Systems Thinking My Way Through: Automation
Automation has kind of become a meme lately perpetually alternating between “The robots are coming for your job” and “No they’ll create jobs” It’s been going on for years, it’s boring, and I’m tired of talking about it… So I’m going to talk about it for about 10 minutes.
First off I want to acknowledge there are smart people who I admire who all have completely different takes on this issue. I’m not here to elevate any one in particular, but I do want to bring up some points that deserve more attention.
First off there’s the notion that this is all the Luddite fallacy, the idea that jobs don’t get destroyed they get replaced so technological unemployment is a wash in the long run. It’s based on the Luddites in the industrial revolution. English mill workers were being replaced by machines, and they protested this new machinery by destroying it. I’d like to use them as a lens to look to the future, because there are some parallels.
There was a neat little piece in the Smithsonian Magazine I used as an undergrad in 2017 called When Robots Take All of Our Jobs, Remember the Luddites. It offers a bit of a different perspective on Luddites who have a pretty bad reputation historically.
When Factory owners introduced machines… “The workers were livid. Factory work was miserable, with brutal 14-hour days that left workers — as one doctor noted — ‘stunted, enfeebled, and depraved’… Poverty rose as wages plummeted.”
“The workers tried bargaining. They weren’t opposed to machinery, they said, if the profits from increased productivity were shared. The croppers suggested taxing cloth to make a fund for those unemployed by machines. Others argued that industrialists should introduce machinery more gradually, to allow workers more time to adapt to new trades.”
It’s interesting to see similar ideas returning now that we are haunted by the spectre of a Fourth Industrial Revolution.
“At heart, the fight was not really about technology. The Luddites were happy to use machinery — indeed, weavers had used smaller frames for decades. What galled them was the new logic of industrial capitalism, where the productivity gains from new technology enriched only the machines’ owners and weren’t shared with the workers.”
With 200 years of extra historical perspective it’s easy for us to say these simpletons should have known that these machines were creating wealth that would benefit everyone in the long run. However it might be good to keep in mind that if you only have a few years left before you die of diarrhea, long term consumer pricing benefits are probably not your first concern.
The article goes on to say “Their trade destroyed, most eked out a living by carrying water, scavenging, or selling bits of lace or cakes on the streets.”
So the concern of the Luddites was that productivity benefits wouldn’t be shared. (1/1) Their wages would go down. (2/2) They might lose their job or get a worse job. (3/3) Ha. Idiots.
So what’s the point? For one we don’t live on an average. If a coal miner loses his job and someone else gets a job in web design… the coal miner still lost his job. That doesn’t make coal good or web design bad it just means that economic growing pains exist. Even if they are short term they still hurt. Acknowledging that does not make you anti-technology. It makes you anti victim blaming. Economic mitigation of technological unemployment is justified.
Let’s take a quick look at what the Luddites were dealing with.
As mechanization increases efficiency increases. Price of a good will usually decrease. I thought about connecting the price of a good to profit with a negative relationship, but I don’t think it works. It would make sense for revenue, but specifically for profit there’s too many lurking variables to make a confident connection.
So, capitalists can be more productive with less resources, which makes them more competitive in the marketplace, and make higher profits. I’d like to point out this is generally not optional. If they don’t do this they may go out of business because someone else will. Losing everything and becoming a worker like everyone else must be avoided at all costs.
Mechanization also means you need less workers to produce the same output, and unless there’s a ton of demand to supply you’ll cut some workers. If all factories are mechanizing at the same time unemployment goes up which means workers are easier to replace so wages go down. Wages go down, profits and poverty both go up. More poverty means more worker anger. Mechanization also obviously leads to alienation since work becomes more and more specialized and repetitive.
Now let’s look at their proposed solutions. (In Green)
First we’ll look at what they wanted the most. (And were never going to get.)
Workplace democracy: Where workers could democratically control what happens with the value they create. Even if their specialization and repetition remained this would reduce alienation in at least one sense. They would likely elect to increase wages (or work fewer hours). Higher wages lead to lower poverty, and less alienation leads to less anger.
Taxes & welfare: Taxes would likely reduce profits and increase the price of the good and they would be used to fund welfare that would reduce poverty and therefore discontent.
Regulation: Slowing mechanization would probably face the most pushback today. That knocks out the issue at the source but you also unnecessarily miss out on the potential benefits of higher productivity and wages, and lowered prices.
Modern Economy
So that was the warm up. A brief glimpse of automation 200 years ago. Now comes the hard part, thinking about the future. To frame it there’s a few things I want to highlight.
First off we live in a precarious system. Unemployment in the Great Depression wasn’t 90%, wasn’t 50%, it was 25%. In the Great Recession it barely grazed 10%. The now ubiquitous Oxford study from 2013 estimated 47% of US Jobs cold be computerized in the next decade or two. I don’t know if that’s true. All I’m saying isif it’s even half true we would be in the most severe economic crisis in almost a century.
When listening to those who are most concerned with automation, their emphasis is often not just on the quantity of jobs that become obsolete. They’re more focused on the rapid, even accelerating speed at which it will happen. It’s more complicated than just just quantity of old jobs vs quantity of new jobs. For example if 50% of jobs are automated in 50 years and an equal number of jobs are created we should make it out OK. If 50% of jobs are automated in 10 years, even if we create an equal number of jobs it’s hard to imagine that going smoothly.
To elaborate on that if we’re turning over not just jobs but entire careers every few years at a time when a 40% of people can’t afford a $400 emergency, education costs are at an all time high, and employers are reluctant to train their own workers, it’s hard to imagine not having a crisis of some sort. We haven’t granted ourselves much wiggle room here.
Of course I’m aware of the absurdity that higher productivity could somehow lead to more poverty, but this is the situation we are in. Stephen Hawking’s last Reddit AMA is evergreen.
I guess you could say “There must be something rotten in the very core of a social system which increases its wealth without diminishing its misery.” If only I could recall who said that. Moving on…
First I tried to make a systems map of some issues facing the modern US economy. After this we’ll look at how automation and potential reforms would fit in.
We’ll start in the middle. Two elements that could theoretically protect Labor power are unionization and labor laws, (the red in the middle.) neither of which are particularly significant factors in the US at the moment. Link
On the other hand, elements that detract from labor power. Capital mobility, offshoring, outsourcing, etc. have taken off in the past 50 years. That’s the cream colored stuff on the left. Link
When the labor market becomes more competitive people want to be more professional and more educated to get better jobs. More people want to go to college, leading to higher tuition and higher personal debt. (Light blue on top) Higher personal debt of course leads to long term declines in consumer spending. Link
Now we’re going to move to underemployment see what that does. That’s the cream at the bottom and moves up to the dark blue or purple. (I’m colorblind and don’t feel like phoning a friend.) Underemployment increases, more people are in part time work or in the gig economy. As a result they likely do not have the benefits they may have had. (They may not have had them anyway.)
When underemployment becomes chronic, we have a labor exodus, as people stop looking for work. Link This labor exodus is contributes to coexisting low unemployment wages since people who have given up looking for work are not counted as unemployed. Link
This exodus is exacerbated by the fact that one thing which actually has gotten cheaper is entertainment. Link Flat screen TV’s and Netflix are tempting ways to pass the time if you have low expectations anyway.
Lower wages, a labor exodus, and more money going to healthcare expenditures, all lead to a decline in consumer spending, which theoretically leads to a decline in corporate profits, although empirically they still seem to be managing for now. Link
So now let’s look at the same map in a future of automation with a few potential reforms thrown in.
I tried to stick mainly to the reforms that are within the realm of discourse today. You can say these aren’t particularly imaginative, and that they only tweak the system, which is true. Radical new systems don’t fit on the same systems map. That’s kind of the point. They’re new systems.
Starting at the top, going counter clockwise.
1. Universal Basic Income: It has become a trendy topic that intersects economists, and futurists. Its main impact here would be maintaining or increasing consumer demand. And I’m assuming it’s being implemented in good faith and it isn’t just an excuse to cut other forms of welfare.
2. Free Public College: It would make attending university more financially attractive, so enrollment would increase. Personal debt from university would go down. Tuition would become functionally cheaper for students, but not necessarily cheaper overall. State funded education could vary significantly based on the retraining needs of workers and employers in a future of rapid automation.
3. Nationalization: It would obviously reduce corporate profits. It’s not really a popular idea at the moment, but I think it is a possibility. It would probably only happen on a large scale, as a response to an existential threat like war or climate change.
4. Global Tax on Capital: In the off chance that Piketty’s idea gains traction, this best fits as a reduction of corporate profits. It’s not perfect but it works well enough here.
5. Worker Hours Reduced: If productivity continues to increase, a potential response could be to reduce the expected working hours of the population as Keynes predicted.
6. Automation: This process is always happening to some extent, but if it happens rapidly on a broad scale, it would have the most impact. The rapid increase in structural unemployment is what would likely send the system into chaos.
7. Right to a Job: the idea has been injected into US political discourse for the first time in decades. If it became a reality, it would increase the power of labor, and reduce underemployment.
8. Single Payer: This change is probably coming regardless of automation, but ensuring healthcare regardless of employment would be an even more significant element in a system potentially stressed by unemployment.
There are plenty of ways to minimize growing pains and each will have their own side effects. (Including effects we can’t see here.) I don’t see any silver bullets here; each has its own potential drawbacks. I think the main takeaway though is that even without much imagination we have enough ways to mitigate whatever economic growing pains we may face to be optimistic about the future of work. And each solution listed fits differently into the map meaning we can be somewhat surgical with our reforms. What concerns me is more what Stephen hawking said. Solutions can be as clever or as ham-fisted as we like. The hard part is actually accomplishing them politically. Each dollar spent on relief for workers has to prove it’s a better cause than more planes that don’t work and ships that keep crashing into each other.
