Crypto exchanges: It’s all scam … or is it? Part I: Fake volumes (I/II)
We have conducted the first publicly available scientific study quantifying the manipulation of volumes by cryptocurrency exchanges.
Full study is available here. For main findings see below.
Blockchain technology is by design meant to bring transparency and self-regulation. Technological limitations have favored the growth of centralized (offchain) exchanges over decentralized (blockchain) exchanges. Most centralized cryptocurrency exchanges are neither regulated neither transparent in the sense that users’ sole source of information is data provided by the exchange itself.
The claim has been made that a significant part of the volume allegedly traded on these exchanges is actually fake.
Increasing volume numbers, increases the exchange’s ranking and popularity among both traders and post-ICO firms looking to list their token. Trading volumes are a major indicator for liquidity, slippage assumptions, risk management models and high volumes is a strong positive factor for a token.
Therefore, reliable volume numbers is essential for making informed decision regarding where to list tokens and which exchange to trade on.
In a scientific study (available here), we have systematically reviewed all the trading activities of more than 50 crypt-exchanges to verify the claim and quantity th extent of the manipulation.
In this first post, we will look at global numbers (which is already pretty worrisome). In part II, we will look more deeply on a coin by coin basis (and it gets scary).
Manipulation has to happen somewhere, we have analyzed the whole reporting process to find out what ‘s going on between the trading tape and the final consumer
Dedicated websites (as coinMarketCap) use the daily volumes provided by the exchanges’ API to report trading volumes. Here is the process from the trading tape (were all trades are being written) and final users.
In order to report fake volumes, exchanges have several options:
1/ Convince reporting websites to report inflated numbers
2/ Report daily volumes that do not reflect the tape (if you think that is it too big: see below)
3/ Print fake trades on the tape (without sending orders on the book)
4/ Add orders on the book and trade with themselves
In our study, we focused on 1/,2/ and 3/ leaving 4/ for a separate in-depth analysis.
Let’s get started
First we have checked numbers reported by dedicated websites against the daily volumes provided by coinMarketCap (the first on Google).
We have computed the ratio of the volume reported by the website by the daily volume reported by the API. Below we show the median ratio per exchange.
As we can see on the graph above, coinMarkCap does report numbers as they are reported by the exchanges’ API (in line with their FAQ). (Exchange 8 being out of range, can be explain by a number not being updated on their side, nothing statistically relevant).
Unfortunately, that’s the only good news of the study.
Too obvious to happen: you bet…
After filtering exchanges which do not report daily volumes or are clearly buggy, we have stored all the trades printed on the trading tapes and compared the numbers of coins printed on the tapes to the daily volumes provided by the exchanges.
Results are shown below (median ratio of numbers of coins printed on the tape by total volume reported by the API per exchange):
We found that 20% of exchanges systematically report between 13% and 100% more trades than what actually happened on the tape.
If we look closer at the results and check on each pair the ratio of reported daily volume Vs. all trades on the tape, we find that most exchanges actually inflate daily volumes on some specific coins. Even though, these specific manipulations do not impact significantly the total volume reported by the exchange (globally), it does matter a lot for people looking at the concerned tokens.
was actually printed on the tape, per exchange
Our study showed that another 40% of exchanges play with numbers (i.e report at least 20% more trades than what it shows on the tape) on a significant number of coins.
Results showed that 60% of the exchanges considered tamper with the daily volumes they report through their API
To wash trade or not to wash trade
If you want to add some fake volume on the tape, you basically have two options. Either you print fake trades directly on the tape or you add orders that you fill against yourself.
The downside with adding orders in the book is that you might get filled against other traders willing to execute (you can always mess up with the matching engine making your orders not fillable, but that adds complexity to your matching engine and traders would notice that they do not get the prices they should).
The downside with adding orders in the book is that you might get filled against other traders willing to execute
If exchanges decide to print fake trades directly on the tape, then we should be able to detect trades that do not update the book as expected. In this study, we have thoroughly checked where trades happen in the order book and have marked obviously fake trades.
In the following graph, we show the median percentage of “fake trades” per exchange (each bar is an exchange).
of the spread
Our results showed that up to 95% of the trading tape of some exchanges is actually completely fabricated.
If we look at individual pairs to find if exchanges temper with the trading tape of specific coins, we find that 28% of the exchanges are involved with wash trading on a significant number of pairs.
Conclusion for today
On this first post, we showed that a large majority of exchanges are inflating their volume numbers.
In our next post, we will show that some coins are more likely to be manipulated than others and that knowing this information is absolutely essential before listing a new token or buying a significant amount of an altcoin.
At an industry level, this study shows the need for increased transparency
in the crypto-trading space. External regulation is far from being the preferred option for industry players and auto regulation seem tough to achieve in the short term. With decentralized exchanges still lagging way behind centralized ones in terms of liquidity, we expect the emergence of some kind of hybrid exchanges using the blockchain technology to add traceability and transparency to their trading activity or some independent label to certify exchanges’ activities.
At an industry level, this study shows the need for increased transparency
in the crypto-trading space.
Follow-us to get updated (next studies to come: Fake Volumes Part II, The exchanges that will have your token crash for sure, Pump & Dump: how to protect your token?, and more…).
For any feedback, comment or inquiry:
