Effects of New Technologies on Labor

January 4, 2019

Previous | Next

David Autor and Anna Salomons, “Is automation labor share-displacing? Productivity growth, employment, and the labor share.” Brookings Papers on Economic Activity, Spring 2018.

Daron Acemoglu and Pascual Restrepo, “Robots and jobs: Evidence from US Labor Markets.” National Bureau of Economic Research, March 2017.

I have been interested in automation’s effects on the labor force for a long time, especially since reading Martin Ford’s Rise of the Robots. Ford raises the specter of a “jobless future” and a massive welfare system to support the unemployed.

Here I discuss two papers representing some of the most serious economic research on this topic.

The questions

To what extent do new technologies really displace human labor and reduce employment? The potential for them to do so is obvious. The mechanization of farming dramatically reduced the number of farm workers. But we can generalize only with caution. In theory, a particular innovation could either produce the same amount with less labor (as when the demand for a product is inelastic, often the case for agricultural products), or produce a larger amount with the same labor (when demand expands along with lower cost, as with many manufactured goods). An innovation can also save labor on one task, but reallocate that labor to a different task in the same industry.

Even if technological advances reduce the labor needed in one industry, that labor can flow into other industries. Economists have suggested several reasons that could happen. One involves the linkages between industries, as one industry’s productivity affects the economic activity of its suppliers and customers. If the computer industry is turning out millions of low-cost computers, that can create jobs in industries that use computers or supply parts for them. Another reason is that a productive industry affects national output, income and aggregate demand. The wealth created in one industry translates into spending on all sorts of goods and services that require human labor.

The point is that technological innovations have both direct effects on local or industry-specific employment, and also indirect effects on aggregate employment in the economy as a whole. The direct effects are more obvious, which may explain why the general public is more aware of job losses than job gains.

A related question is the effect of technology on wages, and therefore on labor’s share of the economic value added by technological change. Do employers reap most of the benefits of innovation, or are workers able to maintain their share of the rewards as productivity rises? Here too, aggregate results could differ from results in the particular industries or localities experiencing the most innovation.

The historical experience

American history tells a story of painful labor displacement in certain times, places and industries; but also a story of new job creation and widely shared benefits of rising productivity. Looking back on a century of technological change from the vantage point of the mid-20th century, economists did not find negative aggregate effects of technology on employment or on labor’s share of the national income. According to Autor and Salomons:

A long-standing body of literature, starting with research by William Baumol (1967), has considered reallocation mechanisms for employment, showing that labor moves from technologically advancing to technologically lagging sectors if the outputs of these sectors are not close substitutes. Further,…such unbalanced productivity growth across sectors can nevertheless yield a balanced growth path for labor and capital shares. Indeed, one of the central stylized facts of modern macroeconomics, immortalized by Nicholas Kaldor (1961), is that during a century of unprecedented technological advancement in transportation, production, and communication, labor’s share of national income remained roughly constant.

Such findings need to be continually replicated, since they might hold only for an economy in a particular place or time. In the 20th century, the success of labor unions in bargaining for higher wages and shorter work weeks was one thing that protected workers from the possible ill effects of labor-saving technologies.

Recent effects of technological change

Autor and Salomons analyze data for OECD countries for the period 1970-2007. As a measure of technological progress, they use the growth in total factor productivity (TFP) over that period.

They find a direct negative impact of productivity growth on employment within the most affected industries. However, they find two main indirect effects that offset the negative impact for the economy as a whole:

First, rising TFP within supplier industries catalyzes strong, offsetting employment gains among their downstream customer industries; and second, TFP growth in each sector contributes to aggregate growth in real value added and hence rising final demand, which in turn spurs further employment growth across all sectors.

To put it most simply, one industry’s productivity may limit its own demand for labor, but its contribution to the national output and income creates employment opportunities elsewhere.

With regard to labor’s share of the economic benefits, the findings are a little different. Here again, the researchers find a direct negative effect within the industries most affected by technological innovation. But in this case, that effect is not offset, for the most part, by more widespread positive effects.

The association between technological change and labor’s declining share varied by decade. Labor’s share actually rose during the 1970s, declined in the 1980s and 90s, and then fell more sharply in the 2000s. The authors mention the possibility that the newest technologies are especially labor-displacing, but reach no definite conclusion. Another possibility is that non-technological factors such as the political weakness of organized labor are more to blame.

The impact of robotics

Autor and Salomons acknowledge that because they used such a general measure of technological change, they couldn’t assess the impact of robotics specifically. They do cite work by Georg Graetz and Guy Michaels that did not find general negative effects of robots on employment or labor share in countries of the European Union. That’s important, since many European countries have gone farther than we have in adopting robots.

The paper by Acemoglu and Restrepo focuses on the United States for the period 1990-2007. (They deliberately ended in 2007 so that the impact of the Great Recession wouldn’t muddy the waters.)

The authors used the definition of robot from the International Federation of Robotics, “an automatically controlled, reprogrammable, and multipurpose [machine].” Over the period in question, robot usage increased from 0.4 to 1.4 per thousand workers. “The automotive industry employs 38 percent of existing industrial robots, followed by the electronics industry (15 percent), plastic and chemicals (10 percent), and metal products (7 percent).”

Adoption of industrial robots has been especially common in Kentucky, Louisiana, Missouri, Tennessee, Texas, Virginia and West Virginia. As Thomas B. Edsall titled his recent New York Times column, “The Robots Have Descended on Trump Country.”

Acemoglu and Restrepo classified localities–technically “commuter zones”–according to their “exposure” to robotics, based on their levels of employment in types of jobs most conducive to robotization.

Their first main finding was a direct negative effect of robotics on employment and wages within commuting zones:

Our estimates imply that between 1990 and 2007 the increase in the stock of robots…reduced the employment to population ratio in a commuting zone with the average US change in robots by 0.38 percentage points, and average wages by 0.71 percent (relative to a commuting zone with no exposure to robots). These numbers…imply that one more robot in a commuting zone reduces employment by about 6 workers.

The workers most likely to be affected are male workers in routine manual occupations, with wages in the lower-to-middle range of the wage distribution

In the aggregate, these local effects are partly offset by “positive spillovers across commuting zones”–positive effects on employment and wages throughout the economy. With these spillovers taken into account, the estimated effects of robotics on employment and on wages are cut almost in half, dropping to 0.20 percent and 0.37 percent respectively.

The authors state their conclusion cautiously, as “the possibility that industrial robots might have a very different impact on labor demand than other (non-automation) technologies.”


While there is little doubt that new technologies often displace labor in particular industries and localities, the aggregate effects on employment and wages are less consistent.  Historically (late 19th and early 20th centuries), employment and labor share of income held up very well. For developed countries in the period 1970-2007, Autor and Salomons found a mixed picture, with robust employment but declining labor share after 1980. With respect to robotics specifically, Graetz and Michaels did not find declines in employment or labor share in the European Union, but Acemoglu and Restrepo found some decline in both employment and wages in the U.S.

It seems fair to say that the jury is still out on the effects of automation on the labor force. It may be that automation has no inevitable effect, but that it depends on how we as a society choose to deal with it. We shouldn’t assume a world of mass unemployment and widespread government dependency on the basis of recent, preliminary results from one country. Authors such as Thomas Friedman, who are more optimistic than Martin Ford about the long-run effects of new technologies, have yet to be proved wrong.

Thank You for Being Late (part 2)

November 15, 2018

Previous | Next

Today I’ll discuss two chapters of Thomas Friedman’s Thank You for Being Late that I found especially insightful: Ch. 8 on the implications of new technologies for employment, and Ch. 9 on the problem of global order.

The future of work

Friedman begins his discussion of work with a bold pronouncement: “Let’s get one thing straight: The robots are not destined to take all the jobs. That happens only if we let them–if we don’t accelerate innovation in the labor/education/start-up realms, if we don’t reimagine the whole conveyer belt from primary education to work and lifelong learning.”

I was pleased to find that Friedman’s position is similar to the one I laid out in my critique of Martin Ford’s The Rise of the Robots Ford predicted a future of massive unemployment, with millions of displaced workers relying on government for a minimal income. We could get a taste of that during a transitional period, but I don’t think that’s a very good description of where we are ultimately headed.

Friedman doesn’t deny that smart machines can now perform many tasks currently or formerly performed by humans. But he makes a sharp distinction between automating tasks and automating whole jobs so as to eliminate the human contribution altogether. The upside of automation is increased productivity. Workers aided by new technologies can produce more per hour, reducing the unit cost of what they produce. That can create a larger market for the product, increasing the demand for labor in a given occupation. A car was an expensive luxury item before the assembly line cut costs to create a mass market and a booming industry. Friedman reports that “employment grows significantly faster in occupations that use computers more,” as in banking and paralegal work.

To give an example from my own experience, financial planning software has automated many of the most tedious tasks involved in preparing a retirement plan, such as mathematically projecting future income from savings rates and asset allocation choices. But that hasn’t resulted in a reduced need for financial planners. On the contrary, it has made the services of a planner affordable for more people. Planners can spend less time doing calculations but more time relating to their clients.

Friedman says, “Jobs are not going away, but the needed skills for good jobs are going up.” What are disappearing are well-paid jobs with only modest skill requirements, like twentieth-century manufacturing jobs.

Retooling education

In general, today’s good jobs require more education; yet it does not follow that a college education necessarily qualifies a person for a good job. That’s not because a liberal education is a waste of time, but because it is only a foundation that must be built upon with lifelong job-relevant learning.

Friedman quotes MIT economist David Autor, who stresses the need for more than one kind of learning: “If it’s just technical skill, there’s a reasonable chance it can be automated, and if it’s just being empathetic or flexible, there’s an infinite supply of people, so a job won’t be well paid. It’s the interaction of both that is virtuous.”

Friedman is a strong believer in a broad, basic education that includes “strong fundamentals in writing, reading, coding, and math; creativity, critical thinking, communication, and collaboration; grit, self-motivation, and lifelong learning habits; and entrepreneurship and improvisation….” Even a robotics enthusiast like Martin Ford acknowledges that humans surpass robots in general intelligence, as opposed to specialized task capabilities.

However, recipients of this basic education will also have to cope with rapidly changing workplace requirements. Technology will play a central role here, both in creating the automated systems with which workers interact, and in enhancing learning processes themselves. Friedman wants to “turn AI into IA,” by which he means turning artificial intelligence into intelligent assistance to support lifelong learning. “Intelligent assistance involves leveraging artificial intelligence to enable the government, individual companies, and the nonprofit social sector to develop more sophisticated online and mobile platforms that can empower every worker to engage in lifelong learning on their own time, and to have their learning recognized and rewarded with advancement.” When the time comes to pick up a new skill, you can probably find an app to help you learn it.

Friedman describes AT&T as one company that is demanding more lifelong learning of its employees, but supporting it with measures like tuition reimbursements, online courses developed in collaboration with online providers, and promotions for those who acquire new skills. This represents a new social contract between employer and employee–“You can be a lifelong employee if you are ready to be a lifelong learner.”

Every major economic shift has involved the rise of a new asset class, such as land in the agrarian economy and physical capital in the industrial economy. The rising asset class today is human capital, and that is where society’s investments must be increasingly concentrated.

The threat of global disorder

In the immediate aftermath of the Cold War, after the collapse of the Soviet Union, The U.S. remained the only superpower and the most obvious model for other countries to emulate. Many thinkers expressed the hope that the world could move faster in the direction of American-style democracy and capitalism. But then the interventions in such places as Iraq and Afghanistan failed to produce stable democracies, the Great Recession called into question capitalist progress, and Americans lowered their expectations for world leadership.

What Friedman calls the post-post-Cold War world is characterized by shrinking American power, especially in the Middle East, and new challenges arising from the accelerations in technological change, globalization and environmental degradation.  In large areas of the less developed world, the danger is that states will fail and societies will sink into disorder, dragging the global political order and economy down. Environmental disasters like deforestation in Central America or drought and desertification in sub-Sahara Africa are uprooting people from their traditional relationship to the land. And while some poorer countries are advancing by providing cheap labor to the global economy, the future may belong to those who can provide smarter labor, and that requires greater investments in human capital.

Friedman says that during the Cold War, superpower competition gave America a reason to assist developing countries, in order to keep them in our camp. The mid-twentieth century economic boom also gave us the means to do so. While many Americans are now inclined to turn their back on the rest of the world, Friedman makes a case for renewed global involvement: “While we cannot repair the wide World of Disorder on our own, we also cannot just ignore it. It metastasizes in an interdependent world. If we don’t visit the World of Disorder in the age of accelerations, it will visit us.” The dislocated people in failed states can become refugees or terrorists. The same technologies that can empower people to learn and produce more can empower them to build improvised explosive devices triggered by cell phones, or perhaps a weapon of mass destruction.

In Friedman’s view, the best thing the U.S. could do to “help stabilize the World of Disorder and widen the islands of decency” would be to help fund schools and universities. He would also like to see us help the poorest people make a living in their own villages by assisting them with their environmental problems. He points out that it costs only 100-300 dollars to restore a hectare of degraded land.

In a world of enhanced interdependence, the haves would do well to invest in the development of the have-nots, domestically and globally. If we do not rise together, we will very likely fall together.


MMT 7: A Full Employment Proposal

July 11, 2018

Previous | Next

This is the seventh in a series of posts about Modern Monetary Theory, based on the text by Mitchell, Wray and Watts. If you have not seen the earlier posts, I recommend that you start at the beginning.

The goal of full employment

The authors argue for full employment on both economic and ethical grounds. Enabling everyone who wants a job to get one maximizes national economic output, providing more goods and services to distribute. Failing to do so not only hurts unemployed individuals and their families, but does lasting damage to economy and society in general:

Persistently high unemployment not only undermines the current welfare of those affected and slows down the growth rate in the economy below its potential, but also reduces the medium- to longer-term capacity of the economy. The erosion of skills and lack of investment in new capacity means that future productivity growth is likely to be lower than if the economy was maintained at higher rates of activity.

The authors are very critical of the dominant trend in recent economic policy, which is to tolerate unemployment while giving priority to fighting inflation. Policymakers came to accept unemployment rates far above the 2% or lower that was normal in the mid-twentieth century. High unemployment has also been accompanied by underemployment, as many workers have been unable to work as many hours as they would like, and also labor force withdrawals, especially by men. The official unemployment rate does not tell the whole story.

The inflation-fighting part has worked pretty well. Sluggish economic growth and high unemployment weaken the bargaining position of labor and help keep wages down. In turn, low labor costs and weak consumer demand keep firms from raising prices. In general, “the use of unemployment as a tool to suppress price pressures has, based on the OECD experience since the 1990s, been successful.”

The authors are troubled by the injustice of making a minority of the population bear the costs of a weak economy. “Joblessness is usually concentrated among groups that suffer other disadvantages: racial and ethnic minorities, immigrants, younger and older individuals, women (especially female heads of households with children), people with disabilities, and those with lower educational attainment.” I would add that the injustice is compounded if those who do make income gains in this economy are mainly the wealthiest 1%. The benefits of price and currency stability are somewhat more widely shared, but “it is doubtful that a case can be made for their status as a human right on par with the right to work.”

The Job Guaranty

Not all countries experienced high unemployment after the end of the postwar economic boom. Some, such as Norway, did more to insure that everyone who wanted to work could find a job.

The idea of the Job Guaranty is fundamentally simple. Since full employment is such a social and economic good, the public sector should take up the slack by employing those who cannot find jobs in the private sector.

“Private firms only hire the quantity of labour needed to produce the level of output that is expected to be sold at a profitable price. Government can take a broader view to include promotion of the public interest, including the right to work.”

The Job Guaranty is also known as the “employment buffer stock approach.” A stock of public jobs provides a buffer to protect the economy from a weak private sector.  Government acts to stabilize employment, spending to hire more labor when the private sector is weak, and reducing spending and public employment when it is strong. That would also have a stabilizing effect on national income and consumption.

The authors suggest that the wages paid in the Job Guaranty program would function like a national minimum wage, since they should be low enough to “avoid disturbing the private sector wage structure when the JG is introduced.” It wouldn’t compete with the private sector enough to drive up wages in general. On the other hand, they also want the wages to express “the aspiration of the society in terms of the lowest acceptable standard of living.” They do not discuss how these goals might be in conflict, but advocates of a “living wage” generally regard today’s minimum wage as too low.

Price stability

Proponents of the Job Guaranty expect it to be less inflationary than traditional Keynesian policies, which recommend government spending in general to stimulate the economy. When government increases its general spending, that runs the risk of driving prices up by competing with private firms for labor and other resources. However:

There can be no inflationary pressures arising directly from a policy where the government offers a fixed wage to any labour that is unwanted by other employers. The JG involves the government buying labor off the bottom, in the sense that employment at the minimum wage does not impose pressure on the market-sector wage structure.

Government would not be involved in a bidding war with private companies for labor, since it would only be hiring labor for which there was no other demand.

The benefits would ramify throughout the economy because of the growth in public works, income, and consumer demand. That should stimulate some expansion in the private sector as well, to meet the increased demand. Private firms could get the additional workers they needed by hiring them away from the Job Guaranty program. That would be fine with the government, which would no longer need to employ them. The program simply absorbs unneeded labor until it is needed again, but does nothing to bid up the price of labor. It supplies a boost to aggregate demand only when there is enough unused capacity in the economy to respond to it. So there is no reason to expect either cost-push or demand-pull inflation as a result of the JG itself.

Effects on public deficit and private surplus

The expected economic effects of a Job Guaranty follow from the macroeconomic relationships described earlier.

GNP = C + I + G + CAB  [see MMT 3]

Gross National Product = Consumption + Investment + Government Spending + Current Account Balance

(T – G) + (S – I) + (-CAB) = 0  [see MMT 4]

These three sector financial balances add to zero:

T – G = Government balance of tax revenue minus spending

S – I = Private sector balance of saving minus investment

-CAB = External sector balance expressed as the current account surplus held by trading partners

Let’s start from the present U.S. situation, where financial surpluses in the private sector and the external sector are balanced by a large government deficit.

Let’s hold the external balance constant, so we can concentrate on the effects of a Job Guaranty on the domestic sectors, public and private.

When the Job Guaranty program starts:

  • G rises
  • GNP rises even more than G, because of the consumption multiplier
  • Government deficit rises
  • Private sector surplus rises

We are assuming that the increase in G is not offset by an increase in taxes. That would keep the increase from showing up in disposable income and block the multiplier effect on consumption. Since G rises but T doesn’t, the deficit (T – G) rises.

According to Modern Monetary Theory, the sovereign government can issue currency to spend beyond its revenue, and this public debt is sustainable. The government can also borrow money by issuing more treasury bonds without “crowding out” private borrowing, as is often alleged. That’s because the private surplus must increase in tandem with the public debt in order for the sector balances to offset. The mechanism by which this happens is the effect of Government spending on Saving due to the saving multiplier. Some of each additional dollar of income is saved, so S rises, and the surplus S – I must rise as much as the deficit T – G, other things being equal.

At the end of MMT 4, I expressed some concern that surplus savings not invested in real productive assets could lead to excess speculation and financial instability. This text does not address that possibility, but it makes me nervous about growing public deficits and private surpluses indefinitely.

Hopefully, the Job Guaranty program stimulates the general economy. As aggregate demand rises, the private sector needs to hire away more of the labor in the Job Guaranty program, so the program can be scaled back. But in order to sustain GNP at a high level, another variable in the GNP equation must increase to offset any reduction in government spending. Presumably that would be Investment, since the firms hiring more labor will also be providing more workplaces, equipment and expanded inventories. That leads to this optimistic scenario:

As private sector demand picks up:

  • G falls, but I rises
  • GNP is sustained at full-employment level
  • Government deficit falls
  • Private surplus falls

Private surplus (S – I) falls because of the rise in investment, which absorbs more of the uninvested saving. I also think that when the private sector is strong, it might be a good time to reduce the public deficit and private surplus by raising taxes on the wealthy, but the text does not get into that.

Necessary but not sufficient?

I like the text’s proposal for a Job Guaranty. I accept the authors’ argument that increasing public debt to fund it is not necessarily bad, since public debt is more sustainable than private debt. I would hope, though, that a period of expansionary fiscal policy might get the economy to a place where public deficits and other sector imbalances could actually be reduced.

One potential problem with the optimistic scenario is that investment in new technologies might displace too much labor, throwing millions of workers back into the Job Guaranty program. As private sector demand picks up and the private labor force moves toward full employment, that would strengthen the bargaining power of labor, according to the author’s conflict theory (see MMT 6). Ideally, investment in new technologies would raise worker productivity and justify wage increases. That would be a long overdue boost in productivity, which has been rather stagnant lately. On the other hand, automated and artificially intelligent systems could replace too many workers, especially those with limited education and technical skills. One can imagine a large underclass of otherwise unemployable workers stuck in minimum-wage jobs in the Job Guaranty program.

In order to develop human potential to the fullest, which is one of the text’s goals, government may need to spend on human capital development as well as the Job Guaranty, although the same program would have some effect on both. General spending to promote education, training, health care, and so forth are also needed.

Writers such as Martin Ford in The Rise of the Robots envision a massive welfare system to support people whose labor is no longer needed. I agree with the authors of Modern Monetary Theory and Practice that paying people not to work is a tremendous waste of human resources. “Providing welfare rather than work to those who want to work is not only an admission of defeat (the labour market fails to provide enough jobs), but also wastes resources and generates social costs.”

I accept the fundamental premise of this economics that “the most important resource in any economy is labour.” I want to enable people to do marketable work of some kind, although new technologies could raise productivity to the point where they wouldn’t need to devote many hours to it. I think that goal is best achieved through a balance of public and private investment. I hardly need to point out that little of this is likely until the present regime is history.

Rise of the Robots (part 3)

May 24, 2017

Previous | Next

Because Ford’s book is focused on the loss of human jobs to robots, he has next to nothing to say about job creation. If, however, a higher level of intelligence enables human beings to do things that machines cannot, as Ford himself admits, maybe we can do more of those things as we turn over the narrower thinking tasks to the machines.

The personalized service frontier

If there is any new frontier in job creation that can escape the rise of the robots, I think it would be in the realm of personalized services, the least routine and predictable things we do. In fact, when a service professional is helping a client, the problem of predictability is compounded.  If you’re a legal professional, artificial intelligence systems that process information about laws, cases and legal documents will be a great help. But lawyers still have to apply the law to the unique circumstances of the client’s case, and that is a more creative task.

Similarly, thousands of students can listen to the same lecture online, but they need a creative teacher to engage with their particular thought processes as they struggle to reconcile new ideas with what they already think. That’s why many educators are talking about “flipping the classroom”–letting students gather more information online while changing the classroom from a lecture hall into a setting for more creative collaboration. If all that students know how to do is take in lectures and cough up the material for the test, they will be at risk of being replaced by a machine. We can give in to the machines, or accept the invitation to take education to a new level that requires smaller classes and more teachers.

In so many areas, people need more personalized services than they are getting. In addition to teachers and financial planners, they need mental health services, legal services, job training, drug treatment programs, child care, and of course affordable health care. The question is whether these services will remain scarce and expensive, or whether we can expand the market for them in the information economy.

Making services more economical

We can be fairly sure that many menial service jobs will eventually be more economically performed by robots than by humans. The days of supermarket checkout clerks are numbered. The problem for aspiring professors, counselors, financial planners, and so forth is a little different. It is not so much that robots will replace them, but that too few people will be able to afford their services, or that they themselves will not be able to afford the price of admission to their desired profession.

I can think of several ways that the information revolution could help. As automation lowers the cost of producing goods and routine services, people can spend a larger portion of their income on personalized services. And information technology should also save labor in the personalized services themselves, bringing costs down there as well. A lawyer assisted by artificial intelligence shouldn’t have to spend as long preparing a case. I know that as a financial planner assisted by sophisticated software, I was able to prepare a financial plan in a very reasonable amount of time and at a very modest cost. My plans always had a human element, with personalized commentary as well as machine-generated tables and charts, but the human-machine collaboration made the service more affordable for my clients.

The technology was also very affordable for me. I did have to rely on a financial software company that no doubt made more money than I did. Ford emphasizes the centralization of information capital, a situation in which a few companies controlling software and Big Data can dominate markets while employing very few workers. But there is another side to that. Information can be duplicated at a very low marginal cost. Software development may be costly, but as the cost is spread over more and more copies, the unit cost keeps shrinking. An aspiring financial planner or other service provider can subscribe to software support for a very modest annual fee. Such easy access to information capital should make it easier to create personalized service jobs.

A big price of admission to many service professions is the cost of education. Education is such a public good that its cost should be widely spread throughout society. Making students go heavily into debt in order to learn what they need in order to be contributing members of society is not a very sensible policy. Ford agrees, and he hopes that new technologies can reduce the cost of instruction. He seems less interested in expanding higher education, since he expects people at all levels of education to have trouble finding jobs. I am more interested in such expansion, because I believe that the jobs we can create will usually require more education than the jobs we destroy.

The role of the public sector

If we agree that education is a public good whose cost should be widely spread throughout society, that suggests a major role for the public sphere in making it more accessible and affordable. The same logic could be extended to other services. Services that contribute to the general health, education and welfare of the population constitute public goods that are worthy of some public funding. Not only do such services create jobs in themselves, but they can help people build their human capital and meet the demands of the advanced economy, keeping them one step ahead of the robots.

Ford isn’t very supportive of this kind of public funding. Here’s what he has to say about elder-care:

The main problem with elder-care robots as they exist today is that they really don’t do a whole lot….The realization of an affordable, multitasking elder-care robot that can autonomously assist people who are almost completely dependent on others probably remains far in the future….It might seem reasonable to expect that the looming shortage of nursing home workers and home health aids will, to a significant extent, offset any technology-driven job losses that occur in other sectors of the economy….[But] by the time the majority of older people reach the point where they need personal, daily assistance, relatively few are likely to have the private means to hire home health aids, even if the wages for these jobs continue to be very low. As a result, these will probably be quasi-government jobs funded by programs like Medicare or Medicaid and will therefore be viewed as more of a problem than a solution.

So here we have a valuable service that isn’t being provided either by robots or enough human workers, and yet Ford rejects the expenditure of more public money to fund it. Once again, that reveals his narrow focus on his recommended basic income guarantee to support consumption. In effect, he would rather have government pay people not to work than to work. We can find more work for robots, but not create more jobs for humans.

Public funding requires some form of taxation. Conservatives often oppose higher taxes, especially on the wealthy, on the grounds that they will interfere with investments by the “job creators” in economic growth. If capital should become as self-serving as Ford expects, with businesses increasing profits by destroying jobs rather than creating them, that argument should become less convincing. One wonders how high unemployment will have to go before people turn to the public sector for job creation, as they did in the 1930s.

A broader moral argument

Ford is concerned about growing inequality, and he does make the argument that as taxpayers who have supported basic technological research, people have a legitimate claim on technology’s benefits. I agree, but I would also ground popular rights in a more basic principle, the dignity of human labor. Let the machines do the work they can do better than people can. But respect people as more than just purchasers of what the machines provide. Help people be as creative as they can be as producers–paid and unpaid–as well as consumers.

Rise of the Robots (part 2)

May 23, 2017

Previous | Next

In Rise of the Robots, Martin Ford describes how smart machines are starting to replace human workers in more and more kinds of work, raising the threat of a “jobless future.” He sees no alternative but a government-guaranteed basic income to support the millions of people who will have trouble finding jobs.

Although I have described Ford’s outlook as bleak, I am not opposed to letting machines do the work they can do better and more efficiently than humans. Nor am I opposed to strengthening the social safety net to assist workers who are most hurt by this transformation. What I do want to suggest is that Ford’s focus on job destruction alone may miss some of the more positive possibilities of the information economy.

Where does displaced labor go?

In the late nineteenth and early twentieth centuries, the United States experienced a dramatic decline in farm jobs, as we transitioned from a predominantly agricultural society to an urban industrial society. That was a hardship for many former farmworkers, who had to move to the city and compete with other displaced workers for whatever wages they could get. The “Great Migration” of African Americans from the rural South and waves of mostly Southern and Eastern European immigrants heightened the competition and provoked racial and ethnic conflict.

Nevertheless, the situation did not result in protracted high unemployment. The expansion of manufacturing and service industries absorbed most of the displaced workers. In addition, a lot of human activity was shifted away from employment by shortening the work week, phasing out child labor, and encouraging retirement. The Fair Labor Standards Act of 1938 reduced the standard work week to 40 hours and established minimum wages and overtime pay for many workers. These measures distributed the available work more widely, while keeping incomes up. The expansion of leisure also helped create job opportunities in leisure industries such as travel.

Ford would say that the ways we have avoided high unemployment in the past are no longer relevant, since we now face worker displacement on a much larger scale. As an artificial intelligence expert, he is focused more on how machines will work than on how humans can create and distribute work for themselves. He is very imaginative when it comes to what robots might do, but less imaginative when it comes to how humans might adapt. He seems to be making a couple of simplifying assumptions: first, that there’s only a fixed amount of work to be done; and second, that whatever work remains for humans after the robots move in must be hogged by the few rather than shared by the many. If these assumptions are wrong, then the job outlook isn’t as bleak as he makes it out to be.

Will humans still work?

IBM used to have a slogan: “Machines should work; humans should think.” Now that the machines are getting smarter, that distinction is harder to defend. Martin Ford’s slogan might be: “Machines should work and think, and a lot of humans should just consume.”

That’s a little unfair however, since Ford himself acknowledges that computers think only in a very limited and specialized sense. “Even IBM’s Watson, perhaps the most impressive demonstration of machine intelligence to date, doesn’t come close to anything that might reasonably be compared to general, human-like intelligence. Indeed, outside the realm of science fiction, all functional artificial intelligence technology is, in fact, narrow AI.” Ford is also more reluctant than other computer scientists to assume that general machine intelligence is inevitable or even possible, or to speculate that robots will soon have the ability to rebel against their creators.

Some brilliant philosophers and scientists have questioned whether a dead machine could ever have the capabilities of any living organism, let alone a human one. (See, for example, Robert Rosen’s Life Itself.) Arguments for continuing to distinguish ourselves from machines are not confined to religious traditionalists or Cartesian dualists who look for an immaterial soul within an otherwise mechanical system. They are made by scientists like Stuart Kauffman, who have a less reductionist and mechanistic conception of nature itself.  Many thinkers reject the idea that the universe in general and human thought in particular is reducible to algorithm (computational procedure). Philosophers in the Whitehead tradition argue that each human experience is a unique creation that synthesizes a multitude of past experiences. We should be careful about extrapolating from what we know well–our own machines–to aspects of the natural world that remain deeply mysterious to us.

Granted that computers can simulate human work by detecting patterns in what humans have already done. That includes existing works of art. A computer has “already produced millions of unique compositions in the modernist classical style.” But humans invented that style by experiencing and expressing modernity. The human creative work is a synthesis and expression of lived experience. The machine simulation is a meaningless exercise because the dead machine has no lived experience to express.

Ford says that narrow intelligence is all that the machines need in order to do most human work, since “the tasks that occupy the majority of the workforce are, on some level, largely routine and predictable.” But maybe the reason why so many jobs have been routine and predictable is precisely because we haven’t had anyone besides humans to do the boring work until now. Now that we can automate more of what we’ve been doing, how do we know there isn’t some new frontier of creative activity for humans to explore?

The distribution of human work

Suppose Ford is correct that half of the existing work could be done by machines. We can still imagine the future economy in more than one way. At one extreme, half of the human workers keep their existing jobs, while the other half become unemployed. At the other extreme, the human work is distributed among the same number of human workers so that each works half as much as before. The benefits of the new technology are taken the same way they were in the twentieth century–in the form of higher productivity and more leisure.

Because Ford is focused on replacement of the human worker, he plays down the possibility of productive collaboration between human and machine. The machines are out to get your job, and if you work with them you will be helping them learn to do so. “If you find yourself working with, or under the direction of, a smart software system, it’s probably a pretty good bet that–whether you’re aware of it or not–you are also training the software to ultimately replace you.” But that is true only to the extent that the work is predictable, general intelligence is irrelevant, and the human touch is dispensable. Ford seems to vacillate between admitting that machines cannot do everything and talking as if they can.

A recent article in the New York Times was titled, “Meet the People Who Train the Robots (to Do Their Own Jobs).” It reported that some companies are asking their employees to train artificial intelligence systems to act more like humans. However, the workers who told their stories did not see their human role as very endangered. A travel agent who used A.I. to book hotels said, “It made me feel competitive, that I need to keep up and stay ahead of the A.I.”  Using the system “frees me up to do something creative.” A customer service rep who was training a system to answer customer questions said that “she doesn’t foresee a future where she’s out of a job. Too many questions still require a level of human intuition to know the appropriate answer. There are also times when rules need to be broken, like when customers ask for an extension on their account because of some circumstances beyond their control.” The executive who developed a system for searching and analyzing legal documents said that he “doesn’t think A.I. will put lawyers out of business, but it may change how they work and make money. The less time they need to spend reviewing contracts, the more time they can spend on, say, advisory work or litigation.” As for myself, I have been using technology all my professional life to become more creative and productive, and I have trouble imagining any occupation where such collaboration couldn’t occur.

If most occupations allow for both human and technological input, the benefits of that collaboration could appear in some combination of higher output and reduced work hours. As with the twentieth-century technological advances, many workers could produce more while working less, and that would spread the available work to more people. How exactly this would be accomplished in our time I don’t know. It probably would not be as simple as legislating a new standard work week. But if the alternative is mass unemployment and paying people not to work, I think society will find a way.

Productivity and income reconsidered

The potential for human-machine collaboration calls into question Ford’s most basic contention, that artificial intelligence is a worker replacement and not a tool for raising worker productivity. In the “golden age” of the American economy, “As the machines used in production improved, the productivity of the workers operating those machines likewise increased, making them more valuable and allowing them to demand higher wages.” But those days are gone, along with a big chunk of the labor force. Are they really, or is it that we have not yet seen the social changes that would translate new technological capacities into worker benefits? Perhaps we are living in a period like the 1920s, just after the introduction of the assembly line but before the New Deal regulated wages and hours and recognized labor’s right to organize.

I return to a fundamental economic problem I raised in the last post. If technology can make us richer in output, then why should people settle for being poorer in consumption? In Ford’s imagined world of massive unemployment, government taxes the winners to provide a basic income to the losers. In an alternative vision, workers are typically technologically assisted, highly productive, and employed fewer hours so that many can work. They can become better off in two ways: winning a fair share of the benefits of their own productivity, and consuming goods and services that have become less expensive. Recall that assembly-line technology led to both higher wages and less expensive cars, a win-win for ordinary people. As in the past people will have to fight for such gains, and they will have to use the tools of democracy to get government on their side.

Another possibility explored by some futurists, such as Jeremy Rifkin, is that people who spend less of their time in paid employment will use new technologies to create goods and services to share for free. (If you find that idea absurd, you can start paying me for my posts right now!) The abundance of goods and services produced in the “collaborative commons” could reduce people’s need for money, softening the economic impact from reducing paid work hours.

In my final post on this book, I’ll explore possibilities for creating new jobs that robots are unlikely to do.