Uber’s impact on Taxi Crime in Chicago
[h2 class=”” span=””]Introduction[/h2]
Unfortunately, taxis have been the venue for thousands of crimes committed in Chicago in the last decade. As of 4/2/2014, 5,295 crime incidents had occurred in a taxicab since 2001, i.e. more than a crime a day during that time.1 Furthermore, OSHA recently reported that taxi drivers are more likely to be victims of a crime than professionals in any other job.2
While there are many explanations for why so many crimes occur in taxicabs, one of the reasons is that neither the drivers nor the passengers in taxis have identities or transparent feedback systems like Uber’s rating and account framework. When a driver on Uber receives a request from a passenger, he/she knows who the pickup is for, what the passenger’s rating is, and that the passenger has provided a valid credit card to the system.
A second, related explanation concerns electronic payments. A recent study by Richard Wright, Erdal Tekin, Volkan Topalli, Chandler McClellan, Timothy Dickinson and Richard Rosenfeld provides suggestive evidence that a move towards an electronic economy and away from a cash economy reduces crimes like robbery and assault.3 Uber offers a transportation service that depends entirely on a cashless, electronic payment system. Therefore, we might expect that, as riders and drivers have joined the Uber network, crimes in taxicabs have been reduced.
In order to study the impact of Uber’s network on crime in taxis, we assembled a dataset of taxi crimes and used a simple econometric model to estimate the impact of Uber on crime in taxicabs. The results are suggestive and potentially large – the entry of Uber in Chicago appears to have significantly reduced the crime rate in taxicabs which was otherwise quite stable. Compared to the 300 days before Uber entered, the rate of taxicab-located crimes decreased by 20% in the 300 days after Uber entered. This result holds true for other windows of time, and it does not simply represent an overall downward trend or seasonality in taxicab crime for which Uber might be a proxy. Furthermore, there is no corresponding bump in crimes committed in non-taxi automobiles to suggest that the crime is simply moving to other kinds of cars.
Future studies need to refine to further explore this result, but, combined with the other research cited above, the crime-preventing benefit of technology cannot be discounted.
[h2 class=”” span=””]Details[/h2]
Uber used the city of Chicago’s crime database to assemble a dataset of crimes that occurred in taxicabs in a 300-day window around when Uber entered the Chicago market on 9/11/2011. Before 9/11/2011, the number of taxi crimes per day was roughly constant – a simple correlation with an intercept between the date and the number of crimes shows no statistically significant relationship between the date and the number of crimes.
What happened when Uber entered the Chicago market? In order to estimate the impact Uber has had on taxi crimes, we estimate a basic regression of the form
daily taxi crimes = a + b * I(Uber has entered),
where I() is an indicator function or dummy variable. This regression indicates that the entrance of Uber is associated with .2 fewer taxicab crimes per day, significant at the 95% level:
You might wonder if there is simply a trend that we are capturing. Including a daily trend:
daily taxi crimes = a + b * I(Uber has entered) + c * day_number
This does not change the estimate by very much (it’s a little smaller, and the p-value is higher):
Similarly, one might be concerned that the effect is seasonal, i.e. that Uber entered at a time when taxicab crime typically peaks. If we run the model on all the historical data Chicago offers (back to 2001), adding dummies for month of the year and year does not change the estimate.
This estimate is also consistent as you vary the size of the data window. The coefficient varies only minimally as the window ranges from 30 days before and after Uber to the baseline 300 days around when Uber entered.
Finally, as a robustness check against the possibility that Uber’s entrance is merely correlated with some other change in Chicago that reduced crime, we applied the same basic model to crimes occurring in alleys and sidewalks. We have no reason to suspect that Uber’s entrance in the market would have an impact on these crimes, and indeed, we do not see a statistically significant and robust effect on these types of crimes.
As a data-driven company, we understand the power of information for communities. Since day one, our mission has been to connect people with reliable rides through the use of data and technology. As our footprint has grown throughout the years, so has our ability to use the Uber network in different ways. Starting today and […]
Today, we are excited to announce that Uber will give $5.5M to support a new robotics faculty chair as well as three fellowships at CMU. This gift is part of a partnership we announced earlier this year. In addition, we’re pumped to be part of a growing innovation ecosystem in Pittsburgh that includes world leading research institutions and companies, as well as an increasing number of start-ups.