As a new species of cryptographic asset based on blockchain technology, the non-fungible token (NFT) became prominent and attracted attention from academia and practitioners. Unlike cryptocurrencies, which primarily serve as a medium of exchange, NFTs are blockchain-recorded digital assets that can be anything digital, e.g., images, videos, and songs. The NFT market reached its peak in 2021 when the total trading volume increased by 21,350%, the average price increased by 1,542%, and market size increased by 4,440% compared to 2020. However, the market experienced a large slump in Q3 2022, with a 76% drop in the average price and 77% drop in volume of dollar traded. The boom-bust cycle in the NFT market is in many ways like historical financial bubbles, e.g., the Tulip Bubble, the South Sea bubble and more recently, the dot come bubbles. As the forerunner, Dowling (2021a) finds pricing LAND, a rapidly growing type of NFT in Decentraland, inefficient. As NFTs generate trivial cash flow, the soaring prices of NFTs are sometimes considered speculative bubbles. In this sense, some argue that participants in the NFT market mainly profit by purchasing overvalued digital assets and then reselling these assets to the next batch of “greater fools” (the example of “greater fool theory” in the cryptocurrency market). An updraft in prices encourages more people to buy assets and draws more media coverage, and in turn, causes even more people to buy, creating considerable profits for an earlier generation of players. This strategy works if enough newcomers are willing to take over the assets at even higher prices. But once people start to question the sustainability of the upward trend of the asset price, the momentum may switch to the reverse direction where everyone tries to escape from the market as quickly as possible to sell the asset before the price becomes too low, creating a catastrophic market decline. Can we ascribe the ups and downs of NFT to market bubbles? Does investors’ herding mentality play a role in this price cycle? This research provides the first set of empirical evidence on herding in the NFT market.
Herding is defined as individuals’ suppressing their own beliefs and basing their investment decisions solely on the collective actions of the market, even when they disagree with the majority of predictions. With herding, market participants’ private information fails to be reflected by market prices, leading to price deviation from the fundamental rational value. A high correlation between investors’ beliefs and decisions magnifies asset price volatility. Herding is a potential explanation for the extreme return and instability in the NFT market.
Although some literature claims that the necessary conditions for the emergence of herding include the uncertain fundamental value and varied private information among investors, which seems contradictory to the zero fundamental value of the NFT according to the standard valuation method, recent studies manage to extend the discussion to the case of NFT. If an agent obtains additional utility from taking identical actions by others (payments externality), then herding could also emerge. Like cryptocurrency, the more widely NFTs are used, the more benefited holders of NFTs would be from the larger community of NFTs (network effects), thus creating a positive payment externality. In this paper, we propose investigating whether herding exists in the NFT market and exploring its dynamics and the possible spillover across NFT submarkets.
Our study contributes to the burgeoning literature on NFT and herding as the first one to examine investors’ herding in the NFT market, different from previous studies mainly focusing on the pricing efficiency of NFT. With data indicating individual investors’ historical transactions, we are the first to examine the role of investors’ trading experience in herding. Meanwhile, our study provides a potential explanation of the cause of the boom and bust in NFT prices.
The main findings
Our empirical evidence supports the existence of herding in the NFT market. There are three waves of herding: from Nov. 2017 (the beginning of our sample) to Oct. 2018, from Dec. 2019 to Feb. 2020, and from Aug. 2020 to Apr. 2021(the end of our sample). Launches of several early submarkets, i.e., Opensea, Cryptokitties, and Godsunchained, and the release of the leading collection of NFT CryptoPunks at the end of 2017 coincide with the first period. The second period is near the time when the other two submarkets, Atomic and Decentraland, were launched, while the latest period is when the media intensely covered NFTs in the second half of 2020 as the NFT market started to proliferate from July 2020. During these three periods, the daily market return became higher and more volatile than usual.
Newcomers and investors’ attention with herding formation
Oh et al. (2022) note that in the NFT market, experienced investors earned higher returns for each unit of Ethereum (ETH) invested than those inexperienced, possibly because inexperienced investors purchase NFT with a higher average price than experienced investors. Most newcomers have traded only once and are unfamiliar with the market situation. They are, therefore, more likely to follow the market consensus and thus fuel the formation of herding. Thus, we conjecture that herding is more likely to occur when newcomers flow into the NFT market. Our empirical evidence supports this conjecture, suggesting the emergence of herding when more newcomers join the market. They bear higher costs of purchasing while pushing up prices.
Fortune-made stories in the NFT market spread to the public, either by word of mouth, news media, or social media, driving people to participate and then fostering the development of the NFT market, a case of Narrative Economics. Herding emerges when investors pay more attention to good/bad news about NFTs, and the pattern, together with the newcomers, configures a story in which newcomers urged by the overwhelmed reporting by media join/leave the NFT market and push up/down the prices, and herding emerges accordingly.
Herding in the NFT market and the performance of cryptocurrencies
While NFTs may be considered alternative assets to Bitcoin, investors willing to purchase NFTs must first convert their fiat money to Ethereum. To buy NFTs, an investor can choose to sell Bitcoin and buy in Ethereum. Therefore, Ethereum, indispensable for trading NFTs, may perform differently from Bitcoin when herding arises in the NFT market. Our empirical results show that while herding in the NFT market tends to emerge as the return on Ethereum increases, it tends to diminish as the return on Bitcoin rises. This result is compelling for investors with portfolios on both NFTs and cryptocurrencies for diversification, as they can adjust their portfolio accordingly to avoid the harm of herding on diversification.
Within- and cross- submarkets herding
Investors trade on several platforms or instruments in the traditional financial market simultaneously. Investors in stock markets usually follow news about the global market. For example, Chiang & Zheng (2010) note that investors in all of the major markets, including those in Asia, Latin America, and Europe, imitate the actions of investors in the US market, a case of herding across markets. Herding across markets is due to the cost of information and the free-of-charge information revealed by investors’ actions in other markets. Therefore, we wonder whether herding exists across different NFT submarkets (OpenSea, Atomic, Cryptokitties, Godsunchained, and Decentraland), in which case investors in one submarket would follow the investing decision by investors in other submarkets. Although herding within each of the five submarkets is prominent, herding across submarkets rarely happens in general. This result concords with Nadini et al. (2021) and White et al. (2022) that investors in the NFT market tend to focus on a limited number of tokens, while the trading counterparties also are in small constrained groups, deterring their attention to other submarkets.
Conclusion
Our study is the first to examine the existence of herding in the NFT market, and the empirical evidence shows that herding is more likely to arise when market returns and the price experience large upward/downward movement. Besides, herding is more likely to occur when prominent events happen, e.g., the launch of the NFT submarkets and the release of the leading NFT collection. We have also investigated the agents whose behavior correlates with herding. As newcomers step into the market, herding becomes more likely, and the emergence of herding correlates with investors’ attention driven by media coverage. Therefore, it is very likely that newcomers were attracted by the reporting in the media and then purchased or sold large amounts of asset the NFT market, pushing up or down prices as they lacked knowledge of the market situation. Ethereum returns are positively correlated with herding in NFT markets, while Bitcoin returns are negatively correlated with herding in NFT markets. Lastly, we find that investors in different submarkets do not herd about other submarkets in general.
Te Bao is an Associate Professor of Economics at the School of Social Sciences and an affiliated Research Fellow at the Joint NTU-WeBank Research Centre on Fintech, Nanyang Technological University, Singapore.
Mengzhong Ma is a Ph.D. student in Interdisciplinary Graduate Program at Nanyang Technological University, Singapore.
Yonggang Wen is an IEEE Fellow and the President’s Chair Professor of Computer Science & Engineering at the School of Computer Science and Engineering (SCSE) of Nanyang Technological University, Singapore.
This post is adapted from their paper, “Herding in the Non-fungible Token (NFT) Market,” available on SSRN.
