Two of the most important transportation-industry announcements of the past century have both taken place in the last five years, and both have been made by complete transportation-industry outsiders. The first announcement, on Oct. 9, 2010, was the revelation that Google (GOOGL 0.55%) had developed self-driving cars that had already made enough successful trips to circle the earth five and a half times. The second announcement, made yesterday evening on 60 Minutes, was Amazon.com (AMZN -1.64%) CEO Jeff Bezos' public reveal of Amazon Prime Air, a development-stage drone system that promises pizza parlor-like delivery speed: 30 minutes or less from the warehouse to your door aboard an eight-rotored automated copter.
Source: Amazon.com.
After the show ended, Amazon added a new page to its site explaining the Prime Air service. In addition to Bezos' on-air explanation, which was that the service would be limited to packages under five pounds (86% of the company's shipments weigh less than five pounds) and to a ten-mile delivery radius from fulfillment centers, the page (you can visit it yourself by clicking here) offers a few rather lofty predictions. Amazon claims that Prime Air drones will be "as normal as seeing mail trucks on the road today," and that the service will be ready to go live as soon as the Federal Aviation Administration adopts its new drone-related rules in 2015. The explainer site also happens to have a video of the system in action, so let's take a look...
Based on Bezos' history of product reveals, we have little reason to assume that this is vaporware. It certainly won't be operating this year, as FAA rules won't even allow unmanned drones to be used for commercial purposes until 2015. Even after the rules change, it's likely to take several more years to put the Prime Air system (and other commercial drone systems) into action. The FAA's own roadmap for integrating these unmanned aerial vehicles into the nation's aviation infrastructure projects that it will take more than a decade to complete its long-term goals. Ensuring that thousands of delivery drones can seamlessly and safely take to the skies will undoubtedly be a terrific undertaking for both government regulators and private enterprise, but Amazon's entry is likely to be only the first surge in a flood of corporate interest. Once enough money enters a new field, government support often follows.
Keep in mind that Google's self-driving cars are only legal in three states, but even this is only on a test basis for the time being. Laws move slower than technology, suppressing the wider development of both drone delivery systems and driverless cars for the moment. Amazon and Google have the resources to develop these technologies with no hope of immediate commercialization, but few other companies can say the same.
But make no mistake, Google's self-driving car technology and Amazon's drone-delivery system are on the leading edge of a post-human transportation network, where no one will be needed to actually drive the cars, fly the planes, and deliver the goods that are so essential to modern human life. That is a big deal, as important to the progress of civilization as was the invention of the locomotive and automobile in the 19th century. The average person's life is mediated through transportation. You go to work, you go out to buy things or you buy things that are delivered to you, you drive or fly or take the bus or the train from place to place nearly every day of your life. Without our vehicles, we'd all be driving oxcarts full of turnips down dirt roads.
These transportation tasks take up an awful lot of human capital to do one of the simplest things you can imagine: get someone or something from Point A to Point B. And that human capital is highly error-prone -- around the world, roughly 1.3 million people die and as many as 50 million more are injured in auto accidents every year. There were far fewer aircraft fatalities, but a total of over 3,000 aviation accidents -- the number that occurred last year in the United States -- is still too many. Delays on highways and tarmacs across the country add up to over $150 billion each year, an amount roughly equal to 1% of U.S. GDP. If we assume a similar impact worldwide, then transportation delays might dent global productivity by over $800 billion each year. All of those deaths and injuries likely add trillions to the tab in treatment costs and lost labor.
Imagine a world where none of this is true. No one gets hurt or dies in auto accidents. No one is stuck waiting on a tarmac. No one waits days to get their packages. It won't arrive in the next year or the next five years, or perhaps even the next 15 years, but when the risk-reward equation becomes so obviously tilted in favor of automated alternatives, even companies that might otherwise want to keep humans at the wheel will realize that they ultimately need to give these tasks over to the machines. And there's no reason to expect that this won't happen.
Technology keeps improving, but human skill sets are biologically limited. Technology also tends to get cheaper -- if the major cost of a fully autonomous transportation network is in the sensors and processors, then we can expect those sensors and processors to follow the curve of Moore's Law and become much cheaper by the time laws catch up to technology. The costs of energy will be harder to overcome, but if Amazon can develop drones with a ten-mile delivery radius, then a 25-mile radius or a 50-mile radius should simply be a matter of scale and better battery technology. This is less of an issue with self-driving cars, which are largely retrofitted versions of hybrid vehicles and could easily transition to electric drive as charging networks proliferate.
Source: Anggoro Prastyo via Picasa Web Albums .
Will the economic benefit of automated networks outweigh the job losses? I took a look at the potential employment disruption from an automated transportation network six months ago, in broad terms, but let's get more specific now that another of the world's most important tech companies has thrown its weight behind an unmanned future. There are about 4.5 million people working in the American transportation and warehousing sector. Many are drivers of all sorts of vehicles, others are pilots, and large numbers do the hard work of sorting, stacking, lifting and packing the millions of things American consumers ultimately buy every day in stores and online. Let's imagine a future where all of them are unemployed -- say, in 2020.
The American civilian labor force stood at 154 million in 2010, and is expected to grow to 164 million in 2020, a relatively modest growth rate by historical standards. The Bureau of Labor Statistics expects the transportation industry to add roughly 900,000 new jobs by 2020, putting its new employment total at 5.4 million jobs. Wildly inaccurate projections are a bit of a hallmark for the BLS in recent years -- its 2012 job-growth projections for construction and manufacturing, made in 2002, were too generous by a combined five million jobs. But let's say that 5.4 million transportation jobs is a baseline. Those workers will make up 3.3% of the entire labor force in 2020, which is a significant number. It's essentially the difference between the unemployment rate at the time of Lehman's bankruptcy in the fall of 2008 and the unemployment rate near its peak in mid-2009.
But transportation is a massive, complex industry that impacts virtually everything else in the economy, from energy production to manufacturing to food service and retail. We can probably estimate that every transportation job ultimately helps to support at least two jobs elsewhere.
The average annual transportation salary is roughly $45,000, which is only $6,000 below the household median wage in the United States, and which would work out to about $250 billion in total industrywide wages for 5.4 million people, assuming that average salaries will be more or less unchanged between now and 2020. If the two jobs supported by each transportation job only make $30,000 a year, that still works out to another $320 billion in lost wages, for a total of $570 billion lost to automation. Impacts elsewhere in the economy will be harder to quantify -- millions of lost jobs invariably leads to a recession, but most of those jobs lost are typically regained later. If two jobs are lost elsewhere for every job lost in transportation, then the total impact will be equal to roughly 10% of the labor force in 2020. That's won't just a cause a recession, it'll likely lead to another Great Depression.
The benefits of automating our transportation and delivery network will no doubt be enormous, but at the same time it's likely to cause a workforce dislocation we're not fully prepared for. Some displaced truck drivers will no doubt find new jobs in drone maintenance or perhaps driverless-car programming, but it's unlikely that these new fields will absorb enough people. The relatively mature American tech industry (computing and information technology) provides fewer jobs today than were lost in the American manufacturing industry over the past decade. Before jumping for joy at the prospect of 30-minute deliveries for everything under the sun, let's pause to consider what we'll have to give up in exchange.
