Robots Won't Destroy Us: How Automation Creates Jobs
One argument against the idea of technological unemployment, offered by many people who today sincerely style themselves as leading defenders of the free market, goes that automation will create more jobs than it destroys but due to the nature of the market, the nature of those jobs is, if not fundamentally unknowable, functionally indescribable for the purposes of the argument over automation. How could a person a hundred years ago, the reasoning goes, predict the existence of jobs like “app developer,” “nuclear engineer,” or “diversity and inclusion consultant”? The jobs of the future will be just as foreign to us as our jobs would be to those of the past, and we do not lament the world’s present paucity of candlestick makers and buggy whip manufacturers. While these people are more correct than their opponents, It must be acknowledged that this argument for the market is so thoroughly uncompelling that you might think it was originally concocted by its enemies. One may as well say that after automation, we will all get well-paying jobs with vacations, pension, and so forth once we enter the New Jerusalem, or after we achieve full communism. In fact, far from being unknowable, the types of jobs created by automation are highly categorically predictable. Automation in the production of higher-order goods directly creates jobs in producing lower-order goods that require those same higher-order goods as inputs. Automation in the production of consumer goods both increases living standards and makes human laborers more price competitive relative to machines.
When a firm or industry is automated, jobs will be created in precisely those industries which use, as a factor input, the good or service whose production has recently been automated. This is the case regardless of the specificity of the input. For example, if there is a breakthrough in the production of semiconductors, more computers can and will be made, whether by existing firms or by new, increasingly niche ones. Likewise with electricity itself. Cheaper electricity means that nearly every firm on the grid faces reduced operating costs. Firms that otherwise would have been unprofitable suddenly become profitable, able to be brought into being by a sufficiently alert entrepreneur. The first place a recently replaced worker should look for work is at his previous employer’s client firms. In economic terms, he should attempt to shift to producing a lower-order good in the same supply chain.
Consider a world in which a full half of working men are employed mining for coal, which provides most of the world’s electricity. One day, a nuclear power plant goes online, more than doubling the nation’s energy production and selling its power at an order of magnitude below the previous price. The plant only employs a couple dozen engineers. Over the course of a year, a supermajority of coal plants and mines shut down, and a mere fraction of the coal miners shift over to mining and transporting uranium, which has over 150,000 times the energy density. Just as it is obvious to any observer that society has plainly been made better off through the proliferation of affordable, emissions-free electricity, it should also be intuitively clear to all but the most stubborn of antifuturists that the loss in jobs amounts to little more than a speed bump in the economic lives of the newly unemployed coal miners. This is because electricity is a factor input in nearly all lines of production in any modern economy. Cheap electricity creates jobs, because it makes previously outlandishly expensive projects suddenly potentially profitable.
Importantly, the productive opportunities created will always exceed the amount destroyed, because technology is only ever adopted, in market economies, when it is profitable to do so. More wealth is created than is consumed, and so there are more resources to be combined in potentially productive ways by any entrepreneur who detects the opportunity. Unprofitable technologies, for readers wondering, are usually adopted when countries are attempting to pursue a policy of import substitution. For example, Nazi Germany’s prewar attempts to wean firms off their dependence on imported oil by forcing companies to use an ersatz oil made of liquified coal. Another example would be American subsidization of “renewable” energy sources like wind and solar.
But what about automating the production of the lowest-order goods, a.k.a. consumer goods? Surely jobs automated out of the yo-yo factory are essentially gone forever, since no businesses, aside toy stores, will see their expenses fall thanks to a reduction in the price of children’s toys. This is correct. However, it is important to remember that consumer goods are a factor input in the production of labor, and labor is a factor input in the production of everything that hasn’t been automated. People are, in terms of opportunity cost, now cheaper than new machines, which produces yet more firms willing to hire them. The automation in the production of pure consumer goods, to the extent that there are such things, represents real wage increases, or an increase in the standard of living, for everybody else in the society.
In other words, automating the production of yo-yos makes all human workers marginally more price competitive relative to machines in other industries, by decreasing the cost of living. Here it is also important to remember that consumption does not create jobs; savings do. If consumers spend less on toys and stow the rest at the bank, this is not in any way a waste. That money enters the loanable funds market. Toys are cheaper, parents have more disposable income, which is more money to put in the bank. More savings lower the natural interest rate, making that money available to spend on other goods and services, in either case raising the demand for labor. Where the decreasing cost of living improves the relative standing of humans in comparison to machines, increased investment raises the total demand for labor, whether in human or robot form.
Importantly, this logic applies to all service sector employment. Among non-Austrians who concede the net positive consumer effects of automation in the past, the service sector is conceived as something of an “employer of last resort.” But, they argue, what happens when waitresses, doctors, garbagemen, secretaries, bankers, lawyers, and economists are also automated? For the survivors of this imagined labor market decimation, this will be an age of unparalleled material prosperity, but what about for everyone else? Will I change my tune then?
One cannot “run out” of job openings, even in the service sector, in a growing economy, no matter how many robots are made and sold. All entrepreneurship in the service sector represents a cost-of-living decrease, or real wage increase, for anybody who consumes the service. Automation of service sector jobs, assuming it is caused by market conditions, will always improve the competitiveness of human labor relative to robotic, and this increase in available resources will raise the number of potentially profitable ventures at every stage of production, increasing demand for labor in all of its forms.
Finally, I should emphasize that this seemingly flawless process I describe is the process of economic development. It seems too good to be true, but it isn’t, and in the end we know it is true because some countries are richer than others by orders of magnitude and people from poor countries risk their lives just to live off the scraps of rich ones. Technology improves living standards, because all development that can be appropriately called development is made with an eye toward final use by consumers.
This should not be taken to mean that all automation is worthwhile. Automation, like any other entrepreneurial process, is useful and good only to the extent that it efficiently allocates resources to satisfy consumer ends. The only way to know if this is actually the case for a given technology is the market test of profit and loss. There are plenty of ways for the government to force or push the adoption of a technology. One of the most obvious is through the introduction of a minimum wage causing the replacement of human labor with self-serve kiosks. Different forms of regulation and subsidy can deform price signals such that more valued resources are squandered in the development and deployment of a technology than were produced or saved through its implementation.
When one encounters something one does not understand, one should not fear it, but neither should one begin to mystify it. And many defenders of technological growth do mystify that which they appear not to understand when they advance arguments involving “fundamentally unknowable” gains from technology. But these gains, categorically, are knowable. Technology creates productive opportunities “downstream” of its deployment. Those opportunities are not, and can never be, entirely taken up by more robots, because machines’ very existence as goods in a market economy increases the capabilities and competitiveness of humans.