Submitted by Mechanism Capital
We propose to distribute 90,000 CREAM tokens in the protocol treasury to users that provide liquidity to CREAM’s capital markets across Ethereum (60,000 CREAM) and Binance Smart Chain (30,000 CREAM). The rewards will be vesting with a cliff, such that new CREAM would only begin to enter the market months after the start of the Liquidity Mining Program and will be released to LPs in a linear fashion.
For: Allocate 90k CREAM tokens from the treasury for a targeted liquidity mining program with a duration of 90 days
Against: Nothing changed
CREAM has a warchest of tokens, and much of these have already been earmarked by governance for Liquidity Mining purposes. Other competing platforms (like Aave) are beginning to consider offering LM incentives in order to attract TVL, meaning that the battle for liquidity in lending/borrowing markets is heating up. Additionally, CREAM lost a good amount of TVL after the Alpha Hack, which many mistakenly believe to have affected CREAM lenders.
A targeted Liquidity Mining program can boost utilized liquidity and efficient use of the protocol, while also helping to bootstrap the BSC arm of the protocol.
From the above, we see that the CREAM has over 773k unallocated tokens, some of which can be put to use as part of liquidity mining programs. For this specific program, we propose using 90k CREAM tokens to be distributed over a period of 90 days. Here is a table of how this proposed program stacks up against Aave’s and Compound’s respective LM programs (note that “annualized rate” for CREAM assumes zero LM inflation apart from this 90-day proposed program).
Although the dollar amount of CREAM distributed daily is lower than the dollar amount of AAVE and COMP distributed, the CREAM platform currently has significantly less liquidity than Aave and Compound, so the rewards are actually proportionally higher (on a TVL basis) than these other LM programs.
In addition to restricting rewards to a brief (i.e. three month) incentive period, the CREAM LM program will also target specific assets. Instead of attempting to grow liquidity on the platform by distributing rewards evenly across assets and chains, we believe that LM should optimize to grow liquidity in the most effective and productive ways possible.
Targeting stablecoin lending
In general, stablecoin lending is less common and less attractive to crypto native lenders, who would rather deposit non-stablecoin assets and borrow stablecoins against them as a form of leverage. In part for this reason, borrow APYs tend to be quite high (upwards of 20% on CREAM) and utilization rates for stablecoin deposits are much higher than they are for other assets like ETH, wBTC, and DeFi governance tokens.
CREAM can help defray the opportunity cost of lending stablecoins by offering CREAM token rewards specifically to stablecoin (i.e. USDT and USDC) depositors. This will create a virtuous fly wheel effect, as more stablecoin deposits will free up more stablecoin borrow liquidity, initially driving down borrow APYs until the additional stablecoin loans are met with borrowers flowing in from other platforms. Moreover, users that prefer to gain extra leverage on their crypto assets will provide more liquidity to non-stablecoin pairs in order to take advantage of additional stablecoin borrow liquidity. In short, incentivizing stablecoin deposits over all other asset deposits (or borrows) will give the CREAM platform the most “bang for the buck.”
Why asset utilization matters
Non-productive deposits—i.e. deposits that sit idle on the platform and are not actually loaned out to willing borrowers—do not produce as much revenue to CREAM holders (in fees). Although revenue declined sharply after the Alpha attack, as a large supplier of FTT removed their collateral, CREAM revenues have since recovered to what they were before the hack, even though TVL remains lower.
As Luciano has pointed out, this is due to the fact that the collateral that was removed was mostly idle collateral that was not being met with equivalent borrow demand, and thus not creating meaningful revenue for the protocol. In fact, although CREAM has less than $200 million (in v1), it is generating over one quarter of what Aave generates in revenue despite Aave having over 20x the TVL. The goal of any liquidity mining program should not be simply to increase TVL, but to increase productive TVL—and targeting stablecoin lending should accomplish just that.
As mentioned above, CREAM tokens will be distributed to liquidity providers in a delayed fashion. The liquidity bootstrapping program will last 90 days, during which no rewards will be distributed. Afterward, there will be an additional 30 day cliff, after which rewards will begin vesting on a weekly basis to be claimed by liquidity providers. Although the tokens will be earmarked for distribution within 90 days, they will be distributed (after the cliff) over a period of 180 days.
The purpose of this distribution schedule is to ensure that the market will be able to absorb any additional sell pressure that results from this bootstrapping program. Liquidity is fickle and mercenary, so we should not delude ourselves about the farm-and-dump potential that programs like these hold. To mitigate some of the worst of the selling effects, this program inflates circulating supply very slowly and over an extended period of time, much like the way Synthetix allows stakers to claim new rewards 52 weeks after their initial deposit.
Why 90k tokens?
Our recommendation of 90k CREAM tokens (or 1k distributed per day for 90 days) is an intentional one. If there are 2.925m CREAM tokens in total, the CREAM from this program represents only 3% of total supply, but of circulating supply it constitutes 14%. Rather than distributing a larger number over a longer period of time, we believe that this aggressive, short-term plan will not set in motion a permanent mechanism that contributes to sell-side pressure (e.g. such as Curve’s token distribution through near-perpetual liquidity mining), but it also provides a substantial amount of new issuance to be claimed by liquidity providers.
On a more granular level, we can map out—using a method we have called liquidity targeting—why this amount of new CREAM is squarely within the optimal range for a liquidity mining distribution like this one. Above, we explored why TVL is a flawed indicator of the health of a lending/borrowing platform (given the large discrepancy between proportional revenue for CREAM and Aave). Instead of starting with TVL as a target metric, it makes sense to target a certain revenue number that would make CREAM even more robust relative to its competitors. Here, we believe that CREAM should aim to close the gap between its and Aave’s daily revenues. CREAM is already undervalued on a revenue basis relative to its competitors, but achieving something closer to revenue parity with Aave would not only increase liquidity on the network but would help to cement CREAM as one of the tier-1 lending/borrowing protocols and would, ceteris paribus, drive substantial value to the CREAM token—potentially even offsetting sell-side pressure that results from liquidity mining inflation.
To achieve revenue parity with Aave, CREAM would need an increase in productive deposits, as mentioned above. An additional $200 million of stablecoin deposits—assuming maximum utilization—would generate an additional 8k in revenue per day. Will 90k CREAM over 90 days bring in this additional $200 million?
A distribution of 1k CREAM per day at a price of $125 per cream means $125k in daily token rewards, or $45 million annualized. If the total assets earning yield were $200 million, that would be 22.5% APR purely on the CREAM inflation alone (i.e. not even factoring in the organic interest); if there were $500 million, that would be 10% APR, etc. Our estimate, given alternatives that are available elsewhere on the market, is that yield has to be above 20% during this period (at static CREAM prices) to be attractive to LPs, meaning that we can expect total yield-earning assets to be around $200 - $400 million. Excluding the Iron Bank, there is less than $50 million in stablecoin liquidity on CREAM across v1 and BSC, so if these rewards were to target only stablecoin deposits on these parts of the platform, we can anticipate additional stablecoin inflows of at least $100-200 million, perhaps even closer to $300 million.
Stablecoin deposits across CREAM v1 and BSC have approximately a 75% utilization rate, which is paralleled on other lending/borrowing platforms like Aave and Compound, so we can further anticipate that whatever additional stablecoin lending liquidity enters CREAM will be utilized at a similar rate (75% is the maximum utilization rate for stablecoin lending on CREAM). If we assume an 75% utilization rate on inflows of $200 million, and we assume that borrowers are paying 20% APR in interest, then that would be an additional $30 million in annualized interest, or $82,000 per day. With a reserve factor of 10% for stablecoins,* that comes out to an additional $8.2k per day in revenue (over $2 million annualized) to the CREAM reserve (and by extension to CREAM holders).
The split between CREAM v1 and Binance Smart Chain
The purpose of this liquidity mining program is to bootstrap protocol liquidity across v1 and BSC, so the question becomes: how much CREAM should be allocated to each?
We believe that bringing more liquidity to CREAM’s BSC arm should be a priority for the protocol. With Ethereum gas fees reaching record levels, BSC recently penetrated through to the crypto mainstream, with leading projects accruing billions of dollars of volume and value locked. Although the migration over to BSC has slowed of late, there remains a great deal of liquidity on this chain, and CREAM should make an effort to establish itself as the leading lending/borrowing protocol on BSC. To that end, we propose allocating one third of LM tokens (30k) to BSC stablecoin lending (with 60k allocated to v1).
This distribution of LM tokens to v1 (Ethereum) vs. BSC is proportional to the current distribution of liquidity across the platforms, since incentivizing liquidity on Ethereum and BSC is the core goal of this program.
*Note that the current reserve factor for stablecoins is 5%. However, this is below the market average (Aave’s reserve factor for stablecoins, for example, is 10%). There is already work being done to revise the reserve factors for the Iron Bank, and we propose that a similar revision in the reserve factors (at least for stablecoins) for v1 and BSC is also in order. However, a revision of the reserve factor may require a separate governance proposal.