Millions of jobs at risk from autonomous vehicles, yet solutions slow to match pace
3 million. The number oft-quoted by Fareed Zakaria, David Axelrod, and a host of others as to the numbers of jobs likely to be lost upon the deployment of self-driving vehicles. A loss of jobs of this magnitude would equate to approximately 1% of the US population — and 2% of the overall labor force. The economic ripples of such a shift, if it happens as quickly and unchecked as analysts warn, could match or exceed the severity of the depletion of jobs in manufacturing of the past few decades.
The oft-suggested remedy: a Jobs Bill. A program to overhaul training and retraining across the country to provide new jobs that could fit into the changing economy. Yet discussion of such a program has largely been limited to the realms of journalism and analysis most concerned about the problem. Discussion on a Congressional or Administrative level has been virtually nonexistent.
However, there are still many questions left unanswered — about both the problem and the solution.
- What is a real timeline for widespread usage of self-driving cars?
Full autonomous vehicles (considered to be Level 4 or above by the SAE) would enable potential replacement of drivers. The earliest Level 4 vehicle scheduled to be available for widespread release is Tesla in 2018. Companies like Audi, Ford, BMW, Nissan, and Volvo all have projected Level 4 roll outs in the early 2020’s. Of course, it remains to be seen how successful these initial efforts will be with the public.
Additionally, automation for larger vehicles such as trucks and buses may not arrive simultaneously with these roll outs. The startup Otto (acquired in August 2016 by Uber) is looking to retrofit existing large vehicles for autonomous driving for an estimated cost of $30,000 per vehicle. They currently plan on launching a fleet of autonomous trucks in 2017.
- Would 3 million jobs really be lost?
Let us first consider how many vehicle-driving jobs (in millions) currently exist in the US today.
This graph does not include the number of ride-sharing (Uber, Lyft, etc.) jobs that currently exist in the US — a number estimated to be over 500,000.
It is difficult to say what the rate or extent of job loss could be. Representatives of trucking organizations have said in the past that a human being would always be necessary to monitor the vehicle in case of a number of emergencies that an on-board computer might not be able to handle. They are likely right in some respect, but it is unlikely that these truckers of the future would be paid as much as the trucker of today (averaging a $40,000 salary according to the Bureau of Labor Statistics).
Furthermore, fewer truckers might be needed even if some oversight is required. Drivers might be able to oversee multiple autonomous trucks in their caravan. Or truckers might be tasked to monitor a given area for emergencies and repairs.
The story could very well be similar in the cases of buses, delivery trucks, and warehouse vehicles.
The Bureau of Labor Statistics also estimated (in 2014) that the year-over-year increase in most of these job sectors would be between 3–4%.
- What would a jobs program look like? What would be needed to make it successful? Are there other options?
There are many questions of what such a government jobs program would exactly do — what jobs would it focus on retraining for? What if recipients are reluctant or unwilling to undergo retraining? From what we have already seen, large numbers of individuals can completely drop out of the labor force in record numbers in the wake of the Great Recession.
It has been often suggested that retraining programs should focus on tech and specifically programming roles. An important distinction should be made here though, as the much of the tech/programming training that is currently offered in both Universities and coding bootcamps focus on the skills needed to become a software engineer at large Silicon Valley tech company (C, Java, Data Structures, Algorithms, etc.). Would individuals in a retraining program find these particular skills helpful? Potentially, but more targeted training programs that could focus on roles within the industries of their communities might also be valuable — learning the DSL used by the local small tech company, the CRM that manages the information of the consulting agency in town, electronics for hardware/autoshops, etc.
A balance between this approach and more broad-based, traditional tech learning might be most optimal given the pace at which software is evolving in these communities.
This does ignore the fundamental question, would individuals want to undergo retraining? A striking assessment from J.D. Vance’s book, Hillbilly Elegy: A Family and Culture in Crisis, suggests that communities heavily effected by the disappearance of industry find it difficult to adapt to change.
On the one hand, this is particularly of concern considering that the current average age of a truck driver is 55 (Bureau of Labor Statistics). On the other hand, participants (in at least the trucking industry) might largely arrive upon retirement age as autonomous trucks become commonplace. While this would be a serendipitous occurrence, it is far from likely… at a moment in the US when retirement ages have been on a steady increase across industries.
Other options to mediate the implosion of jobs in industries being overturned by technological advancement often include UBI (Universal Basic Income). A number of current programs, including one by Y-Combinator, are currently conducting research into what an effective UBI implementation might look like. However, determining the long term effects of such programs will likely take years as well.
Hopefully research and action will be hastened, these jobs are about to start running on fumes.
Updated Feb 28: Chart describing numbers of driving jobs updated