Tuesday, September 12th, 2017 12:56 pm

Robots and Artificial Intelligence

Question A: Holding labor market institutions and job training fixed, rising use of robots and artificial intelligence is likely to increase substantially the number of workers in advanced countries who are unemployed for long periods.

Responses
 

Source: IGM Economic Experts Panel
www.igmchicago.org/igm-economic-experts-panel

Responses weighted by each expert's confidence

Source: IGM Economic Experts Panel
www.igmchicago.org/igm-economic-experts-panel

Question B: Rising use of robots and artificial intelligence in advanced countries is likely to create benefits large enough that they could be used to compensate those workers who are substantially negatively affected for their lost wages.

Responses
 

Source: IGM Economic Experts Panel
www.igmchicago.org/igm-economic-experts-panel

Responses weighted by each expert's confidence

Source: IGM Economic Experts Panel
www.igmchicago.org/igm-economic-experts-panel

Question A Participant Responses

Participant University Vote Confidence Comment Bio/Vote History
Acemoglu Daron Acemoglu MIT Uncertain 7
Recent research finds negative employment effects from industrial robots. Effects of AI and more mature robotics tech could be different.
-see background information here
Bio/Vote History
         
Alesina Alberto Alesina Harvard Did Not Answer
Bio/Vote History
         
Altonji Joseph Altonji Yale Uncertain 4
Bio/Vote History
         
Auerbach Alan Auerbach Berkeley Disagree 3
Bio/Vote History
         
Autor David Autor MIT Disagree 6
It's not impossible, but I'm not so far seeing the evidence that "that this time is different."
Bio/Vote History
         
Baicker Katherine Baicker Chicago Disagree 1
Bio/Vote History
         
Banerjee Abhijit Banerjee MIT Strongly Agree 7
Bio/Vote History
         
Bertrand Marianne Bertrand Chicago Agree 3
Bio/Vote History
         
Brunnermeier Markus Brunnermeier Princeton Uncertain 5
Bio/Vote History
         
Chetty Raj Chetty Stanford Did Not Answer
Bio/Vote History
         
Chevalier Judith Chevalier Yale Did Not Answer
Bio/Vote History
         
Cutler David Cutler Harvard Agree 3
Bio/Vote History
         
Deaton Angus Deaton Princeton Agree 8
Not because it must, but because the policies needed to prevent it will not be forthcoming,.
Bio/Vote History
         
Duffie Darrell Duffie Stanford No Opinion
Bio/Vote History
         
Edlin Aaron Edlin Berkeley Did Not Answer
Bio/Vote History
         
Eichengreen Barry Eichengreen Berkeley Disagree 6
Lots of jobs will be left requiring compassion, communication and intuition.
Bio/Vote History
         
Einav Liran Einav Stanford Uncertain 3
Bio/Vote History
         
Fair Ray Fair Yale Disagree 5
Bio/Vote History
         
Finkelstein Amy Finkelstein MIT Uncertain 5
Bio/Vote History
         
Goldberg Pinelopi Goldberg Yale Agree 5
Bio/Vote History
         
Goolsbee Austan Goolsbee Chicago Disagree 9
Bio/Vote History
         
Greenstone Michael Greenstone Chicago Agree 4
key is the length of adjustment, recent research, including Acemoglu and Restrepo, suggest meaningful periods of time.
Bio/Vote History
         
Hall Robert Hall Stanford Uncertain 7
So far, the effects seem to be small-the labor force has shrunk only a bit. But the future could see much more shrinkage.
Bio/Vote History
         
Hart Oliver Hart Harvard Agree 5
There will be jobs in health and care service.Workers will need training. Unemployment may rise initially, but maybe not long-term.
Bio/Vote History
         
Holmström Bengt Holmström MIT Agree 5
Bio/Vote History
         
Hoxby Caroline Hoxby Stanford Uncertain 10
Bio/Vote History
         
Hoynes Hilary Hoynes Berkeley Uncertain 9
Bio/Vote History
         
Judd Kenneth Judd Stanford Agree 8
Unemployment insurance benefits will be used to facilitate longer job search spells in the job market with fewer opportunities.
Bio/Vote History
         
Kaplan Steven Kaplan Chicago Agree 3
Bio/Vote History
         
Kashyap Anil Kashyap Chicago Uncertain 3
they will destroy a lot of jobs and will create many too. hard to know if the net, long-term effect will be substantial
Bio/Vote History
         
Klenow Pete Klenow Stanford Agree 5 Bio/Vote History
         
Levin Jonathan Levin Stanford Agree 4
Yes, with significant uncertainty. How quickly jobs will be lost and new jobs will be created is hard to forecast.
Bio/Vote History
         
Maskin Eric Maskin Harvard Agree 7
Bio/Vote History
         
Nordhaus William Nordhaus Yale Strongly Disagree 9
For at least the next 20 years. After that, depends on evolution of AI etc.
Bio/Vote History
         
Saez Emmanuel Saez Berkeley Disagree 6
Bio/Vote History
         
Samuelson Larry Samuelson Yale Agree 8
But with appropriate institutions and retraining, the effect on unemployment could be significantly mitigated.
Bio/Vote History
         
Scheinkman José Scheinkman Princeton Uncertain 6
Historical evidence suggests that this increase will be relatively short-run.
Bio/Vote History
         
Schmalensee Richard Schmalensee MIT Agree 4
Bio/Vote History
         
Shapiro Carl Shapiro Berkeley Uncertain 1
Bio/Vote History
         
Shimer Robert Shimer Chicago Disagree 5
Holding fixed institutions, wages and labor force participation will fall for some workers, but unemployment will not increase
Bio/Vote History
         
Thaler Richard Thaler Chicago Uncertain 1
I don't understand the premise. We create tools but no one learns how to use them?
Bio/Vote History
         
Udry Christopher Udry Yale Agree 6
Bio/Vote History
         

Question B Participant Responses

Participant University Vote Confidence Comment Bio/Vote History
Acemoglu Daron Acemoglu MIT Strongly Agree 9
Negative wage effects on at least some workers and productivity improvements are likely. Politics is the real constraint on redistribution
Bio/Vote History
         
Alesina Alberto Alesina Harvard Did Not Answer
Bio/Vote History
         
Altonji Joseph Altonji Yale Agree 5
Bio/Vote History
         
Auerbach Alan Auerbach Berkeley Agree 7
Bio/Vote History
         
Autor David Autor MIT Agree 9
Panelist meant to strongly agree (question misread). Though these techs will almost surely ↑ GDP, losers almost surely won’t be compensated.
Bio/Vote History
         
Baicker Katherine Baicker Chicago Agree 1
Bio/Vote History
         
Banerjee Abhijit Banerjee MIT Uncertain 7
Having a job matters even if it does not change earnings. I wonder what mechanisms will keep the population engaged and emotionally rewarded
Bio/Vote History
         
Bertrand Marianne Bertrand Chicago Uncertain 2
Bio/Vote History
         
Brunnermeier Markus Brunnermeier Princeton Agree 5
Bio/Vote History
         
Chetty Raj Chetty Stanford Did Not Answer
Bio/Vote History
         
Chevalier Judith Chevalier Yale Did Not Answer
Bio/Vote History
         
Cutler David Cutler Harvard Agree 7
Bio/Vote History
         
Deaton Angus Deaton Princeton Strongly Agree 10
Not that they will. "Could be"
Bio/Vote History
         
Duffie Darrell Duffie Stanford No Opinion
Bio/Vote History
         
Edlin Aaron Edlin Berkeley Did Not Answer
Bio/Vote History
         
Eichengreen Barry Eichengreen Berkeley Agree 8
Bio/Vote History
         
Einav Liran Einav Stanford Strongly Agree 3
Bio/Vote History
         
Fair Ray Fair Yale Agree 5
Bio/Vote History
         
Finkelstein Amy Finkelstein MIT Strongly Agree 9
Bio/Vote History
         
Goldberg Pinelopi Goldberg Yale Agree 5
Bio/Vote History
         
Goolsbee Austan Goolsbee Chicago Strongly Agree 10
Bio/Vote History
         
Greenstone Michael Greenstone Chicago Agree 6
"could"-- yes but is a favorite academic ? of unclear real world relevance- benefits from big dislocations are not usually redistributed
Bio/Vote History
         
Hall Robert Hall Stanford Agree 4
There are already some bad signs about the shrinking labor force. Those not in the LF are unhappy and inclined to opiods.
Bio/Vote History
         
Hart Oliver Hart Harvard Agree 5
The pie is bigger and so monetary compensation is possible. But people may lose self-esteem if they no longer have a"good" job.
Bio/Vote History
         
Holmström Bengt Holmström MIT Agree 7
Bio/Vote History
         
Hoxby Caroline Hoxby Stanford Uncertain 10
Bio/Vote History
         
Hoynes Hilary Hoynes Berkeley Uncertain 9
Bio/Vote History
         
Judd Kenneth Judd Stanford Strongly Agree 9
The key word is "could". It is unclear how effective the political institutions will be in making this happen. I am pessimistic.
Bio/Vote History
         
Kaplan Steven Kaplan Chicago Strongly Agree 8
Bio/Vote History
         
Kashyap Anil Kashyap Chicago Agree 5
But i am skeptical this will be pursued much in the US, witness the program on trade adjustment assistance
Bio/Vote History
         
Klenow Pete Klenow Stanford Strongly Agree 5 Bio/Vote History
         
Levin Jonathan Levin Stanford Agree 4
The potential benefits are large. Will they be used to compensate for job loss? May be determined more by politics than economics.
Bio/Vote History
         
Maskin Eric Maskin Harvard Agree 7
Bio/Vote History
         
Nordhaus William Nordhaus Yale Agree 9
Slightly weird question. Impacts are + and -, with net being positive. Hard to know where gross job losses fit into the + and -, however.
Bio/Vote History
         
Saez Emmanuel Saez Berkeley Agree 8
Bio/Vote History
         
Samuelson Larry Samuelson Yale Agree 8
It is (alas) by no means clear that we can muster the political will to make the appropriate compensations.
Bio/Vote History
         
Scheinkman José Scheinkman Princeton Agree 7
Bio/Vote History
         
Schmalensee Richard Schmalensee MIT Strongly Agree 6
But, of course, it is not likely that those benefits will in fact be used to compensate the losers.
Bio/Vote History
         
Shapiro Carl Shapiro Berkeley Agree 6
Bio/Vote History
         
Shimer Robert Shimer Chicago Strongly Agree 8
But the caveat is that compensating these workers will further reduce their employment
Bio/Vote History
         
Thaler Richard Thaler Chicago Agree 3
The key word is "could". In the recent years the gains from tech have not trickled down much in terms of money.
Bio/Vote History
         
Udry Christopher Udry Yale Strongly Agree 9
...but three compensation is very unlikely to be paid for most of those who lose their jobs.
Bio/Vote History
         

10 New Economic Experts join the IGM Panel


For the past two years, our expert panelists have been informing the public about the extent to which economists agree or disagree on important public policy issues. This week, we are delighted to announce that we are expanding the IGM Economic Experts Panel to add ten new distinguished economists. Like our other experts, these new panelists have impeccable qualifications to speak on public policy matters, and their names will be familiar to other economists and the media.

To give the public a broad sense of their views on policy issues, each new expert has responded to a selection of 16 statements that our panel had previously addressed. We chose these 16 statements, which cover a wide range of important policy areas, because the original panelists' responses to them were analyzed in a paper comparing the views of our economic experts with those of the American public. You can find that paper, by Paola Sapienza and Luigi Zingales, here. The paper, along with other analyses of the experts' views, was discussed during the American Economic Association annual meetings, and the video can be found here.

The new panelists' responses to these statements can be seen on their individual voting history pages. Our ten new economic experts are:

Abhijit Banerjee (MIT)
Markus K. Brunnermeier (Princeton)
Liran Einav (Stanford)
Amy Finkelstein (MIT)
Oliver Hart (Harvard)
Hilary Hoynes (Berkeley)
Steven N. Kaplan (Chicago)
Larry Samuelson (Yale)
Carl Shapiro (Berkeley)
Robert Shimer (Chicago)


Please note that, for the 16 previous topics on which these new panelists have voted, we left the charts showing the distribution of responses unchanged. Those charts reflect the responses that our original panelists gave at the time, and we have not altered them to reflect the views of the new experts.

We have also taken this opportunity to ask our original panelists whether they would vote differently on any of the statements we have asked about in the past. Several experts chose to highlight statements to which they would currently respond differently. In such cases, you will see this "revote" below the panelist's original vote. We think you will enjoy seeing examples of statements on which some experts have reconsidered.

As with the 16 previous statements voted on by new panelists, these "revote" responses are not reflected in the chart that we display showing the distribution of views for that topic: all the charts for previous questions reflect the distribution of views that the experts expressed when the statement was originally posed.

About the IGM Economic Experts Panel

This panel explores the extent to which economists agree or disagree on major public policy issues. To assess such beliefs we assembled this panel of expert economists. Statistics teaches that a sample of (say) 40 opinions will be adequate to reflect a broader population if the sample is representative of that population.

To that end, our panel was chosen to include distinguished experts with a keen interest in public policy from the major areas of economics, to be geographically diverse, and to include Democrats, Republicans and Independents as well as older and younger scholars. The panel members are all senior faculty at the most elite research universities in the United States. The panel includes Nobel Laureates, John Bates Clark Medalists, fellows of the Econometric society, past Presidents of both the American Economics Association and American Finance Association, past Democratic and Republican members of the President's Council of Economics, and past and current editors of the leading journals in the profession. This selection process has the advantage of not only providing a set of panelists whose names will be familiar to other economists and the media, but also delivers a group with impeccable qualifications to speak on public policy matters.

Finally, it is important to explain one aspect of our voting process. In some instances a panelist may neither agree nor disagree with a statement, and there can be two very different reasons for this. One case occurs when an economist is an expert on a topic and yet sees the evidence on the exact claim at hand as ambiguous. In such cases our panelists vote "uncertain". A second case relates to statements on topics so far removed from the economist's expertise that he or she feels unqualified to vote. In this case, our panelists vote "no opinion".

The Economic Experts Panel questions are emailed individually to the members of the panel, and each responds electronically at his or her convenience. Panelists may consult whatever resources they like before answering.

Members of the public are free to suggest questions (see link below), and the panelists suggest many themselves. Members of the IGM faculty are responsible for deciding the final version of each week’s question. We usually send a draft of the question to the panel in advance, and invite them to point out problems with the wording if they see any. In response, we typically receive a handful of suggested clarifications from individual experts. This process helps us to spot inconsistencies, and to reduce vagueness or problems of interpretation.

The panel data are copyrighted by the Initiative on Global Markets and are being analyzed for an article to appear in a leading peer-reviewed journal.

chicago booth