Inquiry Into the Potential for Predatory Pricing Strategies by Rideshare Companies In Chicago
Rideshare apps like Uber and Lyft have become a relatively ubiquitous part of our smartphone-enabled world in the United States. It goes without saying that these apps have fundamentally altered human mobility, and made a tangible impact on our society. Sentiments towards these rideshare apps vary greatly depending on who you ask, and what types of questions you pose. Some, for example, will laud the Uber and Lyft platforms as fantastic innovations that not only have been linked to reduced drunk driving fatalities, (1) but also enable a more efficient healthcare system and promote better health outcomes by improving access to care for patients who otherwise struggle to get to their healthcare appointments. (See Cerner’s partnership with Uber, and the existence of UberHealth for some preliminary examples of this). Conversely, many see these rideshare apps as a nuisance that have been linked to decreased utilization of public transportation (and thus, increased greenhouse gas emissions) and increased traffic in urban areas (2). Urbanists have just barely scratched the surface in our understanding of the short and long term implications of rideshare apps; however, it’s important that we continue to be critical and inquisitive of these platforms.
Although I have experienced the convenience of these apps and understand the benefits they provide, I also have developed a cynical attitude towards large corporations — especially technology companies headquartered in the Bay Area. Furthermore, Uber and Lyft’s involvement in attempting to pass Propoposition 22 in California which, had it not been ruled unconstitutional, would have stripped their workers of all employee protections, supports my concern that these companies are more interested in profits than people. My personal bias is relevant, as it has prompted me to question the pricing practices of these rideshare apps. Broadly, I’m interested in seeing if Uber and Lyft deploy any unfair pricing strategies.
When taking a preliminary look at the Chicago Transportation Network Provider (rideshare) data , I was particularly interested in the “Additional Charges” column. I was unable to find any explanation that detailed what these additional charges were associated with, which piqued my curiosity (and my skepticism) and frankly, I assumed the worst: that these additional fees were a way to capitalize on individuals who relied on rideshare services the most (eg. people who can’t afford a car, individuals living in less affluent areas, etc.). Upon reflection, this may be a bit of a leap in logic but, nonetheless, it did inform my analysis. I hypothesized that rideshare apps had more expensive “additional charges” in community areas (neighborhoods) with worse socioeconomic conditions.
As a point of clarification, I operationalized socioeconomic conditions by using the dataset Selected Socioeconomic Indicators in Chicago from 2008–2012. The Chicago Department of Public Health used six indicators of significance (crowded housing, percentage of households below poverty level, unemployment percentages, education rates, dependency rates, and per capita income) to develop a “hardship index” for each community area in Chicago. Below I discuss the results from my analysis.
Preliminary Data Findings
When narrowing the data to rideshare trips in January 2020, there were a total of 7,923,063 trips recorded. These rides, on average, lasted ~966 seconds (~16.1minutes) and were ~5.80 miles. The average “additional charges” per ride is 3.562
Addressing the Hypothesis
In order to address my hypothesis and investigate the differences in additional charges applied to trips originating at different community areas, I grouped the rideshare data in a way that made it easier to investigate differences between the neighborhoods. The first bar graph below shows the average additional charges applied to rides that started in a given community zone. The second bar graph shows the hardship index assigned to each community area..
Generally, we see that there is not an egregious variation in additional charges by community area — the additional charges for almost every community area, with the exception of two neighborhoods, range between $2 and $4. Two community areas, 56 and 76, stick out as having significantly elevated additional charges associated with them. Community area 56 corresponds with Garfield Ridge, which is where Midway Airport is located. Community area 76 is where O’Hare is located, which likely explains the additional charges. From these bar graphs, it doesn’t appear that additional charges are weighted towards less-privileged community areas. However, let’s visualize this in another way.
Visualizing the same data mapped to the community areas, we’re able to see that of the community areas with the highest additional charges (the areas in yellow on the upper map), there is only one community area (#34) that has a hardship index in the highest quintile (the areas in yellow on the lower map).
Lastly, to confirm that I was not missing some sort of pattern or correlation, I ran a bivariate regression, to better understand the strength of the relationship between the hardship index and the average additional charges attached to each ride.
The results from the regression actually indicate a negative linear relationship between additional charges and hardship index, although it’s worth noting the somewhat low adjusted R-squared, meaning there is a very weak relationship between the model and “additional charges”. While we can’t assign too much weight to this regression, because of its weak adjusted R-squared, our findings surely do not support the research hypothesis that rides originating from community areas with higher hardship indexes were subject to more additional charge.
From the overall analysis, it doesn’t appear that rideshare apps are charging higher additional fares in neighborhoods with worse socioeconomic conditions, which is a positive! Although, it doesn’t guarantee that these apps don’t participate in predatory pricing practices. Further analysis that looks at differences in fares (and fee per second information) would be a great way to expand upon this research, to see if there are in fact any unfair pricing strategies to be uncovered.
To close, I feel it’s important to acknowledge some shortcomings in the datasets used for this analysis. I used hardship index data that was developed based on data from 2008–2012, which is pretty far away from the rideship year, and there could’ve been significant changes in a neighborhood’s socioeconomic status during that time frame. Additionally, the hardship index data is based on estimations and does not represent the entirety of a neighborhood’s condition, which is worth acknowledging. Lastly, the transportation network data has a good amount of estimating and rounding to maintain privacy, which could potentially skew any findings from this analysis.
(1) Wang, Xuan, Hassan Marzoughi Ardakani, and Helmut Schneider. “Does Ride Sharing Have Social Benefits? — Core.ac.uk.” Twenty-third Americas Conference on Information Systems, Boston, 2017. Accessed November 23, 2021. https://core.ac.uk/download/pdf/301371727.pdf.
(2) Ovide, Shira. “What We Got Wrong about Uber and Lyft.” The New York Times. The New York Times, March 29, 2021. https://www.nytimes.com/2021/03/29/technology/what-we-got-wrong-about-uber-and-lyft.html.