Automation and Employment--Posner

The question I consider is whether automation, technical progress, outsourcing, and the increasing complexity of products and services—all closely related phenomena—are reducing demand for workers, especially but not only those who do not have a high IQ. 

There is a long history of exaggerated concern, rightly derided by economists, with the replacement and hence displacement of workers by machines. The idea at its simplest is that labor-saving devices whether in home or factory production, or in services, displace labor, leaving workers unemployed. The logic is simple. Over the course of my lifetime I have seen an enormous improvement in the quality of razor blades. A razor blade used to be good for one only shave; now it lasts weeks, even months. So fewer razor blades are made and therefore fewer workers are required in order to make them. The result is a worker surplus—but the logic that predicts this result is flawed. Workers made superfluous by technical advances constitute a pool of labor available to producers of new products and providers of new services. The existence of the pool places a lid on the price of labor, and therefore on the prices of the new products and services (since labor is an input into those products and services). Low prices stimulate buyer demand, which results in greater production, increasing the demand for labor. For centuries incomes have been rising in tandem with technological progress. 

There has long been concern that automation of manufacturing was eliminating high-paying factory jobs, propelling the factory workers no longer needed into lower-paying service jobs. But this concern, too, is exaggerated. It overlooks the fact that higher pay for factory work is compensation for the fact that such work tends to be strenuous, monotonous, dirty or unhealthful, and dangerous. Compensation for disamenities is not a component of real income but an offset to a cost. It is the obverse of long vacations, which are an amenity of teaching and therefore a component of teachers’ real income. 

But there is nothing inevitable about the virtuous process whereby automation and related negative effects on particular jobs merely shift workers to other jobs that are equally or more desirable. Workers may be highly compensated for possessing human capital that is specialized to a labor market that is shrinking. We are seeing this in law today. Because of automation, outsourcing, and more efficient management practices, the demand for lawyers is down, forcing many lawyers either to drop to lower rungs in the profession’s ladder or to leave the profession entirely for work in types of job which their human capital specialized to the practice of law has less, or maybe no, value to employers. 

There is nothing new about this phenomenon, and in the long run labor markets adjust—fewer persons are applying to law school today and so eventually the lawyer glut will end. But at the same time there is no guaranty that labor markets as a whole will adjust smoothly to changes in demand in particular labor markets. That depends on the size and rapidity of those changes, and in turn on the size and rapidity of changes in automation, outsourcing, and management practices, all linked to technology in a narrow or broad sense: automation obviously, but outsourcing (for example of legal research to lawyers in Indiana) because of improvements in communications technology (especially online) and management practices because of advances in psychology, management science, human resources, computers, and communications. 

If technological advance is very rapid, causing in turn a large and very rapid drop in demand in a large labor market, the economy may not be able to absorb the sudden surplus of labor in a short period of time. The result will be soaring unemployment that will retard normal market processes by reducing incomes and in turn production and therefore in the demand for workers. Take the case of the driverless car. This technology is advancing very rapidly and has great promise for reducing labor costs (drivers), traffic accidents (for example by eliminating drunk driving), traffic violations, and traffic jams (by optimizing speed, lane usage, and choice of routes and times). There are approximately 4 million truck, taxi, limousine, and bus drivers in the United States, not to mention gas station attendants and traffic policemen. Not all these jobs will be eliminated overnight, but they could go quite fast. 

Apart from specific job categories, technological advances in products and services, along with greater outsourcing opportunities and free trade with other technologically advanced nations, increase returns to IQ and educational achievement relative to other worker qualifications such as strength, quick reflexes, and physical fortitude. The design, production, and control of robots require intellectual qualities that are not required in factory workers. Though there is evidence that IQs have increased over the last century and may continue doing so, and though some day it will be possible to increase IQ by altering brains, technological advances may continue for some time to increase the wage premiums for high-IQ workers and reduce wages of average- or low-IQ workers, thus increasing the rate at which inequality of incomes is growing. Not that the jobs of high-IQ workers are immune from the effects of technological progress; I gave the example of lawyers. Any job category involving a high degree of specialization is vulnerable. Think of the impact of photography on portrait painters, or of computers on typesetters.

These trends bear on the current debates over the size of government. Technological advances are increasing longevity, and with it an increase in the dependent population. By reducing demand for workers, and therefore employment and wages, in many labor markets, the same technological advances may be creating a second dependent population, consisting of people of working age and their children who cannot support themselves without public assistance that will either replace or augment wages. Republicans may therefore be tilting against windmills in thinking that the size of government can be reduced.

Comments

[Thomas Rekdal]

The idea that ever higher standards of living can be achieved with ever lower contributions of labor is sometimes offered as the centerpiece of utopian revery, sometimes as the beginning of dystopia. Suppose it is true that the pace of automation and technological change is now producing surplus people (economically speaking) faster than the economy can absorb. What would be the likely outcome of such a trend?

It is possible (just barely) to imagine a technological elite ruling over an increasing number of impoverished Helots, but this hardly seems either stable or sustainable. A more likely result would probably be some sort of social insurance scheme, along the lines suggested by Robert Shiller, that would expand the concept of disability at one's job to include the disappearance of one's job.

Yet this would clearly not be sufficient in itself. Something would also have to be done to remove the stigma from having one's life go awry in this way. One can imagine whole new professions springing up whose whole task would be to give meaning and value to lives now economically meaningless.

Here my powers of prognostication fail. This seems like a job for rhapsodes and poets. Perhaps there will be a future for the liberal arts after all.

[D]

@Posner: "Compensation for disamenities is not a component of real income but an offset to a cost."

You're completely missing the point.

The laid off factory worker who takes a lower paying non-factory job may indeed get a better working environment. Their new job may be less "strenuous, monotonous, dirty or unhealthful, and dangerous". But those issues are irrelevant if the new job doesn't pay enough to support them.

Being able to trade cash income for better working conditions is a luxury that many cannot afford.

[Aliveoccupation]

So there's a decline of demand for lawyers... and (according to you) a decline in demand for low-IQ workers. Are you suggesting lawyers have low IQs?

[Terry Bennett]

Thomas - Interesting vision, but let's not totally discount the stigma, which is not without value. In its absence, individuals will languish in self-pity, permanently enabled by the welfare state. It is said that failure consists not of falling down but of staying down. For large numbers of people whose personalities lack a certain drive, the external stimulus of shame is the only thing that will goad them to get off the couch and re-engage.

[Thomas Rekdal]

Terry-I am not at all convinced that the path of automation we are on will really lead to a world without work for most of us. But if it does, I find it hard to see how the Puritan ethos you recommend continues to make sense. What does one do in DisneyLand if one is not one of the few running the trains? Surely our ideals of worthy human endeavor would change under such circumstances, and from whence would the change emerge if not from the arts? Hence my tongue-in-cheek expression of hope for the liberal arts.

[Abhay Saraf]

One of the other things to consider is reduction in working hours for the entire population. The effective number of jobs can be doubled overnight if we move to 20 hrs per week of work instead of current 40 (in developed economies).

Average working hours have been trending down for centuries and there is no need for that trend to stop. Wages will have to adjust accordingly to account for reduced working hours. However, things keep getting cheaper in real terms and if reduced wages can account for a minimal standard of living then it might not be a problem.

There will always be people working long hours but if humans employ machines in general, more time can be devoted to pursuits of intellectual activities or leisure.

[D]

Anyone interesting in (or concerned about) the issue of future technological unemployment, and how the problem might be resolved, should read the book "The Lights in the Tunnel: Automation, Accelerating Technology and the Economy of the Future"

http://www.thelightsinthetunnel.com

the book is focused entirely on this issue...

Here's also a review of the book at the Washington Post:
http://www.washingtonpost.com/opinions/matt-miller-the-robots-are-coming/2013/01/09/caac3e0a-5a57-11e2-88d0-c4cf65c3ad15_story.html

[Neilehat]

Anyone remember the Luddites? Similiar problem then, similiar problem now. Highly skilled Craft workers within various Cottage Industries supplanted by the Automation and scale of the New Industrial Revolution. It took time to integrate these individuals and uplift the Industrial Culture so that it became profitable for most. The Industrial age is now giving way to a "Cybernetic" age. Like the change from the Cottage Industry to Modern Industry, so todays Industry is giving way to a Cybernetic Age. The issue is that we must provide for a smooth transition to those displaced until they can be integrated into the new Economic reality. Just as we did in the past, but not without its fits, starts and failures, which hopefully, we can avoid...

[Terry Bennett]

Thomas, I get your point now, along with Abhay's point. If we evolve technologically to the point that it takes us an hour a day of work each to generate our sustenance (plus enough surplus to buy and maintain the machines that produce that multiplier), I'm fine with everybody working an hour a day (till they die, get rich, or become disabled). At the present time however, I'm not a big fan of government as an enabler for the gumption-challenged.

I also agree with Neilehat, that there is room for us to be smarter about this process, this time around. The people who lived through the Industrial Revolution probably didn't have the perspective to know they were living through what would later be called the Industrial Revolution. They knew something was up, but supply and demand were still essentially infinite at that time, because the world was a much bigger place with many fewer people and so much flux that probably nobody noticed the profound structural shifts as they occurred. We on the other hand have enough recorded history under our belts to put our predicament in a context, and figure out the necessary adjustments to fly more smoothly to the new steady state. Let's keep thinking...

[Necessary Proper]

I wrote a blog post on a related topic last night:

Elizabeth Warren Profoundly Misunderstands Labor Economics

which relates to this quote from Senator Warren:

"If we started in 1960 and we said that as productivity goes up, that is as workers are producing more, then the minimum wage is going to go up the same. And if that were the case then the minimum wage today would be about $22 an hour ... , with a minimum wage of $7.25 an hour, what happened to the other $14.75? It sure didn't go to the worker."

http://thenecessaryandproperblog.blogspot.com/2013/03/elizabeth-warren-profoundly.html

[Dpinsen]

Judge,

This observation of yours is pretty dated:

"higher pay for factory work is compensation for the fact that such work tends to be strenuous, monotonous, dirty or unhealthful, and dangerous."

Have you seen a modern auto factory? They are so clean and safe that Mercedes USA offers tours of them. And working in a factory is no more monotonous than working in a typical service sector job and no more strenuous (probably less strenuous) than lower paid work unloading pallets and stocking shelves in a big box retail store.

Concerns about the loss of manufacturing jobs are not exaggerated. When a worker goes from making $20 per hour in a factor to $10 per hour in a big box retailer, he's not the only one whose livelihood takes a hit. He has less money to spend now, so jobs in other sectors that his spending supported suffer too. In addition, he is less likely to be a net federal tax payer at the lower wage.

[Barrett Mattison]

An interesting TED talk that addresses the topic of technology replacing labor can be found here: http://www.ted.com/talks/andrew_mcafee_are_droids_taking_our_jobs.html

[Nagarajan V]

The bloggers who wrote the above piece as well as the individuals who have commented on it have, one and all, have given little thought to one crucial factor. If they had not read "The Limits to Growth" published in 1972, I would suggest to them to do so now. The concerns expressed in that treatise have not vanished with the passage of time but, on the contrary, with the diminishing and disappearing sources of fuels like coal and petroleum, base metals, Rare earths and even potable water, they have indeed become accentuated, especially with the sudden awakening and rapid industrialization of societies like those in China and India who had been in deep slumber until a couple of decades ago.

[Ben Frigo-Vaz]

True machines are now doing more and more higher IQ work, assuming there is no show stoppers to stop the growth in machine intelligence its only a matTer of time until machines can do everything people can and more, it only a matter of time until machines are smarter in ever why then us. Hopefully intelligence does not mean will, and that no matter how smart they get they won't have any desires of their own. Once machines can do everything we can, all human employment would become obsolete. But long before that a flaw in the poster's argument will become obvious to all: today and in the now economic growth has not been keeping up with automation: automation means more products per worker but in developed countries demand is not going up fast enough to buy all those products, ergo we need less workers. There fundamental limits of energy, materials and human necessity that puts a ceiling on demand, automation keeps increasing the amounts of products per worker but demand is limited therefor we will or are already hitting a point where we need to cut jobs because of automation. US median income in adjusted dollars has remained relatively stagnant for 50 years, most high paying unionized manufacturing jobs have given way to multiple part-time low wage service sector jobs, buying power has dropped and thus demand for products has dropped, automation does promise to make products cheaper but all this has done is equalize out the reduce income of the common consumer leaving our economy with lack luster growth and growing social inequality. Those that own the capital are making more and more from automation, they take this money and generally speculate with it, resulting in our economy become highly unstable with rapid cyclic fits of growth and recessions. In conclusion the anti-Luddite argument that people will just find other work is bunk, today's situation proves it bunk we don't need to wait until machines are smart then us to worry about automation.