🤷‍♀️The Problem

DEXes based on Automated Market Making are a cornerstone of DeFi. DEXes solve several problems associated with centralized exchanges, such as wash trading (e.g., Aloosh and Li (2021), Amiram, Lyandres and Rabetti (2021), and Cong, Li, Tang, and Yang (2021)), exploits (e.g. Mt Gox), outright fraud (e.g. Quadriga and, most recently FTX).

Moreover, CEXes undermine the crypto promise of free, accessible, and open capital markets, as they are based on centralized entities, capable of censoring accessibility or activities at their sole discretion.

However, currently, DEXes are plagued by significant inefficiencies, originating mostly from insufficient liquidity, making trading on DEXes excessively costly. This is especially true for the “99% of assets” i.e. all assets except for 5-7 most liquid ones.

Trades involving such assets suffer from insufficient direct liquidity, which results in indirect (“triangular”) trades. Such trades embed significant costs - duplicate fees, price impact risk, and gas fees. These excessive costs are of tremendous magnitude — based on our research or real trading activity - around $500M were wasted in 2022 just on Uniswap and its forks on Ethereum alone.

Triangular trade illustraion

This problem cannot realistically be solved by adding sufficient liquidity to all pools, even on DEXes with large TVL, as the number of trading pairs is close to quadratic in the number of assets. Adding liquidity into pairs with limited trading volumes is a losing proposition for liquidity providers in the presence of impermanent loss.

An additional problem we solve is dedicated specifically to LP yields and alignment of the economic incentives to support the protocol growth. Most dominant DEX protocols either use DAO governance mechanisms to decide how to allocate LP rewards; or use “Gauge voting” by LPs to decide how to allocate new emissions between pools. These schemes are not efficient enough in our view. The first scheme, i.e. using DAO proposals, is not structured enough and is inherently slow to implement; and the second scheme, i.e. using Gauge voting, has resulted in the past in whales exploiting their voting power to divert incentives in their direction at the expense of all other LPs. Additionally, both schemes suffer from lack of relevant data to make informed decisions for the benefit of the protocol as a whole.

A true Tragedy of the Commons.