Research Projects

Stop the Video

Research Projects

METRANS
STATUS: In Progress YEAR: 2019 TOPIC AREA: Sustainability, energy, and health Transportation planning, policy, and finance CENTER: PSR

Using big data to estimate the environmental benefits of congestion pricing in the Los Angeles metropolitan area

Project Summary

Funding Source: Caltrans
Funding Amount: $99,604
Start and End Dates: April 1, 2020 to March 31, 2021

 

Project Description

Los Angeles is now one of the global leaders in urban traffic congestion. On average, Angelinos spend 104 hours stuck in traffic each year. For a typical worker, this is equivalent to a total loss of 13 working days in a year. And, in total, the estimates of the social cost of traffic congestion in Los Angeles add up to $9.7 billion dollars per year, or $2,408 per driver. These extraordinarily high levels of congestions, and de-facto excessive number of passenger vehicles and heavy-duty trucks on the road at different locations and times, also translate in both local air pollution and greenhouse gas (GHG) emissions concerns. In California, despite incentives for the adoption of cleaner vehicles and increased penetration of electric vehicles in the fleet, GHG emissions from transportation continue to increase. And when it comes to local air pollution, while tough regulations have certainly brought dramatic reductions in air pollution and improved health, Southern California remains the nation's smoggiest region, and continues to fail to meet federal ozone standards. Regulators claim that cleaning the air to federal standards will require a massive transformation of California's transportation sector, but, to date, these proposals have focused primary on technology and ignored increases in vehicle miles traveled. In contrast, economists and some policymakers, argue that policies that promote cleaner technological adoptions need to be matched with pricing policies that control vehicle miles traveled, and encourage drivers to find alternative ways to commute. In turn, for these pricing strategies to be effective, investments in alternative modes of transportation are needed. With increased discussions that may eventually result in the introduction of congestion pricing in Los Angeles and San Francisco, there is still a unique opportunity to inform the design of these programs, so as to also maximize their potential in reducing GHG emissions and local pollution. Suppose indeed Los Angeles and San Francisco move forward with congestion pricing, and remove vehicles and heavy-duty trucks from freeways (and possibly other roads) so that traffic congestion is alleviated. Given typical origin/destination pairs and routes that drivers follow on their journey to work at peak periods, what would be the additional benefits in terms of GHG emissions and local air pollution from removing excessive vehicles from freeways and roads at different locations and times of the day? 


The purpose of this project is to provide a credible empirical answer to this question. Specifically, we will put together the most comprehensive ‘big-data' to study this question. These data include a rich network of detectors located on freeways in Los Angeles and San Francisco that measure speed and flow in real-time, and novel and unexploited data from Aclima that measures in real-time the concentrations of various local air pollutants (including, NO, NO2, Ozone, Black Carbon), by relying on Google cars that drive repeatedly across different locations and time periods in these cities. We then apply standard econometric techniques, as well as recent data visualization and machine learning methods, to develop practical tools for policymakers to infer the relationship between traffic congestion and pollution, and identify hot-spots and targets for policy interventions that maximize both reductions in congestion and pollution. Others have developed simulation models (with require structural assumptions, often unknown) to be able to examine this question. In contrast, our proposed framework offers several advantages: First, because of the ‘big-data', instead of forcing structural assumptions, we rely on modern empirical methods from econometrics and machine learning, to actually learn about the relationships between traffic volumes (and traffic composition) and air pollution. As a result, our approach is substantially more flexible and likely less vulnerable to potential unintentional ‘mistakes' that result from structural assumptions made by the researchers in simulation models. Second, our proposed framework allows for visual display of the relationship between traffic volumes and pollution across different locations and times. The advantage is that it facilitates the development of a tool that policymakers can use to prioritize areas of intervention. In particular, it allows for visual illustrations of the congestion and pollution reductions that result from alternative ways of designing congestion pricing programs. Third, to the extent that pollution patterns do not follow exactly congestion, which atmospheric scientists attribute to thermal inversions, our framework allows for the measurement of the effects of past hours traffic on pollution in the following hours, as well as the effects of traffic on accumulated pollution over several hours of the day. The advantage is that our work will be able to highlight how congestion pricing must be adjusted to maximize pollution reductions.

P.I. NAME & ADDRESS

Antonio Bento
Director, Global Master of Public Policy and Director, USC-Price Environmental Initiative, Price School of Public Policy
Ralph and Goldy Lewis Hall 214
Los Angeles, CA 90089
United States
[email protected]