- An analysis of impermanent loss in Uniswap V3 found that approximately half of users providing liquidity to the protocol are losing money when compared to HODLing.
- While Uniswap V3 generates the highest trading fees of any DeFi protocol, impermanent loss dominated the fee income in over 80% of the pools analyzed.
- Uniswap pools included in the analysis generated $199m in trading fees and incurred $260m in impermanent loss, leaving a net loss of over $60m and 49.5% of LPs with negative returns.
- The percentage of users suffering negative returns was as high as 70% in certain pools; the proportions of users suffering losses in key pools include MATIC/ETH (51%), COMP/ETH (59%), USDC/ETH (62%), COMP/ETH (59%) and MKR/ETH (74%).
- The study also found no statistical evidence that users who adjust their positions more frequently performed better than users who don’t, calling into question the widely held belief that “active” LPs outperform “passive” LPs in Uniswap V3.
- To the contrary, the study found that beyond the 1 hour horizon LPs who staked for longer periods on average lost less than those who were staking only for short periods.
Automated Market Makers (AMMs) have evolved into the cornerstone of decentralized finance. Users have staked over $30 billion in AMMs across every major blockchain, generating billions in trading fee revenue per year. However, the financial risks associated with AMMs remain poorly understood. Often overlooked is the cost of providing liquidity, known as impermanent loss.
A new study peels back the curtain on liquidity provider profits in AMMs today. Analysis of 17,000+ wallets providing liquidity in Uniswap V3 reveals that roughly half of users are suffering negative returns on their staked capital due to impermanent loss.
While Uniswap V3 generates the highest trading fees of any DeFi protocol, impermanent loss entirely wiped out fee income in over 80% of the pools analyzed.
Only 3 of the 17 pools analyzed earned fees that exceeded the impermanent loss.
Seventeen pools were analyzed in the study, accounting for 43% of Uniswap V3’s TVL. Pools were chosen by size (pools with less than $10m TVL were excluded), data availability and token composition (like-kind and stable-to-stable pools like renBTC/WBTC and USDC/DAI were excluded).
Analyzed pools generated an impressive $108.5b in trading volume and $199m in fee income from May 5th to September 20th, 2021. However, during the same period, the pools incurred over $260m in impermanent loss, leaving 49.5% of LPs with negative returns.
In certain pools, the percentage of users suffering negative returns was as high as 55-70%, including MATIC/ETH (51%), COMP/ETH (59%), USDC/ETH (62%), COMP/ETH (59%) and MKR/ETH (74%).
On the right, the percentage of wallets in each of the analyzed pools earning money (green) vs. losing money (red). On the left, the mean returns of losers (red) and winners (green).
After finding that the average Uniswap V3 liquidity provider underperforms a basic buy and hold strategy, researchers sought to understand whether certain groups consistently outperform others. In particular, the study examined whether “active” users who adjust their positions more frequently performed better than “passive” users who don’t.
To do so, the duration of pool positions was compared to the profits earned for their LPs. The assumption was that, on average, shorter-term positions likely belong to more active LPs. Those who go into the market and leave it shortly thereafter are likely to be implementing a premeditated strategy and can be considered more sophisticated, or at least more active, than those who stay longer. By analyzing the profitability of positions by duration, researchers sought to test the hypothesis that active LPs perform better. However, no correlation could be found between shorter-term positions and higher profits. Across all groups, impermanent loss outpaced fees earned, calling into question the widely held belief that active LPs outperform passive ones.
IL vs. fees measured by position duration.
The only group that consistently made money when compared to simply HODLing were just-in-time or “JIT” liquidity providers who provide liquidity for a single block to absorb fees from upcoming trades, then instantly remove their position. This liquidity was provided intra-block and it did not cause any meaningful IL, leaving 100% of the fees as profit. All other segments have an IL/fees ratio that is greater than 1, indicating a net loss of value. The upper bound on this ratio was as high as 1.8, meaning that liquidity providers incurred $180 in IL for every $100 worth of fees, leaving a net loss of $80.
“Our core finding is that overall, and for almost all analyzed pools, impermanent loss surpasses the fees earned during this period,”
the study’s authors concluded.
“Importantly, this conclusion appears broadly applicable; we have collected evidence that suggests both inexperienced retail users and sophisticated professionals struggle to turn a profit under this model.”
Learn more about impermanent loss.