Search the Site

Episode Transcript

I’m Stephen Dubner, and this is Freakonomics Radio. If you are new to the show, you might assume from its name that we deal with straight-ahead economics – employment, cost of living, things like that. But if you’ve been listening for a while, you know we don’t do much straight-ahead economics. We tend to favor stories like the economics of sleep; we do stories about the significance of the lowly pencil; we even made a show about the unlikely triumph of the belt.

This week’s show deals with a straight-ahead economic issue. It’s about an idea whose time finally may have come. An idea that may present a solution to two separate problems. The first problem is that a lot of full-time jobs in the modern economy simply don’t pay a living wage. The second problem is that even a lot of those jobs may not be around for long. Erik Brynjolfsson is an MIT professor who deals with the economics of technology.

ERIK BRYNJOLFSSON: We’re now beginning to have machines be able to augment and automate our brains and replace mental tasks. Machines can do computations and make decisions and we’re still in the early stages of this, but we believe that the implications will be at least as profound as what the Industrial Revolution did for our muscles.

In a book he co-authored with Andrew McAfee, Brynjolfsson calls this technological disruption the “second machine age.” Yes, it will likely produce gains in living standards but it also might lead to a massive evaporation of jobs. This can be aggressively argued in either direction.

BRYNJOLFSSON: There are folks who are confident that technology is going to lead to a bounty for all and make everyone better off. And there are other people who are very concerned that we’ll have a jobless future and a lot of people will be left in poverty.

Take just one example: driverless cars, a technology I’m sure you’ve heard something about. It is not hard to believe that in 10 or 30 or 50 years, many of our vehicles will drive themselves. So what happens to all the people who drive for a living – all those taxis and Uber cars and ambulances and heavy trucks. In the U.S. alone, that’s roughly 3.5 million people. Sure, some jobs will be created by driverless technology. But what are all those drivers supposed to do when their livelihood is obliterated?

BRYNJOLFSSON: There will be winners and losers. Already we’re seeing that some kinds of tasks and skills are much less valuable and people with those tasks and skills have seen their wages fall. And so it hasn’t been a rising tide that lifted all boats.

So, if there’s a massive evaporation of jobs, and if too many hard-working people continue to earn too little to live on, what happens next? Could one solution to both problems be an idea that’s been floating around in economics circles for decades? Here’s the idea: what if everyone automatically received, from the government, a guaranteed basic income?

MUSIC: Jimkata, “Innocence” (from In Motion)

There are of course many forms of government welfare around the world, but universal payouts are rarer. The Alaskan government, for instance, has since 1982 paid a dividend on oil revenues to all residents, about $1,100 a year on average. (Not exactly enough to live on.) But several governments are looking at more substantial payments.

In some cases, everyone would get a payout – rich or poor, employed or not. In others, only low-income people would benefit. Switzerland is planning a referendum – which probably won’t pass, to be honest – that would pay every adult around $2,600 a month, and every child about $650. Similar proposals are under consideration in parts of The Netherlands, Spain, and Canada. And Finland is considering an experiment that would give up to 10,000 people roughly $625 a month, tax-free, which would replace most existing welfare benefits.

How about in the U.S.? The idea has not gotten much traction in Washington, although Bernie Sanders has expressed his sympathy for it — perhaps not too surprising for a self-declared socialist. But that’s the thing about the guaranteed income idea. You can see why a socialist, or just about anyone on the left, might like it — theoretically, at least; it’s a permanent financial safety net. But conservatives should also like it, again, at least theoretically, because it’s a way to shrink the welfare state by folding up a bunch of byzantine entitlement programs into one tidy check.

Now, what would a guaranteed-income program actually look like? Where would the money come from? And if you give everyone in America a check for something like $20,000 every year, wouldn’t inflation render the poor people just as poor as before? These questions are all really, really hard to answer. But we can probably learn at least a little bit by going back in time.

*     *     *

In the late 18th century, founding father Thomas Paine argued for a substantial universal payout based on our collective share of the Earth’s natural resources. It didn’t happen. In the 1940’s, the economists George Stigler and Milton Friedman helped put another version into play. It was called a negative income tax; it was meant to support people in grave need.

EVELYN FORGET: So the idea was that if you had no income from any other source, you’d receive a certain level of income, a sort of socially agreed upon level.

Evelyn Forget is an economist at the University of Manitoba.

FORGET: And as your income from other sources, and in particular from the labor market increased, your benefit would decline, but it would decline less than proportionately.

And in this video, Milton Friedman, in his usual direct style, explains what he saw as the benefits of the idea:

MILTON FRIEDMAN: The proposal for a negative income tax is a proposal to help poor people by giving them money, which is what they need. Rather than as now by requiring them to come before a governmental official, detail all their assets and their liabilities and be told that you may spend x dollars on rent, y dollars on food, etc., and then be given a handout.

The hope was a negative income tax would streamline the burgeoning welfare system without diluting the incentive to work as much as the existing system did.

FRIEDMAN: Our present welfare programs, our present direct relief, and Aid to Dependent Children, programs in effect have a 100% tax rate. Because if a family on relief, if let’s say a woman — as currently is mostly the case under Aid to Dependent Children — if a woman who is on relief takes a job and earns $100 and she’s honest, she is required to have her relief payment reduced by $100. That’s one of the ways in which we’ve been producing poor people.

More than 1,000 economists from 125 American universities signed a statement supporting “a national system of income guarantees and supplements.” The idea had support from many quarters, like the staunch libertarian Friedrich Hayek:

FRIEDRICH HAYEK: I have no objection against a flat minimum security to everybody who cannot earn enough in the market.

To civil-right leader Martin Luther King, Jr.:

MARTIN LUTHER KING, JR.: There is a great deal that the society can and must do if the Negro is to gain the economic security that he needs. Now, one of the answers, it seems to me, is a guaranteed annual income, a guaranteed minimum income, for all people and for all families of our country.

In 1968, the Republican president-elect Richard Nixon, appointed as his urban-affairs counselor Daniel Patrick Moynihan, the liberal sociologist – and future Democratic Senator — who was intensely focused on race and poverty. They created what came to be called the Family Assistance Plan.

RICHARD NIXON: What I am proposing is that the Federal Government build a foundation under the income of every American family with dependent children that cannot care for itself — and wherever in America that family may live.

The Family Assistance Plan was sent to Congress. If it got through, it wouldn’t be a guaranteed universal income, but rather a “floor under incomes” for poor families. Nixon was very, very careful to draw this distinction:

NIXON: This national floor under incomes for working or dependent families is not a “guaranteed income.” Under the guaranteed income proposal, everyone would be assured a minimum income regardless of how much he was capable of earning, regardless of what his need was, regardless of whether or not he was willing to work … A guaranteed income would undermine the incentive to work; the family assistance plan that I propose increases the incentive to work.

But was this balance possible? To find out, the federal government had begun a series of experiments.

FORGET: The purpose of all of these experiments was to look at the labor market outcomes. The big concern everybody had at the time was, you know, if you give people money for nothing, why won’t they just quit their jobs?

Starting in 1968, the federal government ran a series of four negative-tax experiments in different parts of the country, with different demographics: low-income urban couples, rural families, black households — especially those headed by single women. Selection was randomized, which meant there was always a control group. This is thought to be the first time the U.S. government used this kind of experimentation to measure the effects of a policy idea.

Granted, it would be many years before they’d have the data, and Nixon’s Family Assistance Plan was still making its way through the political channels. Soon, a similar experiment was happening in Canada.

FORGET: I was sitting in a first-year economics class at Glendon College in Toronto, and my professor came in and started to talk about this wonderful experiment that he said was being conducted way out west somewhere and it was going to revolutionize the way we delivered social programs.

Evelyn Forget points out that the U.S. movement had a conservative thrust, with Milton Friedman advocating for smaller government. But Canada was different.

FORGET: In Canada, the support tended to come from the left. It was sort of a Social Democratic impulse, and most of the rhetoric was in terms of social justice.

It was called a “mincome” experiment, as in “minimum income.” It was conducted in just a couple of places, including a town of about 12,000 called Dauphin. It’s in Manitoba, a few hours northwest of Winnipeg.

FORGET: Dauphin was a small town like a lot of other small towns in Canada. It was very dependent on agriculture. There was nothing unique about it and that’s what made it such an interesting saturation site.

A “saturation site” meaning that every person who met the income eligibility was offered the deal. In Dauphin, that meant about 30 percent of the town.

FORGET: If your income falls below a certain level you’ll receive so much. And as your income increases, your benefit will be reduced. The nice thing about the guaranteed annual income — because people get to keep a portion of their benefits as they go out to work — is that this becomes a supplement for the working poor as well, and there’s an encouragement to work.

So in the mid-to-late 1970s, if you were unlucky enough to be poor, but lucky enough to live in Dauphin, Manitoba, you would be eligible for a check from the government. Forget says that a family of four would get the equivalent of roughly $15,000 current U.S. dollars a year. This was in lieu of any existing welfare benefits.

FORGET: And it wasn’t enough to give people a comfortable life. When we interviewed a number of the participants one of the participants had this lovely line, “that it was enough to add some cream to the coffee.”

That said, the Canadian government spent millions of dollars on the project, but for a good cause: to find out if it could support its low-income citizens without killing the incentive to work. So did the mincome experiment work well enough to be turned into policy? What did the data show?

FORGET: No one ever wrote a final report. There was no more money forthcoming to do an evaluation of the project and it just disappeared.

Uh-oh. Between an economic downturn and political upheaval, the Canadian mincome project lost its mojo.

FORGET: The researchers were told to archive the data for future research. And effectively what they did was to empty the file cabinets and put all the paper into cardboard boxes.

Back in the States, things weren’t going much better. The Nixon-Moynihan Family Assistance Plan had died in Congress. The right thought it was too expensive; the left thought it wasn’t generous enough. Ultimately, however, there were at least two direct legacies of the plan, which are still around today: Supplemental Security Income, which increased federal support for the elderly, blind, and disabled, and the Earned Income Tax Credit, which is a small negative-income tax, although not a guaranteed income.

By now, the political momentum for some kind of universal guaranteed income had greatly diminished, no matter what the data might have to say about whether such a thing could actually work. But what did the data say? The American experiments were still ongoing, and the final analysis was still way down the road. In Canada, remember, the data had been packed up in cardboard boxes.

FORGET: I finally did track it down. A few years ago, I was doing a lot of work on the relationship between poverty and health outcomes and I remembered this old experiment and I wondered if I could find any link to health outcomes and to quality of life for the participants.

It took a while but she found the boxes of data.

FORGET: I found 1,800 of them in a warehouse in Winnipeg under the authority of the National Archives. But, I mean, I was literally looking at 1,800 boxes full of really badly organized data. No one had ever constructed a database, and no analysis had been done.

She thought about going through the boxes and building a database …

FORGET: And I sort of realized that my life was finite and I wasn’t going to be able to do anything with that particular data.

But thinking about the mincome experiment in Dauphin did give Forget an idea.

FORGET: One of the defining features of this period in Canada was that this was the period when universal health insurance was introduced. And one of the unanticipated consequences of universal health insurance is that you collect quite a lot of data on individuals.

And those data had been sorted and digitized — data on education, employment, and of course health. So what if Forget could look in these data for people who lived in Dauphin, and who presumably got that extra government money, and compare them to other low-income people in other towns who didn’t get that mincome boost?

FORGET: And I was able to find everyone that was living in Dauphin during that particular period. So the first thing I did was to look at hospitalization rates. And I found that hospitalization rates had fallen by about 8.5 percent for the test group in Dauphin relative to the matched control group.

OK, so it looked like the mincome recipients were healthier. What else did Forget find in the data?

FORGET: One of the findings was that high-school completion rates had increased, and we discovered that boys, in particular, in low-income families had been under a fair amount of pressure to become self-supporting as quickly as they could. When mincome came along, some of the families decided that they could allow their sons to stay in high school just a little bit longer. So instead of quitting school at age 16 and getting their first full-time job, these boys stayed in school until they were 18, until they graduated from high school and took their first full-time job a little bit later.

So in the raw data it might look like employment fell, since a bunch of 16- and 17-year-old boys were not joining the workforce. But, in fact, this was good news, since they were staying in school, presumably setting themselves up for a better work life. But what about employment among grown-ups? That’s the big concern of any guaranteed-income plan. Did they stop working once they started getting a government check?

FORGET: Grown-up people with full time jobs don’t actually reduce the number of hours they work by very much.

One exception: women, especially those who wanted more time off after having a baby.

FORGET: At the time I think the legal entitlement was four weeks. And it’s sort of interesting because as a society we’ve already decided that longer parental leaves are something that we want to support. So they effectively anticipated that policy change.

So, the effects of the Canadian mincome experiment seemed to be pretty positive. Now, keep in mind the dollars to do this have to come from somewhere — new taxes or cuts in entitlements or other services. That, of course, is going to be a major part of any discussion about any guaranteed-income plan. But the data seem to show that the biggest concern — that everyone would just quit their jobs — did not happen, at least not in the Canadian experiment. Now what about the U.S. experiments?

FORGET: In general, the finding was that primary earners really don’t reduce the number of hours they work very much when you offer a guaranteed annual income. It’s secondary earners and tertiary earners — that is, second adult incomes and the incomes of adolescents that fall a little bit.

Very similar to the Canadian results. So there too, the news was pretty good, although there was something troubling in the U.S. data.

FORGET: When the findings started to come in, it looked like the families that were receiving a guaranteed annual income might, in fact, have a higher divorce rate. And this became very political very quickly in the U.S. There were all kinds of concerns that this was an attack on the American family.

Daniel Patrick Moynihan, a champion of the guaranteed-income idea, was also a champion of the nuclear family. The divorce findings led him to withdraw his support. But Evelyn Forget says that when the data were examined more carefully over time, the divorce controversy seems to have been overblown.

FORGET: There was a subsequent work about a decade later which suggested it might have been a statistical anomaly.

So the evidence from these experiments, as narrow as it may be, seems to suggest that a guaranteed income can produce positive results. The funding and the politics are, of course, separate, and large, questions. But one reason the idea has gained momentum lately, is that the world has changed a lot since those experiments; our economies have changed a lot. There’s already been wage stagnation; technological disruption keeps on coming. All of which makes at least one economist, Evelyn Forget, think that the timing is right.

FORGET: I do see a fundamental and ongoing change in the way labor market operates, and I do think we need to make some fundamental changes in the way we deliver social programs. I think a guaranteed annual income could do a very nice job of addressing some of these issues.

But economists are not universal in their support. Take, for example, Robert Gordon, author of The Rise and Fall of American Growth:

ROBERT GORDON: As far as a guaranteed minimum income, I’m not in favor of that. I’m in favor of a modest increase in the minimum wage, and a substantial increase in the earned-income tax credit, which encourages low-income people to work.

DUBNER: Why would you not be in favor of the minimum guaranteed income?

GORDON: There’s too much of an incentive, if you guarantee income, to replace work. And if you provide a guaranteed minimum income to everybody, then those with low skills will drop out of the labor force and will no longer work. So I think it’s a matter of incentives. A guaranteed minimum income would put a very high implicit marginal tax rate on going to work, for those with relatively low skills.

Is Robert Gordon right? Are boosting the minimum wage and the earned-income tax credit better solutions than a universal basic income? Or what about wage insurance, where the government supplements the income of people whose jobs just don’t pay enough to live on? Again, these questions are nearly impossible to answer, right now.

*     *     *

But let’s assume the guaranteed-basic-income idea is worth pursuing. How would it work? Would it require a huge tax hike or would the money come from efficiency or productivity gains, or from eliminating other government benefits? There are the metaphysical questions as well:

SAM ALTMAN: Does basic income make people happy and fulfilled? Does it make them less stressed? Does it make them do better long-term planning?

Sam Altman is the president of Y Combinator, in Mountain View, California.

ALTMANT: Y Combinator is a company that funds start ups.

Meaning, a venture-capital firm. For instance …

ALTMAN: We’ve invested in Airbnb, Dropbox, Stripe, Zenefits, Instacart, reddit. We’ve invested in about 1,100 companies in total. More recently we also have started funding basic research.

Which is not what v.c. firms typically do.

ALTMAN: No, but I think we’re a fairly unusual venture capital firm.

To that end, Y Combinator is launching an in-house research project on a guaranteed-basic income.

ALTMAN: The tentative plan is to hire a social-science researcher and to have him or her pick a group of people and give some of them some amount of money, maybe different amounts of money for five years, and observe what happens relative to the control group.

“What happens” would include all kinds of metrics:

ALTMAN: Do their personal balance sheets go up? Do they learn new skills? Do they create new wealth for society? All of these sorts of things.

The research would also try to figure ways to drive down the baseline cost of living.

ALTMAN: Today, maybe it costs — I don’t know — thirty or forty thousand dollars a year to have a really great life in a big city in the U.S., minimum. What can we do to make that number a lot lower? How can we have technology continue to drive that number down?

The biggest problem, Altman says, is housing.

ALTMAN: And that is the hardest thing to fix with technology. It’s the most dependent on policy. Other than housing, I think we can make great progress just with technology. If you think about a world where we get nuclear fusion to work, we make continued progress on agricultural technology and water technology and transportation, or we use technology to really make education a lot more efficient, you can easily see the costs of the baseline come down a huge amount. And I think if you’re not a kind of a crazy techno-optimist, which I’d say I am, you underweight that.

Being a techno-optimist, in Silicon Valley, is why Altman started thinking about a basic income.

ALTMAN: I think there’s a sense that technological job destruction is accelerating more than people outside of Silicon Valley realize. And I think this is mostly narrow applications of A.I., but it’s going happen for lots of other reasons. So every time we’ve had a major societal revolution of some sort — we had the Agricultural Revolution, we had the Industrial Revolution — where anytime you have a significant fraction of human jobs get eliminated in a relatively short period of time, eventually humans do figure out new things to do. But there’s quite a lot of disruption while they’re happening. And so I think people in Silicon Valley are saying, “Man, there are going to be incredible changes in the next few decades. What can we do to make that as smooth as possible?” I do think that as we, as a society, get richer and richer, there should be some sort of floor for everybody. And in a way, basic income is sort of the libertarian approach to doing this, right? It’s actually healthy for a society if some people get incredibly rich. But there should be a floor below which we as a society don’t allow people to fall.

You could argue that Y Combinator’s funding of basic-income research is a guilt play, since it is the kind of company that helps create so much of the technological disruption that may destroy jobs. Altman would disagree with that argument.

ALTMAN: So, we view our job as to increase the amount of innovation in the world. And mostly that means funding startups. But that is the mission, not funding startups in itself.

In Altman’s view, the idea of a basic income isn’t counterintuitive at all, even in a capitalist framework. He sees it as perfectly intuitive.

ALTMAN: I think the default belief of a child is that everyone should get food to eat and a place to sleep, and it strikes most children as incredibly unfair that that’s not the case. In fact, I think that’s so universal that it’s somehow encoded into our genes, this idea of fairness and equality. And I think then people get older and move away from it.

Every conversation you have, with anyone, about the idea of a basic income, inevitably comes around to one point: the value of work. Not just the financial payoff, but what you might call the moral payoff — that if anyone who wanted to could get a basic income without working, that it might destroy our moral fiber, our social fiber.

BRYNJOLFSSON: People get a lot of meaning from participating and contributing to society. And the way we’re set up right now, if we just got rid of the jobs and replaced it with a check, I think a lot of people would be unhappy.

Sam Altman isn’t so sure.

ALTMAN: Maybe 90 percent of people will go smoke pot and play video games. But if 10 percent of the people go create new products and services and new wealth, that’s still a huge net win. And the kind of American puritanical ideal that hard work for its own sake is valuable, period, and you can’t question that, I think that’s just wrong.

FORGET: Well, I think that there are lots of kinds of work. I think that people do need to create meaning in their lives, and for a lot of people that does come through work. I’m not sure that they necessarily need eight hours of work a day and 40 hours of work a week to find that meaning.

Forget, like Sam Altman, sees a potentially huge upside in freeing people of the need to work.

FORGET: If you look at the 18th and at the 19th century, some of the great scientific breakthroughs and some of the great cultural breakthroughs were made by people who did not work. These were gentlemen of leisure, right? These were people who had enough family money to support themselves. They certainly didn’t have to dirty their hands doing the kinds of work we take for granted. I don’t think these individuals felt useless; I don’t think their contribution was negligible. I think it was very important to the development of the world.

That may all be true. But if, for whatever reasons, we all start working a lot less — whether or not there’s a guaranteed income involved – humankind is in for a big change. For most of civilization, work has been a huge part of what it means to be human. So what happens if the techno-optimist dream comes true? What happens if computers and robots can take care of most, or all, of our work? Will our robot overlords even want us around anymore? Or will we just be in their way?

Sure, it’s a troubling thought. But consider one example, which may seem way off-base at first. But, consider another population that used to have to work for a living: the relationship between dogs and humans goes way back.

GREGER LARSON: We know that dogs were not just the first domesticated animal — the first domesticated anything.

Greger Larson is at the University of Oxford in the Department of Archaeology.

LARSON: And for the purposes of this I am studying the origins and evolution of dogs.

The domestication process was long and complicated.

LARSON: So what we think what may have happened is that a pack of wolves rather than following say, caribou or going after megafauna or something else, started following us instead, because the wolves could make a living simply by living off of the back end of us, and in doing so they would put a selection pressure on themselves to become more acclimated to us.

More acclimated and, over time, more useful. Let’s say you’re hunting and you just nick your prey.

LARSON: If you have a dog with you and whatever you just nicked is leaving a blood trail, that dog will be able to sniff out that injured animal and then locate it, and so all you have to do is follow it.

Over time, dogs did all kinds of work for us. They herded cattle; they guarded our property; they even helped cook our food. Seriously. A breed called the turnspit dog had a long body, short legs, and it would run in a wheel that turned a spit over a fire, to make sure the meat cooked evenly.

LARSON: These dogs were bred, or at least the best ones were just able to sit in these wheels and that’s all they did all day was just run and run and run and run.

Eventually, technology put the turnspit dog out of a job. Just as technology may eventually put a lot of people out of jobs. A lot of dog work was replaced with machinery or other technology. Like people, some of them found new and different work, as guide dogs, as therapy dogs; dogs were trained to sniff out drugs or explosives, even illness in a human.

But, let’s face it. Most dogs aren’t that clever. So when their jobs began to disappear, they didn’t find new work. So, did dogs just vanish? No they did not! In fact they flourished. When we no longer needed them as workmates, we turned them into pets. In the U.S. alone, about 54 million households own a dog — easily the most common pet. (Cats are second, in about 43 million households; freshwater fish a distant third, in 12 million.) Our spending on pets has tripled in just the past two decades, to more than $60 billion a year.

Dogs went from being on the production side of the human equation to the consumption side, big time! So who’s to say that we, their masters, won’t find a way to make peace with our robotic masters even when we no longer have any work to do? Personally, I think I would make a great pet for anyone willing to adopt me. Anyone interested? Serious inquiries only, please, to

*      *      *

Freakonomics Radio is produced by WNYC Studios and Dubner Productions. Today’s episode was produced by Greg RosalskyThe rest of our staff includes Arwa Gunja, Jay Cowit, Merritt Jacob, Christopher Werth, Kasia Mychajlowycz, Alison Hockenberry and Caroline English. If you want more Freakonomics Radio, you can also find us on Twitter and Facebook and don’t forget to subscribe to this podcast on iTunes or wherever else you get your free, weekly podcasts.

Read full Transcript