Henning Meyer’s August 13 essay examined the employment implications of computer-guided machines taking over cognitive tasks that previously were the exclusive province of humans. For the foreseeable future, Meyer reasonably assumed that paid work will continue to provide not only money income, but also a sense of purpose and a setting for social interaction. He wishes to support this and proposes better education, more flexible employment relationships, and public guarantees of employment or training to ameliorate the loss of cognitive jobs to smart machines.
However, the essay misses a background issue that advanced affluent societies need to address and that Keynes’ celebrated 1930 essay raised. If work for pay remains the primary venue for purposeful activity, what is it that workers are being paid to produce? If keeping people employed is to be a good thing, then the output resulting from their employment must also be beneficial.
What kinds of consumption will grow as a result of machines’ advancing cognitive abilities? Meyer advocates somewhat more public employment for things that the polity values but profit-driven markets don’t. He singles out the growing need for low-skill, in-person health aides for aging populations in affluent countries. Machines lack empathy and cannot render such high-touch services, which we have evolved to need as social animals. But, private markets never want to pay a living wage for this work, so government directly or indirectly must be the employer.
What other kinds of jobs exist that even very smart machines cannot do, and what is the pattern of household consumption that cognitive technology then ultimately fosters? Many jump to the conclusion that low-skill jobs are most vulnerable to mechanization, but Meyer’s example of home health aides refutes that. And high-skill jobs certainly aren’t immune to displacement either, given that algorithms now read x-rays, do basic legal research and land commercial aircraft.
Cognitive technology may have very uneven effects across both high- and low-skilled jobs, which implies equally uneven effects for the resulting production and consumption. Remember, though, machines may have a difficult time replacing some low-wage workers like hairdressers. They will suffer the usual cost squeeze for high-touch services identified long ago by William Baumol. They will face increasingly tough competition from the kinds of consumption that technology does promote – even though hair will still need cutting.
Of course, trying to foretell the composition of future GDP growth is extremely speculative, just like predicting which jobs will be vulnerable. However, there is one thing that we do know. If paid work remains the primary venue for purposeful activity in affluent societies, then most of the resulting income will be spent in private markets. Growing money incomes from paid work will promote the continued expansion of private markets for spending those incomes.
How Well Are We Doing?
Let’s look then at how well consumption from private markets currently advances wellbeing in affluent societies. What does the current pattern of market consumption tell us about the unending expansion of private market activity that would accompany the preservation of paid work? In particular, do the benefits of market consumption yield diminishing returns to wellbeing as Keynes suspected?
Staffan Linder showed decades ago that in societies with high average income the most important constraint is not money but time. We all get 168 hours a week – no more, no less – in which to enjoy our spending, and no one lives forever. Growth of market purchases arithmetically implies less time on average with any given purchase and getting less satisfaction in turn. Diminishing returns to wellbeing do not result from satiation, but from limited individual time constraining market pleasures.
Consumers try to save time by purchasing services, so that we don’t all have to be our own doctors or dispose of our garbage individually. However, most examples of service jobs that smart machines will enhance are intermediate inputs, such as management or marketing services, hired to boost sales of the final products households purchase. Tech-savvy marketing in particular is the business of synthesizing market desires we didn’t know we had, and the resulting satisfaction is fake. It is the relief that comes from first applying itching powder and then scratching.
Many final services that households do directly buy aren’t particularly amenable to enhancement with smart technology. As mentioned, it seems quite difficult to improve the productivity of some low-skilled services with algorithms. Gardening or domestic work, for instance, may never be mechanized and are simply the purchase of one person’s time by another, with no net time saving for society.
Differences in income and wealth drive the demand for low-skill services that one could just as easily perform oneself – irrespective of a society’s average level of market income. The weeds must be pulled and the beds made; it’s just a question of whose time gets used up, rich or poor. For such services, greater inequality could help preserve paid jobs that computers can’t master, but at the cost of subservience and time fragility for the majority.
Buying Promises We Don’t Understand
Purchases of high-skill consumer services face a different problem. We hire plumbers, auto mechanics, or doctors to save the time we’d have to spend becoming experts ourselves. Such specialized services can be a great benefit of progress.
However, the more we hire specialized expertise, the more we need to understand those same specialties to make smart service purchases, to get value for our money. With greater service specialization, the information gap between buyer and seller widens. For profit-seeking firms, this creates a growing incentive to provide specious services, i.e. money grabs disguised as market exchange. The more complicated cars become, the easier it is for the mechanic to pad the bill or make unnecessary repairs.
I believe that specious services constitute a huge portion of consumer expenditure in affluent countries like the U.S. Here, for instance, the biggest household spending increases in recent decades have been for finance, insurance, and real estate services and for health services. Neither sector has delivered value for money. The income of the financial sector has doubled while the allocation of capital has been poor and financial risks have been magnified, not reduced. Our mostly privatized health system is twice as expensive as other countries but delivers markedly worse health outcomes.
If half of this income stems from providing useless assistance, then the scale of specious services from these two sectors is measured in trillions of dollars. One must add to this everyday cheating, empty business promises, and commercial trespass that monetizes previously informal social interactions like sports. Just because money changes hands and contributes to GDP doesn’t mean that households get value.
Keynes’ Two-Headed Problem
If much of what makes GDP go up in affluent societies is empty or harmful, then preserving paid employment doesn’t resolve Keynes’ 1930 question: What happens when we achieve high productivity while also facing diminishing returns to wellbeing from market activity? Keynes predicted that people would choose less paid work and turn to the erudite non-market pleasures that he favored. People haven’t taken this path because market capitalism is self-reinforcing.
Evolution has programmed humans to strive and acquire. Furthermore, it is in any individual’s interest to get ahead of the next guy, because inequality awards winners exclusive pleasures and dominion over others. Though this is a zero-sum competition with a loser for every winner, it nonetheless fuels market growth. We all endeavor to be the restaurant patron rather than the server, and that keeps GDP increasing.
I haven’t the faintest idea how affluent societies might shift humans’ need for purposeful activity from paid work to something else. I don’t know how one might increase non-market consumption in peoples’ lives, somehow making it, not free, but unpriced. I do believe, though, that continuing to rely on paid work as society’s primary organizing principle will likely lead to more unsatisfying growth for the majority, because the nature of affluence itself limits market satisfactions.