America Responds to Change: Implications for globalization and the future of work

This BBL Seminar is NOT for Quotation.

Date February 27, 2018
Speaker Bruce STOKES (Director, Global Economic Attitudes, Pew Research Center)
Moderator IWAMOTO Koichi (Senior Fellow, RIETI)

The United States has experienced significant cultural and economic change over the last 50 years. Nearly half the public does not like the pace of modern life and three-quarters believe their traditional way of life is getting lost. The percentage of foreign born in the country has tripled since 1967. The non-white portion of the public has tripled. The role of women has changed dramatically and those with no more than a high school education have become marginalized. Wages have stagnated for much of the period. Such changes can, in part, explain much of the uneasiness with globalization, which is change on steroids. And fueled the populist backlash against trade. Moreover, the transformation of American society over the last two generations has set the social, economic and political context for the next major shift: the advent of robots and artificial intelligence and the impact they may have on the future of work.
In this BBL seminar, Mr. Bruce Stokes will share recent Pew survey results and lead a discussion of this topic.



Bruce STOKES's Photo


Globalization was the last huge change our societies went through. The pace of change is now accelerating, with potentially huge changes looming in the near future as a result of automation, robotics, and artificial intelligence (AI). This acceleration of the pace of change—starting with globalization—has led to many problems. We have to assume that the pace of change may actually increase, and we, as a result, should probably anticipate even more problems.

I will mainly be sharing data on American public opinion. Later this year, we will be releasing a survey of public opinion in 38 countries, including Japan, about the pace of change, how people are responding to it and how they feel about it, and about attitudes toward AI and robotics.

The United States has changed very dramatically over my adult lifetime. Over the last 50 years, the percentage of non-white people has tripled. The percentage of foreign-born people has tripled. When I graduated from high school, 75% of Americans only had a high school education or less. Now, only 39% of the population have a high school education or less. This group went from a majority to a minority and now feels demeaned, although widespread consensus exists that a better educated population is a good thing. There have also been many other changes.

The American public on globalization and trade

One thing we know about globalization is that Americans are of two minds about it. They support it in principle but not necessarily in practice. We cannot dismiss this contradiction. Americans also have opinions about trade, but it is not a priority. Every year, we ask Americans to identify the biggest issues facing their society. Every year, trade is in one of the bottom two positions on the list, neck-and-neck with climate change. Relative to other things, people do not think trade is that important, but their concerns need to be put in context. Concern has increased somewhat over the last eight years: 32% said it was a top priority in 2010 and now 38% say it is a top priority. It has moved in a statistically significant direction. The perceived importance of climate change has also changed in striking fashion.

We asked, "What has helped or hurt your job?" Most people don't blame any of these things. Those who do blame, point fingers at outsourcing as the most important issue hurting their jobs. The second most important factor identified is the importation of foreign products. Workers at this point think that automation has helped their jobs. We will see whether that sentiment continues over time.

We ask people every year, "Are trade agreements good for the country?" Americans in principle believe that trade is good for the country: 58% said so in 2015. What happened after 2015? We had a presidential campaign in which all of the major candidates attacked trade. Politics can drive down public support, and it fell to 45% in October 2016. However, as soon as the politicians stopped talking about it, support rebounded. There is a partisan divide on this. Among the general public these days, Democrats are the free traders and Republicans are the protectionists. This has changed because the composition of the two parties has changed: the Democratic Party is increasingly a party of young people, minorities, and women. Those three groups are most supportive of trade, according to most of our surveys. The Republican Party is one of old white men. Why are such people opposed to trade? In the U.S. population, those most likely to have had manufacturing jobs that they lost to trade are men over the age of 50, of whom many voted for Donald Trump. Women didn't have those jobs. Young people were not born yet. Minorities were denied access to many of those jobs. Americans' views on trade are, in part, about life experience.

We asked people, "Does trade create or destroy jobs? Does it raise or lower wages? Does it help or hurt the economy." No matter how you ask these questions, people say it destroys jobs, lowers wages, and hurts the economy. For those of you who are economists, we asked whether trade increases or reduces prices. Economists tell us that we should support trade because it lowers prices. In the 41 countries in which we asked whether trade increases or reduces prices, only the Israelis believed that trade reduces prices. No one else believes that. Economists are telling us that, but the average person doesn't see the relationship between, for example, the reduced cost and greater power of personal computers and trade.

New challenges due to technological innovation

We are about to face another challenge with the advent of robotics, AI, etc. What do we do about that? At the end of the Barack Obama administration, I was invited to help brief the staff of the National Security Council. They were preparing memos for Hillary Clinton, but they turned out to be for Trump. A technology professor from MIT said that even though professors have lost credibility in the past for saying that new technologies were right around the corner when they never came about, he pleaded with us to believe them now because these things that were once seen as science fiction are now right around the corner. We have to prepare policies in response. If it doesn't pan out, we will have lost some time and effort. If it does pan out, we will have let our people down.

The estimates of how many jobs could be lost vary enormously. 47% of U.S. jobs are vulnerable according to a study at Oxford University. Researchers at McKinsey & Company said 50% of jobs are automatable. On the other hand, the Organisation for Economic Co-operation and Development (OECD) said it is only 9%. Even the experts are not sure. When we surveyed AI experts, they predicted that AI will be able to outperform translators by 2024 and will be able to drive trucks by 2027—there are 3.5 million truck drivers in the United States. By 2031, they believe that AI and robotics will be able to do retail work (there are 3.5 million cashiers in the United States). Many jobs will potentially be disrupted.

The cost per hour of manufacturing workers versus robots is striking. One estimate is that robots cost about $4 per hour. A manufacturing worker in the United States costs $36 per hour. Any manufacturer that can automate a job will probably consider doing it with those kinds of differences in cost. Robots are still more expensive than the average labor in China, but it's not that much different. It's already much cheaper to use a robot than a worker in Japan.

This is the most challenging statistic to my mind. Some technological optimists say that we adapted to change in the past so that we will be able to adapt to change in the future. However, this overlooks the fact that between 1900 and 1940, the United States moved from an agriculture-based economy to a manufacturing economy. It took 40 years. Nearly 40% of the labor force was disrupted over that period. 1.2 million jobs were lost per year on average. The problem is that it is estimated that only 700,000 of those people found new jobs. In this period of change to which everyone supposedly adapted, half a million people got lost every year.

One estimate is that over the next 20 years, 2.5 million jobs per year will be disrupted by robotics and AI. That is twice the amount of disruption in half of the time. We have no idea how many of these people can be reabsorbed into the economy. What will be the reaction of the people who lose their jobs to robots? They will look for someone who promises to fix this problem. We all know that a fix is probably not possible, but we should be wary of a populist reaction.

American views on robot/AI revolution

We asked people in the United States whether they had heard about robots and AI, and 85% said they had heard a lot or a little. 77% said they thought it was extremely or somewhat realistic that this is going to happen. They were not enthusiastic about it: a majority said they are worried. The idea that people are looking forward to this is wrong. We then asked people what kind of things they think will happen over the next 20 years, and how it will impact life. People anticipate that doctors will use computers to diagnose and treat, stores will be fully automated, and deliveries will be made by drones.

One of the most interesting findings is when we gave people a list of jobs and asked which they thought would be affected. People had different views, but only 30% of Americans said that they thought it was going to affect their job. That means it is going to affect him and her but never me. I think we have to contemplate what happens if it also happens to "me." Who feels their jobs might be threatened? People with high school educations or less and also young people who probably understand this better. Gallup had interesting results. 73% of blue collar workers said they thought that AI would eliminate more jobs than it creates. When we surveyed experts on this very same issue, with more or less the same question, 52% of experts said it will create more jobs than it destroys. There is an elite-versus-popular gap. That can lead to new political tensions.

We asked the public what they believe the negative outcomes of this will be. We did not find a great deal of belief in positive outcomes. 76% said inequality is going to grow. 75% said they do not believe that this technological revolution will create many new better paying jobs, which is one of the claims made by technological optimists. It is not too early to begin to ask people expressing a lack of enthusiasm and worry about how they want businesses and the government to respond. We offered various alternatives. 85% of Americans say these advanced robotics and AI-driven machines should be limited to doing dangerous or unhealthy jobs. That is not likely. 63% say people should pay an additional amount to have a human being serve them in a store rather than a robot. 61% say the government should offer a guaranteed income to meet basic needs, which may be one of the major policy debates going forward. We asked, "Who has the obligation to care for workers displaced by robots and AI?" Americans are among the most individualistic people in the world, so they tend to say that people should take care of themselves. However, on this question, they were divided over whether the government or individuals have the responsibility. Democrats are most likely to say government has a role to play, while two-thirds of Republicans say no.

Most people say there should be limits on what businesses should do, with very little support for businesses replacing human workers because of cost, etc. We asked about training. It is essential for people to receive skills training. The United States has a very decentralized and ineffective skills training system which the Japanese tried to help us fix at one point, but the effort failed. The Germans are now trying to help us fix it. A Gallup poll asked people whether a person who loses their job to technology would need more training to find another job. An overwhelming majority said yes in many parts of the country, but only 42% of people in the Midwest said so. There seems to be a lack of recognition in areas where the impact may be most likely, which makes retraining even harder. And that is the part of the country that mostly voted for Trump.

We also asked a different question: "Who should have the responsibility for ensuring that the American workforce has the right skills and education?" 72% of people in the United States said it is the individual him/herself. They put the responsibility on the displaced person. This is a political challenge the United States has to face, as some people do not feel they need retraining and seem unsupportive of society providing retraining. Will people actually get the skills and education they need by themselves?

Gallup asked about universal income. People are divided. We found more support for it than Gallup did. Democrats think it is a good idea, Republicans think it is bad.

There are a number of related studies on our website. More U.S. and worldwide surveys will be coming over the course of the year. Our site is free and searchable. Let's open this up to questions.


Q1. Some car and airplane manufacturers have already invented automated vehicles. When do you think automated cars or airplanes will be introduced?

That's beyond my expertise. We asked the elites and the public about automated cars. My personal view is that younger people will be earlier adopters even though older people would benefit more. We took away my mother's driver's license when it seemed unsafe for her to drive, and she felt she lost a lot of freedom. An automated car would solve that. As an older person myself, the idea of getting into a car and letting it drive is scary, but if the automated car has a lower accident rate than humans, it would actually be safer. As for airplanes, maybe people would be more willing to accept it as we don't see the pilot anyway

Q2. On ideology in American politics, the delineation seems to have shifted to one between populists and libertarians. How precise is the distinction? If populism has become a major political ideology, it doesn't seem to be reflected in your data. There could be more support for government intervention if populism grows.

There are libertarian-versus-populist tendencies across a number of subjects, but the issue is that we only have two political parties and will likely continue to only have two. There are populists and traditionalists and libertarians in the Republican Party and, to a certain extent, the same in the Democratic Party, and sometimes they are libertarians about different issues. Is there going to be a political realignment along the lines of liberalist versus populist? That may have been a goal of Steve Bannon, but that has not really worked, and he is gone. The tax bill that passed in Congress was pretty traditionally Republican. However, deficit reduction has traditionally been a major Republican issue, and the Republicans just increased it. Now, the Democrats are worried about that, which they never did in the past.

Our surveys do show a more important trend: the polarization in American society. We ask almost every year about values. The parties have always had different values, but the differences were small in the past. Now they are huge. Self-identified Republicans have moved to the right while self-identified Democrats have moved to the left. There is no middle ground on many issues. What is even more depressing is that when we ask people about non-ideological issues and about more personal issues, they are also divided along conservative and liberal lines. For example, we asked recently, "Do you have many friends that belong to the other party?" A majority of Republicans—about 57%—said a lot of their friends are Republicans. 65% of Democrats said the same about their friends. We are ideologically self-segregating ourselves. That will be hard to fix. We asked people, "How would you feel if your son or daughter married someone of another race?" Many liberals said that would be fine, and many conservatives said that would not be fine. It's not just about big-versus-small government or trade; it's on issue after issue. That will make it harder for other countries to deal with the United States. Who are these Americans?

Q3. What is your perception of China's influence? It is felt everywhere. Open and free societies seem to be in competition with authoritarian societies in the development of advanced technology and innovation. Which type of society do you believe to be friendlier to future innovations? To what extent do Americans perceive China in that way?

One of the surprising outcomes of the 2017 survey was that American views of China actually improved somewhat, despite some harsh criticism of China during the 2016 presidential campaign. I don't know why. In talking to American elites about China, I think the attitudes have gotten much worse. Certain American elites in the business and academic communities who had economic interests in China have become very negative. Young people have a more favorable view of China than older Americans, which has always been the case.

Q4. I wonder about Americans' reaction to Xi Jinping expressing strong support for free trade. It seems a ridiculous statement to me. Second, what is the Democratic Party's policy on job training? To avoid more unemployment, you have to invest in training.

The average American didn't hear about Xi Jinping's statement in Davos and has no reaction to it. He said it, but I imagine most people didn't hear it. Among the elites, people found this to be a joke, but also there is some concern that his vision of what that would be would be some kind of "Chinese version" of free trade. As for retraining, what I have observed over the years is that non-economists will now very often say we need more retraining. They did not say that 20 years ago. The U.S. economy is now doing well, but workers need to be retrained. Other than some German- and Japanese-style programs, not much evidence exists that we know how to do this, and not much evidence that we will invest in it. The Trump administration cut spending on retraining programs. The programs we tried in the past did not work. On-the-job training requires trust between industry, workers, and the government. Very strong unions in northern Europe help retraining to work more effectively. Unions are weak in the United States; no one exists to oversee retraining, there is no trust, and companies think only in the short term. A mix of problems have contributed to the fact that we do not do training very well. Even without these problems, it remains difficult. How do you train someone for the coming future? Things are moving so fast.

Q5. You expect a rapid decrease in demand for human labor in the future. However, in Japan, we are experiencing a rapid reduction in the size of the human labor force. Do you think that these differences in demographic patterns will affect attitudes toward AI and so on?

We don't know, but we will try in future surveys to get at differences in different countries. Looking at the United States, it's going to be a challenge. It may be less of a challenge for Japan or Korea. My only concern is that it could lead Japan and Korea and others to not worry enough about this problem. It may affect you less, but it may also affect you in different ways. It may not involve job losses but something else.

Q6. First, on free trade, I am surprised by the dramatic reduction in support for free trade among Republicans. I would like to know the reason for this and how this will affect future U.S. politics. As for the displacement of workers, what kind of jobs can be created by new technologies and how will that impact workers?

The fact that Republicans are less likely to support trade than Democrats today has a lot to do with the changing nature of the Republican Party: it's more male, white, and older. The rhetoric of the campaign is also a factor: very sharp criticism of trade by Trump, etc. Part of the partisan nature of American public opinion is that people tend to feel better about an issue when their guy is in the White House. They expect that their leader will do a better job of dealing with it than the other side would have.

As for jobs created by new technologies, certainly among these are robotics engineers and programmers. When I talk to technology optimists, they say not to worry, we were worried about the death of the buggy whip industry, and it came out fine. There will be new jobs in the future that none of us can conceive of now. But will there be enough of those jobs? Will they be satisfying? Will they pay enough? At the extreme, technological optimists say we won't need people to work. But human beings have derived a large part of their identity, for thousands of years, from what they do. We understand who we are by what we do. The people I went to school with are not going to paint watercolors and write poetry if they don't need to work. They are proud to be steelworkers. What happens if some people are just not needed? We will find a way to keep them alive, but what are they going to do all day? People think community, family, work, and religious life are all changing. Will this AI/robotics revolution be another blow to people's sense of identity through the loss of their work? Maybe we should worry about that.

Q7. Such great differences exist between the U.S. West and Midwest on the issue of retraining. What factors are responsible for this? Can any public policy or global industrial action be taken that can change the attitude regarding trade policy?

The West is younger than the Midwest. Older people may feel they do not need retraining. As for trade: my hometown has a steel mill which is now Japanese-owned. Its workers know that about 15% of what they make is exported. They are not mindlessly anti-trade. The fact that they export and that the company has explained this to them are responsible for this. Companies should educate their workers in this way. Nevertheless, 60% of my hometown voted for Trump despite this. We have to tell workers that change is inevitable. The government also has to rebuild public trust in the idea that businesses and the government will actually help them. Our surveys indicate that the public doesn't trust government or big companies. The big companies are sometimes the ones with the wherewithal to try retraining. But if you don't trust the companies, it may not work.

*This summary was compiled by RIETI Editorial staff.