Breaking down bias in the NFT marketplace: Study reveals female and Black avatars sell for less—and offers solutions to promote equality

The United States has a long history of gender and racial disparities in wealth, education, employment, and consumption. Research on modern marketplaces, including online rental and labor markets, has identified ways to reduce gender and racial inequality.

The non-fungible token (NFT) market, an emerging online marketplace, plays a crucial role in translating users’ real-world identities into the NFT marketplace. Given recent social movements and the younger demographics associated with these digital spaces, some anticipate that the NFT marketplace and its more liberal and progressive inhabitants will reduce prejudice and social bias.

In a new study, researchers from the Tepper School of Business and other universities examined the impact of avatars’ gender and race on transaction prices in the NFT marketplace. They found that contrary to expectations, there were pronounced gender and racial disparities in NFT prices. They also found that avatars with features associated with high-tech or higher education countered these disparities. The authors offer insights into policy strategies to promote racial and gender equity in the NFT marketplace.

The study, by researchers from the Tepper School at Carnegie Mellon University, New York University, and Harvard University, appears in the International Journal of Research in Marketing.

“Based on the observation that NFT investors often represent a younger, wealthier, and more politically progressive demographic, we sought to determine whether this marketplace might be less susceptible to prejudice and social bias,” says Yuan Yuan, a Ph.D. student in marketing at Carnegie Mellon’s Tepper School of Business, who led the study. “As the NFT market continues to grow and increasingly leverages digital communities as brand assets, such questions will become increasingly important.”

Gender and race are key factors that can lead to discrimination and inequality within fields such as marketing and economics. These disparities impact both the fairness and the performance of the NFT marketplace, where avatar headshots serve as representations of individuals’ digital identities. In this study, researchers sought to identify the effects of avatars’ gender and race on NFT prices, how these effects varied over time, and what providers can do to mitigate disparities.

The study found that female CryptoPunks avatars were sold at a 37% lower price than their male counterparts (i.e., avatars with similar image attributes), while Black CryptoPunks avatars were sold at a 31% lower price than their White counterparts. In a controlled lab experiment, the study causally confirmed both gender and race price premiums, as well as the presence of racial bias in these purchasing behaviors.

Researchers also identified features of the avatars linked to high-tech or higher education (e.g., 3D virtual reality headsets or nerd glasses), finding that avatars with these features countered the disparities. This suggests incorporating such attributes into avatars could counteract bias in the NFT marketplace.

While previous studies have suggested that conservative ideologies may exhibit greater prejudices against underserved groups, including women, and that such biases are more pronounced among older adults compared to younger ones, this study challenges the notion that gender and racial biases are diminished in the emerging, technologically advanced marketplace. The findings reveal deep-seated and persistent societal stereotypes and prejudices in the digital marketplace, particularly in the devaluation of Black avatars across demographics, highlighting the complexity of biases in digital marketplaces.

“By considering our findings, NFT creators and platforms can promote a more balanced marketplace by introducing attributes that counteract bias,” explains Kannan Srinivasan, Professor of Management, Marketing, and Business Technology at Carnegie Mellon’s Tepper School of Business, who co-authored the study. “NFT platforms that want to harmonize efficiency with fairness can also adopt policies that discourage bias, providing more equitable guidelines for the creation and transaction of NFTs.”

The authors note some limitations of their study, including the challenge of distinguishing between factors perceived as more valuable and those that are personally preferred, and the exclusivity of their focus on human avatars. In addition, they note that their database's diversity might not fully represent the general population.