This proposal is for discussion at this time. If sentiment is positive, I will refine and formally propose.
TLDR
This proposal seeks to reduce the vortex on DAI in Bancor3 from 90% to 10% in order to maintain DAI held in Bancor3 when the deficit is zero for DAI.
Abstract
As the $BNT price (in USD) increases, the deficit in the DAI pool will reduce until there is a surplus. At which point, I would expect a number of LP’s to withdraw as they can capture higher yields elsewhere.
Reducing the vortex from 90% to 10% should result in an (~9x) increase in the fees going to LP’s and so encourage AUM to stay in the pool. Higher AUM in the pool is prefered as it results in less price impact for a given trade value. Loosing all the DAI in Bancor3 can be expected to reduce overall volume in the protocol and should be avoided.
Motivation
The Bancor3 master vault contains $4.26 M DAI and has a deficit of $0.63 M USD. This is a 12.9% deficit. If the BNT price increases by 13% to ~$0.53 then the deficit will be zero and DAI LP’s can withdraw with minimal penalty. Such a price increase could be driven by $BNT out performance, or by merely tracking a more general increase in crypto prices ($BNT will tend to follow the average of the B3 omipool: $ETH, $LINK, $wBTC, $ENJ, $ICHI…)
When the deficit is zero, DAI LP’s have a choice:
- stay in Bancor3 earning 0.2% and risk becoming locked in again of $BNT drops in USD terms.
- move to a different protocol with higher yields and different risks:
Note, that as AUM drops there will be an increase in % fees for LP’s. Even so, there could be a significant drop in AUM before fees increase to a level that persuades LP’s to stay in the pool. As an example, dropping AUM by 90% would increase the current LP income from 0.3% to 3.0% and so be comparable with the Yearn vault.
However reduced AUM is likely to result in reduced volume. For a given trade size, the impact on the pool exchange rate increases as the pool AUM drops.
For a $8 M constant ratio (e.g. Uni v2) pool, a $20,000 trade will loose 0.5% to price impact / slippage. For a $800,000 pool, a $2,000 trade would have the same price impact on the trade. As AUM drops, all trade via that pool become more expensive.
Basically, smaller pools are less competitive and capture less volume.
By increasing the income for DAI LP’s before the pool runs into surplus, the intention is to reduce the drop in AUM. A LP getting 3% income is more likely to let the DAI ride than one earning 0.3%.
Impact
The obvious impact of this change is that less fee income would goto the vortex / Bancor protocol / BNT holders.
Over the last 7 days the pool has earn $2,040, so a 80% change in vortex would be ~ $233 less vBNT getting burn each day.
Complications
As DAI is one of the pools in the Almanak optimisation trail, such a change may complicate the analysis. However, as the Almanak experiment is targeting total pool fees, the % sent to the vortex should not impact that trial.
Note: DAI is selected rather than USDT or USDC as the latter 2 have much larger deficits. As BNT approaches $0.87, we may need a similar change for USDT and USDC.
This proposal assumes that pool vortex rates can be set for each pool. If only a single rate can be used for all pools, then this proposal is impossible to implement.
KPI’s
I would propose a number of KPI’s for this proposal. These are that 30 days after the DAI pool deficit is zero:
- DAI balance maintained > $3,000,000
- DAI pool income > $30,000 per day
- DAI LP income > 1%
A stretch KPI would be an increase in $bnDAI issued (i.e. net deposits into the Bancor3), currently 4,920,460 bnDAI are circulating.
For
Reduce the vortex fee on DAI pool from 90% to 10% until further notice.
Against
Leave the vortex fee unchanged.
Links
Bancor v3 Analytics
DAI yield - Bancor3
Bancor3 DAI yield - DefiLlama
DAI yield - DefiLlama
bnDAI - etherscan
example (neglecting price impact on the smaller pools…).
