Two new studies estimate the share of criminal transactions in the Bitcoin ecosystem.A�The answers are so different that presumably, neither study can deliver a valid result.A�From a methodological point of view, however, the works are highly interesting – if not watertight.
SEX, DRUGS AND BITCOIN
The first study presented here is titled “A�Sex, Drugs, and Bitcoin: How Much Illegal Activity Is Financed Through Cryptocurrencies?A�”.A�It was written by three economists from Sydney and Riga.A�The scientists used a wide set of tools to expose all the illegal activity in the Bitcoin universe.A�Their result is extreme: “We found out that about a quarter of Bitcoin users and half of Bitcoin transactions are associated with illegal activity.” This amounts to about 72 billion dollars a year, putting Bitcoin in a league with the US and European markets for illegal drugs.
How do the researchers know this?A�They have used all available data and methods to separate the “illegal” from the “legal” users and to find their share.A�Methodically, this is extremely interesting as it covers almost the full spectrum of what is possible through blockchain analysis.
First, the researchers have linked all the addresses on the blockchain to wallets.A�They used the Union Find algorithm, which analyzes transactions and forms clusters.A�For example, if someone receives bitcoins with two different addresses and then forms a transaction in which these two addresses are inputs, then this shows that the two addresses belong to the same owner.A�This makes it easy to determine which addresses belong in a wallet.
In order to filter the sample of users, the researchers then cut out the transactions of exchanges and miners as well as all transactions with a value below one dollar.A�After these operations, they had a database of about 106 million users, who together executed 606 million transactions worth about $ 1.9 trillion.A�These users then sorted them by specific properties, such as the number of transactions, the average size of the transactions, the number of counterparties, the use of blenders, and so on.
After all, researchers have linked this sample of users to criminal activity.A�They used three methods for this.A�First, they looked for Bitcoin confiscations by the authorities, for example, by searching through court records in the United States.A�Since these files contained references to transactions or addresses, they were able to identify the user.A�Similar results were achieved by analyzing auctions of confiscated bitcoins.A�In this way, they were able to identify 1,016 illegal users.
The second method led to the most successes.A�Researchers have identified the hot wallets of 17 Darknet marketplaces.A�The wallet analysis enabled them to find users’ addresses and wallets, resulting in a little more than 6 million Darknet customers.A�Finally, the third method was based on analysis from online forums.A�In these users have occasionally posted their address, for example, to ask what has happened to a transaction.A�As a result, the scientists have discovered another 446 criminal users.
Thus, the researchers have identified a total of 6,223,337 criminal Bitcoin users.A�These represent 5.86 percent of all Bitcoin users.A�If you extrapolate it to transactions, their share is even bigger – about one third of all transactions.
However, we only have the first part with it.A�In the second, the researchers have tried to make estimates about illegal activities with Bitcoin, which do not depend on already existing connections.
For this they have used two methods.A�First, because of the transaction histories, they have formed network topologies.A�Put simply, if your friends are thieves, you’re one too.A�”If we know that users A, B, and C are involved in illegal activities (for example, because their bitcoins were confiscated by the police), a user X who deals exclusively or predominantly with A, B, and C is also likely to be in illegal Involved in actions. “Using algorithms, researchers have attempted to trace the topologies of illegal networks.
Second, scientists compared the usage patterns of decent and criminal Bitcoin users.A�There are several indications of illegality: Criminal users send out more transactions (while law-abiding users tend to hoard bitcoins), they tend to use methods to disguise the source of their coins, such as blenders, and finally, their activity correlates more with a high one Activity in the Darknert scene, but barely with the Bitcoin-Hypes to be noticed in Google-trends, which express rather the interest of investors.
The two methods allow further estimation of the share of Bitcoin users involved in illegal activities.A�The average between the two estimates is about 25 percent, which leads to about 26.8 million criminal Bitcoin users, which face almost 80 million honest users.A�However, these 25 percent of the users provide for a much higher proportion of transactions, which the researchers specify with about 44 percent or 269 million transactions.
These results, the researchers say, also say something about the intrinsic value of bitcoin.A�”Our results show that Bitcoin is currently used relatively broadly as a means of payment to trade in illegal goods or services, and that the illegal use of Bitcoin is likely to make a significant contribution to the fundamental value of Bitcoin.”
However, this is just the perspective of a study.A�Another study comes to significantly different results.
The second recent blockchain crime investigation is titled “A�Bitcoin Laundering: An Analysis of Illicit Flows Into Digital Currency ServicesA�” (“Bitcoin Money Laundering: An Analysis of IllegalA�Cash Flows on Digital Currency ServiceA�Providers”).A�It was released by the Center for Sanctions and Illicit Finance, a think tank advising the US government, and Elliptic, a company that inspects and sells the blockchain.
The short study published by the two organizations is above all the money laundering of Bitcoins, ie the process that disguises the (criminal) origin of Bitcoins and these change into “clean” Bitcoins or equal to dollars.A�Analysts explored Bitcoin transactions between 2013 and 2016 to chart trends in how Bitcoins, which are clearly linked to criminal activity, are going to be changed or cleaned up.A�”The parameters of the study are deliberately slim to make the data more manageable,” explain the authors.A�They acknowledge that “this is likely to reduce the volume of criminal transactions used for the analysis.” Less than the amount, the study depends on the patterns.
To identify the “dirty” bitcoins, Elliptic’s analysis tool was used.A�This combines the public blockchain data with a private set of data linking bitcoin addresses to known entities, which may be exchanges, but also blackmailers or darknet markets.A�”We’ve seen 214 unique conversion services, including virtual currency exchanges, gambling sites and blenders.” There were 102 illegal entities.
The authors make no secret that the study can only have a limited validity.A�”The study looks at well over half a million Bitcoins that flowed directly from illegal sources to the conversion services in the years 2013 to 2016.A�This is not meant to be an exhaustive overview of this type of activity since we have not tried to identify all the illicit sources of bitcoins.A�However, it covers the majority of known, significant entities of this nature. “However, it does not consider cash flows that go through unknown middlemen before they land on the exchange platforms.A�Therefore, the authors assume that the actual volume is significantly larger.
Despite these limitations, it is surprising to see how far the results differ from those of the first study.A�The overwhelming majority of criminal cash flows come from relatively few entities, especially the big darknet marketplaces.A�Nine out of 102 illegal entities accounted for 95 percent of the washed bitcoins.A�”According to our study, the absolute percentage of ‘dirty bitcoins’ that landed on the convection platforms was rather small.A�Only 0.61 percent of the money that arrived here in the four years could be assigned to an illegal source, with the highest value (1.07%) in 2013. ”
Although the authors assume that the true proportion is higher.A�But do not expect it to be 25x higher than your estimate.
METHODOLOGICAL PROBLEM AREAS
The results of the two studies show a very different picture of the share of illicit trade in Bitcoin.A�AA�A�study by Europol on the darknet markets alsoA�A�suggests that the volume of these markets is very small when compared to general drug trafficking or legal Bitcoin transactions, asA�A�a study by an international research team hasA�A�shown Usage of Bitcoin decreases significantly.
Where do these huge differences come from?A�Therefore, that every statistic above all confirms what their makers want to say?A�Do the researchers from Sydney and Riga want to extrapolate the proportion of criminal activity, while Elliptic tends to downsize it?A�Both studies are likely to have their shortcomings, perhaps showing how far they are wrong.
The study by Elliptic and the Center for Sanction and Illicit Finance itself explains that it is not representative and that it is quite possible that the true share of illegal transactions is considerably higher.A�But is it as high as the other study suggests?
If we look closely, we find some methodological weaknesses in the study of the researchers from Sydney and Riga, which make the results less valid.A�For example, the estimation of the number of users by linking addresses to wallets.A�This “clustering” provides insights, but rather no reliable results.A�Because it is not so difficult to escape the clustering.A�You can form different wallets with Trezor, Ledger or Electrum and separate them consistently.A�In this way, a user can have multiple wallets that are unreachable.A�This is one of the starting points of the study – the number of users – a highly questionable statement.A�One should not only assume that the actual number of users is significantly lower, but also that criminal users invest more energy,A�to prevent wallet connections.A�Thus, not only the absolute, but also the relative number of criminal users should be significantly lower than assumed by the study.
Furthermore, the researchers pre-filtered their set of users.A�All transactions have been cut out of stock exchanges and miners and with a value below one dollar.A�Thus, the study shows only a subset of users.A�The people who send bitcoins from purse to purse, or from mining pool to the savings wallet, or to a stock exchange, or all those who once enjoyed playing Satoshi Dice, fetching bits of bitcoin for free, or otherwise experimenting with microtransactions have – these users count in the study not as a user.A�The fact that the proportion of criminal users is higher, if one reduces the amount of users so, is unsurprising.
It could also be said that connecting darknet hot wallets to users can also hit legitimate users, depending on how deep the algorithms dig – apart from the fact that the number of real users via wallet clustering is not valid.A�Identification through activity patterns provides false positives with relatively great certainty, such as people who use blenders or generally like to use bitcoins to pay or make other (legal) transactions.
So there are many points that can be criticized in the study.A�Such constraints at best make the results estimates.A�The actual values a��a��can be significantly lower.
However, one can hardly say where the truth is.A�Bitcoins are undoubtedly used for illegal things, and to a considerable extent.A�However, whether this covers 2 or 20 percent of the various metrics for bitcoin activity – users, transactions, volumes – is hard to say.A�To some extent, the system is transparent – but not enough to clearly answer the question of the proportion of illegal activities.