Freakonomics Radio - 420. Which Jobs Will Come Back, and When?

Episode Date: June 4, 2020

Covid-19 is the biggest job killer in a century. As the lockdown eases, what does re-employment look like? Who will be first and who last? Which sectors will surge and which will disappear? Welcome to the Great Labor Reallocation of 2020.

Today's show is the first of two episodes about employment and, of course, unemployment, considering what's been happening with the COVID-19 pandemic. It is especially about re-employment, that is, what kind of jobs are coming back and when and which jobs aren't coming back. So we will hear from a labor economist with the Federal Reserve, another economist who used to work in the White House and the Department Federal Reserve, another economist who used to work in the White House and the Department of Labor, another economist who thinks that even before the pandemic, we had automated away too many jobs, and another person who shares that view about jobs and

automation, the former presidential candidate Andrew Yang, whose call for a universal basic income has become a lot more urgent in the past few months. These two episodes are also about prisoners, specifically prisoners and reemployment. And we'll ask whether that research can tell us anything about COVID-19 reemployment generally. We are living through an historic disruption, a jolt to the labor markets that was unimaginable just a few months ago. There will be books and books written

about it, and 50 years from now it'll show up in economics textbooks. Maybe they'll call it then what we're calling it now, the great labor reallocation of 2020. From Stitcher and Dubner Productions, this is Freakonomics Radio, the podcast that explores the hidden side of everything. Here's your host, Stephen Duffner. We'll start today with a particularly interesting researcher. My name is Jennifer Doliak, and I'm an economics professor and the director of the Justice Tech Lab at Texas A&M University.

What is the Justice Tech Lab? It is a research group that focuses on empirical research related to crime and discrimination. So we try to find answers to some of our trickier policy problems. Doliak did not set out to become a crime economist. I kind of assumed I would be an investment banker like every other econ major. But she repeatedly found herself drawn to the topic. I saw a New York Times article about DNA databases and how the laws vary state to state in terms of which offenders are required to provide DNA to our criminal justice DNA databases. And I thought, ah, that's a great natural experiment.

An experiment that could help her answer a key question about criminals and incentives. So if your DNA is added to the database, then your DNA sample is frequently compared with DNA evidence from crime scenes. And so my paper showed that that dramatically reduced recidivism for people in the U.S. Meaning because it acted as a disincentive to commit further crimes because of the knowledge that your likelihood of being caught is higher? Exactly, yeah. So increasing the probability of getting caught then has a deterrent effect on future crime. And then it just turns out crime is really interesting. Now, do you have a lot of personal experience with crime?

Not yet. Maybe one day. It may strike you that crime is an unusual topic for an economist, but if you think about crime as an industry, it's a big one. There are nearly 1.5 million people currently imprisoned in the U.S. It typically costs about $100 a day to lock someone up. So that alone is $150 million a day or some $55 billion a year. Then there are the costs associated with law enforcement, the criminal defense and court systems, and so on, all of which produce countless issues that appeal to an economist. I think the biggest issue is the incredible inefficiency. So we lock a huge number of people up every year. We incarcerate them for a really long time. That has cost to them personally, likely has cost to

the community. There are probably other more cost-effective ways to reduce crime than just locking people up. Policing is another place where we have a bunch of research showing that hiring more police officers on average reduces crime, but we know remarkably little about what is valuable for police to actually spend their time on during the day and what tips into what we would consider over-policing. And there are just a bunch of problems like that in the courts and re-entry and so on. Some of those 1.5 million imprisoned Americans have plainly done terrible things and belong in prison, perhaps forever. But a consensus has begun to emerge that many of those 1.5 million don't necessarily belong in prison, that they and the rest of us

would be better served if they were back in society and employed, contributing. This is one of the rare policy positions these days that has support across political boundaries. This is probably my favorite thing about working in this space is that I frequently wind up talking to people who are far more liberal and far more conservative than I am at the same table, and everyone's trying to work toward the same goal. Jennifer Doliak sees people engaged in criminal justice reform

for one of three main reasons. So you've got the libertarians who look at our criminal justice system as just another big failed government program and they want to shrink it. And then you have the more religious right that believes strongly in redemption and second chances. And then you've got folks on the left who worry about our criminal justice system and the racial disparities involved in it and criminalizing poverty and so on. And all these people, regardless of ideological perspective, are asking the same important question. What's the best way to reduce the prison population

while keeping the public safe? One key number to look at is the rate of recidivism, that is, the likelihood that someone released from prison winds up back in prison. Recidivism numbers in the U.S. are astonishing. In a typical year, roughly 600,000 prisoners are released. Within three years, two-thirds of them will return to prison. If you happen to be released at a time when the local labor market is really bad, like right now, we might expect recidivism rates overall to increase. We're making it really hard for people to kind of get back on their feet and build a stable life free of criminal activity.

A stable life free of criminal activity is dependent on getting a job after prison. How hard is it for a former prisoner to get a job? So this is a much more complicated answer than you might expect. Part of the challenge in working in this space is how little we actually know about this population. So if you think about the census and all of our other big national surveys, none of those surveys ask a question about whether you have a criminal record. That sounds terrible for you. Yes, it is challenging. Why do the kind of surveys that you mentioned not include questions about imprisonment? I would think the roots are

anti-discriminatory, yes. You know, you worry that people wouldn't answer truthfully or that they might be worried about the government, you know, coming after them if they say yes to that question. Imagine for a minute that you are the hiring manager for a company with a few hundred employees. Maybe it's a food packaging company or a construction firm. You are constantly needing to hire new workers, typically young men who don't have a college education. Would you prefer to hire someone who has or hasn't been to prison? The argument that you hear sometimes is that people who've come out of prison and have had a hard time finding a job and really, you know, don't want to go back,

they make great employees because they will work really hard and they'll be really loyal to you. So you do hear that from employers sometimes. Sometimes, but not all the time. Many employers are much less eager to hire a former prisoner. The reasons that employers tend to give are usually along the lines of they're worried about the legal liability that comes with having someone on the job who might commit another crime. There are other potential things like, you know, people coming out of prison on average are at higher risk for substance abuse and that sort of thing. So if you are the hiring manager at that construction or food packaging firm, how do you know whether a potential employee has been in prison. Remember, the census and other surveys don't ask that question.

The internet has helped make criminal background checks more widely available, but that still costs time and money. In the old days, without a centralized database, background checks were slower and less comprehensive. So employers came up with a simple solution. On every job application, they just asked the applicant to say whether they'd been in prison. So traditionally, there's a box you check.

You know, check this box if you have a criminal record, if you have a felony conviction. You could, of course, lie, but this box was considered useful enough that for decades it was included on job applications. In the late 1990s, however, there arose a movement called Ban the Box. So Ban the Box is a policy that aims to help people with criminal records get their foot in the door with employers and get a job. And it works by prohibiting employers from asking job applicants if they have a criminal record until relatively late in the hiring process. In other words, employers could still do a background check before hiring someone, but they'd be less inclined to rule out an applicant early.

Theoretically, this would give former prisoners better chance of getting hired. They wouldn't be dismissed out of hand just from their application. And with a better chance of getting a job, maybe there would be less recidivism. That at least was the idea. And especially after the last financial crisis, when young men without a college education were having a really hard time getting jobs, the Ban the Box movement gained momentum, again, across the political spectrum.

So you have some very left-leaning groups coming at this for more social justice reasons. And then you had, you know, the Koch brothers coming from a more libertarian perspective and also the belief in redemption. They were really pushing for this policy as a way to ensure second chances for people who had paid their dues and served their time and now wanted to get a job. Jennifer Doliak first heard about the Ban the Box movement in 2013 after she finished grad school. As with the study she had done on DNA databases, Ban the Box policy differed from state to state. What winds up getting a lot of media attention is what happens in the federal government. And, you know, nothing gets done in the federal government. Especially in her field.

Criminal justice policy is almost entirely made at the state and local level. In this case, President Obama did ban the box on federal government jobs in 2015. As for private employers, 30-some states had also banned the box. So, inevitably, a diverse array of states. Yeah, so California and Massachusetts implemented ban-the-box policies, but you also have Nebraska that implemented ban-the-box policies. And so you really do have a mix of places that are more, you know, coastal and more liberal and places in the middle of the country that are not.

But also, there were a bunch of states that hadn't instituted the policy. So once again, Doliak had a nice, clean, natural experiment. She could measure the effect of the policy by comparing hiring practices in states that didn't ban the box with those that did. If ban the box works really well, then we should see an increase in employment for the types of people that are more likely to have a recent criminal conviction that might worry an employer. When Doliak says the types of people, that means she couldn't measure the effect on ex-prisoners per se, because remember,

administrative data generally doesn't include criminal records, but she could measure the effect on one particular cohort of the population. So if this policy is working, then we might expect employment to increase, particularly for young men of color who don't have a college degree. That's the group that is most likely to have a recent conviction. So is that what happened? Did Ban the Box legislation accomplish its intended goal and produce more jobs for young men of color without a college education? Before I tell you that, let's hear what Doliak thought when she first

heard about the Ban the Box policy. I remember thinking, well, that could totally backfire. Backfire? Because why? If you remove information about who has a criminal record, which presumably employers are worried about, then they might try to guess who has a criminal record. Because guessing is faster and cheaper than running a criminal background check on every applicant. And then they might simply discriminate against the entire group that contains people who are more likely to have criminal records. So Doliak's concern was that banning the box would not only not help former prisoners, but that it might hurt other applicants who happen to share demographic traits associated with former prisoners. A classic unintended consequence, if true, of course, that was just her concern, her

hunch. She and her co-author, Benjamin Hansen, spent a few years assembling and analyzing the data to see if it actually was true. And what did they find? So we found that on average across the U.S., in places that ban the box, employment fell by 5% for young Black men who didn't have a college degree and by 3% for young Hispanic men who didn't have a college degree, and by 3% for young Hispanic men who didn't have a college degree. And so the next question is kind of like, who are they hiring instead? And we found on average, the people who benefited were older men of color. So if you're

trying to avoid people who are actively involved in crime, older people are a pretty good bet. Criminal activity tends to decline with age. And what about young, undereducated, white potential employees? So for young white men who had the same level of education, we saw employment increase for that group. What seemed to be happening was that employers were substituting from young men of color to young white men. My sense is that you came into this like the person that's watching a horror movie when the character is like going into the abandoned barn and you're saying, no, can't you see what's going to happen? I mean, did it seem that obvious to you? It felt very obvious to me. If employers are statistically discriminating based on who is likely to have a record, which is our hypothesis going into the study, then that's the same group that we might expect to see a reduction in employment for.

Okay, so talk about the response to your paper as word about the paper began to get out in policy circles. I'm sure that everyone said, oh my goodness, we've made a terrible mistake. This policy not only is not helping the people we wanted it to help, but it's actually hurting people who have nothing to do with this at all. Let's roll back these policies immediately. Is that what happened? Yes, exactly.

Ban the box has been repealed everywhere. No. So, no. Unfortunately, we got a lot of pushback. By the time we had enough years of data to actually test the hypotheses and put out studies that we're really confident in, there was a really established ban-the-box lobby. And so at this point, there are organizations that are very wedded to not helping people with criminal records get employment, but passing ban-the-box laws. And you're the pesky economist that says, well, yeah, but you're actually not getting a very good outcome. Exactly. Right. The challenge is that most people who are on the ground working on the policy

aren't thinking necessarily about like, what's next? Like, how do people then respond to this new set of incentives? And that's what economists are trying to think about. The reason I wanted to bring you the story today about Jennifer Doliak's ban-the-box research is to show how hard it can be to create policies that hit their target without collateral damage. Doliak says it was obvious to her, even before she did the research, how the policy might backfire. But you can appreciate how all the advocates and activists thought it was

a good idea. The problem is now it's too late. The policy came faster than the research that showed how the policy could backfire. And if there's anything harder than making new policy, it's undoing existing policy, even when the existing policy appears to be hurting some of the most vulnerable people in the labor market. And that was before COVID-19. That was when the U.S. unemployment rate was under 4%. The official rate has since quadrupled, at least, and the true unemployment rate is even higher. So what kind of employment and reemployment policies are being created now? Will they also hurt some of the people they're intended to help? What kind of lessons should we take away from this ban-the-box story? We'll get into all that right after the break. If you haven't yet checked out

our new spinoff podcast, No Stupid Questions, please do. The psychologist Angela Duckworth and I take turns asking each other questions about uncertainty, happiness, anger, immortality, you know, regular water cooler stuff, even though we don't have water coolers anymore. You can find it wherever you get your podcasts. No stupid questions, it's called. We will be right back. The COVID-19 crisis has scrambled just about every norm and protocol we know. One of the biggest disruptions is in the labor market. The pandemic has destroyed jobs like nothing else in recent history. There have already been a number of policy responses, and there will be many more. But here's the thing about policy, especially when it's created in the midst of a crisis. A policy intended to produce X might produce some X, but it'll often produce a lot of Y, too, and maybe even some Z. Ideas that look great on paper or on a bumper sticker bump up against the real world and turn into something else, like the ban-the-box policy we heard about earlier. So what does re-employment

look like as the COVID-19 crisis starts to ease? Who will be first and who last? Which sectors will surge and which will perhaps disappear? Do we see even more automation or less? Let's start with some numbers, as gruesome as they might be. Before the pandemic, the U.S. economy was enjoying its longest boom in history. In February, the official unemployment rate was a measly 3.5%. This is worth remembering. Our current economic crisis had nothing to do with the underlying economy.

It's all about the pandemic. In any case, the latest unemployment rate for the month of April? 14.7%. The reality is that number is too low. Betsy Stevenson, an economist at the University of Michigan, has served as chief economist of the U.S. Department of Labor, as well as on the Council of Economic Advisors in the Obama administration. Many of the people who reported that they were employed, but they weren't able to go to work,

they were absent from work, not working from home, but absent from work. Many of those people should have been counted among the unemployed, at least among the temporary unemployed. Which means the official unemployment rate is skewed. So what other data sources are there? So the fastest data that comes out is applications for unemployment insurance. Those are called unemployment insurance initial claims data. The CARES Act made a bunch of people who are not normally eligible for UI eligible.

And this is called pandemic unemployment assistance. Since the middle of March, 40 million people have applied for unemployment benefits. Another way is to take a look at our employers and ask our employers how many people do they have on payroll compared to the number of people they had on their payrolls in February. Those losses are around 21 million, quite a bit less than the unemployment applications. There's a third way that we can try to take a look at the state of the labor market. We can survey households and ask people, you know, what were you doing in a particular week? Were you at work or not? All these numbers are in this sort of 20 to 30 to 40 million. That's a huge range, but it all paints the same picture of really a staggering scale of work stoppage. So where does Stevenson put the true unemployment rate for April? The true unemployment rate is probably around 19, 20 percent.

And that number may be low or may rise further. Let's put that in context. During the Great Recession of 2007 to 2009, the unemployment rate reached 10%. Nine million jobs were lost over the course of those two years. With COVID-19, 20 to 40 million jobs were lost in about two months. If we use Stevenson's estimate that the true unemployment rate is now around 20%, that's getting us into the territory of the Great Depression when unemployment hit 25%. Now again, the comparisons are not parallel in both the Great Depression and Great Recession. Job loss was a consequence of economic calamity. With COVID-19, it's job loss because of an economic shutdown produced by a public health calamity. Still, if you are a labor economist,

the current numbers are literally unbelievable. If a year ago, the same person had brought me this table of statistics and said, this is what I calculated, I would have said you did something wrong. That's Abigail Wozniak. She is a research economist at the Federal Reserve Bank of Minneapolis. We have, in the course of a month, taken a turn that really has magnified the losses of the Great Recession many times over. We've eradicated all of the job growth in the last 10 years. Wozniak has spent her career studying recessions. Here's one sobering finding. If you are a college graduate who enters the job market during a recession, you never really catch up. Your lifetime earnings are around 10 to 15 percent lower. With COVID-19, Wozniak has been

surveying Americans across the country about their physical and financial health. Some patterns have already jumped out. For instance, unemployment has risen dramatically for workers who don't have a college degree. And I looked back to the Great Recession numbers, And really, that pattern looks very similar. But she has noticed one difference. In most recessions, Black workers experience more unemployment than whites. In this case, they are about the same. And in this event, unemployment rates for Hispanic workers have shot past those for Black and white workers, and they are experiencing the highest levels of unemployment right now.

So that is really something that's quite different. That's likely because some of the industries most affected by the pandemic, including the service sector, have a lot of Hispanic workers. The same for construction, which has also slowed. So how does Wozniak think about recovery and re-employment? That, she says, depends on what kind of recovery we get. So I think folks might have heard that kind of popular phrase, the jobless recovery. And what that was connecting to was a kind of empirical fact that employment was somewhat

slower to recover in the last couple of recessions prior to the Great Recession. And in fact, there's good evidence that it never quite fully recovered, particularly for some skill groups of workers, that what we had in the more recent recessions was a significant reallocation of what jobs looked like so that when employment started to pick up coming out of the recession, the types of jobs that folks were doing looked a little bit different. That is opposed to an older style of recession where the previous jobs tended to return. If you look at recessions in the 60s and 70s, you would see folks reporting layoffs, this very kind of formal, we're laying you off, but we might call you back

type of separation. So what determines the nature of the job recoveries this time around? Betsy Stevenson again. What ultimately happens to job loss depends on the consumption choices that you and I and everyone else make during and in the wake of this pandemic. Stevenson sees three factors that are potentially changing our consumption choices. Number one, we are learning to use new technology and the pandemic has forced us to make some investments in change that we might not have made before, at least not right away, because change is hard. But with a pandemic, we've been forced to learn how to order things online, how to use Zoom, how to use Google Hangouts. So those investments will make us more likely to continue to use new

technology going forward. Okay, that makes sense. What else? We're also adjusting to a new level of risk that may lead to changes in how we live our lives for a while and the things we want to consume. So people who are avid travelers may not go back to traveling for quite some time. People who love to go out and hear music at concert halls or love to go out to dinner, they may not continue to do those things at the same level or same degree. Also sensible. And let's not forget, there is risk on the labor side of the consumption equation, too. In many states, including my state of Michigan, dentists were told they should not practice during this period of rising contagion.

And this would be a very hazardous job if you were a dentist or a dental hygienist, cleaning people's teeth, being in their mouth, maybe they're asymptomatic COVID person. So there's a risk to the workers. That's going to change whether people want to go back to those jobs. And finally, there's the fact that incomes have changed. We're going to see lower household income.

And when there's lower household income, people buy less. Some households may not even find that their income has declined, but there may be concerns that their income will. That causes a lot of people to slam the brakes on their consumption. You can see how this might snowball. The more that people are worried about earning money, the less they'll spend, which creates less demand for the jobs that people need to get back to work.

But, of course, supply and demand are elastic, especially during the pandemic. Consider the demand for deep freezers. What we see is that people tend to buy fewer durable goods, like a deep freezer, when incomes decline. But people who are adjusting to a new level of risk seem to think that they'd like to be able to buy more of their groceries at once, be able to put more things in the freezer, be better prepared for not being able to get to the shops. Indeed, the sale of deep freezers is way up, as are just about all home and kitchen goods, like pasta makers and soda machines and

water filtration devices. So if you make those products, you still need employees. But if you sell those products in a physical retail store, you don't. Going into a store physically now represents a level of risk that means that many people, even if they prefer to shop in person, will find that the benefits of shopping in person don't outweigh the new costs. Before the pandemic, there were nearly 5 million retail sales jobs in the U.S. They were already under extreme pressure from online shopping and automation. Now that trend has been supercharged.

Just take a look at some recent bankruptcy filings. JCPenney, J.Crew, Neiman Marcus. Macy's furloughed most of its 125,000 employees. How many of those jobs do we think will come back? A recent paper by the economists Jose Maria Barrero, Nicholas Bloom, and Stephen Davis estimated that 42% of recent layoffs will result in permanent job loss. That sounds, well, it sounds terrible. Even if you think of yourself as an optimist, it sounds like the glass is at least 42% empty. I can make that glass a little more half full. Abigail Wozniak, again from the Minneapolis Fed.

So the way they come up with that estimate is that, again, from older recessions, when we had significant amounts of layoffs, it looked like about three quarters of folks who were on layoff were eventually rehired. What we're missing with that is some of these reallocation effects that happen in recessions. That is the newer jobless recovery recessions, not the older ones. So reallocation just means the economy was doing a certain set of jobs or tasks or making a certain set of things. And then they kind of switched to making some different things, right? So you can think of the 20th century as just one big exercise in reallocation. We went

from lots of people producing agricultural goods to almost nobody producing those. And now we produce other things. But that happens on a bit more of a granular scale in recessions. It's reallocation. It's not just straight up destruction. And that process will take time. Wozniak is reluctant to predict which industries or firms might benefit in the long run and which might suffer. This seems wise. We are still in the middle of the crisis, after all. And she works for the Federal Reserve, where guessing about the future isn't strongly encouraged. But we can do some guessing, educated guessing at least. Any industries reliant on people congregating en masse,

those are obviously a pretty bad short-term bet. What about healthcare spending? Hospital systems have taken a huge hit during the pandemic in large part because they had to defer all their money-making activities. But afterwards, you could imagine there will be a massive appetite for public health spending and healthcare infrastructure generally. What about technology?

Already, the pandemic has made some of the rich a lot richer. Big firms like Facebook and Amazon and Google and Apple have not only seen their business models weather the pandemic, for the most part, they've also been able to hire a lot of software engineers and data scientists from startup tech firms that have been crushed by the pandemic. Here's another educated guess. The rise of automation will likely continue. Just to give a sense of where we were before COVID-19, a 2017 McKinsey report estimated that by 2030, roughly half of the work currently being done in our economy would be automatable. Automation is really one of the key technological aspects of the last several hundreds of years.

That's Darren Acemoglu, an economist at MIT. If you think about what the British Industrial Revolution was, it was a slew of automation technologies that took the tasks that were previously performed by skilled artisans in textile and in other factories and found ways of automating them, mechanizing them. Asimoglu studies the disruption that automation can cause in the labor markets. You know, it's most classically summarized by our discussions of the Luddites, you know, the revolters who broke textile machines because they were afraid of losing their jobs to the machines. And we always equated Luddites with irrationality or being anti-technology. But actually, when you

look at the details, the Luddites weren't exactly wrong. They understood that they were losing their jobs, and there's a reason for it. As Asamoglu sees it, automation typically creates one of two outcomes. If automation is bringing significant productivity gains and cost savings, then it will create disruption, but there will be other compensating changes that may undo some of that disruption. So people who lose their jobs in textile factories can get jobs in other factories or other sectors. In other words, the labor is reallocated. But on the other hand, if automation is not very productive, then you get a double whammy. You get the disruption, but you don't get any of the benefits. And especially marginal Marginal automation that takes place because of mistakes or because policy somehow encourages automation at the margin would be of this sort, which Pasquale Restrepo, my collaborator on much of his work, and I call so-so automation.

An example of what Asimoglu calls so-so automation? The automated checkout machines in a grocery store. At least until recently, they've been improving. Also, automated customer service, still pretty bad. Even automation in Tesla's car manufacturing plants, which the company realized was excessive and wasn't even saving much money, so they reversed some of it. But if you go back to the decades after World War II, the story was different. Back then, productivity and labor demand were increasing in parallel. There were technological improvements that did require less labor.

Think of a Ford assembly line. But the gains in productivity meant more jobs in other areas, like auto design and sales and marketing. For labor, this painted a pretty good picture. And then, if you look at the post-1980 period, a completely different picture. Productivity continues to increase, somewhat more anemically, somewhat less strongly than before. But labor demand, especially for workers without a college degree, is stagnant. And in fact, for high school graduates, it's declining. So what changed in the 1980s?

A few things. For starters? In the U.S., we provide huge subsidies to firms to adopt equipment instead of hiring labor. So what that does is that it encourages a lot of so-so automation because a lot of technologies that you would not have adopted and you would have just employed workers, you now find it profitable to do so

because it has tax advantages. The U.S. wasn't alone in writing legislation and tax code that treated capital more favorably than labor, but we were more zealous than most countries. Ronald Reagan got things started. This was the famous trickle-down economic model. And subsequent administrations kept it going, some to a greater degree than others. But in addition, we have also cut corporate taxes, both the actual rate, but also the

effective rate, because we've had an expansion in S-corporations that don't pay the corporate income tax. And then finally, we've cut the income tax rates of the capital owners, the top of the income tax. So all of these have contributed to reducing the effective tax on capital to essentially zero. It's pretty easy to make the argument, in retrospect, that the U.S. overprivileged capital at the expense of labor. Corporations and their high-end employees have continued to earn more and more money. At the lower end, meanwhile, there's been wage stagnation and income inequality has spiked. But you also need to acknowledge that the past few

decades have seen one of the greatest technological explosions in history, much of it emanating from the U.S., and that this might not have been possible without all that cheap capital. So if you look at some of the automation technologies of the last several decades, you know, they're very, very promising. Industrial robotics, I think that's one of the most important technological breakthroughs. And more recently, we have AI. By its nature, AI is very rich. You can use artificial intelligence, machine learning, deep learning as a way of creating new tasks, new functions for humans.

Think about teaching. If you provide more input to teachers in real time about what are the hidden strengths and weaknesses of each student or groups of students, that can really transform the teaching profession. But instead, what is it that we do with AI in the education domain? What we do is that we try to replace the teachers, automated testing, multiple choice questions that can be graded so that the teachers don't have to

interact with the students as much. I think the good type of automation, we're all the beneficiaries of it. You know, robotics, it creates disruptions, but, you know, we don't want to dial the time back and have workers assemble cars and be subject to work accidents and costs of production are much higher. So that encourages offshoring. So there are a lot that we can do with the right automation technologies. But the problem is that we do too much of that and we ignore the humans. The humans, in this case, including as many as 40 million Americans who've already lost their jobs because of COVID-19. Now, the current crisis actually heightens some of these trends. On the one hand, it is deepening our dependence on automation in a good way. I mean, in this time of lockdowns, if we did not have access to digital technologies, to AI technologies, to robotics, we would be in dire straits in terms of weathering

the storm. But on the other hand, this could be an impetus to going to a future of just automation and nothing else, because it's already the case that major companies in the U.S. are automating and are banking on automation. They're becoming stronger out of this. We're becoming more afraid of social interactions and all of the physical contact that are involved. And we're making a lot more investments in automation at the moment. Now, what I'm worried about is that this will go more in the wrong direction. Even if automation doesn't move in what Asimoglu calls the wrong direction, even if we get better than so-so automation, the fact is we are still staring

into what may turn out to be a labor abyss. Let's assume on the bright side that most of the jobs lost from the pandemic do return, even relatively soon. But what if only 10 or 20 percent of them don't? That right there is several million people. How are they going to get by? One idea that's been gaining momentum. So we need to adopt universal basic income today, right now. We'll be back next week to continue this conversation about the great labor reallocation of 2020.

Until then, take care of yourself and, if you can, someone else too. Freakonomics Radio is produced by Stitcher and Dubner Productions. This episode was produced by Zach Lipinski. Our staff also includes Allison Craiglow, Greg Rippin, Matt Hickey, Daphne Chen, Mary DeDuke, and Corinne Wallace. We had help this week from James Foster. Our intern is Emma Turrell.

Our theme song is Mr. Fortune by The Hitchhikers. All the other music was composed by Luis Guerra. You can get Freakonomics Radio on any podcast app if you'd like the entire back catalog, use the Stitcher app or go to Freakonomics.com, where we also publish transcripts and show notes. As always, thanks for listening.