The digital currency bitcoin (BTC) was introduced in 2009. Bitcoin and the many other digital currencies are primarily online currencies. The key currencies are those based primarily on cryptography. Bitcoin is the leading cryptocurrency, but there are nearly 2000 others.
Bitcoin has experienced a meteoric rise in popularity since its introduction. Its success has inspired scores of competing cryptocurrencies that follow a similar design. Bitcoin and most other cryptocurrencies do not require a central authority to validate and settle transactions. Instead, they use cryptography (and an internal incentive system) to control transactions and manage the supply. A decentralised network validates transactions. Once confirmed, all transactions are stored digitally and recorded in a public ‘blockchain,’ which can be thought of as a distributed accounting system.
The proliferation of cryptocurrencies and changes in technology have made it easier to conduct ‘pump and dump’ schemes. Many of the cryptocurrencies available today are illiquid and are characterised by very low trading volumes on most days, with occasional volume and price spikes.
Cryptocurrencies have only recently become a subject of research in economics, but the topic has been of interest for longer in computer science (for early work on incentives by computer scientists, see Babaioff et al. 2012 and Eyal and Sirer 2014). Numerous researchers have conducted studies in order to document and combat threats such as Ponzi schemes, money laundering, mining botnets, and the theft of cryptocurrency wallets. Ron and Shamir (2013) attempt to identify suspicious trading activity by building a graph of bitcoin transactions found in the public ledger.
In the case of economics research, Gandal et al. (2018), showed that the first time bitcoin reached an exchange rate of more than $1,000, the meteoric rise was likely driven by fraud in the form of fraudulent trading activity. Griffin and Shams (2018) found that tether, a digital cryptocurrency that is pegged to the US dollar, likely led to a significant fraction of the increase in the price of bitcoin and other cryptocurrency prices during the meteoric rise in cryptocurrency valuations in 2017.
In our recent work on cryptocurrency pump and dump schemes (Hamrick et al. 2018), we quantify the scope of these schemes on Discord and Telegram, two widely popular group messaging platforms with 130 million users and 200 million users, respectively. Both platforms can handle large groups with thousands of users, and they are the most popular outlets for pump and dump schemes involving cryptocurrencies.
Technologies like Telegram and Discord allow people to easily coordinate such schemes. Telegram is a cloud-based instant messaging service using voice over internet protocol (VoIP). Users can send messages and exchange photos, videos, stickers, audio, and files of any type. Messages can be sent to other users individually or to groups of up to 100,000 members. As of March 2018, Telegram had 200 million active users. Discord, first released in 2015, has similar capabilities and 130 million users as of May 2018. Discord and Telegram are primary sources for cryptocurrency pumps and have been used for pump and dump schemes on a large scale. Perhaps because of the regulatory vacuum, many of the pump groups do not hide their goals.
We identified 3,767 different pump signals advertised on Telegram and another 1,051 different pump signals advertised on Discord during a six-month period in 2018. The schemes promoted more than 300 cryptocurrencies. These comprehensive data provide the first measure of the scope of pump and dump schemes across cryptocurrencies and suggest that this phenomenon is widespread and often quite profitable.
The data collection required for the analysis was substantial. Pump data were gathered by collecting messages posted to hundreds of dedicated Discord and Telegram channels using their APIs and manually labelling messages that signalled pumps.
We also collected price data on nearly 2,000 coins across 220 cryptocurrency trading exchanges from Coinmarketcap.com, the leading website of aggregated data on cryptocurrency trading, during the six-month period from January to July 2018. This gave us a total of 316,244,976 price data points across all of the coins listed. The data collected are at the finest granularity presented by Coinmarketcap.com at the time of collection, that is, five-minute intervals. We then matched the extracted pump signals announced on Discord and Telegram with the trading data.
We next measured the ‘success’ of the schemes, which we define to be the percentage increase in the price following a pump. Ten percent of the pumps on Telegram (Discord) increased the price by more than 18% (12%) in just five minutes. Recall that the January–July 2018 period was a period in which cryptocurrency prices and trading volume were falling significantly. Hence, such percentage increases were ‘achievements’ for the pump schemes.
Finally, we examined what factors explained the ability to increase price. The most important variable in explaining success of the pump is the ranking of the coin, where ranking is based on market capitalisation (which is highly correlated with trading volume). Coins with lower market capitalisation typically have lower average trading volume. Lower average volume gives the pump scheme a greater likelihood of success.
We found that pumps using obscure coins with low market capitalisation were much more profitable than pumping the dominant coins in the ecosystem – the median price increase was 3.5% (4.8%) for pumps on Discord (Telegram) using the top 75 coins; it was 23% (19%) on Discord (Telegram) for coins ranked over 500. (bitcoin is the top ranked coin and has rank #1.) We discuss the effect of other variables on the ‘success’ of the pumps in our paper.
Three other (essentially) concurrent papers also examine pump and dump schemes on cryptocurrencies, but with a different emphasis. Kamps and Kleinberg (2018) use market data to identify suspected pump and dumps based on sudden price and volume spikes. They evaluate the accuracy of their predictions using a small sample of manually identified pump signals. Xu and Livshits (2018) use data on roughly 200 pump signals to build a model to predict which coins will be pumped. Their model distinguishes between highly successful pumps and all other trading activity on the exchange. Li et al. (2018) use a differences-in-differences model to show that pump and dumps lower the trading price of affected coins.1
Our work is different from the other concurrent work in several important ways. First, we have collected as many pump signals as possible from channels on Discord and Telegram. We also evaluate them all, without restricting ourselves to the successful pumps. Second, we investigate reported pumps for all coins with public trading data, not only those taking place at selected exchanges. This enables us to incorporate ecosystem-wide explanatory variables such as the number of exchanges on which a coin is traded in order to assess what makes a pump and dump scheme successful.
Why should we care about pump and dump schemes in cryptocurrencies? Recent trends indicate that bitcoin is becoming an important asset in the financial system. Further, trading in cryptocurrency assets has exploded as the market capitalisation of cryptocurrencies grew stunningly in the past few years. In February 2014, the market capitalisation of all cryptocurrencies was approximately $14 billion. As of February 2018, the total market capitalisation was approximately $414 billion, before falling back to $122 billion in December 2018. Bitcoin itself reached a peak of more than $19,000 before plummeting over the next few months to $6,000. Currently (as at mid-December 2018), the bitcoin price is close to $4,000.
In February 2018, there were more than 300 cryptocurrencies with market capitalisations between $1 million and $100 million. In January 2014, there were fewer than 30 coins with market capitalisations between $1 million and $100 million. The markets for such cryptocurrencies are very thin and subject to manipulation.
As mainstream finance invests in cryptocurrency assets, it is important to understand how susceptible cryptocurrency markets are to manipulation. We have provided the first measure of the widespread scope of pump and dump schemes. We encourage the nascent cryptocurrency industry, regulators, and researchers to work together to try to eliminate manipulation in cryptocurrency assets.
Babaioff, M, S Dobzinski, S Oren and A Zohar (2012), “On bitcoin and red balloons,” Proceedings of the 13th ACM Conference on Electronic Commerce 2012: 56-73.
Eyal, I and E Sirer (2014), “Majority is not enough: Bitcoin mining is vulnerable,” Eighteenth International Conference on Financial Cryptography and Data Security 2014.
Gandal, N, J Hamrick, T Moore and T Oberman (2018), “Price manipulation in the Bitcoin ecosystem,” Journal of Monetary Economics 95: 86–96.
Griffin, J and A Shams (2018), “Is bitcoin really untethered?”.
Hamrick, J, F Rouhi, A Mukherjee, A Feder, N Gandal, T Moore and M Vasek (2018), “The economics of cryptocurrency pump and dump schemes”.
Kamps, J and B Kleinberg (2018), “To the moon: defining and detecting cryptocurrency pump-and-dumps,” Crime Science 7(1): 18.
Li, T, D Shin and B Wang (2018), “Cryptocurrency pump-and-dump schemes”.
Ron, D and A Shamir (2013), “Quantitative analysis of the full Bitcoin transaction graph,” in Financial Cryptography and Data Security, volume 7859 of Lecture Notes in Computer Science: 6–24.
Xu, J and B Livshits (2018), “The anatomy of a cryptocurrency pump-and-dump scheme”.
Yuxing Huang, D, H Dharmdasani, S Meiklejohn, V Dave, C Grier, D McCoy, S Savage, N Weaver, A Snoeren and K Levchenko (2014), “Botcoin: Monetizing stolen cycles,” in Proceedings of the Network and Distributed System Security Symposium.
 There have been media articles about the pump and dump phenomenon as well. Mac reported on pumpand dump schemes in a Buzzfeed article published in January 2018, available here.This was followed by work by Shifflet and Vigna in a Wall Street Journal article published in August 2018, available here.