Low uptake of an optional pricing algorithm on Airbnb by black hosts had the effect of widening racial disparities in earnings on the platform, a new study has found.
The optional “Smart Pricing” feature, which dynamically adjusts the cost of a night’s stay based on current demand and allows hosts to set a minimum price, was launched in 2015 to assist users in securing more bookings.
In instances where hosts chose to opt in to using the feature, the study of more than 9,000 Airbnb hosts by researchers at Carnegie Mellon University found that the algorithm decreased racial disparities in earnings, caused by lower demand for black-owned homes, by 71 per cent.
However, the relative hesitancy of black hosts to use the tool meant the overall effect was to widen the discrepancy.
“The most surprising finding was that even though the algorithm was benefiting both black and white [hosts] more, it led to greater social inequality in the whole population, because of significantly [lower] adoption rates amongst black hosts,” said Param Vir Singh, the paper’s author and a professor at Carnegie Mellon University.
“So even a well-meaning algorithm can lead to greater social inequality . . . there are unintended consequences of algorithms,” he added.
The findings have prompted renewed calls for independent regulators and academics to be given access to internal Airbnb data in order to independently assess racial discrepancies on the platform.
Color of Change, a US non-profit civil rights advocacy group, said: “Discrimination analysis shouldn’t be left to tech companies self-regulating . . . In order to truly root out racial bias within tech companies’ platforms and products, regulators . . . must have access to company data that analyses the experiences of black users.”
A 2016 Harvard study found that guests with distinctively black names were “16 per cent less likely to be accepted relative to identical guests with distinctively white names”.
Singh and his team randomly selected 9,396 Airbnb properties in 324 zip codes, mostly within seven large US cities, to analyse their earnings. During the period of the study — from July 2015 to August 2017 — 2,118 properties made use of the Smart Pricing feature for varying periods of time.
The study observed that before the introduction of Smart Pricing, white hosts on Airbnb made on average $12.16 more than black hosts, when controlling for other factors, such as the desirability of the location.
Among those who adopted the Smart Pricing algorithm, the revenue disparity was reduced to $3.46, researchers said, since the algorithm corrected for the greater price-sensitivity of black-owned homes to fluctuations in demand.
However, adoption among black hosts was 41 per cent lower than for white hosts, which meant that overall, the revenue gap between the black and white hosts actually increased to $14.30 after introduction of the algorithm.
Singh said: “Airbnb could try to get [black hosts] to use the algorithm and cover any losses for a few weeks or months, to give them confidence that the algorithm is actually valuable to them. It can increase adoption and benefit both Airbnb and these hosts.”
In response to the study, Airbnb told the Financial Times that it had seen a 50 per cent increase in hosts opting to switch on the algorithm over the past three years. “We developed our Smart Pricing tool to help ensure every Airbnb host has the tools needed to capture the economic benefits of home sharing and to address the very disparities this study examines,” it said.
The reason for black hosts shunning the use of opaque algorithms could be “because of racism, discrimination, or history”, said Stefan Grant, founder of Noirbnb, an alternative to Airbnb aimed primarily at the African-American community.
“We’re not always the earliest adopters,” Grant said. “This Smart Pricing issue with black hosts is just a microcosm of many bigger racial issues. We have to figure out unique ways to navigate — because the intended target for whatever the product may be, isn’t always us.”
This article has been amended to clarify that the increased racial disparity was caused by differing adoption rates rather than by the pricing algorithm itself