Initial coin offerings (ICOs) have recently emerged as a popular method of financing blockchain-related startups. In an ICO, a startup creates and distributes its digital tokens, typically in exchange for Bitcoin, Ethereum, or fiat currencies such as U.S. dollars to raise capital to fund its operations. The startling growth of this market, coupled with rampant speculation and volatility, has both generated excitement and raised concern about potential exploitation or fraud.
In a new paper, we study “pump-and-dump” schemes (P&Ds) in the cryptocurrency market. P&D is a form of price manipulation that involves artificially inflating an asset price before selling the cheaply purchased assets at a higher price. Once the assets are “dumped,” the price falls and investors lose money. Such schemes are most common with microcap stocks and have recently become popular in the cryptocurrency market. The U.S. Securities and Exchange Commission deems P&Ds illegal in the stock market, but the regulation of P&Ds in the cryptocurrency market is weak or nonexistent. Many tokens are difficult to justify either as investments or consumer products and do not fit neatly into existing securities or consumer-protection laws. The regulation of cryptocurrencies also requires more global coordination than other assets, since tokens typically are traded globally.
In the cryptocurrency market, manipulators often organize “pump groups” using encrypted messaging apps such as Telegram. They create Telegram channels and invite other investors to join. They frequently advertise on social media platforms to attract investors. A Telegram channel operator can post messages for other members to read. For a planned pump, the operator announces the target date, time, and exchange, usually at least one day in advance. However, they do not disclose the identity of the target token until the scheduled time. Members also receive multiple reminder messages before the announcement of the token symbol. As we show in this study, a typical cryptocurrency P&D lasts for only several minutes, leaving little time for non-members to participate. Therefore, it is reasonable to believe that Telegram channel members are dominant participants in P&Ds.
Using 500 hand-collected cryptocurrency P&Ds, we document several stylized facts. First, P&Ds have dramatic short-term impacts on the prices and volumes of most of the pumped tokens. In the first 70 seconds after the start of a P&D, the price increases by 25 percent on average, trading volume increases 148 times, and the average 10-second absolute return reaches 15 percent. A quick reversal begins 70 seconds after the start of the P&D. After an hour, most of the initial effects disappear. The above findings hold for both liquid and illiquid tokens, although they are stronger for relatively illiquid tokens.
We also document that prices of pumped tokens begin rising five minutes before a P&D starts. The price run-up is around 5 percent, together with an abnormally high volume. These results are not surprising, as pump group organizers can buy the pumped tokens in advance. When we read related messages posted on social media, we find that some pump group organizers offer premium memberships to allow some investors to receive pump signals before others do. The investors who buy in advance realize great returns. We deem investors who know the identity of pumped tokens in advance as insiders and the rest as outsiders. Calculations suggest that insiders’ average return can be as high as 18 percent, even after considering the time it may take to unwind positions. This far exceeds the round-trip trading costs that are between 20 and 50 basis points. For an average P&D, insiders make one Bitcoin in profit, approximately one-third of a token’s daily trading volume. The trading volume during the 10 minutes before the pump is 13 percent of the total volume during the 10 minutes after the pump starts. This implies that an average trade in the first 10 minutes after a pump has a 13 percent chance of trading against these insiders and on average the outsiders lose more than 2 percent (18 percent x 13 percent).
The quick reversal means that an outside investor who is unaware of the P&D timing in advance needs to buy and sell very quickly to make a profit. Equally important, liquidity may disappear when the investor wants to sell. We conduct a performance analysis that accounts for real-time liquidity. Our analysis shows that, on average, only investors who buy in the first 20 seconds after a P&D begins can make a profit, and they can do so only if they do not hold their tokens for very long. For example, if an investor buys tokens between 10 and 20 seconds after the P&D starts and begins selling one minute later, he will lose 0.72 percent of his investment on average. An investor who buys target tokens one minute after a P&D starts will lose more than 1 percent, even if he sells immediately after the purchase.
The price and volume patterns around P&Ds are analogous to asset bubbles that have been studied in the literature. However, these cryptocurrency bubbles develop within much shorter time intervals. We conjecture that one plausible mechanism for this is that P&Ds attract overconfident investors who believe that they can time the market more accurately than others can (see Scheinkman and Xiong, 2003; Daniel and Hirshleifer, 2015). Another possible mechanism is that these investors have gambling preferences. The short-term returns on pumped tokens are very high and salient. Investors may overweigh the possibility of these returns in their decision-making or overestimate the skewness of token returns (see Barberis and Huang, 2008).
P&Ds, or other forms of price manipulation, are generally considered undesirable and should be regulated. Besides illegitimate wealth transfers, are there any other detrimental consequences of P&Ds? We investigate this question using a natural experiment. On November 24, 2017, Bittrex, a U.S.-based cryptocurrency exchange, announced that it would ban P&Ds. Although it was not able to eliminate P&Ds altogether, it reduced the frequency of P&Ds on its exchange. This ban affected only the tokens traded on Bittrex, not tokens on other exchanges. We find that the ban increased token prices and liquidity of Bittrex tokens. While there are many aspects of cryptocurrency regulation, our findings suggest that P&Ds should be regulated.
Barberis, N and M Huang (2008), “Stocks as Lotteries: The Implications of Probability Weighting for Security Prices,” American Economic Review 98 (5), 2066-2100.
Daniel, K and D Hirshleifer (2015), “Overconfident Investors, Predictable Returns, and Excessive Trading,” Journal of Economic Perspectives 29 (4), 61-88.
Scheinkman, J and W Xiong (2003), “Overconfidence and Speculative Bubbles,” Journal of Political Economy 111 (6), 1183-1219.
This post comes to us from professors Tao Li and Baolian Wang at the University of Florida and doctoral student Donghwa Shin at Princeton University. It is based on their recent paper, “Cryptocurrency Pump-and-Dump Schemes,” available here.