A prominent economist has highlighted the critical need for congestion pricing to manage the potential surge in traffic induced by widespread adoption of self-driving cars. Arpit Gupta, a researcher whose work often touches on technological impacts on infrastructure, asserted on social media that such measures will be essential to prevent gridlock.
"We will need congestion pricing to deal with self driving cars inducing way more driving," Gupta stated in a recent tweet. This sentiment echoes findings from a National Bureau of Economic Research (NBER) working paper, which posits that without economic incentives, technological advances like autonomous vehicles could lead to zero improvement in social welfare due to increased traffic.
The NBER paper, authored by Mostrovs and Michael Schwarz, emphasizes that "Increased traffic congestion can destroy all efficiency gains from self-driving cars." It suggests that carpooling and congestion pricing are "highly complementary" and more effective when combined than either measure alone. The researchers argue these strategies can improve commuter welfare without relying on government revenue redistribution.
Experts widely acknowledge that autonomous vehicles could significantly alter urban mobility. While some foresee optimized traffic flow and reduced congestion, others, like Gupta and the NBER authors, warn of "induced demand," where the convenience and lower perceived cost of autonomous travel encourage more people to drive, potentially exacerbating traffic. This could include empty "robotaxis" cruising between fares or individuals opting for longer trips.
Cities worldwide are grappling with existing traffic challenges, with some, like London and New York City, already implementing congestion pricing schemes. These systems charge vehicles for entering designated zones during peak hours, aiming to reduce traffic volume and encourage public transport or carpooling. The debate surrounding self-driving cars and their impact on urban infrastructure underscores the growing importance of such policy tools.