Why you should trade on leverage using money markets

Kamel Aouane
Contango
Published in
7 min readFeb 16, 2024

On Contango, the implied funding rate on money markets is cheaper and less volatile than regular perps.

Introduction

Perps are the most traded instruments in crypto. And yet some OGs already grasped the power of DeFi back in 2017 to open leverage positions on MakerDAO through looping strategies (more info here). Since then, several DeFi protocols have allowed users to gain a leveraged exposure, similar to a perp position, by using money markets such as Aave or Compound.

Simtopia, together with the University of Edinburgh, University College London and Contango.xyz, have released the first academic research comparing those two products: the traditional perps you can find on order book exchanges, e.g. Binance or dYdX, and loan positions built on top of Aave v3. The results are staggering: the implied funding rates of loan positions are cheaper (2x to 3x difference), and less volatile (2x to 4x difference), than traditional perps. The intent of this post is to present those academic results in an easy to digest format.

Funding fees

Does it matter?

If you open a leveraged position for a short time frame, funding rates don’t matter much… Right? Think again, my friend. Some trading platforms offer up to 100x, but let’s say you are a “conservative” Degen opening a 20x position when funding rates are at 30% annualized. If price does not move, you will lose your entire initial margin in less than 2 months or 1.64% DAILY. Even a 5x position, with a 30% annualized funding rate, makes you lose 12.5% per month. The table below shows you how much of your initial margin you are losing daily depending on various funding rates and leverage inputs. Yeah, funding rates matter a lot.

Loses on the initial margin depending on funding rate and leverage.
Loses on the initial margin depending on funding rate and leverage.

You may say that an annualized funding rate at or above 30% is not realistic? Well, check this funding rate chart of dYdX during the bull market in 2021. Anything above the red line is above 30% annualized.

BTCUSD hourly funding fees on dYdX during the bull market between March and October 2021. Anything above the red line is above 30% annualized. Extracted from fundingrates.xyz .
BTCUSD hourly funding fees on dYdX during the bull market between March and October 2021. Anything above the red line is above 30% annualized (source: fundingrates.xyz) .

Perps

All you need to know is that perps offer trading on leverage and, by trading those products, traders pay or receive funding fees on top of the PnL due to price movements. The more bullish the market is, the more longs pay shorts. The more bearish the market is, the more shorts pay longs. This behavior is illustrated in the chart below.

Bitcoin price (orange) and funding rates (purple), extracted from Kraken blog
Bitcoin price (orange) and funding rates (purple), extracted from Kraken blog

In the publication, Simptopia has created a financial model for perps which could be simulated following different parameters with their online dashboard. E.g.:

  • The drift μ could be configured to account for market regimes: the more bearish the market is, the more negative is the drift, whereas the more bullish the market is, the more positive is the drift.
  • The volatility σ is used to model the degree of price variability: the higher the volatility the higher the degree of price variability.

The simulation below presents a simulated perp price for μ=0.9 and σ=0.3 where you can observe how the funding fees, or the differences between perp and spot price, are correlated with the market regime (positive funding fee is synonym to market going up and negative funding fee is synonym to market going down). By simulating thousands of those prices, meaningful data could be extracted.

Spot price (in blue) and the difference between the perp price and the spot price for a simulated perp price for μ=0.9 and σ=0.3.
Spot price (in blue) and the difference between the perp price and the spot price for a simulated perp price for μ=0.9 and σ=0.3.

More information on the mathematical models and on the simulations could be found on Simptopia's blog post.

Loan positions

Loan positions are built on top of the 20B$+ of liquidity available on DeFi spot and money markets. An overview of their construction is presented in this blog, from which is extracted the diagram below, where a trader longs 1 ETH with 200 DAI as margin while ETH = 1000 DAI.

Steps to open a long loan position.
Steps to open a long loan position.

The implied funding rate of loan positions is determined by the difference between the cashflow on the lending and borrowing legs of a position. It can be positive or negative and it depends on the leverage chosen by the trader.

The lowest and least volatile funding fees

Historical data

The study takes data from April to October 2023, to compare the 8-hour funding fee of loan positions, built on top of Aave v3, with various leading perp exchanges such as dYdX or Binance. The results show that, on a long position with 5x leverage:

  • Loan position funding fees are the cheapest. They are around 3 times and 2 times cheaper than dYdX on ETHUSD and BTCUSD respectively.
  • Even if Binance has a lower variability than loan positions on ETHUSD, across the 8 instruments analyzed in the publication, loan positions have the least volatile funding fee. Loan position funding fees are 2.5 times and 4 times less volatile than dYdX on ETHUSD and BTCUSD respectively.
8-hour funding rate mean and standard deviation (std) across different platforms from April to October 2023, extracted from the publication.

The plots below show the cumulative funding fees over a 6 month period on both ETHUSD and BTCUSD for loan positions and different platforms. The longer you stay in a position, the cheaper it is to use a loan position.

Cumulative funding fees on loan positions and different platforms on ETHUSD (left) and BTCUSD (right) from April to October 2023, extracted from here.
Cumulative funding fees on loan positions and different platforms on ETHUSD (left) and BTCUSD (right) from April to October 2023, extracted from the publication.

Simulations

Using the dashboard, several simulations were carried out for a period of 6 months (t=0.5) with the same volatility regime (σ=0.3) and different drift parameters: slightly bullish (μ=0.6), bullish (μ=0.9) and highly bullish (μ=1.2). As for the historical data, we find loan positions have the cheapest and least volatile funding rates compared to perps.

Mean and standard deviation (std) of funding rates for loan positions and perps for a period of 6 months (t=0.5) with different drift parameters: slightly bullish (μ=0.6), bullish (μ=0.9) and highly bullish (μ=1.2).
Mean and standard deviation (std) of funding rates for loan positions and perps for a period of 6 months (t=0.5) with different drift parameters: slightly bullish (μ=0.6), bullish (μ=0.9) and highly bullish (μ=1.2).

Below are presented the funding fee distribution with the same parameters. Note how the funding fees are more concentrated around one value, hence showing less volatility.

Funding fee distribution for a fixed σ=0.3 and different drift parameters: slightly bullish (left), bullish (middle) and highly bullish right). Loan position is in red and perps in blue.
Funding fee distribution for a fixed σ=0.3 and different drift parameters: slightly bullish (left), bullish (middle) and highly bullish right). Loan position is in red and perps in blue.

Similar results can be obtained by simulating bearish market conditions, where loan positions have the lowest and least volatile rates for going short.

Analysis

Liquidation times

A direct consequence of entering a long loan position, with the lowest and least volatile funding rates in a bullish market, is a lower probability of getting liquidated. This is observed in the liquidation times of our simulations below where, for the sake of comparison, we have assumed a liquidation LTV of 95% on loan positions and a 5% maintenance margin on perps.

CDF of liquidation time for loan positions and perps with a neutral market (μ=0 on the left) and a slightly bullish market (μ=0.3 on the right).
CDF of liquidation time for loan positions and perps with a neutral market (μ=0 on the left) and a slightly bullish market (μ=0.3 on the right).

Correlations

We have seen that the mechanics of perp and loan position are different. Consequently, funding rates do not require to increase/decrease at the same time. This assumption is observed through the correlations between loan position and different platforms. While funding rates are correlated between platforms that have similar mechanics, implied funding rates from loan positions are little, and sometimes even negatively correlated with perps funding rates. This behavior is observed in the table below.

Correlation of 8-hour funding rate time series from April to October 2023 between various platforms for ETHUSD (upper table) and BTCUSD (lower table). Red denotes positive correlation, blue negative; the stronger the shade, the higher the correlation. Extracted from the publication.
Correlation of 8-hour funding rate time series from April to October 2023 between various platforms for ETHUSD (upper table) and BTCUSD (lower table). Red denotes positive correlation, blue negative; the stronger the shade, the higher the correlation. Extracted from the publication.

Conclusions

As we have seen previously, the more bullish the market is, the higher is the perp funding fee on long positions. In this context, the main force trying to lower the funding rate on perps are traders going short to collect the funding rate paid by the longs. On loan positions, one could expect that the more bullish the market is, the higher are the implied funding rates. After all, the more people go long on ETHDAI, the more DAI are borrowed, hence increasing the utilization rate and borrowing rate of DAI, and the more ETH are lent, hence decreasing the utilization rate and lending rate of ETH. However, other forces and use cases are offered on money markets. E.g. a lower borrowing rate on ETH offers new arbitrage opportunities where traders could borrow cheap ETH to stake them on Liquid Staking Tokens (LSTs), such as stETH where the bearing rate may be higher. These various use cases and arbitrage conditions existing within money markets explain why funding rates do not necessarily increase in the same proportion as perps. However, one could expect the funding rates of perps and loan positions to be in sync with each other due to arbitrage conditions.

Looking ahead

Loan positions have incredible potential and new DeFi innovations are pushing the boundaries of what could be achieved. This study could be extended to more money markets with their unique properties. For example, better cost control could be achieved by using Spark, a money market powered by MakerDAO offering stable borrowing rates with deep liquidity. Higher leverage and lower maintenance margin could be reached by using Morpho which has introduced permissionless markets with custom liquidation LTVs. Visit app.contango.xyz to discover loan positions on 12 money markets and 6 chains.

About Contango

Contango lets you loop anything on-chain. You can create leverage (re)staking positions, arb rates differentials, farm points, or simply go long or short like a perp at low funding. Ape in like a degen with 1-click Strategies, or trade like a pro on the sleek Trade interface.

Website | Twitter | Discord | Docs | Blog

--

--