by Andrew Yang
I don’t think Jeff Bezos has it out for local malls, per se. But there will nonetheless be hundreds of thousands of people who suffer from the demise of retail, driven by e-commerce giants like Amazon: the mall workers, the people who liked shopping at the mall, the county workers who needed the mall’s property tax to pay for their jobs, the property owners near the mall, and so on. Hundreds of our communities are going to have giant holes blasted in them by progress that will disrupt thousands of lives and livelihoods in each. And the victims are likely to be among the weakest in the labor market—retail workers are paid less than workers in most other industries and typically lack a college degree. Where are they going to go?
It’s not just the malls—little shops and restaurants everywhere are closing. You’re probably seeing empty storefronts around where you live and work right now.
A New York Times op-ed by the economic historian Louis Hyman detailed the plight of towns in upstate New York and other places that have seen their retail sectors decline and offered some recommendations for how workers could readjust to the new economic reality:
Main Street… exists, but only as a luxury consumer experience… If the answer to rural downward mobility is to turn everyone into software engineers, there is no hope… Today, for the first time, thanks to the internet, small-town America can pull back money from Wall Street (and big cities more generally). Through global freelancing platforms like Upwork, for example, rural and small-town Americans can find jobs anywhere in the world, using abilities and talents they already have. A receptionist can welcome office visitors in San Francisco from her home in New York’s Finger Lakes. Through an e-commerce website like Etsy, an Appalachian woodworker can create custom pieces and sell them anywhere in the world.
This op-ed is a great summary of the general constructive thinking. It recognizes that the retail sector will shrink and happily debunks the ridiculous “let’s turn everyone into coders” idea, which is realistic for only a tiny proportion of displaced workers. If you dig into the author’s alternative suggestions for workers though, they’re equally unrealistic and could only be offered by someone who hadn’t tried any of them. Upwork primarily finds work for developers, designers, and creatives on a global scale. Asking a retail worker from small-town America to log on and get work assumes they have a skill to offer. These global platforms have people offering their services from abroad, who can price their time at as little as $4 an hour even for a college graduate. Getting work on these platforms is highly competitive, doesn’t pay very well, and carries no benefits.
Offices in San Francisco have either iPads or human beings as receptionists. They don’t have avatars staffed by human beings in small towns hundreds of miles away. Selling woodwork on Etsy is the kind of thing that would work for a handful of humans and is unlikely to feed your family. On average, sellers’ income from Etsy contributes only 13 percent to their household income and is intended as a supplement to traditional work. Forty-one percent of Etsy sellers who focus on their business full-time get their health care through a spouse or partner, and 39 percent are on Medicare or Medicaid or another state-sponsored program.
It’s possible that some workers in towns with dying retail stores could find menial jobs on their computers as telemarketers, phone sex operators, English tutors to Chinese kids, or image classifiers to help train AI. That’s not exactly an appealing future though—and long-distance low-skilled jobs are the ones most subject to automation and a race to the lowest-cost provider. Most retail workers at least had the gratification of leaving home, conversation with colleagues and customers, getting a store discount, and generally being a member of society.
The reason that even well-meaning commentators suggest increasingly unlikely and tenuous ways for people to make a living is that they are trapped in the conventional thinking that people must trade their time, energy, and labor for money as the only way to survive. You stretch for answers because, in reality, there are none. The subsistence and scarcity model is grinding more and more people up. Preserving it is the thing we must give up first.
FOOD PREPARATION AND SERVICE
For the number three job in America, the median hourly wage is $10 an hour with an annual average wage of $23,850. Most of these workers have not attended college. Food service and food prep workers are not in immediate danger of replacement to the same degree as are call center workers and retail workers. Mom-and-pop restaurants are not changing their practices anytime soon, and food service workers are generally so inexpensive that the incentives to replace them are modest. The restaurant industry is facing headwinds due to lower foot traffic in many places, more lunches eaten at desks (the “lunch depression”), high levels of competition, the decline of mid-price restaurants, and the rise of eat-at-home delivery services like Blue Apron, but people anticipate restaurants holding up better than traditional retail.
Still, change is brewing. I had brunch with a venture capitalist friend in San Francisco. She told me an important story: “A company came to me with a software product that helps fast food workers get scheduled for shifts more efficiently among multiple locations. Any given worker could be optimally assigned a shift across several nearby stores. It seemed like a good idea. But when I went to a couple fast food companies and I asked them if they would use this kind of software, their response was, ‘We’re not trying to schedule our workers more efficiently. We’re trying to replace them altogether.’ So I didn’t invest in that company. Instead, I invested in a couple companies that make smoothies and pizza with robots and delivery.”
She’s not alone. There is now a mechanized barista in a lobby in San Francisco. It’s named Gordon. You can text in your order and the robot can be set up in most any location. I tried it and my Americano was delicious, for about 40 percent less than at Starbucks. Gordon provides a more efficient, cheaper, and equally high-or even higher-quality product than a human barista. In the morning, when you are running late to work and all you want is a quick cup of coffee, these pluses will be valuable. After Gordon debuted, Starbucks was forced to issue a statement saying that it didn’t plan on replacing its 150,000 baristas.
Some workers will be easier to replace than others. For instance, we all like fast food drive-thru restaurants for their efficiency and do not mind the limited human interaction. In fact, 50 to 70 percent of fast food sales take place at drive-thru windows in the United States—McDonald’s being the one that most of us know and (used to) love. There are 1–2 workers per location who take the order through the speaker—they wear those cool headsets. These workers will be replaced by software in many locations in the next five years. Publicly traded fast food chains will be among the most aggressive adopters of increased efficiencies because they have the scale, resources, and quarterly earnings pressures to maximize shareholder returns. McDonald’s just announced an “Experience of the Future” initiative that will replace cashiers in 2,500 locations to start. The former CEO of McDonald’s suggested that large-scale automation is around the corner. “It’s cheaper to buy a $35,000 robotic arm than it is to hire an employee who’s inefficient making $15 an hour bagging French fries,” he said while defending the current prevailing fast food wage of $8.90. The robot arm is only going to get cheaper and more efficient, while the fast food wage has no place to go but up. Approximately 4 million workers work in fast food.
If you’ve been through an airport recently, you might have noticed restaurants that have replaced servers with iPads. Eatsa, a recently opened restaurant chain, has a whole row of iPads for you to enter the order, and then a series of lockers where your food appears. They’ve gotten rid of all of the front-of-house workers. Eatsa was recently named one of the most influential brands in the restaurant industry, and it’s here to stay. All it takes is a few chains to bite the bullet and enjoy labor-free efficiencies and the others will follow quickly. McKinsey estimates that 73 percent of food prep and service activities are automatable.
On the production end, you ca
n now use a 3D printer to make hot pizza in five minutes that can be customized to particular orders. BeeHex’s bot, called the Chef 3D, will appear at select theme parks and sports arenas starting later this year. Just like the robot barista, Chef 3D is faster, cleaner, and more reliable than human workers. Only one person is needed to work the machine, which can mix the composition and lay down the sauce and the toppings in one minute. Apparently it tastes great. No more person in the back making pizzas by the oven. There are companies now launching that are essentially pizzerias on wheels, where they make the pizza in special trucks on their way to you in anticipation of your order.
For the last mile, there are now food delivery robots being used in Washington DC, and San Francisco. They are essentially coolers on wheels that deliver food to your door for around a dollar. One company called Starship Technologies has 20 or so robots deployed that are already learning their local terrain in Washington, DC, which has officially made self-driving robots legal on its sidewalks. These robots will eliminate the need for many deliverypeople.
A friend of mine, Jeff Zurofsky, ran a chain of sandwich shops for a number of years. He said to me, “Our biggest operational issue is that sometimes people just don’t show up to work. We pay significantly over the minimum wage, but employee reliability is a recurring problem.”
Food prep and service jobs are going to remain numerous for a while to come because of low costs and industry fragmentation. But fundamentally, most of the tasks are highly repetitive and automatable. Companies with resources are going to continue to experiment with new ways to reduce costs and we will see fewer and fewer workers in many restaurants over time. Also, as regional economies weaken, restaurants in those regions will struggle and close.
Clerical jobs, retail jobs, and food service jobs are the most common jobs in the country. Each category is in grave danger and set to shrink dramatically. Yet they’re not even the ones to worry about most. The single most defining job in the automation story—the one that scares even the most hard-nosed observer—is the number four job category: materials transport, also known as truck driving.
FIVE
FACTORY WORKERS AND TRUCK DRIVERS
You would have to have been asleep these past years not to have noticed that manufacturing jobs have been disappearing in large numbers. In 2000 there were still 17.5 million manufacturing workers in the United States. Then, the numbers fell off a cliff, plummeting to fewer than 12 million before rebounding slightly starting in 2011.
More than 5 million manufacturing workers lost their jobs after 2000. More than 80 percent of the jobs lost—or 4 million jobs—were due to automation. Men make up 73 percent of manufacturing workers, so this hit working-class men particularly hard. About one in six working-age men in America is now out of the workforce, one of the highest rates among developed countries.
What happened to these 5 million workers? A rosy economist might imagine that they found new manufacturing jobs, or were retrained and reskilled for different jobs, or maybe they moved to another state for greener pastures.
In reality, many of them left the workforce. One Department of Labor survey in 2012 found that 41 percent of displaced manufacturing workers between 2009 and 2011 were either still unemployed or dropped out of the labor market within three years of losing their jobs. Another study out of Indiana University found that 44 percent of 200,000 displaced transportation equipment and primary metals manufacturing workers in Indiana between 2003 and 2014 had no payroll record at all by 2014, and only 3 percent graduated from a public college or university in Indiana during that time period. The study noted, “Very few went back to school, and relatively few seemed to avail themselves of a lot of the government programs available to assist displaced workers.
The manufacturing jobs that still exist require more education and technical skills as factories have become more advanced and automated. Jobs in manufacturing for people with graduate degrees grew by 32 percent after 2000, even as overall employment in the sector was plummeting. Of course, as we’ve seen, most people don’t have graduate degrees or even college or associate’s degrees, and it is unrealistic for many to get them.
“The recession led to this huge wiping out of one-industry towns, particularly in those places that were heavily dependent on the industrial or manufacturing economy,” says Steve Glickman, CEO of the Economic Innovation Group. “We’re asking: What’s around the corner for them? And we’re seeing a shockingly low rate of new businesses that can become the new employers for those regions of the country.”
How do the 40 percent of displaced manufacturing workers who don’t find new jobs survive? The short answer is that many became destitute and applied for disability benefits. Disability rolls shot up starting in 2000, rising by 3.5 million, with the numbers increasing dramatically in Ohio, Michigan, Pennsylvania, and other manufacturing-heavy states. In Michigan, about half of the 310,000 residents who left the workforce between 2003 and 2013 went on disability. Many displaced manufacturing workers essentially entered a new underclass of government dependents who have been left behind.
This is a good indicator of what will occur when truck drivers lose their jobs. The average age of truck drivers is 49, 94 percent are male, and they are typically high school graduates. Driving a truck is the most popular job in 29 states—there are 3.5 million truck drivers nationwide.
Trucks that drive themselves are already rolling out around the world. Self-driving trucks successfully made deliveries in Nevada and Colorado in 2017. Rio Tinto has 73 autonomous mining trucks hauling iron ore 24 hours a day in Australia. Europe saw its first convoys of self-driving trucks cross the continent in 2016. In 2016 Uber bought the self-driving truck company Otto for $680 million and now employs 500 engineers to perfect the technology. Google spun off its self-driving car company Waymo, which is working on self-driving trucks with the big truck manufacturers Daimler and Volvo.
Jim Scheinman, a venture capitalist at Maven Ventures who has backed startups in both autonomous trucks and cars, says that self-driving trucks will arrive significantly before cars because highway driving is so much easier. Highways, the domain of semi trucks, are much less complex than urban areas, with fewer intersections and clearer road markings. And the economic incentives around freight are much higher than with passenger cars.
Morgan Stanley estimated the savings of automated freight delivery to be a staggering $168 billion per year in saved fuel ($35 billion), reduced labor costs ($70 billion), fewer accidents ($36 billion), and increased productivity and equipment utilization ($27 billion). That’s an enormously high incentive to show drivers to the door—it would actually be enough to pay the drivers their $40,000 a year salary to stay home and still save tens of billions per year.
Switching to automated drivers would not only save billions, but it also has the potential to save thousands of lives. Crashes involving large trucks killed 3,903 people in the United States in 2014, according to the National Highway Traffic Safety Administration, and a further 110,000 people were injured. More than 90 percent of the accidents were caused at least in part by driver error. Driver fatigue is a factor in roughly one out of seven fatal accidents. Most of us when we were taught to drive growing up were told to avoid trucks on the highway. There’s a reason for that.
So the incentives to adopt automated truck driving are massive—tens of billions of dollars saved annually plus thousands of lives. They are so large that one could argue it is important for national competitiveness and human welfare that this happen as quickly as possible. Adding to the incentives is that many freight companies report labor shortages because they can’t find enough people willing to take on the physically demanding and punishing job of spending hundreds of hours sitting in a confined space. Truck drivers spend 240 nights per year away from home staying in truck stops and motels and 11 hours per day on the road. Obesity, diabetes, smoking, inactivity, and high blood pressure are common, with one study saying 88 percent of drivers had at least one ris
k factor for chronic disease.
Many, however, will argue for the preservation of truck driving because they recognize just how problematic it would be for such a large number of uneducated male workers to be displaced quickly.
Taking even a fraction of the 3.5 million truckers off the road will have ripple effects far and wide. It is impossible to overstate the importance of truck drivers to regional economies around the country. As many as 7.2 million workers serve the needs of truck drivers at truck stops, diners, motels, and other businesses around the country. Over 2,000 truck stops around the country serve as dedicated hotels, restaurants, grocery stores, and entertainment hubs for truckers every day. If one assumes that each trucker spends only $5K a year on consumption on the road (about $100 per week), that’s a $17.5 billion economic hit in communities around the country. Beyond the hundreds of thousands of additional job losses, many communities may risk losing a sense of purpose without thousands of truckers coming through each day. For example, in Nebraska one out of every 12 workers—63,000 workers—works in and supports the trucking industry.
Truck drivers do not see it coming. Indeed, when Bloomberg’s Shift Commission in 2017 asked truck drivers about how concerned they were about their jobs being replaced by automation, they almost uniformly weren’t concerned at all. Let me assure you it’s coming. Elon Musk recently announced that Tesla will be offering a freight truck as of November 2017. Musk also proclaimed that by 2019, all new Teslas will be self-driving. “Your car will drop you off at work, and then it will pick other people up and make you money all day until it’s time to pick you up again,” Musk proclaimed. “This will 100 percent happen.” It is obvious that Tesla trucks will eventually have the same self-driving capabilities as their cars. Other autonomous vehicle companies report similar timelines, with 2020 being the first year of mass adoption. And it’s not just those driving trucks who are at risk. A senior official at one of the major ride-sharing companies told me that their internal projections are that half of their rides will be given by autonomous vehicles by 2022. This has the potential to affect about 300,000 Uber and Lyft drivers in the United States.