Dynamic Equilibrium In A Society Undergoing Mass Automation (Don't Worry - It'll All Work Out)

With the advent of machine-learning, machines will be able to take on increasingly complicated tasks, and this has raised fears of a future where humans need not apply.

In this blog, I will argue that long-term mass unemployment is an unlikely scenario and is nothing to fear.

First, let's paint a picture of the worst case scenario that most people who fear automation would imagine. Machines replace 90% of the jobs, nobody has any money to spend, and the entire economy collapses.

So even in the worst case scenario, an unsustainable system, surprise-surprise, is unsustainable and inevitably collapses, and it will be replaced by one which is sustainable. Like rabbits proliferating out of control, running out of grass to eat and then dying off in mass, so too will the machines out-compete humans, proliferate out of control and end up killing off the cash flow that sustains them. The factory owner would have no choice but to sell his machines at a massive discount or simply leave the machines to be picked off by scavengers, thus levelling the distribution of capital. Industry will begin anew with a more level playing field, and a more sustainable economy will emerge where more people will have access to the benefits of automation.

But that's the worst case scenario. In a free market where prices adjust according to market demands, we will not see a sudden spike in long term mass unemployment. Forget the historical precedent that backs up my assertion. Just think about the dynamics of pricing in a free market economy.

As automation displaces more and more human workers, the price of goods and services will drop. Because of a lower cost of living, humans can now afford to offer their labour at a lower price. Also, more people seeking jobs will further reduce the price of human labour. At the same time, as more and more companies seek to automate their production, the price of engineers will rise along with the price of computers, robotics and other materials needed to build machines. Cost of making and sustaining the machines goes up, cost of making and sustaining human beings goes down, and the cost curves will eventually intersect. At that point of intersection, we have something called dynamic equilibrium, a term borrowed from chemistry. During dynamic equilibrium, the rate of workers losing their jobs to machines will equal the rate of workers getting hired in other industries.

Now imagine yourself an entrepreneur in a free market economy where mass automation is possible but not yet fully implemented. Human workers are expensive, hard to train, and have many drawbacks, and you have the option to purchase an android that's more efficient, won't give you attitude and is willing to work without rest. Of course you'll go for the android.

Fast forward a decade or two. There is now an excess of cheap human labour and a shortage of androids. And no, the androids can't simply make more androids because the cost of their parts will go way up whereas the cost of food will go way down. The rabbits will have eaten up all their grass, and the humans now have plenty of rabbits to eat. In this environment, let's say you want to start a new business installing solar panels. Even though the android would make a better worker, you can't afford the android. You have to hire the human. Even if you can afford the android, hiring the human will give you a better profit margin.

At this point, you might argue that, even though mass automation might not necessarily lead to sustained mass unemployment, the decline in the price of human labour would cause a decline in living standards. Firstly, the decline in labour price would be softened by the decline in prices in general. Whether or not this would be sufficient to offset the decline in labour price, time will tell.

Second, I would point out that the pie is now bigger. From a resource standpoint, our society would become richer due to automation. You might argue that the distribution of resources would be skewed in favour of the few. That's not a given, but let's assume that's true. What will they do with all that money? There are three options: 1) Go on a crazy spending spree, 2) Horde it all in a secret vault, or 3) Keep growing their business. If they go on a spending spree, they're creating jobs and money will flow back to the wider economy. If they horde their money, then it's doing them no good, and by keeping their money out of circulation, they have increased the value of your money through deflation. If they continue to grow their business or invest in other businesses, then this creates more jobs and money will flow back to the wider economy.

Actually, there is a fourth option. The rich could use their money to buy the influence of politicians and radically transform society in such a way that would screw over the people. That's the main reason why people are concerned about 'the few' become richer and richer. Wealth plus politics equals power. But then that's an indictment of human nature, not of automation. Automation is not the problem. Individuals who seek power for the purpose of abusing said power are the problem. Going after the tools they use, whether it be wealth or machinery, does not solve the problem.

In summary, automation will push down the cost of living and wages to the point where humans are able to compete with machines, thus we will not see sustained mass unemployment. Society will become richer as a result of automation. The distribution of wealth is likely to be skewed in favour of the few, but this is consistent with the majority of recorded human history and is not the result of automation. And at the end of the day, I, for one, look forward to welcoming our robot overlords.