People Misunderstand LP Reward Dilution

Frequently I see people with an exactly backwards idea about the effect of additional mAssets and corresponding mAsset-UST pools on LP rewards.

My aim in this post is to prove that not only does adding additional pools not reduce LP reward APR, but also that it is in fact the best way to boost and maintain LP reward APR going forward.

The amount of MIR given to the pools per amount of time is fixed. The rewards are distributed according to weight. Thus the average LP reward APR is a function of only two things:

  1. the total value of all liquidity pools (excluding the MIR-UST pools which are rewarded separately see Mirror Token (MIR) - mirror)
  2. The price of mirror
    The number of pools is irrelevant. There could be 1 pool or 100 pools, the average MIR per LP-value would be the exact same.

The other factor for total LP earning is the volume traded.

I contend that we need to bring up the VOLUME. Not only does volume increase rewards in the form of swap fees, it is also one of the two most important stats people look at for a defi project (the other is TVL:market cap, which is a very nice looking .3 for MIR). So more volume will mean more fees and a higher price for MIR. A continuously increasing value of total liquidity provided without a corresponding increase in volume is the worst case scenario for LP returns. The best thing for LP earnings long-term is the addition of many more mAssets (preferably at a regular interval).

Why would you trade at a market that has 23 options, when there are markets that have hundreds of options?


You make a good point, but there are a few extra factors to consider.

Per LP, the reward is the same, but this makes an assumption that the users are rational and rebalance their holdings across all pools, which makes an additional assumption that users have no preference between mAssets (which is not true).

The second problem (currently) with adding more pools is that the total rewards get split between N+k pools (k = no. of new pools). Thus even if users were indifferent between different mAssets, the liquidity per asset will decrease assuming no additional increases in liquidity at a given fixed point in time. I would argue that a small set of choices with thick liquidity would be much better than 1000 choices with a trade size of 100 UST at spot that is within the best 5%.

Of course this should serve as an indication that the liquidity structure should change in general, which is something we have been continually studying.

1 Like

I see the variance in rewards by pool to be a feature that generally rewards higher risk pools with a better return, so that seems like a positive feature to me.

I couldn’t agree more about deep liquidity being better, but I can’t help but feel like we should be wider than we are ride now.

I agree with the extra factor, but there is an additional point.
From your analysis we can notice that very likely the amount of MIR received will decrease as adding lp and having money inflow directly in the pools will have this effect. On the other hand the MIR price should go up as we would end up with higher TLV and higher minter activity.
A long term MIR exposure makes sense if you think that minters activity trend is upward: this would lead to an high return from staking and incentivize to buy and hold. A second dimension is, ofc, the expansion on other chains of the minted assets: this case is also increasing minting, hence stakers rewards and LP activity (volume), with, very likely, only a little dilution of the lp returns.
My opinion is that we should find a kind of balance between new lps and expansion of markets for the existing mAssets. In this perspective I tend, anyway, to consider still very little the number of massets: I would be very happy of seeing a double number of LPs and I am quite convinced that the effect on minters activity, TLV and MIR price would be good.

Dev team members have said in the past that ideally Mirror should eventually become the liquidity provider of last resort for mAssets.

I think what we really need before we start adding a large amount of mAsset listings is for more capital efficient DEXs to come online and push mAssets aggressively on their platforms. Perhaps the answer is better implementations of AMM DEXs like Uniswap v3, perhaps it’s orderbook DEXs like Serum/Vega/Injective. But it needs to be something not 100% dependent on incentivized liquidity pools since this is not sustainable nor capable of supporting hundreds of assets.