Redliking: When Redlining Goes Online

Abstract

Airbnb’s structure, design, and algorithm create a website architecture that allows user discrimination to prevent minority hosts from realizing the same economic benefits from short-term rental platforms as white hosts, a phenomenon this Article refers to as ‘redliking.’ For hosts with an extra couch, spare room, or unused home, Airbnb provides an opportunity to create new income streams and increase wealth. Airbnb encourages prospective guests to view host photographs and personal information when considering potential accommodations, thereby inviting bias, both implicit and overt, to permeate transactions. This bias has financial consequences. Empirical research on host earning rates found that white hosts earn significantly more than their minority counterparts, even when controlling for location, size, and amenities. Airbnb’s algorithm augments the effects and propensity of individual user bias, creating a system wherein race neutral variables serve as proxies for discrimination. Contemporary redliking parallels historic inequality related to housing wealth. In the early twentieth century, redlining maps were used to justify withholding investments from Black communities. Today, redliking continues the practice of directing wealth to white communities, reinforces systemic real property barriers by depriving minority hosts of important revenue streams, and exacerbates the racial wealth gap. This Article examines the liability of websites like Airbnb for discrimination experienced by minority short-term rental hosts. The ability of anti-discrimination laws originally enacted to abolish redlining and protect minority consumers to combat redliking is complicated by the fact that sites like Airbnb serve multiple purposes; while guests use the platform to identify and book lodging, hosts use the site to advertise available accommodations. Looking to judicial interpretation of online speech and platform operator liability, this Article proposes two approaches – a general function test and a fragmented function test – to determine website liability for discrimination against short-term rental hosts. Noting the limitations of the existing anti-discrimination legal framework, this Article argues that eradicating redliking requires incorporating lessons on platform design from behavioral economics as well as eliminating opportunities for website algorithms to amplify and operationalize user discrimination.