"Kee Points" with Jim Kee, Ph.D.

Concerns regarding automation and worker displacement, which have been around for centuries, are currently making a comeback. Here are the basic economics of the subject (I apologize for the length!):


Most people have heard some version of a Milton Friedman story in which he was visiting an Asian country (it was China) and watching a canal being built. Amazed that the men were working with shovels instead of modern excavation equipment, he asked his government tour guide why, and the response was, “because it is a jobs project.” To which Friedman responded, “oh, I thought you were trying to build a canal. If it’s jobs you want, you should have the men dig with teaspoons.” The story does a pretty good job of making a few often-overlooked points in economics. The first, which goes back to Adam Smith (circa 1776), is that the wealth of a country consists of the output of goods and services that its countrymen have access to. Capital equipment such as excavation machinery allows fewer people to be involved in the building of a canal, which makes them available to produce other things, hence a greater output of goods and services, i.e. greater wealth. Since the world is characterized (by economists anyway) as one of unlimited wants and limited means, more machines assisting labor should be welcomed, because wants are never satiated, but the supply of labor is always limited. More people digging canals means fewer people building, say, aspirin factories. Another way of saying this is that the fear that automation will destroy jobs fundamentally rests upon the assumption that there is a fixed amount of goods that buyers want, so machines that increase the output of goods per worker leads to fewer workers employed. I’ll go with 20th century economist Frank Knight on this one: “The chief thing which the common-sense individual actually wants is not satisfaction for the wants which he has, but more, and better wants.” Figuring out what those never-ending “better” wants are is a key function of capitalist economies. That is, figuring out what to do, according to economists, is at least as important as figuring out how to do it.


Farming is a common but nevertheless relevant example. In the 1870s agricultural workers comprised half of the work force. Today that figure is closer to 1.5% (St. Louis Federal Reserve: US Census). That means fewer people working in agriculture (but greater output) and more people are producing other things (like semiconductors, software, turbine engines, etc). And, we are wealthier (i.e. more of all goods and services). The reduction in manufacturing jobs (but not manufacturing output) over the past 40 years is somewhat similar - fewer people required for manufacturing means more people available to produce other things.


As an aside, there is a human element here as well, because machines relieve human beings from much burden. An example is the horse-drawn plow replacing human-drawn plows, and tractors replacing the human walking behind a plow. Human beings are always in the sights of being replaced by mechanization, as machines operate without breaks and with a high degree of precision. In fact, the “alienated worker” turning bolts on an assembly line (for example) is always in danger of being replaced by a machine to the extent that her job is like a machine’s. In a sense capitalism forces workers to either pursue “humanized” functions (requiring interaction, communication, judgment) or face being replaced by a machine. You probably weren’t taught that in history or social science class but, as economists say (e.g. Nobel Laureate Frederick von Hayek), capitalism’s history was written largely by its critics. In fact, economist Deirdre McCloskey points out that the term “capitalism” itself was coined by critics of decentralized, price-directed market economies. McCloskey prefers the term “market-tested betterment” over “capitalism,” which is worth a few moments of contemplation.


M.I.T economist David Autor has elaborated on automation by asking why, “despite a century of creating machines to do our work for us, the proportion of adults in the US with a job has consistently gone up for the past 125 years?” He uses the example of US bank tellers, the number of who have roughly doubled in the 45 years since the introduction of the first automated teller machine (ATM). That’s because, while the number of tellers per branch fell by a third, it became cheaper to open more branches, so more branches meant more tellers. But they are doing different work than they used to do, becoming less routine clerks and more, “like salespeople, forging relationships with customers, solving problems and introducing them to products like credit cards, loans, and investments: more tellers doing a more cognitively demanding job.” That’s the pattern of thousands of years of history, from an axe replacing people digging/chopping with rocks, to nets replacing catching fish by hand, to traffic lights substituting for traffic cops and smartphones replacing AAA travel route services. These are all “robots,” argues McCloskey, designed to abridge labor: “Any contrivance substitutes for labor, equivalent to a robot.” She argues that if such substitutions resulted in permanent unemployment, then the unemployment rate would be 95 percent and headed up, rather than below 6 percent and headed down. That’s technological advance, and as UCLA economist Armen Alchian argued 50 years ago, more people are alive today because we are not trying to survive on the technology of 1870 or 1970.


Another aside: It was Alchian (who’s economics textbook is almost a cult classic) who pointed out that wage policies tend to hasten automation. Referring to elevator operators, he noted that union demands for higher wages at first (given the current elevators in place) improved the well being of the operators. But as the machines wore out, and given the now higher operating costs with the higher wages, it made sense to invest in more expensive, automated elevators that required no operator. As operators were laid off, their plight was blamed on “automation,” when it was, in fact wage policies. Interestingly, Alchian also pointed out that elevators are faster where people’s time is more valuable (city office buildings) than they are where it is less valuable (e.g. rural parking garages), the idea being that high speed elevators are more expensive to install and have higher maintenance costs.


But none of this helps the person being replaced by capitalism’s innovation. That’s the trade off, or, “the deal”: Higher standards of living that come with capitalist innovation, but at the expense of continually displaced workers. In fact, it is helpful to think of any new automation or technical advance that replaces the old in terms of three groups of workers affected: (1) the group making the new, like televisions and computers (replacing radios and typewriters), who gain because their skills and knowledge are in high demand, (2) the group of workers unrelated to the industry and unaffected in their employment, but who gain because of the newer/lower cost/better performing product or innovation, (3) the group of workers in the displaced sector (e.g. radio/typewriter workers). So the focus should be how to help that last group, the displaced workers, which shouldn’t be rocket science.


One suggestion is a universal minimum or basic income (UI), formerly referred to as a “negative income tax” or NIT (recently resurrected in a 2014 paper by the Federal Reserve Bank of St. Louis). That solution is conceptually easy to administer but ignores the familiar economic problem of how to provide assistance without creating dependency. That’s why economists have favored some form of unemployment insurance, though it is hard to monitor, subject to fraud and abuse, and not really insurance. There are probably 1001 ways to structure it, but I am partial to proposals that are more akin to whole life polices, i.e. that contain an insurance component and a cash value or investment component. For example, workers could pay into such a plan at the beginning of their careers and basically be covered by the insurance component, which might include 3-6 months of unemployment benefits. Over time, as their pay-in increases and if they’ve stayed employed, used savings in tough times, or otherwise avoided making a claim, they start to build a cash value component that can be used any time…to change jobs, retrain, or even take a work sabbatical. The idea is to create better incentives by invoking some form of true ownership of benefits.