The Economics of Automated Market Makers

Robin Fritsch, Samuel A. Käser, Roger Wattenhofer

TL;DR

This paper challenges the "race to the bottom" theory in Decentralized Finance (DeFi) by proving that Automated Market Makers (AMMs) like Uniswap can sustainably charge protocol fees without losing all their liquidity to zero-fee competitors. By modeling "sticky" trader loyalty and liquidity provider (LP) inertia, the authors derive optimal "take rates" that maximize protocol revenue while maintaining a dominant market share.

What Problem Does This Paper Address?

In the world of AMMs, there is a looming existential question: Are these protocols destined to become "common-good infrastructure" with zero profit margins?

Currently, Uniswap—the largest AMM—distributes 100% of its trading fees to liquidity providers. If Uniswap were to flip its "fee switch" and keep a portion (a "take rate") for the protocol, naive economic theory suggests LPs would immediately migrate to a zero-fee fork (like a "Sushiswap" with 0% protocol fees) to get a higher Return on Investment (ROI). Since liquidity follows ROI and traders follow liquidity, the original protocol would theoretically collapse. The authors investigate whether this "perfect competition" model holds true or if factors like brand loyalty and trust allow for sustainable protocol monetization.

Methods

The authors combine game-theoretic analytical modeling with empirical simulations:

  1. The Analytical Model: They introduce two key variables to represent market friction:
  2. Sticky Trade Volume ($s$): The fraction of traders who use a specific pool regardless of price optimality (due to UI preference, habit, or integration).
  3. Sticky Liquidity ($d$): The ROI premium LPs require before they are willing to move their capital to a competitor (representing trust or perceived risk).
  4. Equilibrium Derivation: They solve for the liquidity distribution where the ROI (adjusted for the "stickiness" factor $d$) is equalized between two competing pools.
  5. Trade Simulation: To validate the math, they built a simulator that routes trades through pools using the Constant Product Market Maker (CPMM) formula, accounting for arbitrage opportunities that arise when pools become imbalanced.

Datasets

The study uses real-world market data to ground its simulations:

  • Uniswap v2 ETH-USDC Pair: The authors extracted 10,000 consecutive trades starting from August 2021.
  • Liquidity Levels: They used the actual total liquidity present in the Uniswap v2 ETH-USDC pool during that period as the baseline for their two-pool competition model.

Key Insights and Ideas

  • Loyalty is a Buffer: The "aha moment" is that even a small amount of loyal (sticky) trade volume acts as a floor for liquidity. If 10% of traders are loyal to Pool A, Pool A can charge a non-zero fee because LPs will still earn enough from those loyal trades to justify staying, even if Pool B has zero protocol fees.
  • The Optimal Take Rate Formula: The authors derive a closed-form solution for the optimal take rate ($t_1^*$) when competing against a fork: $$t_1^* = 1 - \frac{1}{1+d}(1-s_1)(1-t_2)$$ This shows that the optimal fee increases linearly with trader loyalty ($s_1$) and the competitor's fee ($t_2$).

  • The "Fork" Defense: Even against a newly created fork with 0% fees and 0% loyalty, an established protocol with 10% loyalty and 10% liquidity stickiness can sustainably set a take rate of roughly 18% without losing its entire liquidity base.

Key Results

  • Sustainability is Possible: Both the analytical model and the simulations show a clear "peak" in protocol revenue. This confirms that there is an optimal take rate that is strictly greater than zero in most realistic scenarios.
  • Simulation vs. Theory: The simulations closely matched the analytical results, though they found slightly higher revenues for smaller pools. This is because "sticky" trades create price imbalances that attract arbitrageurs, which actually generates more fee volume for the smaller pool.
  • Case Study (Sushiswap): When modeling a pool against an established competitor like Sushiswap (which has a 16.7% take rate), the optimal take rate for the primary pool often lands between 20% and 26%, depending on the loyalty of the user base.

Limitations and Future Work

  • Two-Pool Limitation: The model currently only considers competition between two pools. Real-world DeFi involves a complex web of many competing protocols.
  • Static Parameters: The "stickiness" of traders and LPs is treated as a constant. In reality, if a protocol charges exorbitant fees, loyalty ($s$) would likely decay over time.
  • Nash Equilibrium: The paper focuses on one pool optimizing its rate against a static competitor. The authors suggest that future work should explore a Nash equilibrium where both pools dynamically adjust their take rates in response to each other.

Summary updated

Abstract

This paper studies the question whether automated market maker protocols such as Uniswap can sustainably retain a portion of their trading fees for the protocol. We approach the problem by modelling how to optimally choose a pool's take rate, i.e\ the fraction of fee revenue that remains with the protocol, in order to maximize the protocol's revenue. The model suggest that if AMMs have a portion of loyal trade volume, they can sustainably set a non-zero take rate, even without losing liquidity to competitors with a zero take rate. Furthermore, we determine the optimal take rate depending on a number of model parameters including how much loyal trade volume pools have and how high the competitors' take rates are.
BibTeX
@inproceedings{fritsch2022economics,
  author = {Robin Fritsch and Samuel A. Käser and Roger Wattenhofer},
  title = {{The Economics of Automated Market Makers}},
  booktitle = {AFT},
  year = {2022},
}
Figure