For challenge #3 of the Uniswap Community Research Program, the community explored the latest studies and research related to Loss Versus Rebalancing (LVR). This is a large, highly topical area that we will look to breakdown in multiple parts. This challenge will include the top submissions from our community, including a comprehensive list of resources and links with which to dive into LVR.
LVR is a crucial metric for understanding the efficiency of liquidity provision in AMMs like Uniswap. This analysis synthesizes insights from UCRP submissions to quantify LVR's impact on Uniswap v2 across various networks, including Ethereum mainnet and Layer 2 solutions.
LVR Quantification and Methodology
The analysts used different approaches to quantify LVR, providing a rich perspective on this complex topic.
Our challenge #3 winner, Respired.eth, used the formula LVR = σ²/8 for continuous-time models, where σ represents price volatility. This approach aligns with the research by Columbia University that first termed LVR. The analysis focused on calculating LVR for the top 40 pairs on Uniswap v2 across different networks.
Other submissions leveraged alternative methods to estimate LVR, focusing on transaction data and price movements. Prs1205 calculated LVR as the difference between AMM execution prices and market prices, while Mawutory used a method based on daily price volatility.
Cross-Network Analysis
ETH Mainnet
Respired.eth's analysis showed that on Ethereum mainnet, stablecoin-volatile pairs accumulated the most LVR, with the top five pairs (including DAI-MKR, DAI-WETH, sUSD-WETH, USDC-WETH, and Okinami-WETH) accounting for a significant portion of the total LVR.
Mawutory's analysis provided additional granularity, showing that:
The weekly average LVR on Ethereum mainnet was highest at about 110 basis points.
The DAI-WETH pool had the highest average daily LVR of nearly 200 basis points.
LVR exhibited a weekly pattern, with lower values over weekends.
L2s
Arbitrum:
marutory's analysis revealed:
Weekly LVR on Arbitrum was highest at about 155 basis points.
The WETH-DAI pool showed the highest LVR at about 115 basis points.
Thursday had the highest LVR at nearly 200 basis points.
Optimism:
Weekly LVR was highest at about 65 basis points.
The USDC-sOPWETH pool showed the highest LVR at about 155 basis points.
Wednesday showed the highest LVR at nearly 55 basis points.
Base:
Weekly LVR was highest at about 380 basis points.
The WETH-DAI pool had the highest average LVR at about 500 basis points.
Saturday showed the highest average LVR at approximately 320 basis points.
Respired.eth's analysis corroborated these findings, showing that Ethereum mainnet accumulated significantly more LVR than L2 solutions. This could be attributed to higher trading volumes and more established liquidity pools on the mainnet.
Total Addressable Market (TAM) Estimation
Mawutory provided a comprehensive TAM analysis across networks:
Arbitrum: Average daily TAM of $11.15 million
Base: Average daily TAM of $10.37 million
Ethereum mainnet: Average daily TAM of $10.31 million
Optimism: Average daily TAM of $0.22 million
The total average daily TAM across all networks was estimated at $32.06 million, highlighting the significant economic impact of LVR in the DeFi ecosystem. Unit Zero Labs proprietary LVR analysis demonstrates slightly lower
Key Insights and Patterns
Asset Pair Influence
All three analysts noted that certain asset pairs consistently showed higher LVR:
Stablecoin-volatile pairs (e.g., DAI-WETH, USDC-WETH) generally exhibited higher LVR than other pairs.
Pairs involving major assets like WETH, DAI, and USDC tended to have higher LVRs, possibly due to higher trading volumes or more significant price movements.
Temporal Patterns
Mawutory's analysis revealed interesting temporal patterns in LVR:
Weekly patterns: LVR tended to be higher on weekdays and lower on weekends across most networks.
Monthly patterns: The third week of the month often showed higher LVR compared to other weeks.
Network Comparison
Respired.eth's analysis showed that Ethereum mainnet accumulated significantly more LVR than Layer 2 solutions. However, when normalized for trading volume and liquidity, Layer 2 networks like Arbitrum and Base showed competitive LVR levels, indicating efficiency improvements in these newer networks.
Volatility and LVR Relationship
All analysts noted a strong correlation between asset volatility and LVR. Prs1205's analysis specifically highlighted how periods of high trading activity, such as the surge in March linked to Bitcoin ETFs and Uniswap V3's launch, led to notable increases in LVR.
Implications and Strategies for Mitigation
Based on the combined analyses, several strategies for mitigating LVR emerge:
Asset Selection: Liquidity providers could focus on less volatile pairs or those with lower historical LVR to minimize potential losses.
Timing Strategies: Given the observed temporal patterns, LPs might adjust their strategies based on the day of the week or time of the month to optimize returns.
Network Selection: While Ethereum mainnet shows higher absolute LVR, Layer 2 solutions like Arbitrum and Base offer competitive environments with potentially lower LVR relative to trading volume.
Dynamic Fee Models: As suggested by Prs1205, implementing higher fees during periods of high volatility could help offset LVR for liquidity providers.
Improved Oracle Systems: Faster and more accurate price oracles could help reduce the lag between AMM prices and market prices, potentially reducing LVR.
This comprehensive analysis of LVR in Uniswap v2 across different networks provides valuable insights for liquidity providers, protocol designers, and DeFi researchers. The significant variations in LVR across networks, time frames, and specific pools underscore the importance of creating sound methodologies for which to build analysis off of.
The total addressable market of LVR varied across all studied networks, and a conclusive figure is still elusive. However, the work done in this challenge by the community helps us to establish a solid foundation for further analysis. As the landscape continues to evolve, these insights can guide the development of more efficient AMM designs and help liquidity providers optimize their strategies to mitigate LVR while maximizing returns.
Future research could explore how these findings translate to newer AMM models, investigate strategies to minimize LVR while maximizing returns for liquidity providers, and examine the long-term implications of LVR on the sustainability and growth of decentralized exchanges.
Resources on LVR:
Columbia University paper that first termed LVR:https://arxiv.org/pdf/2208.06046
CowSwap’s batch trading solution: https://arxiv.org/pdf/2307.02074v3
CowSwap docs: https://cow.fi/learn/what-is-loss-versus-rebalancing-lvr
a16z (must read!) https://a16zcrypto.com/posts/article/lvr-quantifying-the-cost-of-providing-liquidity-to-automated-market-makers/
