This post was originally published on Angry Bear. I wrote this primarily in response to a comment on The Economist’s Free Exchange blog which picked up my original post on the likelihood of structural unemployment due to accelerating job automation technology. Free Exchange basically said we don’t have to worry about a serious unemployment problem because the principle of comparative advantage will insure that people will be able to find jobs.
Comparative advantage is an economic concept which says that individuals (or countries) will always be better of trading with each other even if one party has an absolute advantage in everything. (See the Wikipedia entry on comparative advantage for some examples.) Comparative advantage is usually attributed to David Ricardo’s work in the early 1800s and is generally considered to be one of the biggest ideas in economics.
The point I try to make in this post is that comparative advantage may not be all that helpful when people have to compete directly with machines that hold a substantial absolute advantage, and the primary reason is that machines (or software automation applications) —unlike people or countries—can be replicated on demand.
More on the Looming Structural Unemployment Crisis, and on Comparative Advantage
In my previous post, I suggested that job automation technology might someday advance to the point where most routine or repetitive jobs will be performed by machines or software, and that, as a result, we may end up with a serious structural unemployment problem. I’d like to respond to some of the objections that were raised regarding that idea.
I thought I would start with a response at the Economist’s Free Exchange blog, which said:
… in general I am pretty sanguine about the long-term prospects for continued voluntary employment of humans. Technology isn’t free, and even if we arrive at a world where some pieces of technology are better at everything than humans, the principle of comparative advantage nonetheless suggests that people will find work.
The idea is that, since everyone has a comparative advantage in something, just about everyone should be able to find some sort of a job. Thus we can be “sanguine.” Nearly every explanation of comparative advantage I have seen involves either individual people or countries. I haven’t seen examples where machines or automation technology come into play, so I thought I’d take a shot at it here.
Suppose we have a tractor and a team of oxen. Both can be used to plow fields, pull wagons or do other things around the farm. Clearly, the tractor out-performs the oxen in every task. Still, there ought to be some area in which the oxen don’t perform quite so badly relative to the tractor. Maybe the tractor is a little less efficient at plowing smaller fields since it has to make many turns. Or maybe fuel for the tractor is much more expensive in some regions, and so the oxen ought, in those cases, to have some sort of comparative advantage. So why have oxen been completely put out of work in developed countries like the United States?
It seems to me that there are two reasons. First, there is the magnitude of the absolute advantage that the tractor has. A tractor is a disruptive technology relative to the oxen. In order to have a meaningful comparative advantage, it’s probably helpful if you can get fairly close to the competition in at least one area.
The second reason is, perhaps, even more important: tractors, being machines, can be replicated on demand. If we imagine that a shortage of tractors existed, then comparative advantage would work. The available tractors would be deployed in their most productive uses, and the remaining work might well go to the oxen. But, in reality, the farmer can acquire as many tractors as he needs to do all his work, and in fact, he has no choice but to do so in order to remain competitive with other farmers.
As another example, suppose you are a brain surgeon who is also an excellent cook. Now, you might choose to employ a cook who is not quite as good as you are because doing so would free up your time and energy to do more brain surgery. So comparative advantage works there. But suppose you develop a machine (or two machines) with a dramatic absolute advantage in both cooking and brain surgery. Then, you could replicate your machine, and pretty soon there would be no jobs for cooks or brain surgeons.
So it seems like that might be a rule: If an affordable machine (or software algorithm) achieves a dramatic absolute advantage in a job or task, it will most likely be replicated and deployed until all competitors are eliminated. Comparative advantage is not much of a defense against that.
It seems to me that over time (not next week, but over years and decades), machines and software automation applications are likely to achieve that type of dominance in a great many areas, and they will be replicated until they consume all the available work. Any enterprise that failed to deploy this new technology would be less competitive.
All of this, of course, really amounts to nothing more than a restatement of the principle of obsolescence: in the long run, disruptive new technologies don’t find an equilibrium with old technologies. Old technologies get replaced. This applies equally to biological technologies like oxen—and perhaps it will someday even apply to human workers.
That’s an idea that economics is probably not ready to accept. Interestingly, other disciplines like biology or physics don’t give any special status to people. We are assumed to be subject to the same overall rules of nature as anything else. No so, with economics. For economists, people are very special; people are labor, and people get a special “L” in all the equations. Economists assume that people—and not just a few people but the vast majority of available workers—are indispensable to the production process. That has been true historically, but will it always be true?
Then again, maybe I’ve missed something. Maybe there is an area where human workers will always have an absolute advantage: in jobs that require uniquely human qualities or creativity, artistic ability and so forth. A lot of the conventional wisdom seems to suggest that we simply need to retrain, re-educate and redeploy workers into these areas, and everything will be fine. Is that likely to be the case? I’ll look at that idea in my next post.
Martin Ford is the author of The Lights in the Tunnel: Automation, Accelerating Technology and the Economy of the Future and has a blog at Econfuture
DeLong is echoing a post by Mark Thoma and says:
It’s very scary: long-term unemployment has a way of turning into structural unemployment…
Yes it is scary. My new book, The Lights in the Tunnel, is all about how advancing automation technology is going to lead to structural unemployment. I’m concerned that structural unemployment is going to get a lot worse. Economists are saying it will take years for the job market to recover—and during that time, technology will continue to advance. If consumer demand remains depressed (which seems likely), there’s a good chance that technology investment will flow disproportionately into labor-saving technologies such as automation software, robotics, expert systems, etc. That, of course, will result in even more unemployment.
First, I want to stress that I do not believe this is a discussion about a far future, science fiction concept. In my book, I make the point that technology is someday likely to advance to the point where the majority of the routine jobs held by average workers will be automated. That is a lot of jobs—probably most jobs. Two thirds of our population does not have a college degree, and even many college graduates have jobs that can broken down into relatively routine tasks that will be susceptible to software automation algorithms and expert systems. This is something that I think could potentially happen in the next couple of decades.
I am not talking about true machine intelligence. Theoretical physicists would still have jobs. Private sector economists whose jobs consist largely of plugging data into forecasting models and writing formulaic reports might well have to worry.
The point I make in my book is that relentless automation will ultimately result in massive unemployment and extreme and unsustainable income inequality—to the point that mass market production would be numerically undermined because there would simply be two few consumers with viable discretionary incomes. Yes, of course, automation would also make stuff cheaper. Would that make up for the low wages and unemployment? What fraction of total income does the average person spend on manufactured products and services today? You could make shopping at Wal-Mart free, and a great many lower wage workers would still struggle to cover housing and health care costs.
In The Lights in the Tunnel, I argue that we will ultimately have to provide supplementary income to the majority of the population; if we don’t do so, we won’t be able to sustain consumption. That type of scheme, obviously, would have to be supported by some type of taxation, and Hanson, no doubt, finds that highly objectionable.
Dr. Hanson rejects my arguments on the need to support consumers, saying:
Ford’s mass-market theory of production is nothing like standard economic theory. Sure high income inequality might be ethically bad, and threaten political instability, but it does not at all threaten economic collapse – producers can focus on giving the rich what they want, and innovation and growth is just as feasible for elite products as for mass products.
In other words, creative destruction is going to do its thing, and the mass market industries we now have are going to be destroyed and presumably replaced with new industries that focus on producing very high value, customized products targeted at a tiny economic elite—and that’s going to drive the entire economy. That really doesn’t sound like the world I want my children to live in, but leaving that aside, I still wonder if it works in the real world.
It’s very possible that a major breakthrough in an area like machine learning could cause all of this to unfold quite rapidly. Hanson notes this in his paper, saying at one point that machines could go from performing 25% of jobs to 75% within four years. So we are potentially talking about more than half of all the jobs in the economy. Let’s try to imagine an actual scenario:
A major technical breakthrough occurs and is widely publicized. Businesses rapidly deploy the technology and the unemployment rate rises past 15%. Consumer confidence falls to unprecedented levels. Political debates rage on extending unemployment and on revoking the minimum wage. Tax revenues are plunging inline with incomes. As consumer spending falls, businesses either adapt by laying off workers and automating still more jobs, or they fail entirely.
Unemployment climbs to 20%, and the automation of jobs appears relentless. Mortgage defaults are now at an unprecedented level. As job losses mount, at some point, homeowners make a collective and rational calculation: Why keep paying my mortgage? Housing values are plummeting and I’m almost certainly going to lose my job eventually. It’s better to hoard the money; I may never get another job. Besides, if everyone defaults, they can’t evict us all, so there may not even be any consequences. So people stop paying their mortgages, and then, of course, renters quickly see that the same logic applies to them.
How does the financial system and the overall economy survive that in the short run? I may be an economic rube, but I don’t get it. Dr. Hanson says, don’t worry, be happy:
The fraction of production that is given to capital vs. labor depends on the marginal productivity of capital, times the quantity of capital, vs. the marginal productivity of labor, times the quantity of labor. If capital and labor are the only owned factors of production, then if the fraction of income going to labor falls, the fraction of income going to capital must rise. That income flow goes to capital regardless of what assets are used to represent the stock of capital.
Well, ok then. So does that mean all those people will be able to pay their mortgages?
Leaving aside government (which Dr. Hanson surely does not want involved) and exports, production is equal to consumption plus investment, and those are both going to be in free fall, given a scenario like the one above. The primary problem is not with fractions of production—it’s with how much total production is going to occur.
Finally, as Hanson says, “Sure high income inequality might be ethically bad, and [might] threaten political instability.” That’s not something that we can simply dismiss. All of this, if it happens, is going to happen in the real world—where economics cannot be divorced from political and social ramifications. Income Inequality is already at a historically extreme level, and there is little reason to believe the progression won’t continue relentlessly. Bruce Judson has a new book out called It Could Happen Here: America on the Brink, which suggests that the possibility of revolution is not unthinkable if the trend continues. That’s something worth thinking about. No one who has enjoyed any measure of success under our current system would look forward to a world without free markets or where basic property rights were threatened.
Update: Prof. Hanson responded by adding this to his post:
Yes, a sudden unanticipated change would be disruptive, but no disruption does not imply falling production. Ford needs to learn some economics, or listen to some economists.
Well, we’ve been having a pretty good disruption lately. Production has not fallen? Here’s a graph of real v. potential GDP from the Federal Reserve Board of San Francisco. Look’s to me like production fell…
Source: Economist’s View/FRBSF
Robin Hanson, a professor at George Mason University (who also blogs at Overcoming Bias), is one of the few economists who has given serious thought to the potential economic implications of intelligent machines. In Dr. Hanson’s 1998 paper, “Economic Growth Given Machine Intelligence,” he suggests several variations on a growth model which assumes that machines achieve sufficient intelligence to become complete substitutes for, rather than complements to, human labor. Dr. Hanson’s conclusions are very optimistic, and to me, quite counterintuitive. His models “suggest that wholesale use of machine intelligence could increase economic growth rates by an order of magnitude or more.” At the same time, however, he notes the obvious reality that as machines become affordable, and very likely more capable, substitutes for human workers, “wages might well fall below human subsistence levels. “
My immediate reaction to this is that economic growth at any level—let alone of an order of magnitude beyond what we are accustomed to—is fundamentally incompatible with wages that are falling dramatically for the vast majority of workers. We might, perhaps, have vigorous economic growth if the falling wages applied to only a minority of human workers, but it is very difficult for me to conceive of a way in which such growth would be compatible with wages falling across the board—or even for the bulk of workers. The reason is simple: workers are also consumers (and support other consumers). If wages fall dramatically, then consumption must likewise fall because the majority of personal consumption is supported by wage income.
Additionally, I would make the point that as intelligent machines overwhelmed the labor market, the psychological impact on consumers would almost certainly amplify the fall in consumption. As the number of viable consumers in the market rapidly diminished, the mass market business models of most industries would be numerically undermined: there simply would not be enough willing and able buyers to support the high volume production of goods and services that characterizes most of the major industries that make up today’s economy. The result would not be extreme economic growth, but instead falling revenues and, quite probably, widespread business failures. It seems to me that the current economic situation offers a fair amount of support for my position.
In his paper, it appears to me that Dr. Hanson makes two separate assumptions to get around this basic problem of a reasonable balance between production and consumption. In his initial overview of his growth models, Dr. Hanson writes: “We assume that the product of the economy can be either consumed or used to produce more of any kind of capital (i.e., human, hardware, software, other).” I read this to mean that Hanson is assuming that private sector investment might “pick up the slack” left by diminished consumption. This strikes me an unsupportable assumption for the simple reason that business investment is not independent of consumption—or, at least, current investment is clearly a function of anticipated future consumer spending.
Businesses invest in order to better position themselves to reap the fruits of future consumption. As a business owner myself, I can really think of no other good reason for a business to make substantial investments in “human, hardware, software, or other” capital. In a world in which most workers’ jobs are being automated away and will never return, there would be very little reason to anticipate that future consumer spending would be anything but anemic. Therefore, there would be no reason for the private sector to invest. To me, it seems intuitively obvious that overall private sector investment would not increase, but instead would fall in line with plummeting consumption.
Dr. Hanson does, however, offer a second assumption that might help get around this problem:
Everyone a Capitalist
Dr. Hanson suggests that his results “may be compatible with a rapidly rising per-capita income for humans, if humans retain a constant fraction of capital, perhaps including the wages of machine intelligences, either directly via ownership or indirectly via debt.” In other words, he seems to be saying that if consumers have an ownership interest in the economy of the future, then the resulting investment income will be sufficient to make up for the precipitous decline in wages. Presumably this would allow the population to continue consuming. Dr. Hanson fleshes out this view in another article on “Singularity Economics” that was published in IEEE Spectrum in June, 2008:
…human labor would no longer earn most income. Owners of real estate or of businesses that build, maintain or supply machines would see their wealth grow at a fabulous rate—about as fast as the economy grows. Interest rates would be similarly great. Any small part of this wealth would allow humans to live comfortably somewhere…
In other words, everyone (or at least most people) will have a piece of the action, and the returns on that ownership will be so fantastic that almost everyone will have a reasonable discretionary income—with which they can then go out and consume.
I’m going to leave aside the obvious problems with the distribution of wealth and income, as well as with any hope of social mobility, that this scenario implies, and instead focus on a more basic question: would asset values really increase at the extraordinary rate that Hanson suggests? Would they increase at all?
Dr. Hanson seems to be assuming that the stock market (and other productive assets) would increase dramatically in value because investors would recognize that businesses now have a fantastic new technology (intelligent machines) which will enable extraordinarily efficient production. The problem I see with this is that, according to modern financial theory, asset values are not determined by investors’ perceptions about technology. Asset values are defined by investors’ expectations for the future cash flows that will be associated with the asset in question. It seems clear to me that, in the midst of across the board job automation and plunging consumer demand, those future cash flows would be looking pretty minimal.
In order for asset values to increase, investors would have to reason that, because asset values would increase dramatically, nearly everyone would have access to an investment income which could then be used to consume—and thereby create the future cash flows which would justify the current value of the asset.
That strikes me as both circular and unlikely. Dr. Hanson seems to be assuming a perpetual asset bubble that somehow gets going even though it is not even remotely initially supported by fundamentals. In fact, the initial fundamentals would point—quite dramatically—in exactly the opposite direction.
Where would these courageous initial investors come from? During the height of the financial crisis last year, I happened to see a report on CNBC which noted that extremely wealthy people were buying gold and having it shipped directly to their estates. They wanted their gold in their homes, behind their gates and walls, buried in their underground vaults. Are those perhaps the investors that Dr. Hanson is expecting to step up to the plate and begin driving up asset values when intelligent machines arrive and destroy consumer demand?
It seems to me that Professor Hanson’s views are really quite unsupportable. Nonetheless, it is of course possible that I have made an error somewhere or I have misunderstood Dr. Hanson’s arguments. I look forward to Professor Hanson’s response to my thoughts.