a straight rewards proportional to duration of vBNT.

TL;DR fully agree except instead of admin rewards to vBNT based on their length of staking, should be given the max admin reward after a specfic time period.

2 Likes
10

Hi, great proposal. It made me understand few things better and I’d like to share my thoughts:

The main point of the Placeholder’s solution is that increasing scarcity by burning TKNs doesn’t add value to TKN, because it decreases liquidity of the TKN. Their Buyback Machine was designed with Balancer smart-pools in mind, where they create a smart-pool and the fees generated are used to buy back TKN (from ETH/TKN pair) – also considered multi-asset pools (up to 8 TKNs)

Bancor with the launch of LM program guarantees large BNT issuance, which has the following consequences:
– BNT needs to be staked to increase TVL
– LM program incentivizes more BNT staking; larger farming rewards (70%) and the multiplier
– There will be ~800K BNT released weekly (rough guess); BNT price ~$1
– LM is enough of an incentive for staking BNT; ~300% APR on deposited BNT; idle BNT misses out
– % of BNT held outside the Bancor ecosystem will gradually decrease due to the new BNT issuance and migration from exchanges

My concerns with this proposal are:
– It adds complexity to the contract, where CRV’s solution is on Balancer, not directly in their contract
– It also adds complexity to the system. If I’m providing liquidity, I want it to be as simple as possible and I don’t want to find out later that the protocol is seeping revenue out of me. Don’t penalize LPs!

There is the protocol co-investment, which is generating and burning fees. This somewhat counteracts the LM reward emission. We should wait for the LM reward distribution to begin and see how will the LPs respond as they get more and more BNT. I’m of the opinion that the protocol co-investment cap should be higher on large pools, because it’s deflationary and net positive for the long-term aspects of the system and it’s a better option, than taking revenue from LPs. In the later stages (after LM program) the fee burning could be used for buy backs. There must be always BNT inflation to support the growth of TVL in the system.

End game
– After the LM program is over the TKN liquidity will migrate from Bancor and there will be abundance of BNT in circulation. This is the reason why I think BNT burning is necessary, for now.
– after LM, if the TVL will gradually grow and BNT becomes scarce again by co-invested BNT burns, the generated fees could be distributed to BNT stakers (increases supply), or buy backs (increases value)

Final thoughts
– There is enough inflation to increase TVL over time, once the LM rewards get distributed
– Taking away a % of fees from LPs will decrease the LP retention after LM is over
– Why not use asymmetric fee distribution? i.e. 0.11% pool fee; 0.05% goes to TKN and 0.06% to BNT side
– If you distribute % of LPs revenue somewhere, it feels like you’re taking from them, but if it’s a percentage difference in rewards payed out it’s just that (change in incentive).
– If we want to increase the deflation, we should increase the pool contract co-investment on the capped pools, but this could be a non issue when the LM rewards get distributed and restaked into the pools
– I would seriously consider to change the co-invested fees generated to be used for buy backs. My argumentation is that there must be as much BNT in circulation as possible to support TKN deposits. This would increase its value and mitigate the IL coverage cost this way (I don’t know how it would be implemented though)

11

I responded to these points on Discord. Copying them here for completeness:

Thank you for your considered response. I agree with almost everything you have written. The rest of my post (below) is not intended as a counter-argument; I am interested in exploring some of these points further.

I’m of the opinion that the protocol co-investment cap should be higher on large pools, because it’s deflationary and net positive for the long-term aspects of the system and it’s a better option, than taking revenue from LPs.

I have been engaged in several discussions surrounding this idea, including a limitless cap for some pools, and certainly there are community members who see value in doing this (@coiner, @Rof are superbly qualified to continue this part of the conversation). To my mind, the ability of the protocol to mint its own BNT to support the TKN side is problematic, because it devalues the role of BNT holders in the system. The fact that there is little room in the pools is a direct consequence of the demand for BNT being high. The scarcity of BNT, and widespread interest in its use case (as evidenced from the lack of room in the pools), has the hallmarks of a moonshot asset. I think it is better to let the price increase, than the supply.

There must be always BNT inflation to support the growth of TVL in the system.

This is one of the few points we disagree on, and I think this statement is demonstrably false. An appreciation in the price of BNT will also support growth of the system, and is not contingent on inflation. For clarity - I am not saying that deflation, or a lack of inflation, are prerequisite to success of the protocol or the BNT token . However, I do want to point out that minting additional tokens does not create more value, but it simply spreads the existing value thin over a larger supply. Therefore, increasing inflation does not support growth of value (the ‘V’ in TVL). That sort of value comes entirely from utility and speculation. It doesn’t devalue the TVL, either. The inflation process form the view of the protocol is value-neutral.

After the LM program is over the TKN liquidity will migrate from Bancor and there will be abundance of BNT in circulation. This is the reason why I think BNT burning is necessary, for now.

There is no doubt in my mind that some dumping will occur. But, if anything, this is a good reason to not burn tokens . Burning tokens right now is concentrating value on the remaining supply - including the LM rewards. A helpful way to think about this is that some proportion of the value of burned tokens is being attached to tokens that are destined for a post-LM period dump. It doesn;t have to be this way. I have thinking about proposing putting a halt on token burning until after the LM dumping has occurred in ~18 months. This would mean that the tokens being dumped are fully-diluted, and therefore represent a diminished fraction of the BNT market cap, compared to a situation where token had been occurring over the same time period. To my mind, token burning is a bad idea until after LM is done. I’m not particularly concerned either way. I just want to bring your attention to a potentially flawed assumption about the role of token burning during the LM phase.

after LM, if the TVL will gradually grow and BNT becomes scarce again by co-invested BNT burns, the generated fees could be distributed to BNT stakers (increases supply), or buy backs (increases value)

I think the TVL will continue to grow during, and after LM. The idea that fees could be distributed to stakers is worth exploring, but it may be better to introduce it during LM, as opposed to afterwards. I think BNT holders are the least likely to dump, and single-sided TKN holders are almost guaranteed to dump. Therefore, distributing fees as BNT tokens to the BNT holders protects them from dumping; burning, on the other hand, siphons value to the BNT awarded to the TKN community during LM, and is destined to end up in circulation.

there is enough inflation to increase TVL over time, once the LM rewards get distributed.

tl;dr the LM rewards do not dramatically increase the TVL - at best it increases the BNT stake as a % of total BNT by diluting out non-participants. But LM does not create new value, it just dilutes existing value.

If I’m providing liquidity, I want it to be as simple as possible and I don’t want to find out later that the protocol is seeping revenue out of me. Don’t penalize LPs! Taking away a % of fees from LPs will decrease the LP retention after LM is over

I agree with this 100%. Taking a fraction of the LP’s profits, and then redistributing it to them, is a highly inefficient way to do things. A lot of value will be lost in gas payments, so it is also wasteful. This isn’t so much robbing Peter to pay Paul, as it is robbing Peter to pay Peter . Your asymmetric fee distribution idea is a good one, and could potentially skirt around this waste issue.

Why not use asymmetric fee distribution? i.e. 0.11% pool fee; 0.05% goes to TKN and 0.06% to BNT side

There is another way this could be done: the admin fee could come entirely from the TKN side. I have an idea to set this up in a way that will make the single-sided TKN providers happy. This is how it could work. The trade fee would be split in 3 parts: 50% goes to the BNT side, 45% goes to the TKN side, and 5% is used to buy-back BNT. The bought BNT is then distributed to the BNT holders in the same pool, and resulting in burning of minted BNT. This also has the effect of creating more room in the pool, allowing the TKN side to increase their staking amounts. In this way, the ‘tax’ revenue extracted from the TKN side is used predominantly to the benefit of the pool from which it came . This strategy works for federal tax districts, so it stands to reason it could work here, too.

If you distribute % of LPs revenue somewhere, it feels like you’re taking from them, but if it’s a percentage difference in rewards payed out it’s just that (change in incentive).

I think you are (correctly) calling attention to the semantics of how these issues can be handled. I think the best we can do during our discussions is to focus on the bottom line: who pays, the destination of that value, and the gas cost.

If we want to increase the deflation, we should increase the pool contract co-investment on the capped pools, but this could be a non issue when the LM rewards get distributed and restaked into the pools

Increasing the co-investment reduces demand for BNT. Think of it as the protocol diluting BNT holders. I am strongly opposed to increasing the caps further.

I would seriously consider to change the co-invested fees generated to be used for buy backs.

I would also consider this. If you have an idea of how this could operate, I would be interested to hear about it. Overall though, I think the burning mechanism is still a highly elegant solution, and we should think carefully about the consequences of supplanting it.

TL;DR version: @Majo raises some valid objections, and proposes some sound alternatives. Our major disagreement is the effect of inflation and token burning on TVL. Inspired by the criticisms raised here and on the TG group, I think there may be a more efficient implementation, where single-sided TKN holders contribute a small (5%) amount of fees to buy BNT tokens from the supply, which are distributed to the BNT holders on the opposite side of their own pool. This will cause minted BNT to burn, and also open space for additional TKN deposits. Everyone wins!

12

Great proposal,

I would like to add some of my thought and concerns:

Reasoning behind burning/revenue sharing.

Initially, I was in favor of distributing some fees towards vBNTs holders, but after making some research I am not sure if it is the best approach.
I would suggest here to go towards burning only model because;

  • If we do revenue sharing that would make the similar token as every other in the space (sushi, curve) that distribute tokens towards the network token holders.
  • We would not be competitive enough from the revenue sharing perspective, because competition does not burn tokens
  • We need to burn as many tokens as possible to achieve network inflation equilibrium (or close to eq), therefore revenu redistribution would decrease the burn rate.
  • I agree that might be better to introduce it after LM program is over.
  • Our unique value proposition is that you can choose which BNT pool you are able to provide the BNTs and benefit from the current higher volume.

I am not 100% against the idea to distribute fees towards vBNTs holders since I was advocating it from the beginning, however I would like you guys to take a look on it from the perspective I mentioned above.

On the other hand:

  • Even if we have less revenue because the portion of the fees are burned, the market will eventually find an equilibrium in regards to P/S ratio.
  • The average BNT fees index is more attractive for sure, however also gives exposure to low volatility assets pool as well

Admin fees structure.

  • Since we have problems with BNTs being staked, the admin fees should be charged only and only on the TKN side so the TKN providers would pay for their insurance.
  • BNT holders would be rewarded in this way via deflation or revenue sharing.
  • I would aim towards 10-20% of TKN fees to be collected and burned.

I would suggest adjusting the fee either dynamically or providing fixed numbers for big/small caps based on the average/expected IL.

Smaller caps pool should have higher % of fees on the TKN side being confiscated because:

  • Smaller caps are more volatile and have higher IL
  • Smaller caps pools are more shallow, therefore the APR for LPs is higher so they can share more from it without feeling the difference

Big caps shalound have lower fees because:

  • They are more competitive, everyone is looking for a single asset exposure for ETH or WBTC.
  • The are deeper, so the APR is lower for LPs
  • They are less volatile so less APR and lower IL

On the other hand, a stable 15% (for example, to be researched) is straightforward and easy to implement.

The fee level will be based (subject to research, currently in development):

  • Average impairment loss cost, (we are currently researching it)
  • Adjust to the current inflation rate to eventually maintain the stable 5% annual level.
13

Distributing fees is always superior to fee burn IMHO because fee burn has limited impact on token price whereas having a yield on vBNT will allow people to do a discounted cash flow analysis on vBNT thus giving vBNT an objective value based on actual usage and trading fees that accrues from usage.