The Automation Talk

Posted on Posted in Economics, Machine Intelligence, Machine Learning, Public Policy, Technology

“A study at Ball State University’s Center for Business and Economic Research last year found that trade accounted for just 13 percent of America’s lost factory jobs. The vast majority of the lost jobs — 88 percent — were taken by robots and other homegrown factors that reduce factories’ need for human labor.” – AP

Canada’s labour force is around 19.6 million people, of which 18.2 million people are employed. Together, they worked something like 2.4 billion hours that month. In December 2016, something like 1.7 million Canadians worked about 240 million hours in manufacturing.  Roughly. Because of seasonal adjustments and different data at different times. And error.

In terms of our working lives in Canada, collectively, manufacturing is about 10% of us, and 10% of our time. And 10% of our GDP. Again, roughly.

1.3 million Canadians wanted to work, but didn’t have it. 3.5 million Canadians had part time work. It’s hard to tell how many wanted more hours. It shouldn’t be assumed that all 3.5 million Canadians who had part time work wanted full time work. The total number of hours that could have been worked, by those who wanted to work, could be on the order of something like 240 million hours a month. Something like the total number of hours in manufacturing.


I’m trained to think in terms of social cohesion, deficits, and full employment in a service economy. I’m  also trained to think in terms of automation, machine intelligence, and the creation of value. Software eats the world.

As a heuristic, as a point of public policy, we seek to minimize structural unemployment while tolerating frictional unemployment. We want as many people employed, productively, as possible. It’s an unquestioned heuristic. It’s just part of the paradigm.

For a bunch of reasons I won’t explain in painful detail, it’s very, very tough to have low interest rates, balanced budgets, and full employment in a service economy. You can have two out of the three. You can’t have all three.

Most societies chose deficits, even at the expense of damaging social cohesion by creating insider and outsider groups. The way American society chose to justify it to itself is different from the way that French society justified it to itself. Or the Dutch. Or the Germans. Or the Japanese. Everybody society has a weird myth about itself. Each Canadian province did the same.

But deficits can’t run for periods exceeding a few generations’ lifetime. A string of bankruptcies and debasements can’t last forever. There’s a supercycle to these things dating back centuries. There’s no fighting it.

We also want economic growth. We want physical living conditions to be better in the future than they are today. Again, sort of unquestioned. Clothing. Hammocks. Flush toilets. Enough to eat. These are all things everybody wants. Beef. Cars. A really nice place to call your own. These are all things many people really want. There are good debates about what social living conditions should be…the more sustainable the better.

One of the key ways we make our physical futures better is by making more output for every unit of input. Productivity growth is a key factor in making tomorrow better. Fewer human input hours to produce more goods is the essence of productivity growth.

Productivity gains through automation tends to destroy working hours in the sector getting disrupted.

Big disruptions are associated with sharp increases in income inequality. A representation of such inequality is represented by the Lorenz Curve.

We become, in aggregate, better off so long as the Lorenz Curve doesn’t distort too much. An extremely curvy Lorenz Curve is associated with social instability that, if left too long, results in insecurity for all.

You’ll hear a lot about mincome because it’s an answer to alleviating that pressure.

If this combination of aims – equal opportunity, managed social cohesion, free trade, and low deficits sounds familiar, it’s because it was the British answer to managing their productivity disruption in the late 1800’s. Outside of the United States, liberalism isn’t so bad. It’s quite a bit more right of center than socialism, and many of the left refer to classical liberalism as a right-of-center stance.

The Canadian answer to social disruption and a pressured middle class is nearly always ‘increase exports and increase development’. A massive middle class, especially one returning from a major war, can be effectively managed with a smooth Lorenz Curve and a basement to work on over the weekend.

The Disruption

A lot of people, at least in North America, just want to be left the hell alone. They don’t enjoy disruption.

Technological advancement doesn’t leave anybody alone. Ever.

So while stories of massive economic growth and the freedom from drudgery may be fantastic for some, it brings fear, uncertainty, and doubt to large swaths of the population.

More opportunities, more development of social capital, more technological advancement is embraced by a subset of the population, in part because it represents greater chances for specialization and progress.

Specialization is fantastic because it’s a natural form of scarcity. And making yourself scarce is, in general, a pretty good heuristic. It’s a good way to compete globally. Social capital development, the generation of scarce skills that go with you wherever you go, is a fantastic answer to globalization. (So much so that it’s been a backbone of Canadian policy for decades). It’s great for the individual.

It might not be great for communities that are getting hit the hardest by disruption.

That isn’t completely evident in manufacturing towns. The development of social capital means that young people take their skills with them into major centers which favour that specialization. Social capital generation results in youth drain and brain drain. Not great for some communities.

When factories leave for cheaper jurisdictions, or, when the total number of hours available in a community because more machines are doing more of the work, they leave a lot of disrupted people behind. They’re not being left alone. They’re getting harmed.

From a Canadian perspective, this is very, very bad. Because disrupted people tend to get organized and they start doing incrementally more disruptive things. If some people can’t find legal work, they’ll find illegal work. Others will organize in other ways.

Canadians can be wonderfully passive, polite, people. But if they’re disrupted too much, they turn aggressive. 1837 was the last time it went really, really sideways, and it remains deep in the ancestral memory. Canada managed to get through industrialization and de-industrialization relatively intact, in spite of its very deep cleavages, through quite a few very clever (if very deeply imperfect) policies.

Manage the disruption

Technological disruption is inevitable. The desire for economic growth is paradigmatic. Full employment is desirable.

From a macro perspective, what a fantastic opportunity to transition a half a billion hours a month to intensely interesting uses in a global economy. So many people. So much potential. If we could work out the infrastructure to connect smaller centers to larger centers, we really could ease the disruption, and get through it without the worst of the consequences. If we can manage the Lorenz Curve, good things can happen. If we do nothing about the Curve, it’ll makes us all the poorer.

From a micro perspective, you have some great opportunities and choices.

You got the ballot box. That’s your bluntest tool. You can choose to elect people who have a plan for addressing the disruption. If you’re finding yourself really disrupted, you could run for elected office.

You got choices on where you can invest your time, and how you choose to specialize.

It’s not anybody’s fault why disruption happens. It happens. How we choose to manage it, that’s up to us.