Workin' This Job

I recently found myself in a conversation about how the minimum wage in the United States is likely increasing to $15 per hour. Most conversations about minimum wage laws tend to focus on the moral implications of what workers are expected to earn. That a full-time worker should put in forty hours a week for earnings that are right around the poverty line represents a failure of our society. Or, that a worker who lacks the initiative to progress beyond an entry level position expects the relative comfort of lower middle class earnings represents a sense of entitlement that is neither realistic nor sustainable. A few will bring up economics, pointing out that higher minimum wages will tend to reduce employment.

I was thinking in a somewhat different direction. "I think you're gonna see a lot more automation," I said. "You know those touch screens that are starting to show up at Mickey D's so you can punch in your own order? They'll be in every fast food joint. The back of the store will be have a lot more robotics, too".

The reason I say this this is that at the moment cheap labor is a barrier to greater adoption of automation. It's much cheaper and simpler to hire, fire, and tweak the schedules of workers who make $7.25 per hour than it is to commit to the capital expenditure of the machinery that would replace those low-wage workers. A higher minimum wage radically changes the economics, and makes automation a more attractive alternative. There won't be any burger flippers making $15 per hour; burger flippers will be replaced by machines. In a way that will require a lot of caveats, that doesn't have to be a bad thing.

For starters, it's not exactly relevant whether I or anyone else believes that increasing levels of automation in the economy is a good or bad thing. It's inevitable. If the technology is there, if the economic incentive is there, there'll be no way to stop it. And the economic incentive will, at some point, be there. A minimum wage increase will likely accelerate this process, but low wages would only put it off a few years. 

There is also an overall economic benefit to automation. Thanks to the automation that we already have, the economy is more productive than ever. Higher productivity means more value is delivered for the same amount of money, or a comparable amount of value is delivered for less money. Those lower costs and higher outputs have to end up somewhere, and so they result in some combination of lower prices, higher wages, and larger profits which, in the long run, makes everyone better off.

There's another benefit to automation, one that's difficult to talk about, particularly in the United States where we take it as an article of faith that work itself—any work—intrinsically confers dignity onto the person performing the work. What's hard to talk about is that while many of us find our jobs to be stimulating and rewarding and challenging and a well-earned source of dignity and self-worth, there are some jobs that are demeaning. Some jobs just just don't have any intrinsic dignity, and often these are the easiest jobs to automate.

I briefly worked for a small manufacturing company in the '80s. We had an incident where a client rejected a shipment. The product was grossly defective, and had no business being shipped out in the first place. What led to the defective shipment was what I would later learn is an engineer's worst nightmare: a cascading failure. A defect was introduced fairly early in the manufacturing process. Work areas performing subsequent operations, trying to work around the original defect, only managed to make matters worse. 

To understand what went wrong, to prevent similar mishaps in the future, to regain the trust of our client base, would be a daunting task. Each micro-failure needed to be well understood, as well as the relationships among the individual micro-failures that ultimately propagated into a complete failure to deliver value. There were multiple technical points of failure to understand but, even more challenging, there were organizational failures. The work had passed through several departments and yet not one single employee—whether because of interdepartmental rivalries, a sense of time pressure, or just plain fear—felt empowered to speak up, pause the process, and correct the defect. 

Fortunately, management knew just what to do. They called a mandatory meeting of all entry level and semiskilled employees company wide and spent well over an hour berating, blaming, and threatening the most powerless people in the entire organization. At one point one of the senior managers—I'll call him George, which isn't his real name—took the floor. His opening sentence provided the most memorable moment of the entire day. "For the hours that you're here," George told the assembled group, "we own your ass."

I went on to work in the printing industry and at mid career found the industry going through its automation revolution. I quickly realized that I could either become one of the people implementing the automation or I could become one of the people put out of a job by automation. The former seemed like more fun than the latter, and it has led to a highly satisfying second career in software development. The transition was not without its own crisis of conscience. Like many of my age cohort, I had grown up with cautionary tales about the rise of the machine and the demise of humanity, like the Twilight Zone episode in which a cold-hearted factory manager gradually replaces workers with machines, gleefully firing employees by the thousands until no one is left and ultimately, in a cruelly ironic twist, he too is replaced by a machine. It worried me that I might be—almost certainly would be—doing work that would ultimately result in staff reductions. I thought about the inevitability argument that I outlined earlier. If it weren't me doing the work and helping to support my family, it'd just be someone else and the result would be the same. And if it weren't my company it'd be a competitor; we'd be out of business and everyone would be out of work. I thought about productivity and the long-term benefit to the economy as a whole. I had already just about convinced myself that I was doing the right thing, and then I thought about George. Bullies like George build their power base by intimidating vulnerable workers, by pushing people around who have no other place to go. The kinds of jobs that generally get automated away are the kind that are performed by the recipients of George's abuses. I realized that in some small way I could have the opportunity to deprive the Georges of the world of asses to own. I never looked back.

One could argue that working even for George is better than not having a job and, the way things are currently set up, that's quite true. This is where the caveats come in. While it's true that productivity increases generally make everyone better off in the long term, expressions like generally and long term are stand-ins for the fact that a lot of displacement occurs on the way to realizing those benefits. To not properly address that displacement is to risk a level of social cost that prevents us from fully realizing the benefits of greater productivity.

How we address economic and social displacement is an open question that will require considerable expertise to solve. We know a couple things that don't work. A kind of phony populism that plays on people's resentments doesn't work. What went before our experiment with phony populism, which was essentially decades of ignoring the problem or somehow thinking that it would just magically solve itself, didn't really work either. We'll need other ideas. A redefinition of how education is delivered and consumed is likely part of it. A way of decoupling medical insurance from one's employment, to create a more agile workforce made up of workers more willing to give a new job or a new career a try is almost certainly a requirement. A market for so-called livelihood insurance, where a worker can pay a premium on a policy that will pay out benefits if that worker's chosen specialty is ever made obsolete by technology changes, is an interesting idea. A somewhat more unconventional idea but one that is gaining traction is the idea of a universal basic income. It may turn out that we'll need to readjust our expectations to an economy that's not as reliant on human labor as it once was.

In a way, it's a good problem to have. The task we're facing is to figure out how best to manage enormous improvements in productivity, how to ensure that the benefits are arranged so as to encourage innovation and entrepreneurship without incurring the social cost of leaving a large segment of the population behind. Along with climate change, this will be the defining challenge of the first half of the twenty-first century. I think we're up to it.