By: Brandyn Hamilton
Attribution
Executive Summary
- Similar to the findings from the previous challenge examining the impact of gas subsidies on Ethereum Layer 1 (L1) activity, this research suggests that gas subsidies might not significantly influence user activity on Layer 2 (L2) networks. While L2 networks generally cater to more retail-trader activity, the already low fees on these platforms lead to more gas-agnostic behavior compared to L1. The weak statistical correlations observed between gas fees and target variables such as protocol fees and volume indicate that other factors, such as market sentiment, overall trading trends, and broader economic conditions, may play a more significant role. Further research with more extensive data and a longer timeframe may uncover additional insights.
Introduction
Scope & Purpose
- The purpose of this research is to explore the potential impact of introducing gas subsidies into several aspects of the Uniswap ecosystem, specifically L2 networks. The specific networks chosen for analysis were Arbitrum, Base, and Optimism.
- This study concentrates on the period from July 14, 2024, to July 21, 2024, utilizing hourly data. This timeframe was selected to comply with the Dune API's limitations on the maximum number of retrievable data points for my account. Although the period is relatively brief, the high data granularity allowed us to gather over 100,000 data points combined.
- For each L2, ten token pairs representing a sampling of volatile-volatile, volatile-stablecoin, and stablecoin-stablecoin token pairs were chosen.
Methodology
Dune Queries
Pool Data
- Data for Uniswap V3 data focused on Arbitrum, Base, and Optimism L2s, and was collected using custom queries based on @coindataschool's dashboard (n.d.) (see references).
- Specifically, the data was modified for time series analysis by aggregating it to hourly intervals for the period from July 14, 2024, through July 21, 2024. We filtered the data to include 10 token pairs per network, ranked by volume, representing volatile-volatile, volatile-stablecoin, and Stablecoin-Stablecoin pairs. For each token pair, the dataset encompasses fee tier, number of trades, volume (USD), fees generated, total value locked (TVL), volume-to-TVL ratio, and average liquidity per hour. This process yielded 38,148 data points for Arbitrum, 39,017 data points for Optimism, and 36,454 data points for Base.
Gas Fee Data
- Continuing with our selection of L2 networks, we gathered gas fee data from custom queries based on those created by @bnbchain (n.d.) (see references). While the original queries provided daily averages and medians for gas fees in USD across L2 networks, our custom queries were designed to extract this data on an hourly basis, aligning with the pool data. The data spans from April 22, 2024, to July 21, 2024.
Flipside Queries
Trader Classifier Model
- Examining trader activity on Uniswap V3 on Arbitrum, we analyzed transaction count, volume, order size, and the number of unique token contracts for both professional and retail traders. A professional trader is defined as an address with a total trading volume exceeding 1,000,000 USD, an average order size greater than 10,000 USD, or involvement with more than 20 unique token contracts. This analysis covers the period from July 1, 2023, to July 21, 2024.