Automation makes things cheaper, so why doesn’t it feel that way?

by William H. Davidow, Apr 3, 2017

Many Americans have been experiencing massive economic insecurity and want to return to a time when they were able to earn a good living and create a decent life for themselves and their families. President Trump promised to deliver on that wish, to “Make America Great Again,” by reducing business regulations, fixing trade deals, and stopping the loss of manufacturing jobs.

Some of those policies will help, but they’re unlikely to fully address the challenges faced by the middle class.

The problem is that we are now living in a bipolar economy, with a traditional economy at one end and an autonomous economy at the other. The traditional economy is prone to inflation; the autonomous economy toward deflation.

Many middle-class Americans are earning their money in a deflationary economy where wages are falling, and spending it in a traditional economy where prices are rising. Those Americans trapped between the poles of the bipolar economy were among Trump’s most enthusiastic supporters.

How it will impact business, industry, and society.

The traditional economy provides us with most of the basic necessities of life: food, shelter, health care, clothing, transportation, and energy. Money spent on these necessities account for 70% of middle-class expenditures. The prices of most of those expenditures are rising faster than inflation.

Property managers forecasted an 8% increase in rents in 2016. From 2012 to 2014, the median home price rose by 17.3%. The U.S. Department of Agriculture reported that food prices rose 31% from 2005 to 2014. The cost of health insurance has risen by over 44% in the past five years.

The autonomous economy is where intelligent machines, robots, artificial intelligence, big data, and high-speed digital communication power production. Those processes produce high-quality goods and services, frequently with little human involvement. The autonomous economy is extremely labor and capital efficient, sometimes so much so that the prices of its products decline to close to zero. Often the markets they supply shrink in size. Most concerning, the value of human work declines as workers compete with robots for jobs.

The transformation of the movie rental and physical media businesses are examples of what happens when businesses move from the traditional to the autonomous economy.

At one point, Blockbuster Video had 60,000 employees and 9,000 stores. Then it got absorbed into the virtual economy, replaced by Netflix and by Redbox kiosks. Netflix has around 2,450 employees.

This has been good news for consumers. Many movie watchers can now view their movies for free on ad-supported networks like Crackle, which is owned by Sony Pictures Entertainment.

In the meantime, the manufacturers and distributors of physical media are struggling. Blu-ray and DVD sales have suffered declines of more than 10% for the last two years. Even Redbox’s revenue plunged by almost 19%.

If you work in an industry transitioning from the traditional to the autonomous economy — say, if until recently you worked at Blockbuster — but most of what you buy is in the traditional economy (rent, food, health care), you’re probably not doing very well.

The shift toward the autonomous economy is happening in the services sector, too, which makes up 80% of the workforce. (For all our talk about manufacturing jobs, they account for less than 9% of the workforce). More and more services are delivered by computers transacting with other computers, such as when we make travel reservations, when machines manage our investment portfolios by trading stocks on electronic exchanges, when we shop online, and when systems generate our news feeds and negotiate with other computers over what ads will be presented to us.

While we have no good way of determining the precise amount of goods and services produced autonomously, we know it is very large. Brian Arthur, in his article on the Second Economy, or what he now calls the autonomous economy, estimated that, by 2025, autonomous production would produce goods and services equal to the country’s GNP in 1995, about $7.5 trillion.

The shift to the autonomous economy and the pressure it places on workers are likely to only get worse. For an idea of the effects that autonomous production might have, consider the potential impact of the Level 4 Autonomous Vehicle, a car that drives itself and never needs human help.

Imagine that many consumers got rid of their cars and opted for driverless shared car services — a driverless Uber. If that displaced 20% of the 250 million cars Americans own, consumers would save $150 billion, $3,000 for every car displaced, and 1 million jobs would disappear in the automotive related industries — manufacturing, insurance, finance, etc.

There are 3.5 million professional truck drivers in our country, and about 8.7 million people are employed in the trucking business. If autonomous trucks displaced just 20% of them, another 1.75 million jobs would vanish.

The jobs of 200,000 cab and limo drivers, as well as those of 400,000 part-time Uber drivers, are at risk as well. If we could get by with 100,000 of them, the equivalent of somewhere around 250,000 full-time jobs would vanish.

Automated delivery vehicles would facilitate the deliveries from the Amazons of the world. Online retailers, like Amazon, employ about one-third the number of employees per dollar sold as conventional retailers do. If online retailers gained an additional 10% in market share, that would eliminate another 1 million jobs.

Under those assumptions, 4 million jobs would be displaced by the Level 4 Autonomous Vehicle.

Lost jobs have a multiplier effect. If jobs decline in the auto industry, jobs vanish in the automobile supply chain. People who lose jobs purchase fewer retail goods and eat out less often, so other jobs are lost as well.

Numbers vary dramatically about the size of the multiplier effect, but two is a pretty conservative guess. If a multiplier effect of two is used in the case of autonomous vehicles, a total of 12 million jobs would vanish in the scenario outlined above.

You might quibble with these numbers, but that is missing the point. The effects of the Level 4 Autonomous Vehicle will be very large. I personally believe these are conservative guesses.

Autonomous production exerts downward pressure on wages, as workers compete both with machines and with the new influx of job seekers whose jobs have already been replaced. And as we have seen, this process shifts many full-time jobs into the gig economy — places like Uber, TaskRabbit, etc. — where individuals work as independent contractors, without benefits.

It is easy to conclude in response to all of this that there will be a shortage of work in the future, and that we will have to consider other previously unthinkable solutions. Commentators across the political spectrum have proposed guaranteed annual incomes as a potential solution. Others have suggested that shorter workweeks would create more job opportunities. If coupled with aggressive earned income tax credits, these proposals could perhaps allow workers to earn a decent income. Some have proposed offering tax credits to people who perform socially useful work for which they receive no pay, such as someone who stays home from work to care for a child or elder.

Most of us, including myself, feel extremely uncomfortable with these types of ideas. But the autonomous economy is rewriting the economic rules, and so we shouldn’t constrain our thinking.

It is comfortable to stick with the old solutions, ignore the likelihood that they will be insufficient, and just hope. But hope is not a strategy. It will only make the long-term problem worse.

Autonomous production is already unleashing a bipolar economic system, and with it some very tough social challenges. If we are really serious about change, we will have to do radical things to address inflation in the traditional economy and deflation in the autonomous one, in response to a world where the value of human work is under pressure.