The Reign of the Robots: Economists getting it badly wrong
from Dean Baker
A standard fear raised in Washington policy debates is that the development of robots and other forms of technology will displace tens of millions of workers, leaving much of the workforce without jobs. This is remarkable story both because it is not supported by any evidence, but also because it goes in the opposite direction of virtually all the main concerns raised in debates over economic policy.
The story of the rising robots should mean that we are seeing a rapid increase in the amount of output per hour of work. The logic is that the robots are doing work that humans used to do, so we have more output for every hour of human labor.
In fact, we are seeing the exact opposite. The rate of productivity growth, which measures output per hour of work, has slowed sharply in recent years. In the last five years productivity growth in the United States has averaged just 0.4 percent annually, the slowest five year stretch on record. This compares to a rate of close to 3.0 percent annually from 1995 to 2005, which was also the rate of productivity growth during the long post World War II Golden Age, from 1947–73.
There also is no evidence of any uptick of investment in the sectors that would be fueling a robot driven productivity boom. Real investment in information processing equipment has increased by less than 5.0 percent over the last year. That’s hardly the basis for a productivity boom.
While the robots taking over story lacks any evidence in the data, it is also 180 degrees at odds with the country’s current policy agenda. This is perhaps clearest with Federal Reserve Board policy. The Fed opted not to raise interest rates at its most recent meeting, but it did hike rates in December and has indicated that further rate hikes remain in its plans for 2016.
The logic of raising interest rates is a concern that we have too few workers. The concern is that too much demand for labor relative to the supply will push up wages, which will in turn lead to higher prices, leading to a scenario of spiraling inflation. The purpose of raising interest rates is to reduce consumption and investment demand and housing construction, thereby reducing employment so that the country does not run out of workers.
This is the same argument for cutting government spending and bringing down the budget deficit. The case for this policy is that too much demand in the economy is pushing the economy against its limits, most importantly the limited supply of labor. In this context we again would see a story of either rising interest rates and/or higher inflation due to an excessive budget deficit.
And the often told demographic catastrophe story of too few workers and too many retirees is quite explicitly a story where we have too few workers. The problem is that a large population of retirees will be making huge demands on the economy at a time when there are not enough workers to meet this demand. It’s rare that a week passes without some prominent politician urging the country to reduce its retirement benefits to cope with the looming demographic crisis.
The fact that concerns about robots taking all the jobs can persist side-by-side with concerns that we are running out of workers says a great deal about the state of debates on economic policy at present. Apparently, few of the participants in these debates even recognize the fundamental contradiction involved. This is like worrying that we will simultaneously be afflicted with record rains and a severe drought. Either is in principle possible, but they cannot occur at the same time. That shows a great deal about the quality of modern policy debates.
In terms of the underlying issue, whether we face a problem of robots taking all the jobs or too few workers, I come down firmly in the middle. There is little reason to think that rapid productivity growth should pose a problem for workers. We had rapid productivity growth in the Golden Age and it was associated with low rates of unemployment and rapid wage increases, as workers shared in the benefits of productivity growth. Given the slow current pace of productivity growth it is difficult to believe that it will quickly accelerate to a point that is so much faster than in the Golden Age, so that it is actually difficult to combat the job displacement with new jobs in other sectors.
On the other hand, those concerned about labor shortages also seem wrong in every respect. The U.S. labor market is still far from recovered from the full impact of the Great Recession. Employment of prime age workers (ages 25–54) is still almost three full percentage points from its pre-recession level. Given the realities of today’s labor market, there is no reason for the Fed to tighten up with higher interest rates or Congress to reduce already modest budget deficits.
On the larger demographic story, we have been seeing the same picture for the last 70 years. The population is getting older as has been the case for decades. The impact of population ageing will be swamped by the impact of even very low rates of productivity growth.
In short, there are not unsolvable real world economic problems out there. The big problem is a policy elite that doesn’t seem to understand economics.
Leave a comment
—– look inside —– $5.94 / $20.00
—– look inside —– $4.90 / $8.00
—– look inside —– $15.99
—– look inside —– $5.99 / 12.99
—– look inside —– $5.93 / $12.99
—– look inside —– $4.97 / $9.90
—— Ugarteche, Puyana and Madi ——
Gerson Lima / Maria Alejandra Madi
Edward Fullbrook and Jamie Morgan
————— Michael Hudson ————–
Maria Alejandra Madi / Jack Reardon
————- Edward Fullbrook ————-
—————— Steve Keen —————–
————— Richard Smith —————
————– Gustavo Marques————
– Victor Beker and Beniamino Moro –
————– Lars Pålsson Syll ————-
—————– Stuart Birks —————-
Edward Fullbrook and Jamie Morgan
———— Armando Ochangco ———-
Shimshon Bichler / Jonathan Nitzan
————— Mauro Gallegati ————–
————— Herman Daly —————-
————— Asad Zaman —————
—————– C. T. Kurien —————
————— Robert Locke —————-
Real-World Economics Review Blog 
It’s very good that you show the contradiction between opposite worries, those from too few workers vs. those from too many workers. On the other hand; is it possible that low productivity growth comes from the turn of economy into a services economy while a few decades ago manufacturing was a much higher proportion of the whole?
Seems to me services are less responsive to robotization in respect to use of manpower than goods production.
Respect elites not understanding economics…let me respectfully disagree: they understand it very very well. Of course they make it play in their own interest.
“Seems to me services are less responsive to robotization in respect to use of manpower than goods production.”
That used to be the case. By 2050 a robot will be able to do anything a human can do.
“There is little reason to think that rapid productivity growth should pose a problem for workers.” Somehow I am not reassured. This prognostication is built on a stack of if’s any significant technological breakthrough could reverse. Sooner or later robotization is going to ravage labor. Under the current economic structure it can’t do anything else. We should start preparing for it now by changing that structure so that productivity gains are shared by the society at large rather than piped to a wealthy few.
” We should start preparing for it now by changing that structure so that productivity gains are shared by the society at large rather than piped to a wealthy few.”
I agree.
This article is basically spot-on.
I would add that obviously you can have too many workers and too few workers at the same time — if you have a problem with the income or property distribution.
So there can be too many workers assigned to the rich, i.e. only the rich can pay them but they don’t need them. And at the same time there can be too few workers assigned to the poor, i.e. the poor could use more workers to produce more consumption goods (including health care, education, whatever) but they are too poor to pay them.
“I would add that obviously you can have too many workers and too few workers at the same time — if you have a problem with the income or property distribution.”
Or if you have an economy that includes both software developers and baristas.
Well that would only concern unemployment in some particular sectors and would assume that people can’t change their profession.
We’re already confronted with this issue in New Zealand. Elder care here is already a growth industry, but its also known for its dismal trackrecord in poor employment conditions and pay, because it primarily staffed by women with limited education and the companies involved often import qualified nurses trained in India and Philippines who jump at the chance to earn the New Zealand minimum wage which is many times what they’d earn back home. I assure you that your hope that elderly care will be a panacea in addressing the forthcoming employment will be horribly misplaced.
On the matter of you questioning the apparent inconsistency between predictions of an employment crisis due to automation of many occupations and the much touted impact of an aging workforce and population. I actually see the purported shortage of workers to be a poorly thought out hysteria, because those workers (especially women) displaced by robots and A.I. from declining manufacturing workforce and junior clerical and information workers will find themselves with little option but to take positions in elderly care which will constrain workers bargaining power and their resulting wage demands. They will be supplemented by migrants, both legal and illegal on the same basis that agriculture labour has been for many years.
For me its more a matter of the affordability of elderly care, because the primary institution responsible for financing elderly care has been the public purse. Even in the United States Medicare and Medicaid devote vast sums of money to nursing homes to care for elderly social security recipients, because people lack the financial wherewithal to support their own elderly care needs. There is a crisis in affordability, because the primary sources of revenue for governments around the world is in decline, because of dismal economic conditions, stagnant incomes, ageing populations, and the costly burden of past wasteful spending on bank bailouts and military spending and none of this will be addressed, because of the broken nature of modern politics, whether it be in Europe, the United States, South East Asia, or Australasia.
The problem is real but not new. With robots the advantage is not production rate but unit product cost. They are cheaper than human workers, and they don’t pay income tax. However when the reduced product cost is eventually shown in the lower price of the goods produced, there will be more money available for other kinds of goods, so the situation is not all bad and the demand for labor will not suffer too badly. Exactly the same situation occurred during the past rises in technology and somehow the new generation of workers managed.
“However when the reduced product cost is eventually shown in the lower price of the goods produced, there will be more money available for other kinds of goods”
I agree that this trend exists. I disagree with the implication that it will necessarily be sufficient to counter the damage.
“Exactly the same situation occurred during the past rises in technology and somehow the new generation of workers managed.”
They did in countries that implemented the New Deal or its likes. Countries that followed free trade religion did not fare so well.
«With robots the advantage is not production rate but unit product cost. They are cheaper than human workers»
Oh please, there are hundreds of millions if not billions of human workers that can be hired for less than $1,000/year. No robot can compete with that. Robots have to eat too: some form of fuel, and they need maintenance too.
There is no “law of Economics” that says that human workers are entitled to wages that allow their families to have 2 cars and a 3 bedroom house with a garden, and indeed the overwhelming majority of human workers worldwide don’t.
«somehow the new generation of workers managed»
And for “main demographic” workers hourly “real” wages in many first-world countries have been going down since around 1970. That’s how they managed to stay “competitive”, and not all of them, as participation rates have gone down too.
David Chester. Yes the military and the financial cartel will continue to be reliable sinks to absorb surplus capital produced by the monopoly capitalist economy. Even nutcase Tea Partiers are only too happy for the Military Industrial Complex Ponzi scheme to continue to run.
“Increasing spending for the military does a couple of things. It not only not only stimulates the economy, it protects our nation.”
https://philebersole.wordpress.com/2011/08/08/can-we-do-without-military-keynesianism/
I have a problem with any article like this one, that doesn’t even mention that over 20% of American people of working age are unemployed. (Most of them are officially classified as “discouraged”, which means that they have given up all hope of ever getting a job). Surely a few percentage points of productivity either way are less significant than having over a fifth of your work force permanently unemployed.
I agree but this is not even the biggest issue. The productivity of a country is determined by the top 20% of its labor force and unemployment tracks the bottom 20% of its labor force. It is magical thinking to assume that if you bring the unemployment rate to 4%, then this somehow ensures that the top 20% will have the economic systems, inertia, and customers to maximize their output. If this were the case, then Nigeria would have the same output as England.
I basically agree with this, but articles like this—on growth, productvity, employment—leave out half the story in my view. That is how one defines ‘productive work’. In my area alof of the employment for educated people consists of doing ‘start ups’ which make computer apps to make life conveniant (eg where is the closest dentist) , or sort of luxury (microbrewing companies are replacing the mass market beer companies as a form of production and consumption). Political lobbying is also huge—i get maybe 50 soliciatiions a day via computer or ‘junk mail’ for political, social justice, ecological and other causes. Mass media on both left and right are also big buerocracies. (On the ‘right’ they solicit for gun rights organizations, training for investing or ‘flipping houses’, and all sort of questionable health and charitable groups).
“Exactly the same situation occurred during the past rises in technology and somehow the new generation of workers managed.”
What does productive work mean? Is creating an app that provides just a tiny increment above zero utility for hundreds of thousands of people more or less productive than being a pizza cook?
I don’t know how to define productive work, so as a cop out, I just approximate with “work someone is willing to pay for.” It is more a function of wage history than it is a function of intrinsic value. If everybody were middle class then the best burger in the world would probably be worth 10 or 12 buck. If you are a billionaire, then the best burger in the world might be worth hundreds of thousands of dollars if you are willing to be honest to yourself about true utility.
Productive work is simply when the exertion of labor produces something (be it goods, services or documents), that is of value and can fetch a price when sold to a consumer or other kind of user.
“The story of the rising robots should mean that we are seeing a rapid increase in the amount of output per hour of work.”
“The rate of productivity growth, which measures output per hour of work, has slowed sharply in recent years. In the last five years productivity growth in the United States has averaged just 0.4 percent annually, the slowest five year stretch on record.”
These are measuring two completely different things and many of the problems with modern economic theory come from conflating the two. The first reflects “Ability to produce efficiently”. In other words, how much has the economy been walmartized. The second measures “How efficiently are we producing what is actually sold?” In other words, how good are we at matching up production capability with consumption capability based on the production labor costs we are using? It is silly to then write “Therefore robots are not creating a problem on the consumption side. QED” below this. A simple example to illustrate this is, your local retail shop is serving a local community with decreased buying power. It gets more worker for less dollar as desperation grows and greater actual efficiency as it upgrades it’s ability to produce for lower labor dollars because it has to to survive, but it’s actual economic productivity drops because it has more time of employees sitting around in the back waiting for customers to come in.
“The purpose of raising interest rates is to reduce consumption and investment demand and housing construction, thereby reducing employment so that the country does not run out of workers.”
I know some claim this, but there are several different explanations for why we should raise rates:
1) Labor shortage (presented above)
2) Business cycle (There is some magical equilibrium value you should chase)
3) Keep your powder dry version 1 – We are unable to tax or print money to continue stimulus
4) Keep your powder dry version 2 – We are unwilling to tax or print money to continue stimulus. (Note that I am deliberately avoiding discussing the money creation complexities here as they don’t really change the outcome, in my opinion.)
5) We don’t know what we are talking about but raising rates feels prudent.
Regarding 1 – the country would never run out of workers. Instead it would be forced to put them to better use (I know this isn’t necessarily disagreement). Prices might come up, but until and unless hyperinflation comes into the picture (something which as far as I know is never driven by expensive labor), less important than the actual inflation is the distribution of revenues within the company under this situation. To the extent that this is true and based on accurate assessment of outcomes, it is more about the pay of CEOs and stockholders going down comparatively than it is about inflation going up.
Personally I like 4) and 5), but I know that every vote can potentially have a different reason for each voter.
When the productivity of robot s rises, there will be less need to employ people. So how will we be able to earn sufficient money in order to purchase these robot-made goods? By introducing too many robots, nobody will be able to afford to live at a decent level of life, let alone at a raised one..
«So how will we be able to earn sufficient money in order to purchase these robot-made goods?»
In part if the robots are indeed much cheaper producers of goods than human workers then the good produced should be much cheaper too. if the owners of the robots feel like producing them for the poor.
«By introducing too many robots, nobody will be able to afford to live at a decent level of life»
The owners of the robots will obviously be able to buy those goods and enjoy a very «decent level of life».
There is nothing different there between the owners of a gold mine or of an oilfield also being able to enjoy a much more «decent level of life» than those who don’t own a gold mine or an oilfield. Also nothing different with countries that have factories that make airplanes or computer chips versus countries that don’t have them.
It is unfortunate for those who are hardscrabble farm labourers in Somalia versus those who are middle managers at Boeing in Seattle, or unfortunate for the latter compared to members of the Kuwaiti royal family.
The issue is the distributional impact.
Some of this seems easy. For example, as a simple case one can consider baby sitting. (krugman has some story on this, if i recall as an example of market failure, tho that is not my case). In my case we have 2 families, both unemployed people who stay home and take care of babies. To increase GDP and productivitiy, each family sets up a baby sitting business. Every morning the families get up and take their babies to the other family. So they do child care, all on the books. GDP goes up. Unemployed people now are in the ranks of the employed. (it may be basically the same except babies get a new environment, new set of caregivers and playmates, maybe a new playground).
Alternatively one could have a construction business. Get up early, go to the site, collect rocks and move the rocks to another site. Take a lunch break. In the afternoorn move all the rocks back to the original site. Its alot of work. Adds to GDP, and lowers unemployment.
Later on can hire skilled workers to create ‘killer apps’ so using GPS and other techno-utopian methods, one can move even more rocks back and forth and hire more technologically sophisticated people.
On the general subject of robot fantasies – and their irrelevance to most things- Post Post Work Post is worth a look. I give a link (at the end of the comments) to George Caffentzis’s essay and book that Sandwichman refers to.