Sure 95% of the bitcoin market is fake, but the 5% that isn’t is really, really efficient.
That’s the thesis of a new report Bitwise put out last week showing the anomalies on fake exchanges and putting forth new evidence to suggest regulated bitcoin markets are more efficient and mature than popularly believed. The firm recently shook the crypto world when it shared a report with the U.S. Securities and Exchange Commission (SEC) showing how the overwhelming majority of volumes in crypto are fake or non-economic.
Spotting fake exchanges 101
Bitwise’s study of fake exchanges starts with a set of comparisons between reported bitcoin volume on CoinMarketCap and the trading volumes of other assets. Gold, for instance, has a trading volume about 3 times higher than Bitcoin’s reported volume but with a market cap 80 times larger than bitcoin’s. This means that while CoinMarketCap statistics suggest that 12% of all bitcoin gets traded every day, just 0.43% of gold’s total supply changes hands daily.
As Bitwise points out in the report, “it makes little sense to argue that bitcoin is being used as a long-term store of value when its reported turnover is nearly 100X that of gold.” Volumes that appear astonishing but are in fact fabricated will only cost investor confidence in the bitcoin market.
To discern fake volumes from real, Bitwise collected data from 83 exchanges over a week from April 28 to May 5, and charted out their respective trade size, trading time, and spreads patterns.
In the trade size histograms, Bitwise shows most trades on regulated exchanges such as Coinbase are under 3 BTC.
However, trade sizes on fake exchanges completely diverge from this pattern. It does not take a very sophisticated understanding of statistics to notice their trade size distribution appears unnatural. Some only have trade sizes between certain ranges, while others have evenly distributed trade sizes that lack the behavioral spikes as observed on the regulated exchanges.
As Bitwise notes in the report, “there is no economic reason that the trade size distribution of these exchanges should defy the patterns and trends that are present in traditional markets,” concluding that exchanges such as Coinbene and OEX are providing artificial volume through fraudulent printing or incentivizing unfair trading.
The consistency (or lack thereof) seen in graphs showing hourly trading volumes can be another dead giveaway. On a “real” exchange, hourly trading volumes do not form any discernible pattern day to day. However, when compared across exchanges, they do display similar recurring trends and align closely during a volume spike on May 3 between 4-7 a.m. ET.
In comparison, fake exchanges show repeated trading volume patterns throughout the week regardless of market movement. The trends seen typically bear no resemblance to those displayed by more regulated exchanges Bitwise uses for reference. Notably, none of these exchanges witnessed the same May 3 volume spike as the reference exchanges.
Lastly, Bitwise charted out the bitcoin spread changes throughout a week for each exchange. Spreads are the price difference between the highest price someone is willing to buy bitcoin and the lowest price someone is willing to sell bitcoin. The report notes that spreads on real exchanges tend to be very low, typically under $5. On Coinbase pro, the median spread during the week the firm investigated was $0.01. In general, Bitwise found that the average median spread on referenced exchanges was $1.31 in April.
As Bitwise noted, “this reflects a healthy market, where multiple market makers are competing to drive the spread as low as possible, only widening their quotes in response to heightened market risk.”
In comparison, the spreads on fake exchanges are eye-opening. Some of them maintain a spread as high as $110 for the entire week, while others even struck a $700 spread. Again Bitwise concluded here that these exchanges must be manipulating their volumes, since “there is no economic reason that bid/ask spreads should be anchored on high dollar amounts or oscillate in artificial patterns if there is true liquidity on the exchange where traders repeatedly cross the bid/ask spread.”
What about Huobi, OKEx, and HitBTC?
The white paper singles out Huobi, OKEx, and HitBTC as three of the most notable firms that do not make the cut as real exchanges, despite a reported combined average daily volume of over $480 million. In taking a closer look, Bitwise concluded that all three faked nearly all of their volume.
How and why to fake volumes
Bitwise lists five ways exchanges fake bitcoin volumes and explains their possible motivations.
Fake volumes can be produced through:
- Posting non-existant trades on their exchanges
- Wash trading: simultaneously buying and selling a single asset to generate artificial volume, or paying market makers to do it for them
- Bribing traders to trade by paying them or offering lower fees
The major benefits of faking volumes are:
- Win attention on CoinMarketCap and drive traffic to the exchanges
- Attract projects to conduct initial coin offerings (ICO) on their exchanges and earn their listing fees
Bitwise found just 10 out of the 83 exchanges it investigated to meet its criteria for a real exchange, accounting for a cumulative $554 million bitcoin spot volume in April.
A regulated and efficient market
So far, the white paper hasn’t radically departed from the company’s first report on fake volume. However, the second half of the white paper goes a step further to explore the status of the real bitcoin market by taking a closer look at the 10 real exchanges it identifies.
Out of these 10 exchanges, nine hold U.S. Department of Treasury Money Services Business licenses, and six have acquired BitLicenses from the New York State Department of Financial Services.
This does not mean that there is no outlier. Kraken, ranked fourth in daily trading volume on Bitwise’s 10 real exchange list, did not pursue a BitLicense. Meanwhile, Binance, the largest real exchange in volume per Bitwise’s report, is the only exchange among them not registered with the U.S. Department of Treasury.
However, despite their varying regulatory statuses, the 10 exchanges demonstrate rather unified bitcoin spot price. The average bitcoin price deviation for any of the exchanges ranged from 0.06% to 0.20%, with an average of 0.12% across exchanges. The deviation rate has also dropped consistently since the beginning of 2018.
When compared to the bitcoin price on regulated futures exchanges such as Cboe and CME, the spot price shows a deviation of less than 0.25%. Together, CME and Cboe form a Bitcoin futures market equivalent to 48% of the size of the real spot market.
Meanwhile, pricing discrepancies between spot and futures exchanges are arbitraged with impressively fast speed. According to the white paper, more than 50% of 1% pricing discrepancies were arbitraged away within 1 second, and more than 90% within 49 seconds.
As such, Bitwise concludes that fake exchanges have little impact on the price discovery of bitcoin. The company explains that this is because real investors and derivatives products only draw bitcoin price from the 10 real exchanges. Prices displayed on fake exchanges are usually not taken as real market signals by these investors.
As Bitwise states in the white paper, “these results demonstrate that prices between the 10 exchanges not only trade closely together but also have their disparities rapidly arbitraged away, meeting both criteria set forth by the Commission for demonstrating effective arbitrage.”
Bullish for ETF?
The market, if it is indeed as regulated and mature as Bitwise claims, may ease the SEC’s worries on bitcoin trading and prompt the regulator to greenlight the firm’s exchange-traded fund (ETF).
Bitwise filed an ETF application in January, but the SEC has avoided making a decision citing worries over price manipulation and fraud. The SEC wants proof of transparency in bitcoin price discovery and strict oversight over bitcoin-related investment products. Therefore, if Bitwise can convince the regulator that the bitcoin market “is both significantly smaller and significantly more efficient than commonly understood,” it might also put itself on the fast track to provide the first bitcoin ETF in the U.S.