Motivation

Liquidity mining has been a popular method to bootstrap liquidity for decentralized exchanges and lending protocols, where the protocol offers token incentives to people who provide liquidity. However, the liquidity incentivized in such a way is often mercenary, with the farmers immediately dumping the reward tokens.

Why wouldn’t they? It’s free money after all.

The interests of the farmers are not well-aligned with those of token holders: the farmers are being rewarded at the token holders’ expense. Furthermore, once the token incentives are reduced or terminated, the farmers will take their liquidity and leave, leaving the protocol with nothing and making the token’s price trend to zero. It is clear that naive liquidity mining is not well suited for the long term growth of protocols that use it.

This document proposes a tokenomics system called Fungible Ownership Optimization (FOO) which combines several existing tokenomic innovations into a single mechanism. FOO achieves four goals:

  1. Incentivize liquidity
    • This is the bare minimum and achieved by naive liquidity mining
  2. Disincentivize mercenary farming
    • i.e. prevent farming-and-dumping as much as possible
  3. Provide the protocol with a large treasury of cash
    • Funds the growth of the protocol
    • Provides a floor for the token price
  4. Ensure high liquidity for the token
    • Decreases the volatility of the token price
    • Enables a higher incentivized TVL

It does so by aligning the interests of farmers, token holders, and token LPs.

Starting point: Curve vetokenomics

Curve uses a gauge system for incentivizing liquidity:

  • There’s a stream of reward tokens
  • The reward tokens are distributed across different liquidity pools
  • Voters vote on the distribution weights

In addition,

  • Voting power is obtained by locking the CRV token and receiving vote-escrow tokens (vetokens)
    • The longer you lock, the more vetokens you get
    • Your vetoken balance decreases over time and reaches zero when the CRV tokens are fully unlocked
  • Vetoken holders receive up to 2.5x boost in CRV rewards
    • The boost goes up from 1x to 2.5x as the share of the vetoken supply owned by the farmer goes up from 0% to their share of the liquidity pool
  • Vetoken holders receive fees generated by the protocol
    • Curve takes a protocol fee on swap fees

This is the starting point of FOO. The rest of the document will use LIT as the example project token.

Balancer LP token as vote-locked token

Instead of locking LIT, we can instead require voters to lock Balancer LP tokens of LIT to obtain vetokens.

The reason Balancer is used instead of some other DEX is that Balancer supports arbitrary weighting between the tokens in the liquidity pool, so that we can use something like a 80-20 LIT-ETH liquidity pool that ensures the LP token’s value is sufficiently correlated with the price of LIT.

Since the farmers will likely also hold vetokens, either for boosting or for voting on the gauge weights, the farmers are now also the LPs for LIT. This makes it less likely that the farmers will dump the LIT rewards they receive, since they would be dumping into their own LP position.

However, it is still possible for farming-and-dumping to happen, since a farmer without any vetokens can still earn LIT rewards. It is 2.5x less capital efficient, certainly, but a sufficiently capitalized actor (e.g. Alameda) can still be sufficiently incentivized to farm-and-dump by simply throwing more liquidity in. This leads us to the next modification.

No vetoken? No rewards

To fully disincentivize farming-and-dumping, we will change the boost mechanic such that a farmer without any vetokens will receive zero reward tokens no matter how much liquidity they provide.

Curve modifies a farmer’s weight in a staking pool using the following formula: $$ b^* = \min (b, 0.4b + 0.6B\frac{w}{W}) $$ Where $b^*$ is the weight, $b$ is the liquidity provided by the farmer, $B$ is the total liquidity of the pool, $w$ is the amount of vetokens the farmer has, $W$ is the total vetoken supply.

Basically, what this means is that if a farmer has no vetokens, their liquidity is multiplied by 0.4x when deciding their weight in the staking pool. When they have enough vetokens (at least $\frac{b}{B}W$), their weight goes from 0.4x to 1x, which translates into $\frac{1}{0.4} = 2.5$x boost.

For FOO, we will change the formula to the following: $$ b^* = \min (b, B\frac{w}{W}) $$ This means if a farmer holds no vetokens then their weight in the staking pool is zero. To illustrate the forumla better with some examples, let’s say that a farmer Alice has a 60% share of a liquidity pool and 10% share of the vetoken supply. Their weight in the staking pool is: $$ b^*_{Alice} = \min(0.6B, 0.1B) = 0.1B $$

Say that another farmer Bob has a 20% share of a same liquidity pool and 20% share of the vetoken supply. Their weight would be: $$ b^*_{Bob} = \min(0.2B, 0.2B) = 0.2B $$ So even though Alice has 3x the liquidity of Bob, she only gets half the rewards Bob gets, because she doesn’t have enough vetokens. She’s incentivized to either increase her vetoken balance or to decrease her liquidity balance, until the two share values are equal.

This modification forces the farmers to also be token holders and LPs for the token, which further disincentivizes farming-and-dumping.

Moreover, since all farmers are LPs and all LPs are probably farmers, the incentives going towards the farmers also incentivizes LPing, so the staking reward serves two purposes simultaneously: incentivize protocol usage and incentivize LPing.

Call option as reward token

Instead of using LIT as the reward token, we can use call option tokens for LIT as the reward token. This has the benefit of enabling the protocol to accumulate a large cash reserve regardless of market conditions, as well as letting loyal holders buy LIT at a discount.

It’s best to illustrate this mechanism with an example. Let’s say the price of LIT is \$100, and there is a call option token oLIT that gives its holder a perpetual right to buy LIT at 90% of the market price. The protocol issues 1 oLIT to a farmer Alice, who immediately exercises the option to buy 1 LIT for \$90 and sell it on a DEX for \$100. The tally of gains & losses are as follows:

  • The protocol: -1 LIT, +$90
  • The farmer Alice: +$10
  • The DEX LPs: +1 LIT, -$100

Compare this to regular liquidity mining where the farmer doesn’t pay anything to the protocol:

  • The protocol: -1 LIT
  • The farmer Alice: +$100
  • The DEX LPs: +1 LIT, -$100

We have the following observations:

  • Reallocation of cash: Using oLIT instead of LIT as the reward token effectively transfers cash gains from the farmers to the protocol, and the LPs for the token are not affected.
  • Trading off incentivization efficiency for protocol cashflow: In our example, for each LIT token issued by the protocol, the farmer Alice only gets \$10 of rewards instead of \$100 in the case of regular liquidity mining, which is less efficient. The higher the discount is, the more efficient the incentivization is, but the less cash the protocol gets.
  • Effectively a continuous token sale: Instead of giving away tokens for free in regular liquidity mining, we effectively turn incentivization into a continuous token sale at the current market price, which enables the protocol to potentially capture a lot more cash compared to a one-off token sale since the protocol would be selling tokens at a higher price when the market price goes up.

Furthermore, when option reward tokens are used in FOO where the farmers are the same people as the token LPs, the tally becomes:

  • The protocol: -1 LIT, +$90
  • The farmer-LP: +1 LIT, -$90

which means that as the farmers receive oLIT rewards, they get the right to buy tokens from the protocol at a discount and increase their ownership. Over time, the protocol ownership will be transferred away from holders who aren’t providing liquidity and to the farmers who are providing liquidity, which optimizes the protocol ownership.

The tally also stays the same regardless of whether the farmer dumps the LIT gained from excercising the option on the market, since the farmer and the LP are one and the same. Because of this, we can assume that a lot of the farmers will not sell the LIT but lock it into vetoken and improve their yield.

Vesting period? Nah

The addition of a vesting period doesn’t actually make sense, even though ostensibly it might. FOO already incentivizes the farmers to keep their reward tokens instead of dumping them, so adding a vesting period to the option decreases the value of the option token and thus the incentivization efficiency, without necessarily decreasing sell pressure.

Protocol owned liquidity is detrimental to FOO

As we can see from the reasoning above, merging the farmers and the LPs into the same entity makes farming-and-dumping far less likely and can potentially decrease sell pressure significantly. This means that somewhat paradoxically, having liquidity outside of the vetokens actually increases sell pressure, since the farmers now have someone else to dump on. Thus, it wouldn’t make sense for the protocol to provide large amounts of protocol owned liquidity, as it incentivizes farmers to dump on the protocol.

Cash reserve sets floor for token price

Having a large cash reserve that acts as indirect backing to the protocol token sets a rough price floor for the token and decreases downside volatility. The more the token price goes up during the initial bull run, the bigger the cash reserve is, the less the downside volatility is when the bull run ends and the more prepared the protocol is for the next bull run.

Conclusion

Like many things in the crypto space, FOO is an experiment, and our assumptions & reasoning may very well turn out to be erroneous. The endless possibilities for experimentation is what drew us to the space in the first place.

If we are lucky, FOO will improve upon existing vetokenomics systems and better align incentives between LPs and token holders, and may even lead to a powerful tool for bootstrapping long-lasting DEX liquidity. Only time will tell.