Understanding Automation

The United States was clearly slammed by global competition, particularly in the 1990s and for twenty years thereafter. Clothing, appliances, furniture, televisions, cars, and many other durable manufactures could be made overseas for a significant reduction in price. Overseas labor in third world and developing nations often worked long hours under horrible conditions, even when corrected by higher shipping costs and even some tariffs, undercutting American blue collar workers left, right and center. Too many American workers were displaced, never to earn much again. Nobody seemed to care about them.

American companies could not compete with these foreign manufactured goods, so they joined in and move their operations overseas. Marginally-skilled American workers lost their jobs in droves, and there were few economic support systems to replace their lost earning power. Too old or unwilling to go back to school to learn a new trade, many of these workers permanently dropped out of the American workforce, no longer included in the unemployment statistics because they gave up what they believed to be a hopeless job search.

These people still lingered, hoping for magic… and their hero appeared. Donald Trump was the master of blame: globalization was evil, high taxes and regulations prevented money trickling down to increase hires. He would pull out of “the worst trade agreements ever negotiated,” would cut “job-killing regulations and taxes,” and “Make America Great Again.” He would create “jobs, jobs, jobs.” Just one tiny catch: globalization had settled into the new normal. Jobs were no longer being lost to cheap overseas labor (Chinese labor isn’t even cheap anymore). 87% of recent US manufacturing job-loss surrendered to automation, not globalization.

Yet as Donald Trump’s November 10^th speech before APEC members in Da Nang, Vietnam suggests, globalization, not automation, is the main source of American job loss… a message that is most certainly “yesterday’s news.” On the other hand, Chinese leader Xi Jinping, addressing the same body, touted the irreversibility of global trade.

As I wrote in my September 11^th Stitches blog, blaming global trade for today’s employment issues is yesterday’s news. The promise of “reshoring” (bringing manufacturing back to the U.S.) under the notion of U.S. job creation is beyond empty these days. No business with an ounce of economic understanding is going to take cheap-labor overseas manufacturing and bring it back into the United States with the same extent of vastly more expensive labor. And with Congress pretty much telling Mr. Trump that they are not open to punitive duties on American companies who manufacture abroad, the only kinds of manufacturing returning to the United States is driven by the elimination and substitution of labor by automated machinery. And more and more, that automation includes the ability to “learn” and improve the process without human intervention. Artificial Intelligence (AI).

In other words, income inequality fans, the folks who benefit most from reshoring are those who own those sophisticated and automated manufacturing systems – wealthy owners of companies and their financial support communities – and those who lose the most are the displaced workers. Sure reshoring creates a few jobs, a couple of supervisors, monitors and maintenance workers (and even those jobs are getting reduced as artificial intelligence grows), but the number of people required in the manufacturing process is dropping like a stone.

Stitches describes an automated clothing-making machine that can produce a sophisticated finished article of clothing for a per-unit cost much cheaper than the least expensive assembly line in Bangladesh. They’re called “SewBots,” and they are built and programmed by an Atlanta, Georgia, SoftWear Automation. It took ten years to develop, primarily through Georgia Tech’s Advanced Technology Development Center(with a Department of Defense DARPA innovation grant), but the system is stunning in what it can do… and for the longer-term impact on the global notion of labor itself. The folks who own those SewBots will make the money, but the workers who used to make that clothing will not. A parallel question could easily be, what happens truck drivers when we have self-driving trucks?

To understand how fast automation is replacing workers, take a really good look at the oil and gas drilling industry as reflected in the above chart. Even as we are building more drilling rigs, which you would think would generate lots of new jobs, the number of oil and gas workers has plunged. The motivation for bringing so much automation onto drilling rigs came when the price of oil fell precipitously and stayed there. Forced to reduce the cost of fossil fuel extraction, the industry embraced massive job-killing automation to replace expensive labor.

“Thanks to automated drilling, a once dangerous and very laborious task now requires fewer people to accomplish. Automation of oil rigs means that one rig can do more with fewer workers. In fact, it’s expected that what once took a crew of 20 will soon take a crew of 5. The application of new technologies to oil drilling means that of the 440,000 jobs lost in the global downturn, as many as 220,000 of those jobs may never come back.

“Now look back at the chart again, and notice how quickly this all happened. It took TWO YEARS. How did it happen so fast? Because the oil industry didn’t really need the workers it lost in the first place. It’s the oil industry. It’s used to making lots of money, and when you’re making money hand over fist, you don’t need to focus on efficiency. Being lean and mean is not your concern. However, that changes when times get tough, and times got very tough for the oil industry as oil prices plummeted thanks to new competition from yet another technological advancement — fracking.

“So once it became important to increase efficiency, that’s exactly what the oil industry did. It let people go and it invested in automation. In the summer of 2016, oil prices were no longer under $30 per barrel, and had gone back up to around $50 per barrel where they remain. That’s half of the $100 per barrel they’d gotten used to, which is fine as long as they’re able to produce at twice the efficiency. As a result, like a phoenix rising, they emerged transformed. Oil rigs returned to drilling, but all the rig workers didn’t. Those who were let go became simply unnecessary overhead.

“Companies are actively investing in automation because it means they can produce more at a lower cost. That’s good for business. Wages, salaries, and benefits are all just overhead that can be eliminated by use of machines.

“But hey, don’t worry, right? Because everyone unemployed by machines will find better jobs elsewhere that pay even more… Well, about that, that’s not at all what the history of automation in the computer age over the past 40 years shows. Yes, some with highly valued skills go on to get better jobs, but they are very much the minority. Most people end up finding new paid work that requires less skill, and thus pays less. The job market is steadily polarizing…

“[Technology] is only getting cheaper, so each successive drop squeezes out more human labor, and is able to automate more lower-skill labor that is newly more expensive than machines. Expect the next recession to put over ten million of people out of work, and for the economy to realize they didn’t really need those people as workers after all to produce what was being produced. Where 79% of eligible workers aged 25–54 were employed, expect that to fall to 69% or below. The economy simply doesn’t need the number of people it currently employs with the technology we already have available. To add insult to injury, it’s taking longer for the unemployed to find employment, so those suffering next will suffer longer.” Scott Santens writing for the October 24^th Medium.com.

So we now come back to the basic conundrum. Globalization ain’t it. Companies, including some that base their operations on educated workers with white collar skills, cannot compete unless they reduce expensive labor and replace them with machines. As artificial intelligence grows, the level of work that can be delegated to machines rises exponentially… the unemployment reaches higher and higher up the labor economic ladder, eventually replacing even surgeons, lawyers and financial advisors. But as the employment picture declines, as earning power for the labor force declines, as wealth continues to concentrate in the hands of those at the top of the food chain who own the machines at the expense of human workers, where are the consumers to buy those cheaper products? See a problem for everyone here?

What do you do with a highly-educated work force than cannot get a job or can’t earn enough to have sufficient discretionary income to make a difference? Do we tax the machines and spread the wealth to displaced workers under a notion of “universal basic income”? How do we decide who gets what? That’s socialism on steroids. It may be inevitable, but there isn’t a major developed nation with a social-political-economic system that is remotely prepared for this onslaught of “automation replacing humans” at the scale that is well underway.

I’m Peter Dekom, and as politicians – particularly Donald J Trump – focus on an issue with decreasing relevance (globalization) for American workers, nobody is minding the back door as automation will decimate and reconfigure our entire job market.