Proposing Bancor v2.1: Single-Sided AMM Exposure with Elastic BNT Supply

We are excited to share the first proposal in the new Bancor governance framework: Bancor v2.1!

  • The proposed upgrade introduces single-asset exposure & impermanent loss protection to AMM pools
  • Initially, more than 60 ERC20 tokens will be supported as “protected” pools
  • vBNT, Bancor’s new governance token, can be generated by staking in a protected pool
  • If v2.1 is approved, Liquidity Protection will be awarded to eligible LPs retroactively
  • Protect your liquidity, earn fees and generate vBNT on swap.bancor.network
  • Voting will take place October 15-18, 2020, on app.bancor.network

Learn more about Bancor v2.1 mechanics:

Bancor v2.1 was created in direct response to the risks and complexities of providing liquidity to automated market maker (AMM) pools.

Existing AMM pools require LPs to forfeit their long position on their tokens and take on exposure to other assets in the pool. Protocols compensate for these undesirable aspects of liquidity provision by incentivizing LPs with airdrops and liquidity mining rewards. However, these incentives only temporarily mask the issues.

As a result, liquidity provision has progressively turned into a game of “hot money” where short-term LPs hop from pool to pool trying to extract enough rewards to compensate for the risk of impermanent loss. These days, many token holders either don’t participate or limit their participation in AMMs to avoid such risks.

Meanwhile, competition for liquidity providers between AMM protocols is fierce. In order to build a sustainable and profitable protocol, Bancor needs to offer something unique and non-replicable that liquidity providers value more than just plain subsidies.

Incentivizing long-term liquidity

Bancor v2.1 introduces two key features to AMMs:

  1. Liquidity Protection (i.e., impermanent loss insurance)
  2. Single-sided exposure

For example, if you stake 100 of a given token (“TKN”):

  • The protocol will protect the value of your 100 TKN, regardless of their price.
  • Meaning, you can enter a pool with TKN worth $100 and withdraw tokens worth $200 if the TKN price has doubled in the market. Plus swap fees + any liquidity mining rewards.

Liquidity providers accrue Liquidity Protection over time while collecting fees from swaps. The longer you stay in a pool, the more protection you earn until 100% coverage, increasing the ROI on your collected fees.

While AMMs today are chaotic in nature, Bancor Liquidity Protection is designed to provide order by making liquidity provision more deterministic for LPs.

Show me how it works

Contrary to other AMM protocols, Bancor uses its protocol token, BNT, as the counterpart asset in every pool, which allows for novel features such as Liquidity Protection and single-sided exposure.

At a high level, v2.1 proposes an elastic BNT supply to co-invest in pools alongside LPs and cover the cost of impermanent loss with fees earned from the protocol’s co-investments.

Let’s consider an example where 1 TKN = 1 BNT:

  1. A user deposits 100 TKN in a pool
  2. 100 BNT are minted by the protocol to match the user’s deposit
  3. From the issued pool tokens, half are associated with the user’s stake and half are associated with the protocol
  4. Volume goes through the pool
  5. Protocol pool tokens are now worth 110 BNT (due to fees)
  6. An outside LP deposits 110 BNT, which are burnt in exchange for the protocol’s pool tokens.
  7. Note that 100 BNT were originally minted and 110 BNT were burned, removing 10 BNT from supply.
  8. This burned BNT offsets the cost of impermanent loss when a future LP withdraws and spreads gains across all BNT holders.

What’s in it for BNT holders?

BNT holders stand to benefit from the v2.1 system in several ways:

  • Novel AMM features: Impermanent loss protection & single-sided exposure become available to both BNT and non-BNT liquidity providers.

  • Loyal liquidity: Liquidity Protection accrues over time and discourages opportunistic extraction of protocol rewards by “mercenary” LPs. Instead, incremental insurance incentivizes LPs who want to “set it and forget it”, driving sustainable liquidity, fees and demand for BNT.

  • Protocol revenue: BNT holders receive fees either directly when they stake BNT in a pool, or indirectly when the protocol burns fees it earns on the BNT it stakes.

  • Governance power: BNT holders who provide liquidity to a whitelisted (protected) pool receive vBNT, which represents their protected BNT stake and can be used to vote on upgrades to the protocol. This ensures that LPs and governance participants are one in the same.

How to stake, protect and vote

Since Bancor v2.1 changes the protocol’s tokenomics, it is necessary to gain community approval via Bancor governance in order to push the upgrade live.

BNT holders can immediately start earning Liquidity Protection by providing their tokens to a 50/50 whitelisted pool, protecting their stake and generating vBNT, the new Bancor governance token.

If v2.1 passes, Liquidity Protection will kick in retroactively from the time the LP first deposited and protected their stake.

Note that before the proposed upgrade, LPs can stake with Dual-Sided Protection (i.e., with two tokens), while after the upgrade, LPs gain the ability to stake with Single-Sided Protection (i.e., with one token). In both cases, the BNT share of their stake will determine the amount of voting power they receive, with 1 vBNT always equal to 1 BNT.

From there, vBNT can be used to vote on proposals on vote.bancor.network.

Below we propose an Initial Whitelist of roughly 60 pools, which can be changed in the future by Bancor governance. In general, it is recommended to only whitelist pools with some amount of liquidity and conversions. While some of the pools listed below already have sufficient activity, others still require it. We have indicated in bold which pools are already in the whitelist, meaning they are immediately available for Liquidity Protection. All other tokens below will be included once sufficient liquidity is added.

Proposed whitelist:
AAVE, ALEPH, ANT, BAL, BAND, BAT, BNB, BUSD, BZRX, CEL, CHERRY, COMP, CRO, CRV, DAI, DXD, ELF, ENJ, ETH, EWTB, FTT, GNO, gUSD, JRT, KNC, LEND, LINK, LRC, MANA, MATIC, MKR, MLN, MTA, NMR, OCEAN, OMG, pBTC, RARI, RCN, REN, renBTC, renZEC, RPL, RSR, SNX, SRM, STAKE, sBTC, sUSD, SUSHI, SWRV, SXP, TRB, TOMOE, UNI, USDC, USDT, WBTC, wNXM, XDCE, YFI, UMA, QNT, ZRX.

To get most updated list of whitelisted pools, go to Bancor Network, click “stake” and wait for the full list of pools to load.

Economic considerations

The purpose of Liquidity Protection is to attract liquidity providers to Bancor’s pools. Liquidity providers are necessary for Bancor pools to “do business”, and LPs have a choice of where to stake (or not stake) their liquidity. Therefore, the cost of Liquidity Protection is the cost the pools pay for doing business.

The key economic assumption of the v2.1 protocol design is that additional trading fees can both attract large numbers of LPs and, over time, exceed the cost of Liquidity Protection. If the Bancor protocol does not offer sufficiently attractive terms, liquidity providers won’t come, if the terms are too attractive, they will come but they will take the lion’s share of the profits.

Ultimately, like with any business, the key here is to have a robust cost and risk control framework in place that allows Bancor governance to fine-tune the offer to market conditions, and to find a spot that is attractive enough to generate sales and to still maintain a healthy profit margin.

We have been working with a team of financial researchers to study Bancor v2.1 and run economic simulations on the model – and the results are encouraging. Below we’ve used data derived from real AMM pools in the market and excluded incentivized pools. We show two fee-versus-cost scenarios. The two fee scenarios are, in our view, an optimistic scenario of 40% annual linear growth and a pessimistic view of 7.5% annual linear growth. For volatilities, we use annualized 50% and 100% volatility as reference, which aligns with data we’ve observed on mainnet.

fees2

fees1

The key here is that while the cost of Liquidity Protection grows over time, the fees from the system are expected to exceed it. In both scenarios, the break-even point (where fees outweigh the cost of Liquidity Protection) is after about 3 months. This of course strongly depends on the scenarios considered, and different scenarios can have different break-even points.

See further details in Bancor v2.1 Economic Analysis

Ultimately, once the protocol is launched there must be a careful analysis of whether the cost to the protocol of providing its services to LPs is justified by the fee earnings potential, and/or how it should be tweaked by adjusting protocol parameters and proposing design changes through governance.

Risk parameters

Especially during the initial phase of v2.1, it is important to have a robust system of limits in place that allow monitoring while minimizing risk. Specifically, the protocol has three lines of defense, each of which can be updated by Bancor governance:

  1. An overall limit of BNT that can be minted across all pools to match token-only contributions
  2. A per-pool limit of BNT that can be minted to match token-only contributions
  3. The rate at which LPs accrue Liquidity Protection

Dive in

This discussion is a starting point for this BIP and the governance vote, and it’s based on rigorous research on the dynamics of AMMs, economics and staking models. Our goal is to have the community discuss and decide on this critical next step of the Bancor Protocol.

In our view, Bancor v2.1’s value proposition to potential liquidity providers can be very powerful. The fee-earning-potential that this liquidity generates might very well exceed the cost of providing liquidity protection, especially as more liquidity providers are acquainted with the system. Also, contrary to the “we pay you tokens to provide liquidity on our platform” business model so often seen in the AMM market currently, the Bancor protocol seems to have a defensible moat. Even a developer copying the entire code of the protocol will not have the risk-taking Bancor protocol in the background, and this risk-taking is, combined with the BNT community, is the core of Bancor’s unique value proposition.

So what happens next?

Community Discussion October 12-14, 2020: Bancor’s governance platform (Discourse) will be the central platform for discussing and posting responses to proposals, with chat taking place on the Bancor Discord and Telegram channels.

Voting period October 15-18, 2020: Vote on Bancor v2.1 - app.bancor.network

A quorum of 20% of staked vBNT is needed to pass the Bancor v2.1 proposal. If passed, the upgrade will be pushed to mainnet in the following days.

We would like to thank the teams and individuals who helped develop and research Bancor Governance and the v2.1 upgrade, including Keyko , Topaze Blue and Michal Herzyk .

12 Likes

Finally a proposal that overcomes current AMM problems…

5 Likes

Bancor v2.1 is a must for the current Defi ecosystem, i definitely support this upgrade.

4 Likes

like ampl?it’s cool, i definitely support this upgrade.

3 Likes

great proposal, supporting the upgrade

3 Likes

Anybody who is willing to summarise the proposal? We’re not all first language English and/or just investors, not too tech savvy.

Would really appreciate.

2 Likes

when to vote? can not see on https://app.bancor.network/eth/vote/

1 Like

Voting should start today and will last for 3 days.

2 Likes

The vote will begin today and last for 3 days. We will post an announcement on Twitter when it’s live.

3 Likes

Voting on BIP1 is now live: https://twitter.com/Bancor/status/1316736339585495040?s=20

1 Like

:partying_face: Thanks to everyone who voted.

Space at YFI/BNT pool is limited at the moment. Is it possible to run a vote to increase YFI cap and deposit couple of YFIs at once?

It is possible - however, we are trying to keep the minting limits reasonable.

1 Like