A response to
Arjun Jayadev’s argument about the impact of automation on our work and life
”Robots will save us!” — Owners
”Robots will kill us!” — Workers
Arjun has written and posted an article on robots. What caught my immediate attention was the implied optimism: The robots won’t kill us and might save us.
I would like to take the discussion further because i think the issue is so important and we need a narrative about automation and employment that is the best guidance for society given current knowledge.
”…….some optimism that jobs are not disappearing, although we could be working less if we prefer to.” This implies that jobs are not disappearing (nor the income) and leisure is near by if we want it .
But what if jobs are disappearing or hours cut? Working less means trouble for most people if they are locked into a mortgage or costs of children and health. Only a different kind if culture could make this negative of declining income a true positive, and not just for owners, (Arjun raises this question at the end.)
Several questions economists should be clear about:
- Are jobs disappearing?
- Could we work less if we preferred to? Is personal choice actually plausible?
So: jobs disappearing. We have the standard idea common in economics that employment is cyclical. Excess labor will generate opportunities that will lead entrepreneurs to create new business, so work and capital will rebalance.
The standard example is the shift from agriculture to industry.
“So hey, something new will lead to the reemployment of all the displaced workers.” Why assume this will happen again? “Well, because it happened last time.”
That is not a very good argument. It implies Hume’s induction argument for believing the sun will come up tomorrow. The implication in standard in economics is that the economy is a physical, natural, system that seeks equilibrium. But it is not. Economy is the place where entrepreneurs are trying to change it, not to seek equilibrium.
Lets use a little economics. Under digital conditions, if demand goes up, assumed as necessary to create jobs, the typical manager is incentivized to find the cheapest adequately trained employees. But now that is going to be a robot, an algorithm. Increased demand incentivizes a search for automated “workers”. In digital society elites will replace workers with robots. The worker owned, no salary. back to slavery but now 24/7. The possibilities of a more leisure society were co-opted by the 1%.
In agricultural society elites needed agricultural workers.
In industrial society elites needed industrial workers.
In digital society do elites need workers? It Is a partly open question, but trends suggest we have a problem.
We see it coming rapidly in transportation with autonomous cars and trucks. Fewer drivers, fewer repairers fewer police, fewer hospitals, and the insurance companies are in trouble, as is the blue of roadside business.
We see it in agriculture where drones are already examining the condition of the soil and autonomous vehicles delivering seed and fertilizer based on automated soil analyses.
Law firms are not hiring interns. Case analysis and reference gathering already highly automated.
Software is on top of creating software that can write software. Robots repair robots. Managers, used to coordinate, are replaced by scheduling software.
Some argue that it is already menial work that will be replaced by machines but not mind work. Others argue the opposite, that the sophisticated skills of modern education are in fact the easiest to automate, to write down in software. The truth is both: any work that can be routinized, mental or physical is subject to software and automotive solutions.
Ah, what of service? Why have my computer through me talk to a technician who looks at computer screen to figure out what’s wrong with my computer? Just have the computers talk to each other. Service work is coming to be contracts by the day or even the hour. Workers are being forced to put together their income by the day.
The article says “labor and capital have been complementary over the long run as output growth has absorbed more labor.” Which long run? We only have two possibilities. The first is the shift within the industry over the time of industrial society. As population grew, demand group in the workforce could grow. The kinds of jobs reminder relatively constant. The other “long-term” is either the shift from agriculture to industry, or from industry to digitalization. The first shows the effect of the absorption of labor but the second is still an open question, though logic suggests we are in trouble for employment.
We do have the problem of productivity which complicates the analysis a bit. Productivity smells like a great thing: let’s have a more productive society! But the word productivity has been co-opted by management measures which basically measure productivity as an increasing percentage of the process going to owners and the smaller part going to workers.
Which raises the issue, why did most of the increase in productive capacity go to owners and not to workers? Who decides who gets the benefits? I believe the answer is relatively mundane: middle level managers are much happier getting approval from higher up managers by delivering a constrained budget and not an expanded budget which includes increased salaries for workers. The result is that any gain in the organization moves upwards, no downwards. The result over time is the shift of wealth upwards, which of course is the result we see empirically. Of course other things like anti-union, or how much more legal it is for capital to organize than labor.
”Average of real wages of US workers has increased six fold over the 20th century.” But that was mostly in the post World War II period until about 1970. Certainly wages did not rise much through the depression until the war. There is also a problem with “average real wages” because of inflation and the shift of the average towards the high-end with the higher salaries at the top and lowered salaries at the bottom. If we accept these numbers at all, how many fold did the incomes at the top increase?
The article says “We have yet to see the robot revolution displacing workers on a macroeconomic scale.” This may be true but I doubt it. Since we (taking the US) have relatively high unemployment among youth who should have gotten old style jobs, among the middle class which has been hollowed out, and those who have jobs are more part time and lower wages. Is this not in part technological unemployment? It started with the back room accounting offices, telephone operators, managers who did low level coordination of people and numbers.
Then we get “small businesses have found that a reduced work-day may have resulted in less absenteeism, reduction in staff turnover and increase in productivity sufficient to offset the potential cost of hiring more employees. As one business owner put it ‘We thought doing a shorter workweek would mean we would have to fire more, but it hasn’t resulted in that because everyone works more efficient;y.” Note how this is all good from the owners view. But the workers? Fear of losing the remaining job? All negative. In visiting a restaurant or store or service place I try to do a little interview with some of the workers. Insecurity has increased as has stress because the trend is lost jobs and working harder to keep one. All good for “productivity” and the boss.
Listen again to the article, “Increased output can and should produce a decline in work hours. If leisure time is desirable, then over the long run as output and income grow…” Or is it as output grows and income declines? His sentence then ends with “working hours should fall.” OK. Now what is there in the economic or political process that will push income to rise as hours work fall? So far, nothing. We will live with declining hours, much tech induced, and declining income for workers and rising income for owners, for the foreseeable future, and that may just turn out to be in the long run.
Article says “Why do we not take leisure time .. Can we work less, enjoy life more and not suffer large declines in income and output?” Climate urgency may force us to part of this: rethinking output. But we also need to rethink how income is distributed. Other than in the 1%, who in current circumstances can “take leisure time?”
Of course all this might be upset by war, plague, earthquakes, agricultural failures. Coping with the future will require a smart use of new technologies, but if those technologies are used to get rid of workers, increase profits and wealth concentration, then we will not realize the positive possibilities of tech.
Economists educated ad nauseum to the idea of equilibrium have a bias to believe new tech, destabilizing, will re-constitute itself with what is called in physics a “coefficient if restitution” of 1.0
P.S. another example of the march of robots into new territory:
Sewer robots sampling human waste may track drugs, disease through cities
By David Common, CBC News Posted: May 27, 2016
“Soon enough, robots may wander permanently in our sewers, just below our homes and neighbourhoods, analyzing our diets and our health as they suck up what we flush down. It’s already started under the streets of Cambridge, Mass., venerable home of the Massachusetts Institute of Technology. Canadian architect Newsha Ghaeli oversees a team of lab-coated MIT research fellows staring down an open manhole cover into the oozing sludge below. Down there are the answers, and the hope of detecting viral outbreaks long before doctors are even called, using a system of sensors that quickly and inexpensively monitors public health.”