How will the introduction and diffusion of autonomous vehicles (AVs) affect U.S. workers? This highly fraught question promises soon to loom large in hometowns and policy realms across the nation. Given Americans’ current reliance on cars and trucks for most of our transportation, the transition to self-driving vehicles will change many lives and livelihoods, likely for the better for the vast majority. But it will be costly for some. This study advances the national conversation about how to cope with the effect of AVs on workers in three ways: by setting forth a framework for discussion, presenting quantitative simulations and qualitative scenarios to help assess key impacts, and providing policy recommendations for mitigating negative impacts while also setting an agenda for research on policy. We hope that our report will help motivate policymakers and stakeholders to take steps now to reduce the likely negative effects of AV on U.S. workers.
To preview our results, we find that the introduction of autonomous cars and trucks could directly eliminate 1.3 to 2.3 million workers’ jobs over the next thirty years, depending on the adoption scenario followed. While near-term effects are limited, the maximum impact we simulate (which occurs during the 2040s) could raise the overall annual unemployment rate by about 0.1 percentage points and lower labor force participation by about 0.1 percentage points for a number of years, with stronger effects in hard-hit communities or during a recession.
Using evidence from previous dislocations we find that each laid-off worker would likely lose on average about $80,000 in lifetime income due to the disruption, for a total loss of about $180 billion for U.S. workers. These adjustments take into account the probable age of workers being dislocated; on average these workers would have about 16 years of labor force participation left in their careers.
Most of the affected workers will eventually find new jobs or retire. However, this process will take time and may lead to wage increases or decreases. We do not provide specific estimates of job creation associated with AVs; instead we discuss the three general sources of these new jobs: growth in overall transportation, new labor inputs for the AV sector, and increased purchases of other goods and services by consumers who spend less on transportation.
In addition, when driving is no longer a requirement, the duties in many other jobs (such as many home health aides, building contractors, visiting nurses, real estate agents and other sales people, regional supervisors, automobile and vehicle insurance workers, and taxi dispatchers) will change substantially. In total, the jobs whose duties are very likely to change with the adoption of AV employed 7.7 million people in 2016. The change in duties could be associated with better or worse jobs from the perspective of pay or skills; we explore a range of possibilities.
Complacency, fatalism, or ignoring these serious consequences would be a mistake. There are numerous policy options for assisting workers in the affected occupations. As a country, we need to plan now so that the promise of AVs (cheaper and more efficient transportation, dramatic reduction in deaths and injuries from accidents, greater mobility for those who can’t drive, freedom from tedium for those who can) either does not impose huge costs on those directly affected or compensates them for their loss. The economic benefits of AVs, which some have
estimated at $800 billion to $1 trillion per year should provide adequate resources for such policy intervention.1 With advance planning, the task is manageable; according to our scenarios, employment disruptions won’t start in large numbers until after 2030, and will be gradual (about 100,000 jobs disrupted per year, or 0.1 percent of the work force, at the time of peak impact).
Our estimates, like all estimates, have limitations. Our simulations for displacement and unemployment include adjustments for turnover (workers leaving occupations for reasons other than the rise of autonomous vehicles). However, we do not include multiplier effects (the fact that when people lose jobs, their spending drops, leading to additional temporary unemployment from reduced sales from those jobless workers), a clear source of underestimation. On the other hand, if the pace of AV diffusion is slower than our scenarios anticipate (e.g., due to delays in achieving fully autonomous capabilities, consumer hesitance about adoption, regulatory constraints, partial (vs. full) automation that preserves a role for a human driver) or if employers retain and retrain many workers in the eliminated jobs, we may have overestimated.
Chosen excerpts by Job Market Monitor. Read the whole story at “Preparing U.S. Workers and Employers for an Autonomous Vehicle Future” by Erica L. Groshen, Susan Helper et al.