Written by: X-explore, Wu Talks about Blockchain

 

Since 2023, the trading volume of cryptocurrency exchanges has increased significantly. In January this year, the derivatives trading volume of mainstream exchanges increased by 47.6% compared with the previous month. And this soaring trading volume is likely mixed with fake transactions, as the article said.

Wash trading on cryptocurrency exchanges is not a new topic. Due to the lack of regulation on centralized exchanges, some exchanges tend to fake volume to conceal the real trading volume and thus improve their trading volume ranking. Therefore, we constructed a volume analysis method to identify and explain fake trading volume.

We have obtained historical K-line data, real-time depth data, and real-time transaction data from several major exchanges through open APIs. These exchanges include Binance, Bybit, OKX, Bitget, Phemex, Kucoin, and dYdX.

 

1. Overview of trading volume of each exchange

 

From the figure below, we can see that the trading volume of cryptocurrency exchanges has increased significantly since 2023; however, the trading volume pattern of the Phemex exchange is abnormal.

The chart below shows the price and volume chart of the uBTCUSD trading pair on Phemex. It can be clearly seen that starting from August 19, 2022, the volume increased rapidly, and from September 16, 2022 to November 19, 2022, this crazy volume was repeated. After that, without any signs, the volume dropped sharply and dropped to the level before August 19, 2022. Starting from January 2023, the volume seems to have rebounded again and continues to remain at a higher level.

 

2. Research on the relationship between quantity and price

 

Normally, price fluctuations in the market will lead to a surge in trading volume as investors close their trades or open new ones. In our experience, periodic large trading volumes often appear after a large price drop/rise, especially in the futures/derivatives market.

Considering the leveraged trading characteristics of cryptocurrencies, any violent movement in the market may be amplified and lead to stop-loss and take-profit. This will be an important reason for the surge in trading volume. Therefore, we will take the changes in Phemex trading volume when the price fluctuates sharply as one of the targets of our investigation.

Our analysis of the (Pearson) correlation coefficients of major exchanges shows that dydx’s correlation coefficient has a cyclical relationship with trading mining revenue activity. Binance, Bitget, OKX, and Bybit have a strong linear relationship, while Kucoin has a weaker linear relationship, and Phemex is at the bottom.

Therefore, we can infer that Phemex has engaged in false volume manipulation, which occurs when market prices are stable, and abnormally enlarged trading volumes often occur during this period.

 

3. Trading order flow

 

We also analyzed recent transactions on several exchanges.

Large market fluctuations often lead to large trading volumes. This is generally due to:

  • A sudden increase and influx in the number of transactions

  • Large transactions

On the Phemex exchange, we did not find obvious correlations between the raw order flow and the K-line volume, and the size of the order did not show randomness. Compared with Binance, the size of the trades was relatively stagnant. What changed was the density of the trades, not the size of individual trades.

​

4. Order Book Depth vs Trading Volume

 

Depth is considered an important metric for measuring exchange liquidity. Traders prefer markets with better depth because transaction costs are lower. We investigate order book depth across exchanges to determine if there are any observable patterns or trends.

Overall, Binance and dYdX seem to have very good depth and prices. KuCoin and OKX come in close behind, while Phemex has the largest slippage.

 

5. Estimating fake trading volume

 

We construct a model to estimate the true proportion of reported trading volume.

Observing that the correlation between price fluctuations and trading volume on Phemex is weak, we hypothesize that wash trading on the exchange occurs primarily during periods of small price fluctuations. In other words, trading volume during periods of large price fluctuations is real. This is a conservative assumption for estimating the proportion of fake trading volume.

As can be seen from the above figure, there are some obvious discrepancies. The estimated trading volume is completely lower than the reported trading volume. It is estimated that the real trading volume may account for between 5% and 25% of the reported trading volume.

Using the same method to estimate the real volume ratio of Binance and Bybit, the results are very close to the reported volumes.

​

Summarize

 

We used various public data from mainstream centralized exchanges:

  • The trading volume on Phemex shows a different pattern than other exchanges. The sudden increase in trading volume on the Phemex platform seems interesting and it is not normal.

  • Phemex’s volume is seen to have a significantly lower correlation with price candlesticks than any other exchange. This suggests that Phemex’s volume is not entirely dependent on market movements and there may be other forces driving their volume.

  • Phemex appears to have the highest spreads and slippage among most exchanges. This should deter some customers from trading.

  • A ratio analysis of some of the spikes in volume estimates the estimated volume to be around 5% to 25% of Phemex reported volume.

  • By digging deeper into individual trade data, Phemex’s trade size does not change much even when the market is volatile, compared to Binance, where large trades occur when market volatility leads to huge K-line volumes.