Summary of Findings - Liquidity Research
I’ve recently investigated the liquidity in different perpetual swap markets, looking for differences between CEX liquidity (Binance & Huobi) and DEX liquidity (dYdX). The ambition for this research is to help inform how Tracer DAO will engage liquidity providers for its initial perpetual swap markets.
I looked into the monthly volume across markets, the depth in USD for a 1% deviation in market price and the depth in USD for a 5% deviation in market price. I have also calculated some key ratios to reveal how the liquidity providers are making decisions about the amount of liquidity they provide markets.
The first ratio looked at was the Depth at a 5% price deviation : Depth at a 1% price deviation. This ratio was calculated to reveal how concentrated liquidity providers aimed to be around the market price. A narrow, highly concentrated approach would indicate a low tolerance to order slippage whereas a broadened, less concentrated approach would indicate a higher tolerance to order slippage. The median ratio across dYdX markets was found to be 1.757, implying a relatively concentrated, narrow approach by liquidity providers with ~55% of total depth within 1% of the market price. Comparatively, the median across centralised exchange markets was found to be 2.373 (~45% of total depth within 1%), indicating a more broadened liquidity providing approach. The graph below illustrates the different tolerances for order slippage for a narrow and broadened liquidity providing approaches.
The other ratios analysed were the Monthly volume : Depth at a 1% price deviation and Monthly volume : Depth at a 5% price deviation . These ratios were calculated to reveal how tolerant liquidity providers are to liquidity risk. Low ratios would indicate that liquidity providers have a low tolerance to liquidity risk as they maintain a high amount of order book liquidity (depth) for the amount of volume in the market. The opposite is true for high ratios. The median ratios across dYdX markets were found to be 73x and 39x respectively. This implies that for every 100 USD of monthly volume traded there is on average 1.37 USD of depth within a price deviation of 1% and 2.56 USD of depth within a price deviation of 5%. Comparing this to centralised exchange markets, dYdX is seen to be very conservative. The median ratios of the centralised exchange markets analysed were 7741x and 2902x respectively.
In terms of the variance, the key finding was that liquidity is too differentiated across markets. To calculate this, the coefficient of variation was taken for all three key ratios. In this context a large coefficient of variation suggests that liquidity providers take a loose approach across markets whereas a small coefficient of variation suggests a more targeted and deliberate approach.
Specifically, for the first ratio, dYdX had a coefficient of variation of 63.15%. This was more than double the coefficient of variation across markets in centralised exchanges, with a value of 28.48%. Essentially, this indicates that a liquidity structure is not being targeted across markets on dYdX and therefore depth structures vary significantly. For example, LINK has a much narrower depth structure (1.01x), compared to SUSHI (5.14x).
With regard to the second and third ratios, the variance across markets was found to be very significant, with coefficients of variation exceeding 80%. This further indicates that liquidity providers lack a targeted liquidity approach across markets.
Centralised Exchange Ratio Statistics
|Centralised||Ratio of 5%:1%||Ratio of Volume:1%||Ratio of Volume:5%|
Statistics are calculated as an average via the depth charts available on Binance and Huobi for a variety of coins as of 6/05/2021
Decentralised Exchange Ratio Statistics
|Decentralised||Ratio of 5%:1%||Ratio of Volume:1%||Ratio of Volume:5%|
Statistics are calculated via the depth charts available on dYdX for a variety of coins as of 6/05/2021
Tracer DAO needs to ensure that there is an appropriate amount of liquidity provided by market makers such that the trading experience is favourable for users. This means minimising liquidity risk across all Tracer markets. To do this, Tracer should utilise the same, conservative and narrow liquidity providing approach that other DEXs have employed to ensure low liquidity risk across markets and to enforce a low tolerance to order slippage. Further, Tracer can improve upon other DEX offerings by implementing a more targeted liquidity approach, aiming for consistent key ratios across markets.
From the data collected, we believe that Tracer DAO should initially incentivise market makers to provide liquidity within the following general bounds:
|Ratio||Lower Bound||Upper Bound|
|Depth at a 5% price deviation : Depth at a 1% price deviation||1.00||2.00|
|Monthly volume : Depth at a 1% price deviation||50x||100x|
|Monthly volume : Depth at a 5% price deviation||25x||100x|
For more supporting analysis please checkout this document: Liquidity Model.xlsx