Discussed this briefly in Discord, but moving the discussion here.
With the current boosted farms ending in 8-days and with the intention to add USDT this month, we need to think longer-term about how we want to use TRU to incentivize different lending pools.
Some facts:
“Normal” TRU incentives for lending had been 145,467 TRU/day in the TUSD pool.
Per the original allocation, we allocated 195,097,500 total TRU to lending pool incentives and estimated these would go for 1140 days (~3 years). This means our original expectations were that we’d distribute 171,138 TRU/day to lenders across all pools.
The 2x’d currently ongoing rewards for the USDC and TUSD pool are 290,934/day per pool for a total of 581,868 TRU/day.
If we stayed consistent at ~600,000 TRU/day, we’d only have incentives left for 270 days (ending March 7, 2022)
Points to Discuss:
When we add USDT, wETH, and wBTC (which I think are the next three assets we will add), how will we incentivize them?
How long do we want the lender incentives to last? I think another 1.5-2 years is probably conservative with longer preferred. As an example though, if we wanted lender incentives to run for another 2-years, we could only distribue 222,630 TRU/day across all pools. You can see the number of remaining TRU here allocated towards lender incentives here and play around with your own assumptions of TRU/day.
This is probably the first compelling argument I’ve come across for not burning excess Uniswap incentive tokens, but instead reallocating them towards lender incentives to keep a large buffer.
What about tapering TRU rewards with utilisation in the same way that over collateralised platforms incentivise deposits? That way you don’t spend as much on pools in general and directly incentivise the pools that are being used by borrowers.
Incentives for assets like Eth and BTC should be small, if any. With USD, there is competition over opportunity cost of what you could be doing with that money, but there is not much you can do with Eth and BTC. I think Eth in particular is going to be a tough market because people could just stake it instead and have much lower risks. Borrowers have not demonstrated they will pay the 10%+ that would justify lending over staking.
Going back to those assets, I would want to start short because its not clear if there is actual demand to borrow them. Blockfi has been dropping its interest payments for Bitcoin, which suggests there isn’t much. I would not want to commit to a 2 year incentive run, then a few months in we realize there is little demand.
The emission of TRU should be connected to the utilization of the specific pool. For example right now having the TUSD at <20% but same emission as the USDC pool which sees a lot more demand is just senseless.
USDT and USDC should have the largest demand , looking at their marketcap compared to TUSD. Therefore these 2 pools should see higher TRU emission than the TUSD pool. The TUSD pool seems obselete cause USDC/USDT are the most widely accepted stablecoins on every exchange.
if we can link incentives to utilization, we would make a much better use of our emissions… i personally feel 2 years down the line is very limited… i would rather aim for a daily average that gives us 3 years with a degressive model where we gradually reduce the daily emission rate over time… protocol growth should drive price increase which should compensate for the lower emission rates and increased size of the pools…
if we include the uniswap incentive tokens, we have approx 200k/TRU per day available over 3 years… maybe a model that would make sense is one where we average 250k/tru per day in year 1, 200k/day in year 2 and 150k/day in year 3… just an idea to start with… without the uniswap tokens remove roughly 50k per day over 3 years…
also agree with the fact that non stablecoin assets should have very limited or no incentives at all.
I support creating a formula to lower TRU emissions over time. This would prevent reducing rewards too quickly and losing TVL due to a sharp drop in emissions. We have a staking gauge that we want to use which just needs a 3rd party audit.
In the staking gauge, TRU is distributed per block, and split between pools based on allocation points. We could create a distribution model which slowly decreases over time, allowing us to extend the remaining TRU rewards over several years
If we wanted to incentivise a specific pool, we could re-balance the allocation points. Much like how SUSHI or CRV rewards currently work
Another thing which we might want to work on Unlock 5 is requiring TRU to be locked in order to boost rewards. That way only people who are locking TRU can get access to farm rewards. This is what CRV does and it seems to work very well.
If we launched full governance and overhauled TRU incentives at the same time that could be a knockout move. Ideally after lines of credit the smart contracts themselves will be closer to a final state, where we are mainly focusing on how to automate generating a credit score rather than actually changing how the protocol works.
So discussing with Hal just now and he suggested that we bring total daily lending pool emissions down by 10% to 523,681.2 TRU/day, launch USDT, and split incentives equally across the three pools (174,560.4 per pool).
Broadly speaking, agree with the proposal (lowered rewards, split across pools) as stated @ryan.rodenbaugh@hal
I’m looking at two things as I offer that reply:
1. How’s our TVL tracking against market cap?
TVL has more than doubled (100m → 250m), while market cap has remained relatively stable (~100m). With the ongoing assumption TVL is used as an industry standard way of gauging growth/progress (generally leading to growth in market cap), the downward pressure on market cap in the face of rapidly growing TVL is, in my opinion, the rate of TRU outflows.
My sense is that the amount of TRU we’re spending to attract TVL is suppressing growth in market cap, and we’d get similar results in terms of stablecoin deposits with growth in market cap by offering smaller TRU incentives.
2. What is the utilization rate of our existing pools?
Thanks for large loans from a few lenders, our utilization has grown - but before this time, we had a lot of idle capital (in both TUSD and USDC), which we were paying top dollar for in TRU incentives. I think it’s worth decreasing incentives in line with utilization rates: the more borrower demand there is, the more valuable new deposits are - and right now, capital is cheap.
That doesn’t follow what has been suggested at all. Equally splitting the incentives across the pools sounds like a terrible idea. Is this due to a development restriction that requires a lot more time to implement it based on utilisation and thus it’s easier to just drop the incentive rate?
I’m a fan of requiring TRU to be locked in order to boost rewards though, that would remove some sell pressure of the token.
also, do we have the confidence that we can secure a significant volume of new loans to get usdt utilization rate at a reasonable level quite rapidly? if not we will just be spitting out TRU for free… im asking since someone mentioned our existing borrowers are pretty much maxed out for the moment…
For short term -
What if we continue the 2x boost for tfUSDC and let the 2x boost on TUSD expire (i.e. tfTUSD revert to 145,467 TRU/day)?
Given 67% utilization on the USDC pool, I’m happy to continue boosted incentives there. If we end the TUSD boost, the tfTUSD farm APY would drop to ~17% and I’d expect some TUSD to move to USDC. Given TUSD’s low utilization and borrowers’ preference for USDC, I think this would be a good thing.
This would reduce our total lender emissions to 436,401 TRU/day.
edit: seeing as we want to launch USDT this week, I support the 500k TRU/day total emissions proposed here.
For the long term –
I like @carlj’s idea of having 3 years of runway for lender incentives in a degressive model. I tried making a rough draft below:
How the emissions curve (for lender rewards) could look:
Start with current emissions to lenders, decrease emissions by 25% every 90 days (Google Sheet)
How rewards would be split between pools:
Let TRU stakers set % of rewards allocated to each pool via allocation points in the staking gauge. Governance may decide to replace manual gauge staking via a performance-based mechanism in the future.
I also support @hal’s idea for incentivizing long-term TRU staking w/ boosted farm rewards. This could be a good mechanism for encouraging lenders to participate in staking and governance as the protocol matures.
