The Best Methods For Measuring Cryptocurrency Market Liquidity
- By Sam van de Schootbrugge
- Sign-up to discuss on Slack
- 5 min read
Summary
- A new Journal of Banking and Finance paper explores how best to measure liquidity in cryptocurrency markets using only low-frequency data.
- Low-frequency data includes data on open, high, low and closing prices as well as the number of transactions and the dollar trading volume for each subinterval.
- The Amihud illiquidity ratio is good for investors seeking the most liquid crypto exchanges. Meanwhile, measures that use high, low and closing prices are better at estimating liquidity across exchanges.
Introduction and Motivation
The lack of a regulated data feed and the abundance of crypto exchanges make it hard for investors to access the high-frequency bid-ask spreads used for calculating liquidity. As a result, most traders rely on lower-frequency transaction-based data, such as daily high, low and closing prices to evaluate market quality.
A recently published paper from the Journal of Banking and Finance explores which transaction-based liquidity measures investors should use when (i) calculating the level of liquidity, and (ii) comparing liquidity across exchanges. Comparing these measures against a set of benchmark liquidity estimates based on order book data, they find the following:
- Measures that use high, low and closing prices (B and D in the subsequent section) best capture variations in cryptocurrency liquidity.
- The relative performance of liquidity measures does not depend on the liquidity regime, but the two best measures outperform most during periods of high volatility and high volume.
- Investors should use different measures for estimating the level of liquidity vs comparing liquidity across exchanges.
Methodology – Measuring Cryptocurrency Liquidity
The authors compare the performances of some commonly used transaction-based measures of liquidity. These are calculated from data on open, high, low and closing prices as well as the number of transactions and the dollar trading volume for each subinterval. These are the key ones:
- Amihud (2002): The illiquidity ratio calculates the absolute intraday return divided by the dollar trading volume for each subinterval (e.g., a day). This is then averaged across a given interval.
- Corwin and Schultz (2012): This estimator is calculated from the high and low prices of two adjacent subintervals.
- Kyle and Obizehaeva (2016): They derive an illiquidity index based on the ratio of volatility to dollar volume of an asset within a given interval.
- Abdi and Ranaldo (2017): This estimator is based on the natural logarithms of high, low and closing prices in subinterval.
To assess which transaction-based cryptocurrency liquidity measures work best, the authors create four benchmark measures of liquidity derived from high-frequency order book data:
- The percentage quoted spread: This is the percentage difference between the best ask price and the best bid price of each order book snapshot, divided by the quote midpoint and then averaged over all observations in the interval.
- The percentage effective spread: They combine order book snapshots with the first transaction that occurs after the snapshot. The spread is the percentage difference between this transaction price and midpoint, divided by the midpoint and then averaged over all observations in the interval.
- The percentage price impact: They use data sequences consisting of an order book snapshot, the first transaction after the snapshot and the subsequent order book snapshot. The percentage price impact is then calculated as the percentage change in the quote midpoint from the pre-transaction order book snapshot to the post-transaction snapshot, averaged over all data sequences in the interval.
- The cost of a roundtrip trade: They calculate the weighted average prices for a market buy order and a market sell order of a given size, and then express the difference between the two prices as a fraction of the quote midpoint. They then calculate an average across all order book snapshots in the interval. The round-trip measure provides estimates of the execution costs for large trades.
The authors score their performance across three dimensions:
- Their ability to capture the time-series variation in liquidity.
- Their ability to capture the level of liquidity.
- Their ability to capture differences in liquidity across exchanges.
Data
Underpinning the liquidity measures is data on bitcoin and ethereum from three exchanges – Bitfinex, Bitstamp and Coinbase Pro – compiled over a two-year period (2017-2019). All three operate an electronic central limit order book, from which the authors collect information on transactions and order book snapshots comprising the 50 best bids and asks (Chart 1).
The transaction data used for the low-frequency measures includes the price and the corresponding dollar trading volume for each transaction, a UNIX time stamp, a unique exchange-specific ID and a trade indicator which indicates whether a transaction was buyer-initiated or seller-initiated. For the main results, data is collected at one-hour frequencies to get an estimate of daily liquidity.
Source: Kaiko
Results – Exchanges
The descriptive statistics show trading activity is markedly higher for BTC/USD than for ETH/USD on all three exchanges, especially in terms of USD trading volume. Coinbase Pro is the most active exchange, but Bitfinex has the highest average daily USD volume.
On liquidity, the percentage trading costs in the cryptocurrency markets are very low. Average quoted and effective spreads on Bitfinex and Coinbase Pro are below 1bp. The price impact on these two exchanges amounts to approximately 40% of the effective spread, implying that the suppliers of liquidity earn a very small positive realised spread on average. This is not the case on Bitstamp, where liquidity suppliers appear to earn significant realised spreads.
Overall, the authors show that all three exchanges reveal strong similarities in the liquidity measures, both within and between the three exchanges. Bitstamp is, however, substantially less liquid for ETH/USD than the other two exchanges.
Results – Time-Series Variation in Liquidity
An accurate transactions-based measure should capture the time-series variation in liquidity and should therefore be positively correlated with the benchmark measures. Comparing the accuracy of the liquidity measures at a daily frequency, the Abdi and Ranaldo (2017) and Corwin and Schultz (2012) estimators perform best, especially against the effective spread (Chart 2). The effective spread is also highly correlated to the number of transactions and the dollar trading volume. This implies that higher trading activity is associated with higher execution costs.
Source: Paper, page 24
The poor performance of the widely used Amihud (2002) measure is justified as follows: ‘The illiquidity ratio is based on the presumption that, in a less liquid market, a given dollar trading volume will have a larger impact on prices and will thus result in a larger price change’. Given that dollar trading volume is positively correlated with higher execution costs in crypto markets, the Amihud (2002) measure has broken down.
In addition to this main result, two others stand out. First, a hybrid measure consisting of all four transaction-based liquidity estimators does no better than Abdi and Ranaldo (2017) and Corwin and Schultz (2012). Second, the two best measures outperform most during periods of high volatility and high volume, and their performance does not change during periods of high and low returns.
Results – Estimating the Level of Liquidity
The level of liquidity is important for trading strategies and portfolio allocations. The authors determine which of the measures can most accurately assess the level of cryptocurrency liquidity by estimating the prediction errors between the benchmarks and the transaction-based measures.
By this metric, the Amihud (2002) and Kyle and Obizhaeva (2016) measures are best at capturing the level of the effective spread. They have the lowest RMSE and MAE for all data frequencies, i.e., hourly, daily, and 15-daily. The Abdi and Ranaldo (2017) and Corwin and Schultz (2012) estimators perform poorly but improve at lower frequencies.
Results – Estimating Liquidity Across Exchanges
The paper only uses data from three crypto exchanges. Nevertheless, the authors determine which of the transaction-based measures can most accurately rank these trading venues in terms of liquidity. On this front, two general patterns emerge:
- All four measures replicate the ranking of three out of the four benchmark measures (all except the price impact). For the price impact, the percentage of matching rankings is close to 50% and therefore provides little information on the ranking of price impacts across different trading venues.
- The higher the data frequency, the better the liquidity measures are at capturing the rank of different exchanges.
Overall, the Corwin and Schultz (2012) estimator is the most accurate measure of cross-exchange liquidity.
The Bottom Line
Investors seeking the most liquid exchanges should use the Amihud (2002) illiquidity ratio or the Kyle and Obizhaeva (2016) estimator. These measures are also good at estimating the level of execution costs (the cost of a roundtrip trade).
The Corwin and Schultz (2012) or Abdi and Ranaldo (2017) liquidity measures best serve investors hoping to time the liquidity of cryptocurrency markets and enter or exit when markets are liquid. These estimators best capture time-series variation in liquidity. The former is also good at estimating liquidity across exchanges.
Citation
Brauneis A., Mestel R., Riordan R., Theissen E., (2021), How to Measure the Liquidity of Cryptocurrency Markets?, Journal of Banking & Finance 106041, https://www.sciencedirect.com/science/article/pii/S0378426620303022
Sam van de Schootbrugge is a macro research economist taking a one year industrial break from his Ph.D. in Economics. He has 2 years of experience working in government and has an MPhil degree in Economic Research from the University of Cambridge. His research expertise are in international finance, macroeconomics and fiscal policy.
(The commentary contained in the above article does not constitute an offer or a solicitation, or a recommendation to implement or liquidate an investment or to carry out any other transaction. It should not be used as a basis for any investment decision or other decision. Any investment decision should be based on appropriate professional advice specific to your needs.)