A 75-minute roundtable discussion with Niall Ferguson (Senior Fellow, Hoover Institute), Michael Faye (Co-Founder, GiveDirectly), Matt Krisiloff (Researcher, YCombinator), Roy Bahat (VC, Bloomberg Data)
Hi, this is Chris McCoy, the Executive Director at Data4America. Welcome to Unplugged, where we explore the intersection of data, government, technology, and the future. Today's session is recorded live from an event we cohosted with Lincoln Network titled "Man Vs. Machine: The Potential of Artificial Intelligence and Universal Basic Income."
It was the third in our series. This was recorded on December 12, 2016, in San Francisco. It was moderated by Data4America board member Kim-Mai Cutler. The panel included Niall Ferguson, renowned author and Senior Fellow at the Hoover Institute at Stanford, Roy Bahat, venture capitalist at Bloomberg Beta, Michael Faye, cofounder of GiveDirectly, who has run applied basic income experiments in Africa, Matt Krisiloff, who helps lead the basic income research as part of the Y Combinator fellowship. This, of the three we recorded, was the most interesting to me personally. It was a very spirited discussion with plenty of differing opinions, I think you'll enjoy it. If you do, please share it on Facebook, Twitter, and email. As we move into 2017, let me know who you'd like to see on the show. Give me an email at chris at data4america dot org. You can help our work with a donation as well - we have a big year in store. With that, let's go in.
This conversation, Man Vs. Machine: The Potential of Artificial Intelligence and Universal Basic Income, is part of a bicoastal conversation that we are having. We had an event last week in New York in partnership with Manhattan Institute and the Charles Koch Institute focused on innovation, automation, and the future of work, and that conversation was with Charles Murray of AEI, Robin Hanson of George Mason University, and Paul Martino, who's a serial entrepreneur and investor. It will be interesting to see how this conversation unfolds, and we will have the video to compare. You can watch that video of the conversation that happened last week. We are excited to have you all here for this final event of the year, and so, again, as demonstrated on the screen, thank you for your support in the past. We're excited about what 2017 has in store.
With that said, please support Lincoln. We are a nonprofit 501c3 again focused on getting the tech community together to talk about how we can use technology to advocate for more economic and individual liberty. You can go to joinlincoln dot org, make a contribution, and in the back of the room, my cofounder, Aaron Ginn, has some t-shirts that are on sale. Please check out some Lincoln swag that we have, and you can follow us on Twitter and all the other social channels with the hashtag #joinlincoln.
My colleagues from Lincoln are going to get the stage ready for the panel. One of our panelists, Michael Faye, is the chairman and founder of GiveDirectly. GiveDirectly has been doing cash transfers in different countries in Africa since 2011. They've raised over $138 million, they have a global team, and so they arguably have some of the best recent data as it relates to a UB-like cash transfer system. Before the panel takes the stage and Joe gives the intro to the panel, please welcome some remarks by the chairman and founder of GiveDirectly, Michael Faye.
Excellent. I come at UBI from a slightly different place than a number of other folks. I'm a development economist by training. Development economics is the study of how countries grow, why people are poor, and what we can do about those two things. I picked an interesting time to be in graduate school. I was in graduate school in the early 2000s, and there was a bit of a revolution happening in development. What we as development economists started to do is something that everyone in Silicon Valley takes for granted, which is A/B testing. We tested two things against each other and asked the most simple question: what works, and where is there evidence?
What we found was that a lot of what we were doing and had been doing in development was a whole lot less effective than we had hoped and imagined. Things like training programs, microfinance, didn't live up to a lot of what we thought. On the other hand, the simplest thing - giving poor people capital - and something that everybody thought was crazy and silly - worked remarkably well. It turned out poor people didn't all go out and get drunk, they didn't stop working with the money. In fact, people invested it. Incomes went up, education attendance went up, health outcomes went up. That was the state of development in the early 2000s. I had a front row seat; I got to sit in a lot of really good seminars and see some of the papers coming out of Mexico and Brazil on cash transfers, so I did what any academic grad student that didn't want to go into academics did, which was to start to think about what we did with this information.
So we went out and looked for places to give capital grants or cash transfers to the extreme poor. We talked to all the usual suspects - big NGOs, donor agencies - and said, "What do we do? I want to give money to a poor person, can I just give it to you and you transfer it?"
And they said no. There are a lot of reasons we can go into why, but what came out of that was we decided to found GiveDirectly. We started working on it in 2008, we launched in 2011. What GiveDirectly was was an opportunity for the public to give directly and provide capital grants to the extreme poor. For every dollar you give, depending on the country, we give out 90 cents to the recipient. As Garrett said, we've raised about $138 million since launching in 2011, and that's a testament to the public funding it. We're really just the service provider of getting cash to the poor. While this has been happening with GiveDirectly, a lot of governments have been shifting their social programs to cash, which is what we're talking about today. There are now about 130 different cash programs in low and middle income countries, I think it's something like 40 out of 48 African countries have a social safety net program that involves a core cash component. Kenya, which is where we started to work, has multiple cash transfer programs, so this is nothing new.
While there's a big debate happening in the media and cash is still crazy and everybody is going out and getting drunk, the emerging market governments have shifted. Developed country governments have also been doing cash. Social security is a form of cash transfer. Now, while we know a lot about cash - it's been called the most thoroughly researched intervention by Diffid, which is a UK development agency - we don't know a lot about how to do cash. Do we give a little bit of money to a lot of people? Do we give a lot of money to a small number of people? Do we give it monthly? Annually? All these questions around how we actually do a cash program are largely unanswered. So we do what we did for a long time and use our gut and tell anecdotes and stories.
One specific form of cash, universal basic income, has become pretty popular over the last 24 months, which has been a revival of sorts, because it was popular about 40 years ago. Universal basic income is universal. Everybody within some group gets it. It's basic, so it's enough to live on, it's not a huge capital investment, and it's not a tiny bit, and it's an income, so it's over a long period of time. This is not a cash infusion, this is over a long time. As it turns out, we've never actually done that globally, so we decided last year to launch a universal basic income. It is the largest crowdfunded social experiment ever, it'll be about 30,000 households in Kenya, which will be about $30 million dollars. We've raised about $24 million of that. We'll test it, and we'll ask the same questions we've asked before: what do they spend it on? Where do they spend their time? But we're also going to ask how it compares to giving people money for two years, because we know a lot about how people spend it when we give it over a short period. So let's make that direct comparison. We're also going to compare it to giving people a lot of money one time, up front. We're going to answer those questions.
There is a debate: is universal basic income affordable? I think it is a good debate to have, but you need to get much more specific than debating it globally. It depends on what country you're talking about. Scandinavian have a much bigger tax base relative than the US does. In Kenya, we're providing a basic income for 70 cents a day. That is the food security line. That is enough money to get people above a caloric minimum to survive. You couldn't do that in the US for 70 cents a day. The money goes quite far in Kenya. The last thought, which is simply meant to be provocative: it was an exercise that Brookings did which asked, "How much money would it take to get everybody in the world above the poverty line?" Let's say I live on 60 cents a day, the poverty line is a dollar. It would take 40 cents to get over it. So, how much would it take globally? The answer is in the $70-80 billion ballpark. That's not a lot of money. It turns out we spend about $130 billion a year on foreign direct assistance. $80 billion is smaller than that, and we can probably go bigger. So, at least as a first step, we have to ask the question of if we transferred capital directly, if we found more efficient ways to do it, which we now have with things like mobile money and mobile payments, could we actually end extreme poverty today? I'll leave that as the question to discuss basic income.
Thank you Michael. [applause] Now to introduce the additional panelists, Joe Malchow is the CEO of founder of Publir. He is an advisor to Lincoln, and he is responsible for confirming Niall Ferguson, so I'm eternally grateful for his support. Please welcome Joe Malchow.
Thanks for being here. Garrett and I sort of conceived this event a year ago, and about a year ago, the labor force participation rate was hitting its 30-year low. It's since rebounded just a little bit since then, but not too much. It's my pleasure to introduce a couple of panelists. Roy Bahat is with us, he runs Bloomberg Beta, the corporate venture fund associated with Bloomberg LP. Bloomberg Beta investments include CodeAcademy and FlexPort among many others. Roy was previously president of IGN and has been associated with the renegade gaming startup OUYA with McKinsey and the Office of the Mayor of New York, Michael Bloomberg. Roy has graciously agreed to join our event, so thank you very much, Roy.
Niall Ferguson - historian, author, and now in a coup for Stanford University Senior Fellow at the Hoover Institution. The subjects of Niall's study are world-beating people and institutions: Rothschild, Kissinger. In Niall's work, he is also beating. The history departments of the country do not lack for so-called experts in every little nook and cranny of history, but Niall has always been fair without favor to his subject or his field in why things happen or where things are going. We're very grateful to Niall for being here.
Kim-Mai of course, formerly of TechCrunch, now of Initialize Capital, moderating.
I'm not sure if everyone has bought into this theory that there's a real problem ahead, nor particularly am I, but I do offer this as the short, firm introduction. A quick survey of Bloomberg before the market day ended, between 2006 and today, the market cap of Best Buy has decreased 49 percent. JC Penney, 83 percent. Macy's, 46 percent, Sears, 95 percent, Nordstrom, 21 percent. The market cap of Amazon has increased 1,910 percent, and that's a little sampling of what's really happening in the economy. With that, I think we should start our panel, and Kim-Mai, take it away.
So, I'd actually like to start with a little bit of a historical question, which is that this concept of universal basic income and direct transfers is not new. At the end of the late 1960s, you could see ideas coming from both the left and the right around this concept, where, for example, Martin Luther King, Jr., in the last year of his life, focused on a poor people's campaign that would address economic issues for low income communities regardless of their racial background. On the right, the Nixon administration explored a family assistance program under the guidance of Daniel Moynahan. I was wondering if Michael, or Niall, if you wanted to talk about the genesis of this idea, where did it come from? Why did it fall out of favor? Why is it coming back now?
Milton Friedman, who in the later part of his life was a fellow at the Hoover Institution, was a proponent of a basic income model long before Martin Luther King, Jr. came on the scene. What's fascinating about the idea, and what I think is worth spending a little time on, is that it has proponents on both the left and the right. I think the most eloquent advocate in recent years has been my old friend Charles Murray at the American Enterprise Institute, and I think anybody who wants a quick guide to the idea could do a lot worse than read Charles' book or even the recent Wall Street Journal article that he wrote. So what's fascinating about the idea of universal basic income is that it's not just some mad leftist utopian vision. People with impeccable conservative credentials, from Milton Friedman to Charles Murray, have argued for it. That must make you ask the question: if the left and right agree on it, why hasn't it happened?
Great question. I think one thing that happened in the seventies and eighties was known as the negative income tax experiments. These were a series of experiments, rigorous evaluations, of what happens when you have a negative income tax. It's exactly what a lot of us are arguing for now. These experiments happened in Manitoba, in Bend, and in a lot of other places, and then they stopped and universal basic income lost a lot of steam.
Why did they stop? The headlines from these projects was that people stopped working. What we know after the fact, and after that initial wave of negative publicity, was that from self-reported data, people did work less, but they did it to do three things: they went back to school, they started looking for another job, or they stayed home to take care of their kids. We could all argue that those things are pretty good things, but the more interesting thing about these experiments is that when you actually look back and use real administrative data, a lot of the effect on people not working disappears. People didn't work less. This is actually consistent with a lot of other evidence about what happens when you give people cash. One of my favorite papers on the topic is called Debunking the Myth of the Lazy Welfare Recipient, and it is a myth, at least from the emerging market evidence - we have a lot less evidence from developed economies. The thing I think we should take away from all of this is the importance of well done, well communicated studies, so the anecdote of the person who lives in Berkeley and decides to smoke up and play video games with cash doesn't become the story that kills what could be a much more serious policy discussion. At GiveDirectly, we've brought in leading development economists Abhijit Banerjee and Tavneet Suri from MIT, and on the US side, Alan Krueger, who is the former lead of the President's Council of Economic Advisors. I think it's important to have the leading thinkers on the topic, and to be really careful about how you design it, which is why we're having a massive study.
It's worth adding that one reasons ideas never get turned into policies is that just as they're about to be, they turn out to be redundant. Remember, when Milton Friedman and others began talking about this in the 1940s, they were fully expecting that after World War II there would be tremendous unemployment. The phrase "secular stagnation" dates back to that period, too, and throughout history, people have been haunted by the specter of massive structural unemployment. This goes right back to the origins of the industrial revolution, and when they see this catastrophe looming, of structural unemployment, they start to cast around for solutions, and universal basic income comes up, and then what happens? It doesn't happen. There isn't massive structural unemployment.
On the contrary, as happened in the 1950s and 1960s, the US economy boomed, there was no problem of unemployment, and so the whole need for universal basic income seemed to fall away. When did it come back? The 1970s, when unemployment returned. One has to recognize that there's a cycle with this idea, and it correlates pretty closely to the cycle of unemployment. I suspect the same will happen once again. We're haunted in this moment, particularly in this neck of the woods, by a vision, which is deeply unhistorical, that technological innovation is suddenly going to make the human race redundant, or at least the unskilled members of it unemployed. People have been worried about this since the industrial revolution began, and it never, ever happens. Maybe this time is different, but those are the most dangerous words in economic history.
I think the second most dangerous thing is listening to the predictions of economists. This is a profession that has an acknowledged deep failure to predict the future, but I think that is the question that is at the heart of this whole ...
I'm a historian, not an economist.
Of course. Historians' predictions about the future I'm actually much more interested in than economists', but "Is this time different?" is the crux of the technologist-economist debate about this. If I can interject for one second, can I say that you don't even have to care about artificial intelligence and automation in order to believe that a universal basic income makes sense?
One lens is to look at it through the lens of unemployment, but for those who saw the Raj Chetty data - he published in the New York Times last week - we are now living in the first generation in the United States where statistically speaking, a child is less likely to make more money than his or her parents. That has not been true historically - quite the reverse, if you went back to the beginning of where he got his data, which was just pre-World War II, your odds were in the high 80s, low 90s, of making more than your parents. If we want to have a free society where people are able to participate in fully, vote, act without fear of what they would do - this is the third of Roosevelt's Four Freedoms - there is a case that for the first time in the United States, we have not only an un-free society for many people in general, and if not unemployment, at least sub-living wage employment for many. The fact that Wal-Mart's huge victory was to raise wages to $10 an hour illustrates that. It means that there's a powerful non-technological case right now for UBI. That might fade as employment comes back. I'll say one other thing, which is that it seems to me, from reading the history, that there were always hypotheses, going back to Hamilton in the United States for what the growth industries were, and the government acted to support those growth industries. It was early mercantile activity, then it was infrastructure, and right now I think one of the things we are missing in the United States is a good hypothesis for what is the growth industry that has good, well-paying, career path jobs. I literally know of no good hypothesis for what that might be. So that does feel different to me.
Maybe technological unemployment didn't happen at this massive dystopian scale that some people are concerned about, but maybe it did start happening in the 70s and 80s to, say, urban communities of color and low income communities. Here in San Francisco, if you look at the history of people experiencing homelessness, in the 1940s, 50s, and 60s, it's largely not a documented problem until around the recession of 1982, 1983, and then it appears en masse, at scale, and our politicians like Dianne Feinstein and Ronald Reagan start addressing it. If you look at what happened to urban black communities with deindustrialization and the response of the Californian, and then American government, in terms of not addressing investing in education and skills change, but rather looking at expanding the prison system instead, is the reason it's coming back because it's happening to a larger pool of people that have the political mass and majority to vote at scale to say that this economic stress is happening to them?
I actually think the data suggests that the very people who would be recipients of universal basic income - first of all, we have very poor data on this, there's been almost no political research done, a group of us funded a local study about attitudes toward universal basic income in this particular climate of San Francisco, and our venture fund invests in the future of work, so we do a lot of research on this. We did some focus groups with truck drivers in Ohio just to get a sense of what they were thinking. If you talk to the very people who would be recipients of universal basic income today, they will mostly tell you that they don't want one. They will say, "I don't want to owe the government something. I don't want to feel like I'm on a handout," because, of course, nobody wants to feel that way. So I think we have this question as a society of how we support people living free, vibrant, sustainable lives without diminishing their self esteem.
I think Michael has some learnings on this from the experiment.
I can share some personal stories about how hard it is to give people cash with no strings attached.
I did. Every development program has this challenge: nobody believes that you could possibly give them cash with no strings attached. So, political rumors start: you're going to take their baby, you're going to owe the government something. Something has to happen. They're going to force you to work a terrible job. And that's an interesting one, because there's this question of what are we trying to achieve? Is it giving people the economic freedom to make their own choice, or is it giving people jobs? That debate happens in the US, it happens abroad, and a couple countries have actual experience doing this. India has something called the Rural Employment Guarantee Scheme. They don't give people cash, per se, they guarantee them enough work to make a certain amount of money. If you read the press, you can get stories from both sides, and you have these awful pictures of people digging ditches or laying bricks or other menial labor that may not be the most productive use of our time. We have this deep bias that people must work for the cash, and that's been the consequence and history of cash.
We can all talk about the history of cash in one sitting, right?
So the history of cash - not money, totally different - the history of cash transfers is that a lot of the programs that we've seen around the world are what's called conditional. So, I will give you money if you send your kid to school, if you go to the health clinic, and so on and so forth. Okay, that's interesting, did it work? The reality is that the reason we did a lot of these programs as conditional is because it was politically feasible. I can never, as the government, say I'm going to give free money, but if I say I'm going to make sure they sent their kids to school, it becomes a whole lot more palatable. Now, the real genius is if I pick something that most people are already doing. I am basically just giving people cash with no strings attached, even though it's branded as a conditional cash transfer program, which is something to think about.
There's one important distinction that we need to bear in mind when we're talking about these schemes. In the original video presentation by Matt, he made clear in the experiment by Y Combinator that basic income is an add-on, or rather, it does not in any way displace the benefits. He made that clear. That is not at all what Charles Murray is advocating. Charles Murray's argument is that everyone gets $13000 a year baseline, you can earn up to $30,000 with no deduction, but it replaces the existing welfare programs of the federal government lock, stock, and barrel. They are therefore, in that sense, two different ways of thinking about this. One is, and I'll call it the universal basic income of the left, that we are just going to add another benefit and give it to everybody indiscriminately. The UBI of the right says we are replacing the entire federal welfare system - state if you prefer - with universal basic income and that will be it. Now, plan B is financible, because it ultimately solves a very real problem that we should all be aware of, and it's that the existing system of welfare is not salvageable. It is accumulating unfunded entitlements of mind-boggling proportions, so something needs to be done about that. If you're talking about universal basic income and not proposing to replace the existing welfare scheme, you're smoking weed, because you're simply creating a new unaffordable set of entitlements on top of the existing ones.
I think there's also a gap between the Replace Everything and the Replace Nothing camp - there is a lot of silly stuff out there today. Indonesia was doing fuel subsidies and has recently replaced that with cash transfers. That is less distortionary, there are less intermediates, it is a much more efficient way to provide a welfare program. There are other countries that are thinking about these sorts of things. One of my favorite studies is a food stamp study. India gives out food stamps, and there was an experiment that was being run where the government allowed people to buy food stamps for cash. What happened? The vast majority of people sold their food stamps at a big discount, the government saved money, and it was a more popular program. I think there's a big gap between zero everything, and a lot of silly stuff being done that could be replaced by some form of cash transfer.
As a counterpoint to both those arguments at the CBPP, which is a left-leaning think tank in Washington, their head, Greenstein, ran an analysis of universal basic income and he said, "Suppose it provided everyone with $10,000 a year. That would cost more than $3 trillion a year, $30-40 trillion over ten years, and that single figure equals more than 3/4 of the entire federal budget - double the entire budget outside Social Security, Medicare, defense, and interest payments - and is also close to 100% of all tax revenue the government collects." He went on to talk about how some of those more targeted programs, how wiping them out would be very detrimental to people who are in the high need end of that. I focus a lot on local issues, and I can say in San Francisco that when you look at our homeless budget, for example, a lot of it is spent on high-needs users who end up costing $80,000 a year in emergency medical care, so a flat transfer might not cover all of the true cost or quality of life issues around that, so I was wondering if anyone wanted to respond to that.
You have a group in Washington that spends time researching the status quo and it's almost unimaginable to many of us that something so wholesale different from our current welfare state setup might be possible, which is why no one is talking about introducing a bill in Congress this year. What we really need is a 10 to 20-year effort to understand whether we ought to reorient our culture, our economic structure, in order to be able to raise the floor. I think those kinds of analyses assume status quo constraints, tax levels, and I think we can have that debate in 15 years.
What's remarkable, and it goes back to your hired drivers … if this were put to the electorate, it would be rejected. It was rejected in Switzerland this year - it didn't even get that far in Germany because the parliamentary commission rejected it. It's an interesting idea in the sense that something that you would've thought would be very popular turns out not to be very popular. Why is that? I think what's fascinating is that voters, particularly in the Rust Belt, decided they would rather gamble on somebody who was offering them jobs than consider entitlements as a solution in any form. What we're witnessing from the island of California looking at an experiment on the mainland - Trumpland - you're laughing because you think it's all idiotic, but if you look at a paper that's just been published by Otter and Hanson, they show that this is not quite as crazy as it might seem.
Most of the job destruction they show has been the result of free trade far more than technology. They've done some extremely impressive research on this to show that the effect of new technology, in technology centers in parts of the country, was to displace jobs, not to destroy them. It didn't lead to less employment. What was really impressive was the impact of more free trade from the 1980s, which really did reduce amounts of employment. There is a distinction that we need to draw here that isn't nearly drawn often enough between the causes of job loss insofar as that people this year voted for an explicitly protectionist candidate who said he would tear up NAFTA, wouldn't go anywhere with TPP, would engage in a trade war with China and may well follow through on that. Voters on the mainland - particularly in the Rust Belt - were behaving in line with good economic research. They weren't voting to smash the machines, to rampage toward Silicon Valley and burn down the technology companies' headquarters. It's a correct diagnosis that free trade has been very destructive to the employment opportunities of less skilled Americans. In some ways, this debate that we're having is academic in the worst sense, because we should really be debating the new policies that are about to be implemented. This is the biggest policy regime change that the United States has had in economic policy in a generation. We're sitting here talking about universal basic income...
This was planned before the election.
It was planned when we knew the election was coming, but nobody in this room, I will predict, except for me, thought that Donald Trump might win and embark on a protectionist agenda. We'll see, it's going to be a fascinating and possibly very painful experiment whether that can generate the kinds of jobs that people in the middle of the country...
Because we have you on the panel as the UBI skeptic, what kind of alternatives are attractive to you in terms of kind of shaping or reforming the anti-poverty system that we have? Is it in the earned income tax credit, something like trying to make basic jobs better quality or higher wage? What else is on the table of options here?
The reason I think the universal basic income option is not viable is political. Ideally, one would strip away the existing schemes of the legacies of the 1940s and the 1960s and replace them with something very elegant. The appeal of UBI is it's transparent, it's simple, it's cheap to administer, and I'd love to do that, but I live in the real world. You are not going to get any congress in any imaginable state of the universe to say, "We love UBI, first we'll get rid of the existing systems of Social Security and Medicare and Medicaid, and then we'll introduce UBI." It isn't going to happen. What we know from abundant experience is if you come up with a new policy, Congress will graft it on to the existing policies to create yet another tier of government entitlement, because that's politically expedient.
Nobody, even those truck drivers, likes it when their entitlements are taken away. That's what's so fascinating about the Obamacare pivot that Trump did the day after he won. "There are parts of affordable healthcare that I actually like, i.e. the parts that are popular." It's not plausible that we are actually going to get rid of the old welfare system. The alternative is what in fact happened the last time we did serious welfare reform; you take the existing system and you reform it. That's what Reagan did. You have to address the fact that it is completely unaffordable given our existing demographics. It was not difficult, for example, to change the age of entitlement in the case of Social Security in the 1980s. It was politically tricky, but they did it. The same went for the tax reforms that accompanied those changes.
Bottom line, the most successful welfare reform is that which raises growth and job creation. Everything else is essentially a palliative if you do not achieve growth and job creation, and we should not lose sight of that. I will say one more thing, which is the historian's view: the history of debates on poverty resembles the swinging back and forth of a pendulum, and this has been true since the 1830s in England, when the original debates took place about whether the poor should simply be handed money, which they were under the old poor law, or whether they should somehow be made to work for it in workhouses, which was the 1834 system then introduced. When I read the existing literature on universal basic income, I feel like I'm in a time machine that's taking me back to early 19th century England. "Will they just squander the money on drink?" That's exactly what people used to say in the 1830s, arguing for a system of workhouses. That's what makes this discussion for a historian strangely dispiriting. It's like we just keep having the same argument: "Will they waste the money? Can they be virtuous? If only we could make the poor virtuous..." But this is an old Victorian debate, and it's going nowhere. If you give them jobs, they aren't likely to work.
I'm going to give a run at the "This Time Might Be Different" argument just to play to character. Although I run Bloomberg's fund, and we learn in the finance industry that past performance is no indicator of future results, so I'm going to apply that here, which is to say that if you think about the advances that the computing revolution has introduced into the modern economy for the past 25 years, the reality is it's been marginal. You walk into an office, you have effectively the same work experience, only now instead of somebody writing on transparencies with a marker and putting it on a bulb with a mirror, they use PowerPoint or Google Slides or something like that. If you talk to the technologists, speak to software engineers and ask what they're seeing, and I'm happy to have people disagree with me, many things in technology take the effect of the ever-receding horizon. I remember for years and years, people were saying, "Oh, the explosion of mobile is just around the corner, it's two or three years away," and every year it was two or three years away. But in the last five years, in artificial intelligence, we have seen a pace of advancements that honestly the practitioners in the field didn't expect. Everything from a computer being able to tell images, so, for example, we invest in a company that will read satellite images from space and be able to calculate the growth rate in different markets - for example, the Chinese construction market - by counting the shadows on buildings. Something that sounds completely science fiction - this stuff is starting to happen. The telling fact, and we're going to conduct some research on this, is that the estimated arrival dates for some of these new technologies keep getting pulled closer by the technologists.
Can you give me some examples?
Self-driving cars. Our fund is focused on the future of work, and my partner, Shiobhan, came up to me three years ago and said, "We should start investing in artificial intelligence." I thought, "You have watched too many science fiction movies this weekend, let's go back to enterprise software or whatever we usually do. But if you want to work on it, hey, it's your time, be my guest." She came back three months later having looked at 2200 artificial intelligence startups and the message was that every single domain is under attack. Now, three years later, I believe the exact same thing, only more so. A very simple example is the self-driving car. Good estimates for when components of the self-driving car would arrive, if you looked two years ago they were seven years out, and if you look now they're three years out. We lack crisp industry definitions for what exactly does it mean because we already have cars that can park themselves and do highway driving on their own, but I think by a whole cluster of measures - image recognition, purchase of semiautonomous robots for factories, the ability of a computer to play a video game it's never seen before - from the silly to the economically meaningful, there is a case to be made that machines are developing capabilities that compared to human beings, they simply didn't have. I bet the horses went around for a long time saying, "The humans will never find a way to make a horseless carriage." Until one day we did. The self-driving car I look at as an economic fact, but the self-lending mortgage officer, the self-auditing accountant, one of these will be a reality, and then the political landscape will shift. I completely agree with you if we're talking about status quo politics, but give us 20 years, and we can change a lot of things.
I was trying to find common ground but I think it might be a fool's errand. Kim-Mai said given today's federal budget in the U.S., basic income is unaffordable. I think most people would agree that's true. Niall said given today's political climate, we're not going to universal basic income through Congress. I think most people would think that's true. Now, I think there are two debates to be had: one is the very specific full universal basic income - should we implement it today? I don't think it's possible for all of the reasons we've discussed. The second is that the core conceptual elements behind a UBI are the things that actually impact welfare reform.
What are those? Delinking benefits from work. I think people across both sides of the aisle probably think that's a good idea. Giving more cash as opposed to in-kind goods, and getting rid of that waste or inefficiency. I think a lot of people across both sides think that's a good idea. The third piece of this is coordinating social programs, so rather than get housing vouchers and food vouchers and every different kind of program, which is costly as it passes through many layers of intermediates and is not coordinated, can we coordinate it?
You can look at those three concepts and say even that's too much. We're not going to make any progress. On that, I'm probably more optimistic than many because I've seen it happen not in the U.S. in the last decade, but around the world. Many countries have replaced fuel subsidies. Many countries are moving food stamps and vouchers to cash programming. The Kenyan government, where we spent a lot of time, is working to coordinate their various cash programs. India is asking these questions in their annual economic survey, which is one of the most important policy documents in the country. UBI will be in there as part of the discussion, and they're thinking pretty seriously about moving some of their subsidies over. There has been tremendous progress over the last decade.
But it's still a side issue. It's not distracting meaningfully from discussions over what trade policy should be...
Of course not.
The risk to distracting us from the issues at hand, I think presently, is pretty low. There were a hundred of us who signed this economic security project, which is a document asserting that economic security is a human need and we ought to explore cash programs basically to genuflect before the complexity of the problem. For Y Combinator to do a study and a few of us to talk about it here, I don't think is distracting meaningful public attention, but if it turns out 20 years from now, on the chance that history doesn't repeat itself and it turns out to be different, my basic view is that we will wish we had spent the 20 years before that doing economic research, preparing politically, figuring out how to talk about this. Universal basic income certainly is not going to be what we call it in public - we just don't know how to talk about it yet, and it's a mouthful.
I'm wondering whether you're actually going to get a meaningful answer from full-scale experiments. I know we're all supposed to believe passionately in fieldwork, yet I suspect the problem here is that it's almost impossible to create in an experiment what would be the reality in your dystopian future of mass permanent unemployment. It's a dystopia to imagine people with nothing to do except wait for the giants of Silicon Valley to mandate them universal basic income. I want you to understand how Marie Antoinette it seems for people who've made billions of dollars in technology investments to speculate about just how little they're going to dole out to the poor suckers who aren't software engineers. I'd worry a bit about the optics if I were you, because ultimately this is what you're proposing: doles for the mass unemployment. If that's your model of the future, it's dystopian in a way that sort of recalls Blade Runner or Terminator more than the happier episodes of Star Trek. I'm here to cheer you up. The truth of the matter is that every single wave of technological innovation, which enhances productivity by removing cumbersome and inefficient ways of doing things by hand, is set to presage mass structural unemployment, and it never happens. It's quite clear in economics that it's augmenting to have these productivity enhancing innovations. It augments the labor supply, and what most people do when they benefit from the increased productivity when something is available that used to be really expensive and now is really cheap, they say, "Awesome! I'll take most of that as additional consumption. I won't stop working, just knock off." The leisure preference is really sort of one-seventh of what people do with the benefits of any technologically driven increase in income. I have to say that economic history is really against your nightmare scenario.
Let's be clear: this is a Pascal's wager. I think the entire model of "Let's predict what the economic outlook for employment looks like 20 years from now and plan based on a point prediction" is a foolish way for humanity to think about itself. My only argument is that there is tremendous uncertainty about what that future might be and we need to prepare. In a situation where there is technological unemployment, we should be ready. If you look at America today, there are millions of people for whom the game is rigged, they cannot earn more than their parents, it is an unfair and unfree system, and UBI - a friend of mine who is a billionaire investor who shall not be named - said, "I think universal basic income is plan Z. The problem is that I can't think of what plans A-Y might be." I am kind of in the same place, which is groping for some alternative but thinking this might be the best way out.
What is the path forward in your mind for creating the kind of cultural change and coalition building that would be necessary to respond to his critique of the political viability of this idea?
I think we have years of local, grassroots organizing and cultural change ahead of us. We have not seen a version of The Jungle for this era - of what labor laws ought to be called for - so I don't think we know yet. The analogy I draw is to the marriage equality movement. For a number of us, and Jim Pugh has been involved in a lot of the community organizing, for a number of us involved in discussions about this issue, we look at legalization of marijuana and marriage equality as potentially relevant analogues. That was called the gay marriage movement for a long time. It polled not so well, and after many years of work, finally there was the breakthrough that the right way to think about it was marriage equality, and then all of a sudden the political framing changed. In the handout scenario, it's very easy to go along with that dystopian view. What if I told you, "You know what we should have in the United States? We should have a mandatory year of national service. We have very few experiences we share as Americans, and that maybe that ought to be the ticket in to a dividend of UBI for the rest of time." All of a sudden it takes on a different pallor of national unity and contribution. I think we don't yet know what frame to fit it into. All we know is that there is a tremendous risk of large scale structural unemployment caused by technology, and we might even be seeing the beginnings of it today, although I take your point that it may have been caused by trade.
If Plan Z ends up being the plan in your scenario - if we think about what the American dream was in the post-war period in terms of buying a house, getting employment within a vast corporate conglomerate and having a pension and benefits, how do you define the American dream in the scenario you're describing? How would you express that to the American public?
I think this is a place where we need the writers and filmmakers who professionally imagine futures for us to think about it. One of the thoughts I have on this is that careers today look much more like a portfolio of activities than they do a linear succession of jobs, and we may see much more of that. It may not be that different from hunting in the morning and fishing in the afternoon, but we may live in a world where your affiliations to work are not a permanent, full-time thing but a shifting set of projects that you recombine, much like the way a Hollywood film director works today, but just for more of us. I don't know.
I think we've also evolved to a very specific definition of work and job. If you stay home and take care of your kids, it's not work to a lot of people. It's not a job. But it arguably is one of the most productive and satisfying things you can do with your time. As we reformulate our conception of the American dream, we need to think much more broadly than work. I think words like choice and dignity come to mind. One thing Niall said, which I agree with, is as we become more efficient, we're moving from a capital-poor society - at some point, we didn't have enough money for everyone - my initial stat I had about closing the poverty gap was not a possibility two decades ago. We're moving from that world to a world where we will have enough capital, and we will have a surplus of time. We're all becoming more efficient - we can create the same amount with less time. There's a real question of what we do with that time. Who gets it? Do we make the extreme poor work these really meaningful jobs while we get to write poetry and play guitar and take care of our kids, or do we allocate that time across and think about this much broader conception of what that dream is?
Let me add one additional thought to that. If one's talking about American dreams, one has to be careful here not to come up with a distorted image of the past in which everyone was working for the man - some sort of large organization - a characteristic feature of the American past was a very dynamic entrepreneurship and a very dynamic small business sector. One of the disturbing features of the last eight or so years has been that there really hasn't been as much new business creation across the board, either in Silicon Valley or on the mainland. That's partly because the regulatory burdens on small businesses have never bene higher. When it comes to development economics, Hernando De Soto guides that the poor have their labor, cheap though it may be. What they lack is capital, and it's actually making sure that our capitalist society has as many capitalists as possible - that is to say that people have, if not their own business, a stake in business. The truth about Silicon Valley is that it talks about a democracy of netizens but, in reality, the ownership of all these networks is incredibly highly concentrated. Most people are users. They're not owners. If we could widen the distribution of ownership of capital, then all the problems we are talking about would significantly be reduced. That should be the agenda for the future. Not just wondering how little can we pay these people for it to be viable, let's ask, "How can we ensure that people are genuinely stakeholders in our capitalist system and not just facing a simple choice between a really crummy job or universal basic income?"
Thank you. I'd like to open it up to audience questions.
Thank you for an excellent conversation. Professor Ferguson, you talk about the cultural issues. You've only referred to them lightly with a universal basic income in the United States, at least. I'm intrigued that the panel hasn't really grasped onto that. Bertrand Russell wrote an essay called In Defense of Leisure where he visualized a utopian society where nobody actually worked, but everyone was productive. I still have not been able to figure it out. I'd like the other panelists to comment, if they would, on Professor Ferguson's assessment of the political and cultural viability, in this country at least. I appreciate the emerging markets context, but I think this is an American conversation, and, from my perspective, it seems that there are some deep-seeded cultural issues.
As I said, we've done research, and today, it is not culturall viable. When we asked the truckers about universal basic income - and, look, focus groups are not data - but we heard things like, "I'm still waiting for my 40 acres and a mule!" and "I don't want a handout from the government. I don't want to owe the government something." Today, in the current framing of a handout, I don't think it is culturally viable. I also think it's a caricature to say we're solving for how little can we get away with paying people. As an alternative view, many of us are involved in researching this and are curious about the baby bonds idea, which is how you can start life with a stock of capital that would be yours to begin with? That's really a question about timing of payments more than anything else. There are so many variations on this idea, and so many ways to bring it to life, and so many decades of American life we still need to see, but you know what? I could see us saying that we created an American society where the starting line isn't fair. I fundamentally believe that if you hold these two views: one, which is life should be a pretty fair game, two, if I work hard I can give my kids an advantage over other people, if you believe both of those two things, you are contradicting yourself on some level. I think those are both very American beliefs, and there's a contradiction at their core. There are variations on this theme that are not a monthly handout we could still explore that might deepen our conviction that America has spoils we should all share.
Culturally, welfare has gone through different waves. A lot of our views on poverty and handouts is colored by the welfare queen, the drunk person on the corner with the brown bag. One of the interesting things to me about this election is that welfare benefits went from being seen as something that weak people do - "I need welfare, I'm weak" - to something, and Trump came out and said it: "I have not paid taxes. I am strong for not paying taxes, that's what strong, robust people do: they take from the government." And I actually think even in the last six months, the tone around what it means to take welfare has changed. I think it was partly the election. We don't know, and that's why we should test it, btu we've never given universal benefits.
The closest you'll get is the Alaska Permanent Income Fund, or when we got the checks in the mail from Bush. If everybody gets something in the mail, it may change the way we talk about it. It's impossible to imagine what our reaction will be after the fact. We just don't know. I'll just add one thing, which is that I could not agree more that, in a way, the business model of America is entrepreneurialism, and we've seen a dramatic decline in the number of new businesses as we've seen the percentage of Americans employed by large corporations go up and the size of the largest corporations go up. If you want to promote risktaking among the everyday American, make their floor higher so they're not afraid, as many Americans are now, that if they don't have $400, that if they run into an unexpected health issue, or car trouble, they won't be able to pay for it. You could make a case that the single most entrepreneurship-promoting thing that America could do would be something like a universal basic income.
I also think it's dependent on how the program is framed. I look at a lot of housing structures and housing markets, and in a way, assistance exerted through tax subsidies or deductions is much more culturally accepted in America than "handouts." I could say the mortgage interest deduction in the United States, plus non-taxation of impuded rental income for homeowners is anywhere from four to ten times the size of rental subsidies that we offer to low income families, which only cover a fraction of the families that actually qualify for it. One is perceived as welfare and the other is not. There are programs that already exist, like the EITC, that could assist low income families which already fit within the American framework without being something radically new.
The Golden Gate Bridge is a beautiful bridge - does anyone know where the steel was made? To my understanding, it was made at Bethlehem Steel or U.S. Steel, one of the two. Does anyone also know where Bethlehem Steel's plant in Pennsylvania does now? It's a casino. I want to ask all of you: does that really look like the America we should be proud of, where we're spending time gambling, or an America where we're building great things? In the Great Depression, when we were out of work, we decided to invest in two amazing bridge projects in San Francisco that still to this day attract tourists: the Bay Bridge and the Golden Gate Bridge. Steel is more commonly made in China, for example, and other countries. Our new bridge was made in China. It took twice as long as the estimates were expected to have it built by and cost twice as much. There's actually a pipe that sits right above me from AB&I Foundry in Oakland, and I've talked to the CEO, and he tells me that they're actually under tremendous pressure from Chinese imports right now. Niall, to your point about the trade question, I think that is a very pertinent question we haven't addressed fully. I'd also like to ask all of you, what do you think about public works?
You're going to get your wish granted very, very soon. It's clear that the new administration is going to embark on an anti-China policy that could include heavy tariffs on Chinese importants and be an infrastructure investment program that will be building bridges over every available river and bay. The fiscal consequences are sort of eye-watering, since you're adding a minimum of $1.7 trillion to the ten-year deficit at a time when you've already seen bond deals go up. Oh, it's going to be great fun. One thing's for sure: America has chosen a more interesting future than if it had elected Hillary Clinton.
I didn't hear anything about providing an outstanding education on a K-12 basis as a starting point, rather than giving people cash that have come out of failed schools.
I was just about to say that, and I thought, don't be boring, but it's obviously the case. Ryan Chicchetti's name has been mentioned here. He's best known for his work on education showing that if you only have good quality education in public schools, the impact on social mobility is really striking. If you had to pick a single thing to fix in this country, it wouldn't be the infrastructure, with all due respect to the president-elect, it would be the educational system, which is doing dismally. The latest scores underline that. Universal basic education would be my preference, because that seems to me where the United States is failing relative not just to the rest of the world, but to its own history. That's what's so tragic.
No question, we're at an affliction point with technology, but I get a sense that at least three of the four panelists are ready to write off a vast majority of the people in this country. Wouldn't a sensible approach be to blow up an obsolete public education system and start educating kids in public schools so they could compete and create the next generation of jobs and our future, as opposed to being on the receiving end of this?
With all due respect, I'm not sure what you mean by writing people off. Of course we need to invest in education. Now here's the skeptic question: imagine you did, and somebody who has an equivalent level of education in another country is able to provide the same service at a much lower price because of a lower standard of living. It's not clear to me that even education solves the problems that trade - free trade of goods, services, and ideas - creates. I'm not saying it's a bad thing, it's a wonderful thing, and we should all be educated for a wide variety of reasons, but I don't think it's a panacea. My reading of the protectionist history of the United States, and the United States has had great success being a protectionist nation over hundreds of years, was that it was always to protect nascent growing industries and not to protect the decline of an industry that had a lower-priced competitor overseas. I don't think that is our way of talking about protectionism now, and maybe I'm reading the history incorrectly, but I think we need a view on what is the growth industry? China, by the way, writes twice as many AI papers in their academic institutions as the United States does. If you believe artificial intelligence might be a lens for the US to grow GDP, do you have a guess on how much the US spends on AI research every year? Not enough: a billion dollars. A drop in the bucket. The paperclip budget for the Department of Education is probably bigger than the entire amount we spend on artificial intelligence research, and we should be ashamed as a country by that. If that's the growth industry, we should invest much more heavily.
Out of obligation, let me give the economist's answer to a number of these questions. I think we need to divide social spending and welfare into two buckets: one is what is known as public goods - what people need to coordinate on - roads, vaccine discovery, public education, defense, all these sorts of things. Then there is what is known as private goods - stuff like food stamps, housing vouchers, things that you give directly. In no world do I believe we should be replacing public goods with cash. But we also spend a heck of a lot of money on inefficient, private good stuff that we give people that can be replaced in cash. When you create this dichotomy, I think it's the wrong dichotomy to be asking. The second thing I'll say, which is a global perspective, but it's important to look beyond our boundaries, is that training programs have been the staple of a lot of development interventions around our world. They've also become the hallmark of failure. We do not know how to train people in agricultural stuff globally, there are a lot of things we have done a historically bad job on. We've spent hundreds of billions of dollars with almost no impact. We need to be really careful before we assume that training - not public education, but training programs - are the solution to a lot of these problems.
You're talking about a program which will have profound impact on what the future of what the United States may be. You should give some consideration to what kind of a society you want it to be. You want one that's based on gratitude, or one based on entitlement? Because that's going to distinguish what people think about this program. You better decide that.
I'll take gratitude.
If you read 19th century Russian literature, much of it deals with what it calls the Era of the Superfluous Man. Books like Eugene Onegin by Pushkin, Rudin by Turgenev deal with people that don't have much to do in their lives. I'm wondering to what extent is history a guide for the future we might imagine? Can the experience of nobility and gentry in Europe in the 19th century be paralleled to the future of the 21st century?
As a longtime student of Russian literature I'm delighted to end our conversation on that subject. 19th century Europe was bedeviled with a problem of a surplus of educated young men. It was one of the sources of revolutionary instability that the Tsarist regime worried about because it saw what happened in 1848 in Central and Western Europe. The problem of educating people who then can't be employed was a great one in those days, but that's not our problem. Our problem is not overeducated people who can't find jobs; the overeducated people can find too much work, and people like me work 17 or 18 hours a day, seven days a week, and wonder what the hell all the education got them into.
Our problem is the unskilled, uneducated people who don't have anything to do. 19th century literature has an answer to that problem, too, and the answer is a depressing one, but one that we should contemplate: in the 19th century, and most previous centuries, unskilled surplus labor was absorbed in domestic service on a massive scale. Those Russian country houses that you read about in Turgenev or Chekov are staffed by nameless and numerous domestics, who in the pre-1860 literature are serfs. One of the dystopian futures that I contemplate, and I think it accounts for the popularity of Downton Abbey in the United States, is ultimately that the 1% will just end up employing more and more people to do more and more menial tasks that, frankly, you don't really want a robot to do. Fundamentally, the problem with the vision of everybody left redundant by artificial intelligence and robotics is that most people don't really want a robot to do a lot of fairly private, intimate things. They'd really rather have a human being. Most of the economic growth, the employment growth in years to come is going to come from healthcare as more and more of us remain alive but degenerate mentally. That appears to be our basic game plan for healthcare. As we go ga-ga and live to be 99, we're just not going to want robots to look after us, just lots and lots of flunkies. The nightmare vision I have is for a vast American Downton Abbey with hordes and hordes of people employed as nurses and chambermaids.
It's scary, but I'm more worried about the dystopian present than the dystopian future. On one note, yes, people would like care from a machine, but I've watched my kids speak to Siri and Alexa, and when they're old, they'll be perfectly happy to have lullabies sung to them by Siri at night, and they'll find that very soothing. I think we're in for more change than you may think.
Thank you so much, everyone.