[Gauntlet] - Jupiter Perpetuals Optimization: Borrowing Rate Reduction and Competitive Analysis vs. OKX and Bybit

Executive Summary

1. Borrowing Rate Recommendations:

  • Uniform decrease from 0.01% to 0.008% per hour for SOL, ETH, and BTC.

2. Competitive Analysis:

  • Jupiter remains attractive for large traders across the leading CEX exchanges - Binance, OKX, & ByBit. For longer holding periods, centralized venues remain more competitive from a financing perspective through funding rate behavior. These conclusions help inform our view pertaining to the direction of borrowing rate parameters on Jupiter.

  • vs OKX:

    • Jupiter has a clear advantage for short-term trades (2-12 hours) across all markets and trade sizes.
    • OKX becomes more cost-effective for longer holding periods (24-120 hours), especially for smaller trades.
  • vs Bybit:

    • Jupiter shows a slight edge in short-term trading, particularly for larger trade sizes.
    • Bybit gains an advantage for trades held beyond 24 hours due to lower funding costs.

Jupiter Borrowing Rates Recommendations

Market Current (bps/hr) Recommended (bps/hr)
SOL .01% .008%
ETH .01% .008%
BTC .01% .008%

Recommendations Summary

Based on the current market dynamics and asset performance, we recommend the following adjustments to Jupiter’s borrowing rates:

  • SOL: Decrease from 0.01% to 0.008% per hour
  • ETH: Decrease from 0.01% to 0.008% per hour
  • BTC: Decrease from 0.01% to 0.008% per hour

Detailed Analysis and Rationale

Methodology Overview

Gauntlet utilizes asset volatility measures, realized utilization, and a target utilization level of 80% to derive recommendations.

Capped Based Implementation

Gauntlet’s methodology for deriving borrowing rate parameters can result in large recommended changes. To ensure parameter changes do not create unintended side-effects, we cap all changes by 20% between parameter recommendations. Gauntlet plans to re-assess conditions weekly and continue processing reductions if the market sees fit.

Asset-Specific Analysis

Solana (SOL):

Unlike its counterparts, SOL has shown increased volatility, with a current blended volatility of 82.46%. The asset’s utilization has reached approximately 45% by the end of the observed period, signaling moderate demand. Given these factors, we recommend decreasing the borrowing rate for SOL from 0.01% to 0.008% per hour, aligning with the other assets while still accounting for its higher volatility.

  • Recommended hourly borrowing rate: 0.008%
  • Recommended borrowing rate APR (cap): 70.08%
  • Recommended borrowing rate APR: 46.35%
  • SOL Volatility: 82.46%
  • SOL Utilization: 44.97%

The updated capped recommendation for Solana will improve the experience for traders, making trading cheaper, given the current utilization levels. It should be noted that an increase in Ut is expected since borrowing costs will be closer to Binance funding rates.

In general, funding rates on Binance for SOL have been significantly lower on Binance vs borrowing rates on Jupiter. It’s important to gradually shrink this gap by reducing borrow fees by 20%, evaluating the impact of trading volume, and creating training data, and if volume increases by more than 20%, we can proceed with another borrow fee cut until we get to the target rates of 46.35%.

Ethereum (ETH):

Ethereum has shown a decrease in utilization and volatility. The current utilization for ETH is around 23%, indicating a decrease in borrowing demand. Volatility for ETH has also decreased, with the current blended volatility at 64.58%. Given the decreased utilization and stabilizing volatility, we recommend reducing the borrowing rate for ETH from 0.01% to 0.008% per hour.

  • Recommended hourly borrowing rate: 0.008%
  • Recommended borrowing rate APR (cap): 70.08%
  • Recommended borrowing rate APR: 18.49%
  • ETH volatility: 64.58%
  • ETH Utilization: 22.90%

The updated capped recommendation for Ethereum will improve traders’ experience and, given the current utilization levels, make trading cheaper.

Ethereum funding rates on Binance are closer to the current Jupiter borrow fees but still need to be reduced.

Bitcoin (BTC):

Bitcoin has shown similar patterns to Ethereum in terms of decreased utilization and volatility. The current utilization for BTC is around 49%, which is higher than ETH but still moderate. BTC’s current blended volatility is 48.58%, lower than both SOL and ETH. We recommend reducing the borrowing rate for BTC from 0.01% to 0.008% per hour, consistent with the other assets.

  • Recommended hourly borrowing rate: 0.008%
  • Recommended borrowing rate APR (cap): 70.08%
  • Recommended borrowing rate APR: 29.68%
  • BTC volatility: 48.58%
  • BTC Utilization: 48.88%

The updated capped recommendation for Bitcoin will improve traders’ experience and, given the current utilization levels, make trading cheaper.

These adjustments aim to enhance the attractiveness of these assets for borrowers while maintaining appropriate risk-adjusted returns for lenders, considering the current market conditions and utilization rates.

Market Dynamics and Competitive Analysis

This analysis combines both the detailed heatmap review and the broader comparison of all-in costs between Jupiter and Bybit across three cryptocurrency markets: SOL (Solana), BTC (Bitcoin), and ETH (Ethereum). The primary focus is on how trade size, holding period, and funding rates influence the overall cost advantage or disadvantage when trading on these platforms. Data is from August 1, 2024.

SOL (Solana) Market:

Heatmap Analysis:

  • Short Holding Periods (2 to 12 Hours):
    • Trade Size of $70,000: Jupiter shows a slight cost advantage over Bybit for short holding periods, with differences ranging from -0.01% to 0%.
    • Trade Size of $420,000: The cost advantage is more noticeable, ranging from -0.08% to 0.01%.
    • Trade Size of $2,500,000: The advantage is most pronounced at the largest trade size, with a difference starting at -1.47% and narrowing to -1.43% over 12 hours.
  • Mid to Long Holding Periods (24 to 120 Hours):
    • Trade Size of $70,000: Bybit becomes more cost-effective after 24 hours, with Jupiter’s costs rising to 0.45% higher by 120 hours.
    • Trade Size of $420,000: The cost disadvantage for Jupiter increases to 0.39% by 120 hours.
    • Trade Size of $2,500,000: While Jupiter maintains a small advantage up to 72 hours, Bybit becomes more cost-effective as the holding period extends, with Jupiter’s costs narrowing to 0.01% by 120 hours.

All-In Trading Fees: Jupiter has lower trading fees compared to Bybit, which is favorable for shorter holding periods.

Long Funding: Despite lower trading fees, the higher funding rate on Jupiter leads to increasing costs over time, which explains the heatmap’s shift towards a disadvantage for longer holding periods.

Overall Insights:

  • Short-Term Trading: Jupiter is advantageous for large trades held for short periods due to its lower trading fees.
  • Long-Term Holding: As the holding period extends, Bybit’s lower funding rates outweigh the initial fee advantage, especially for smaller trades, making it more cost-effective.

BTC (Bitcoin) Market:

Heatmap Analysis:

  • Short Holding Periods (2 to 12 Hours):
    • Trade Size of $70,000: Jupiter offers a small cost advantage initially, but the difference is minor (-0.00% to 0.05%).
    • Trade Size of $420,000: The advantage for Jupiter diminishes quickly, with costs nearly equal by 24 hours.
    • Trade Size of $2,500,000: Jupiter starts with a small advantage, but it dissipates by 72 hours, and by 120 hours, Bybit is more cost-effective.
  • Mid to Long Holding Periods (24 to 120 Hours):
    • Trade Size of $70,000: Bybit takes the lead after 24 hours, with a 0.47% cost advantage by 120 hours.
    • Trade Size of $420,000: Jupiter’s advantage disappears after 24 hours, and Bybit becomes more cost-effective by 120 hours.
    • Trade Size of $2,500,000: Jupiter’s initial advantage fades, and Bybit becomes the more economical choice by 120 hours.

Overall Insights:

  • Short-Term Trading: Jupiter’s lower trading fees give it a slight edge for short-term trades.
  • Long-Term Holding: Bybit becomes the better option for trades held beyond 24 hours due to lower funding costs.

All-In Trading Fees: Similar to SOL, Jupiter’s trading fees are lower than Bybit’s, benefiting shorter trades.

Long Funding: The higher funding rate on Jupiter causes the overall cost to increase over time, especially for longer holding periods, as reflected in the heatmap

ETH (Ethereum) Market:

Heatmap Analysis:

  • Short Holding Periods (2 to 12 Hours):
    • Trade Size of $70,000: Jupiter has a negligible advantage for holding periods up to 12 hours, with differences ranging from 0% to 0.04%.
    • Trade Size of $420,000: The initial advantage is slightly stronger (-0.03% to -0.01%) but diminishes quickly.
    • Trade Size of $2,500,000: Jupiter starts with a more substantial advantage but it fades as the holding period extends.
  • Mid to Long Holding Periods (24 to 120 Hours):
    • Trade Size of $70,000: Bybit overtakes Jupiter after 24 hours, with a 0.23% cost advantage by 120 hours.
    • Trade Size of $420,000: Bybit becomes more cost-effective after 24 hours, with Jupiter showing a 0.20% higher cost by 120 hours.
    • Trade Size of $2,500,000: Jupiter’s advantage diminishes by 72 hours, and Bybit is slightly more cost-effective by 120 hours.

All-In Trading Fees: Lower fees on Jupiter remain an advantage, but the impact is reduced as holding periods lengthen.

Long Funding: The higher funding rate on Jupiter leads to an increase in costs, which diminishes the initial cost advantage seen with shorter holding periods and larger trades.

Overall Insights:

  • Short-Term Trading: Jupiter’s lower fees provide a modest advantage for short-term trades.
  • Long-Term Holding: Bybit’s lower funding rates make it the better option for longer holding periods, particularly for smaller trades.

Combined Key Takeaways:

  1. Short-Term Trading (2 to 12 Hours):
  • Jupiter’s Advantage: Jupiter offers a consistent cost advantage across all markets and trade sizes due to its lower trading fees, especially for large trades.
  • Best for Larger Trades: The benefit is more pronounced for larger trades ($2.5 million), where Jupiter’s lower fees substantially reduce overall costs.
  1. Mid to Long-Term Holding (24 to 120 Hours):
  • Bybit’s Advantage: As holding periods extend, Bybit becomes more cost-effective across all markets due to its lower funding rates. This trend is particularly strong for smaller trade sizes and holding periods beyond 24 hours.
  • Long-Term Strategy: Traders planning to hold positions for more than 24 hours should consider Bybit, especially if trading smaller amounts.
  1. Trade Size Considerations:
  • Larger Trades Favor Jupiter: For very large trades, Jupiter’s fee structure provides a significant advantage, though this diminishes over longer holding periods.
  • Smaller Trades Favor Bybit: Bybit tends to become more competitive as trade sizes decrease, particularly for longer holding periods.

Conclusion:

Jupiter is generally more cost-effective for short-term, high-volume trading. However, for trades held over longer periods, particularly with smaller amounts, Bybit’s lower funding rates make it the more economical choice. To improve its competitiveness, Jupiter should focus on optimizing its borrow rates.

Comprehensive Analysis of Jupiter vs. OKX Trading Costs Across Different Markets

SOL (Solana) Market:

Heatmap Analysis:

  • Trade Size of $70,000:
    • Short Holding Periods (2 to 12 Hours): Jupiter has a slight cost advantage, with differences ranging from -0.01% to 0.02%.
    • Long Holding Periods (24 to 120 Hours): As the holding period increases, Jupiter becomes less competitive, with costs rising to 0.34% higher than OKX at 120 hours.
  • Trade Size of $420,000:
    • Short Holding Periods (2 to 12 Hours): Jupiter maintains a cost advantage, ranging from -0.07% to -0.04%.
    • Long Holding Periods (24 to 120 Hours): OKX becomes more cost-effective, with Jupiter’s disadvantage increasing to 0.29% by 120 hours.
  • Trade Size of $2,500,000:
    • Short Holding Periods (2 to 12 Hours): Jupiter offers a substantial cost advantage, up to -0.50%.
    • Long Holding Periods (24 to 120 Hours): The advantage diminishes over time, with a smaller difference of -0.15% by 120 hours.

Trading Fees and Funding Rates:

  • All-In Trading Fees: Jupiter consistently offers lower trading fees compared to OKX, especially as trade sizes increase. This is a significant advantage for traders executing large transactions.
  • Funding Rates: Jupiter’s higher funding rates, particularly for long positions, gradually erode its initial cost advantage over extended holding periods. OKX’s lower funding rates become more advantageous as the holding period extends beyond 24 hours.

BTC (Bitcoin) Market:

Heatmap Analysis:

  • Trade Size of $70,000:
    • Short Holding Periods (2 to 12 Hours): Jupiter shows a minor cost advantage, with differences from 0.01% to 0.03%.
    • Long Holding Periods (24 to 120 Hours): OKX overtakes Jupiter in cost-effectiveness, with Jupiter’s disadvantage reaching 0.48% by 120 hours.
  • Trade Size of $420,000:
    • Short Holding Periods (2 to 12 Hours): Jupiter maintains a slight advantage, with cost differences from -0.04% to 0.08%.
    • Long Holding Periods (24 to 120 Hours): OKX becomes more cost-effective, with Jupiter’s disadvantage reaching 0.46% by 120 hours.
  • Trade Size of $2,500,000:
    • Short Holding Periods (2 to 12 Hours): Jupiter offers a clear cost advantage, ranging from -0.10% to -0.06%.
    • Long Holding Periods (24 to 120 Hours): OKX becomes more competitive over time, turning Jupiter’s initial advantage into a 0.37% higher cost by 120 hours.

Trading Fees and Funding Rates:

  • All-In Trading Fees: The fee structure is similar to the SOL market, with Jupiter providing a clear advantage in lower trading fees across all trade sizes.
  • Funding Rates: The funding rate chart shows that Jupiter’s higher rates for long positions become a significant cost factor over time, diminishing its initial cost advantage for trades held beyond 24 hours.

ETH (Ethereum) Market:

Heatmap Analysis:

  • Trade Size of $70,000:
    • Short Holding Periods (2 to 12 Hours): Jupiter holds a minor advantage, with differences from 0% to 0.02%.
    • Long Holding Periods (24 to 120 Hours): OKX becomes more cost-effective, with Jupiter showing a cost disadvantage of 0.21% by 120 hours.
  • Trade Size of $420,000:
    • Short Holding Periods (2 to 12 Hours): Jupiter maintains a small advantage, with differences from -0.03% to -0.02%.
    • Long Holding Periods (24 to 120 Hours): OKX overtakes Jupiter, with a cost difference of 0.18% by 120 hours.
  • Trade Size of $2,500,000:
    • Short Holding Periods (2 to 12 Hours): Jupiter offers a more substantial advantage, ranging from -0.15% to -0.13%.
    • Long Holding Periods (24 to 120 Hours): OKX becomes more competitive as the holding period extends, with Jupiter’s cost difference reducing to 0.06% by 120 hours.

Trading Fees and Funding Rates:

  • All-In Trading Fees: Jupiter continues to offer lower trading fees, which is advantageous for larger trades.
  • Funding Rates: As with SOL and BTC, Jupiter’s higher funding rates, particularly for longer holding periods, lead to a higher overall cost when compared to OKX, diminishing the initial advantage gained from lower trading fees.

Combined Key Takeaways and Conclusion:

Jupiter offers a clear cost advantage for short-term trading (2 to 12 hours) across all markets and trade sizes due to its lower trading fees, especially for larger trades ($2.5 million). However, as holding periods extend (24 to 120 hours), OKX becomes more cost-effective across all markets due to more favorable funding rates, particularly for smaller trades and positions held for more than 24 hours. Jupiter’s higher funding rates for long positions erode its initial cost advantage over time, making OKX the more economical choice for longer-term trades.

While larger trades favor Jupiter in the short term, this benefit diminishes over time due to higher funding rates. Conversely, OKX becomes more competitive as trade sizes decrease and holding periods increase. To improve its market position, Jupiter should consider adjusting its funding rates and exploring tiered cost structures to better cater to traders with different holding periods and trade sizes, which would help maintain its initial cost advantage across a broader range of trading scenarios.

JLP perspective

The markets for SOL, BTC, and ETH are currently experiencing inefficient liquidity utilization, with trading volumes concentrated in various utilization (Ut) buckets, often below optimal levels. Let’s examine each asset:

SOL:

  • Highest trading volume in the 20-30% Ut bucket (633.057070 million USD)
  • Significant activity in higher buckets, especially 80-100% (533.428873 million USD)
  • Wide spread of activity across various utilization levels
  • Most frequent trading (12 days) in the 45-50% range

BTC:

  • Peak trading volume in the 40-45% Ut bucket (98.679015 million USD)
  • Consistent activity in lower buckets, especially 55-60% and 30-35%
  • Most frequent trading (15 days) in the 45-50% range
  • Limited activity above 70% utilization

ETH:

  • Highest trading volume occurs in the 50-55% Ut bucket (102.739524 million USD)
  • Significant activity in the 60-65% range (62.291673 million USD)
  • Most frequent trading (25 days) in the 20-30% range
  • Wide spread of activity across various utilization levels

This distribution across different Ut buckets, particularly the concentration in lower and mid-range utilization for BTC and ETH, is below Gauntlet’s target of 80% utilization. SOL shows more activity in higher buckets, but still has significant volume in lower ranges.

Supporting Data

7D Average Utilization of JLP assets:

Current Utilization of JLP assets:

Solana Utilization:

Volatility of SOL Returns:

Volatility of SOL Utilization:

Strong positive correlation (0.81) between SOL trading volume and 24-hour rolling volatility:

ETH Utilization:

Volatility of ETH Returns:

ETH Utilization Volatility:

BTC Utilization:

Volatility of BTC Returns:

BTC Utilization Volatility:

USDT Utilization:

USDC Utilization:

Conclusion

Jupiter demonstrates a competitive advantage in short-term trading, especially for larger trades held 2-12 hours. The recommended decrease in borrowing rates from 0.01% to 0.008% per hour for SOL, ETH, and BTC aims to enhance this edge. While centralized exchanges become more cost-effective for longer holding periods, Jupiter’s strengths in short-term, high-volume trading position it well in the market. To maintain and improve its standing, Jupiter should implement the rate change, focus on its short-term trading advantages, and explore ways to increase competitiveness for longer-term positions.

10 Likes

Thank you for the thorough analysis!

My observations maybe lead to conclusion that i would not compare JUP vs any other CEX despite the fee adjustments could potentially help drive users from CEX to DEX since currently DEX-es have less barriers of entry (no KYC) and offer on-chain permissionless trading, while CEX always exercise a bigger liquidity layer and more pairs. As you mention JUP already has the edge on the short end, the question would be if JUP becomes the cheapest on the whole spectrum if KPI’s will show there was an impact on the user base. You can always test that for a certain period and compare the results and see if you had a decent amount of new users that would cover for that reduced fee. :eyes:

5 Likes

Great point, @macximus!

Our borrow fee recommendation methodology actually uses asset volatility measures, realized utilization, and a target utilization level of 80% to derive recommendations. It’s not limited to comparison with CEX exchanges only.

Our methodology for deriving borrowing rate parameters can result in large recommended changes. To ensure parameter changes do not create unintended side-effects, we cap all changes at 20% between parameter recommendations. We plan to reassess conditions weekly and continue processing reductions if the market sees fit. This way, we can gather useful training data and observe if the decrease has helped growth, allowing us to push it further if needed.

We agree with the notion of different users, but our broader use case is making Jupiter as liquid and competitive as possible compared to the large CEX players. This approach aims to encourage the world to move on-chain long-term, which is how we grow as a defi ecosystem.

6 Likes

It looks like we are having quite weak volumes last weeks. How about increasing the % that JLP holders get? In the current state of affairs, Jupiter keeps accumulating way more money that it can possibly spend for running the protocol, and JLP holders are not properly being rewarded for the risk they are taking.

I imagine Gauntlet can also play around with the ratio of fees that JLP vs DAO get? It would make sense that JLP holders make more % when not a lot of volume, and less % when a lot of volume.

Or perhaps can use some of that 25% fees that the DAO is getting to incentivize traders, for example, keep 10% in treasury which is a lot, and use 15% to incentivize growth running promotions like leaderboards etc

2 Likes

Thank you @fanatik0. You’re thinking about a lot of the same things we are.

When looking at volumes, we look at notional amounts as well as marketshare within the Solana ecosystem, to abstract away the impact of market dynamics. You are correct in saying that 30d average volumes have decreased in the last week (volume peaked August 16th); this trend was not unique to the Jupiter protocol but was observed on other perps DEXes as well. However, Jupiter actually gained marketshare during this time period. To put this another way, Jupiter captured a larger share of the trading demand in the last week, despite softer overall demand.

With respect to the revenue share, this parameter was recently revised down from 30% going to the DAO to 25% back in June. We continue to gather data on the impact this change has had.

Your question around incentives is valid and we will circle back with the team on considerations there. We’ll provide more thoughts next week.

3 Likes