The development and testing of automated driving technology, particularly for long-haul trucking, is rapidly progressing in the United States. A common deployment model being explored is the "transfer-hub" model, where automated trucks handle highway driving, while human drivers manage the more complex urban segments. This raises significant concerns about potential job displacement in the long-haul trucking sector, a prevalent profession in the US, particularly for men with high-school educations. Contrasting claims exist: while some predict significant job losses, others, often technology developers, argue that new short-haul jobs will offset this. This study aims to clarify the impact of automated trucking on the long-haul trucking labor market by analyzing data and interviewing stakeholders. We leverage the 2017 Commodity Flow Survey (CFS) to estimate operator-hours for different routes, considering technological constraints and various deployment scenarios. This quantitative analysis is complemented by semi-structured interviews with industry professionals, including long-haul truck operators, to understand the feasibility and potential challenges of the transfer-hub model. This mixed-methods approach allows for a more nuanced understanding of the potential impacts of automated trucking, addressing the gaps in existing literature.

Existing literature on the job impacts of automated trucking is limited. Studies like Viscelli (2018) and Gittleman and Monaco (2020) provide estimates of job losses, ranging from several hundred thousand to 400,000, largely focusing on the impact on long-haul highway driving. Groshen et al. (2019) and Waschik et al. (2021) use simulations to project job losses and broader macroeconomic effects. However, these studies often lack detailed analysis of trucking routes, specific technological limitations (e.g., weather restrictions), and direct input from truck drivers. This study aims to address these gaps by utilizing trucking route and shipment data and incorporating perspectives from industry stakeholders, including long-haul operators.

This study uses a mixed-methods approach, combining quantitative analysis of the 2017 Commodity Flow Survey (CFS) with qualitative data from semi-structured interviews. The CFS data provide information on trucking shipments, including origin and destination, distance, weight, and mode of transportation. The researchers used the Google Maps API to estimate highway and urban segments for each shipment, categorizing them into intrastate and interstate journeys, with varying methods based on the availability of Metropolitan Statistical Area (MSA) data. Algorithms were developed to calculate operator-hours for highway and urban segments, considering the 11-hour daily driving limit imposed by Hours of Service (HOS) regulations. The quantitative analysis considers several deployment scenarios: restricted deployment to sunny southern states, deployment across all states during favorable weather months, deployment for journeys over 500 miles, and widespread deployment. Semi-structured interviews were conducted with stakeholders including automated trucking startup representatives, truck drivers, logistics operators, and labor union representatives, using purposeful and snowball sampling methods. The interviews aimed to assess the feasibility of the transfer-hub model and gather insights into the perspectives of long-haul truck drivers.

The analysis reveals that the impact of automated trucking on operator-hours varies significantly depending on the deployment scenario. Widespread deployment across the continental US in all conditions could automate up to 94% of operator-hours. However, restricting deployment to currently active testing regions (primarily southern states) limits the impact to only 10%. Deployment limited to favorable weather months (Q2 and Q3) could impact over half of operator-hours. Even if automation is initially limited to journeys over 500 miles, approximately 43% of operator-hours would be impacted. The study also finds limited evidence supporting claims that increased demand for trucking services due to automation-driven cost and time reductions will compensate for the job losses. Even a hypothetical 50% increase in demand would only offset 5% of the at-risk operator-hours. Interviews revealed that long-haul truck drivers generally see few operational barriers to highway automation, but emphasize weather conditions, lack of lane markings, and potential GPS signal loss as challenges. Drivers largely do not perform significant maintenance, suggesting that automated trucks would need reliable self-diagnostic and remote support systems. A significant portion of the interviews also highlighted a prevailing negative sentiment towards partial automation systems among drivers.

The findings challenge the common narrative that automation will seamlessly create new jobs to offset job losses in long-haul trucking. The significant potential for automation of operator hours, even under limited deployment scenarios, highlights the need for proactive policy interventions. The limited potential for short-haul jobs to compensate for the loss of long-haul positions also adds urgency to this need. The qualitative findings reinforce the quantitative analysis, revealing potential challenges related to weather conditions, infrastructure limitations, and driver perspectives on partial automation systems. This study emphasizes the importance of considering a wide range of deployment scenarios and the human element when assessing the impacts of automation on employment.

This study provides a comprehensive analysis of the potential impacts of automated long-haul trucking on operator-hours in the US, incorporating both quantitative and qualitative data. The findings demonstrate that the extent of job displacement will strongly depend on the pace and scope of automation technology development and deployment. While automation offers economic benefits, it poses substantial challenges for the long-haul trucking workforce. Policymakers should focus on strategies for workforce transition and mitigation of potential negative impacts on employment and the affected communities. Further research should explore the effectiveness of different transition policies and the long-term implications of automation on the broader transportation and logistics sector. The cultural significance of long haul trucking jobs and the potential displacement of workers from rural areas also warrants more attention in future work.

The study's sample size for interviews was limited, potentially affecting the generalizability of the qualitative findings. The accuracy of the operator-hours calculations depends on the assumptions made about routing, average speeds, and HOS regulations. Furthermore, the study focuses primarily on operator-hours and does not directly quantify job losses due to data limitations on the precise number of long-haul truck drivers. Future research could address these limitations by expanding the interview sample, refining the operator-hour calculation methods, and using alternative data sources to estimate job losses more precisely.