Understanding Inequalities in Ride-Hailing Services Through Simulations

Eszter Bokányi,Anikó Hannák

doi:10.1038/s41598-020-63171-9

Abstract

Despite the potential of ride-hailing services to democratize the labor market, they are often accused of fostering unfair working conditions and low wages. This paper investigates the effect of algorithm design decisions on wage inequality in ride-hailing platforms. We create a simplified city environment where taxis serve passengers to emulate a working week in a worker’s life. Our simulation approach overcomes the difficulties stemming from both the complexity of transportation systems and the lack of data and algorithmic transparency. We calibrate the model based on empirical data, including conditions about locations of drivers and passengers, traffic, the layout of the city, and the algorithm that matches requests with drivers. Our results show that small changes in the system parameters can cause large deviations in the income distributions of drivers, leading to an unpredictable system that often distributes vastly different incomes to identically performing drivers. As suggested by recent studies about feedback loops in algorithmic systems, these short-term income differences may result in enforced and long-term wage gaps.

Highlights

As they grow in popularity, ride-hailing and food-delivery services such as Uber, Lyft, Ola or Foodora are quickly transforming urban transportation ecosystems[1,2]
As documented in case studies[9,10,11,12,13], workers are struggling to obtain remedies through official channels[12,14,15], and strikes have become common in the past years with drivers of Uber, Lyft, Ola, Foodora demanding higher fares, job security, and livable incomes all over the world[16]
Most existing literature in the area of taxi matching algorithms is concerned with optimizing aggregate outcomes for the whole system[25,26,27,28,29,30,31]. Such approaches aim to maximize the benefits for the company or to minimize the adverse effects such as CO2 emissions, overall distances driven, or the passenger waiting times

Summary

Introduction

As they grow in popularity, ride-hailing and food-delivery services such as Uber, Lyft, Ola or Foodora are quickly transforming urban transportation ecosystems[1,2]. As described by Rosenblat and Stark in their article “Case Study of Uber Drivers”, Uber-like modern systems subject drivers to algorithmic management methods coupled with an almost compulsory blind acceptance of destinations This results in a hierarchical information flow, in which the company decides the content and the means of disclosing information to the drivers[12]. Most existing literature in the area of taxi matching algorithms is concerned with optimizing aggregate outcomes for the whole system[25,26,27,28,29,30,31] Such approaches aim to maximize the benefits for the company or to minimize the adverse effects such as CO2 emissions, overall distances driven, or the passenger waiting times. Following the line of fairness measurement literature[32,33,34,35], we instead focus on the fair distribution of income from the drivers perspective, because current systems do not guarantee the same income for the same amount of work, neither across workers nor over time[9,10,19,21]

Objectives

Methods

Findings

Conclusion