In a write-up this week by Tim Swanson, Director of Market Research at R3CEV, he published some eye opening information into how bitcoin transactions flow from bitcoin exchanges using blockchain intelligence software Chainalysis.
Chainalysis is one of a handful of companies in the bitcoin space that are focusing on analysis of bitcoin and blockchain data and tracking, mining as much information as possible while connecting the dots. Using this data and working with Chainalysis, Swanson took a granular look into specific transaction “corridors” and movements between miners, exchanges, darknet markets, payment processors, and coin mixers.
In reviewing the data, there were several bitcoin exchanges that were tracked including Bitfinex,Bitstamp, BTCC, BTC-e, Circle, Coinbase, Huobi, itBit, Kraken, LocalBitcoins, OKCoin, and Xapo. All the exchanges were included in the data but only the identities of BTC-e and LocalBitcoins were revealed in the charts and data. The chart below is a backwards looking view into transaction flows between exchanges, for example you can see BTC-e and LocalBitcoins clearly and the other exchanges are unidentified. The thickness of the bands signifies the volume of transactions. The data and charts are for the 2015 year.
Chainalysis’ clustering algorithm for this data was intended to isolate sub-economies, and not see each individual trades (“hops”), so the number of hops between exchanges are removed and you end up with the start and finish of a transaction, and a smooth chord that shows you where the transaction began and ended; which is supposed to give you a better idea of the sub-economy that bitcoin has.
For example, in this chart below you can see LocalBitcoins isolated with different chords of transactions going to BTC-e, SharedCoin, and Agora. The thickness of the band would indicate more volume going to BTC-e.
We’ve known about blockchain intelligence companies for a while now, and there seems to be a market for it as there are more and more that are appearing as the bitcoin network grows. We also know that bitcoin transactions aren’t completely anonymous but using coin mixers we can obfuscate transactions to the point where they are no longer traceable, however, does this analysis now show that possibly the tracking algorithms are more advanced than people realize?
In this post I discuss bitcoin privacy and the blockchain, using bitcoin mixers, and what methods may be best to achieve privacy on the blockchain. It’s possible that this may not be enough, or is it that this data presented by Chainalysis is only capturing those with poor privacy habits (i.e., address reuse, not using coin mixing services, using bitcoin wallets with bad privacy behaviors, etc). It’s hard to know given what has been presented.
However, what is most troublesome is that Chainalysis writes when asked what exchange activity they can actually identify:
It varies per service but Chainalysis (and others) have access to some “full wallets” from clients. Also newer deposits are often not known so the balance in a wallet will be underestimated due to how the current algorithms work.
It begs the question, how and why does Chainalysis have access to “some full wallets” in the first place? Clearly they are using this data for their blockchain intelligence for data mining and tracking. Why would a wallet provider give them “full access.” Just a few days ago a redditor saidthat they spoke with BitPay’s general Manager Marcel Roelants in Amsterdam asking if they actively cooperate with Chainalysis, and he said that they do. BitPay runs the open source bitcoin wallet Copay, and is the biggest bitcoin payment processor.
They check incoming payments for suspicious transactions. They also collect consumer data, certainly for some kind of transactions like buying / selling gold and actually are in favour of putting identity in a bitcoin wallet.
It’s no surprise that BitPay would work with a company such as Chainalysis, but the claim that Chainalysis has access to “full wallets” in order to track and mine data, possibly BitPay’s (this isn’t certain yet) or others, is concerning.
The exchange transaction data shows some ‘sub-economies’ as suggested, such as possible arbitrage opportunities between exchanges or darkmarkets using bitcoin, but with so many hops in-between and users possibly mixing their coins it’s completely plausible the trades in the middle are for sub-economies that aren’t as transparent to what we have been made aware of yet; or maybe bitcoin is being used as it should be, a currency to pay for things on the web.