Loading summary
Transcript121 lines
- [00:10]
A
Hello, and welcome to the Data Detective episode of Slate Money, your guide to the business and finance news of the week. I'm Felix Salmon of Axios. I'm here with Emily Peck.
- [00:25]
B
Hello.
- [00:27]
A
Hello, Emily.
- [00:27]
C
Hi.
- [00:28]
A
How's your employment situation these days?
- [00:30]
B
I am a little less employed these days. Today was my last official day at HuffPost, which just did a lot of layoffs. So we don't know what I'll be doing next yet. But it'll be something cool. I'm sure.
- [00:42]
A
It'll be awesome. We look forward to finding out what it's going to be. No one more so than Tim Harford. Welcome.
- [00:49]
C
Hello, Felix. Hello, Emily.
- [00:51]
A
Tim. I don't know. I've lost track of how many times we've had you on this show. It's always fantastic to have you on this show. You have a new book out called the Data Detective, which we are going to talk all about. It's really fascinating about statistics and how to think about them, how to use them. We're going to talk about the stimulus, of course, the $1.9 trillion that has just been passed, signed into law by President Biden. And we'll talk about the effect of all that. But kind of the thing that we've all been waiting for and that we are going to go into enormous detail about is Beeple. Of course, we have a whole section on Beeple and we have a Slate plus on Florence Nightingale. It's an awesome show. So stay tuned for all of that coming up on Slate Money. So, Tim, we need to kick off by talking about your book, which is statistics are not quite as terrible as maybe they have a reputation for being. Is that the TLDR here?
- [01:58]
C
Yeah. Corey Doctorow, the blogger and science fiction author, said it should have been called how to Truth with Statistics. It's actually called the Data Detective in the US and it's called Other Things in Other Parts of the World because publishing. But it's about making the case that statistics are extremely useful. They're a sort of radar that show us vital truths about the world and that ordinary people asking fairly straightforward questions can make sense of them. You don't need to be an expert. It doesn't need to be too complicated and you can make sense of them.
- [02:32]
B
I love that message because basically I felt like, oh, I don't need to actually understand statistics to understand statistics, because a lot of it is feeling and thinking, and I find I can do both of those things fairly well. You have this knack for explaining things where everything just sort of falls into place.
- [02:52]
C
Thank you, Emily. So kind of you to say.
- [02:53]
A
It's what you're saying here that I don't need to go to school to understand regressions and stuff in order to understand statistics.
- [03:02]
C
Well, if you want to understand all of statistics, then understanding what a regression is will probably help. But actually, you can get a very long way to making sense of the kind of statistics you see on social media or in newspaper headlines or on the TV by asking really straightforward questions like, well, the first question I say, this is going to be the really weird one. This is the one that baffles the statisticians. But the first question is actually, how is this statistic making me feel? Because so much of what we believe or choose to disbelieve is, is about our emotional reactions. So really, the very first thing before you hit, retweet or like or tell your friends about this number, just observe your own emotional reaction, calm down for a second, and then go back and have another look. And you'll be surprised that when you take that, just that couple of seconds to calm down and you look again at the statistic and you're no longer angry or afraid or feeling vindicated or smug, it looks a bit different. And then other questions are simple things like, is this number going up or down? Is this number big or small? Can I compare this number to other numbers that I understand? I mean, I get towards the rocket science later in the book, but it isn't, in fact, as complicated as people make out.
- [04:17]
A
And I think this is incredibly important to anyone who pays too much attention to polling, especially during elections, which is basically everyone in America, because polls are a form of statistics. Right. And any one poll is, you know, because you get emotionally invested in one candidate rather than another candidate, you start feeling too much about this poll rather than that poll, and you can miss the first of the trees.
- [04:47]
C
Yeah. And there's enormous amounts of noise in polling, and people need to take a step back and say, well, what in fact, is an opinion poll? One of the rules in the book is ask who's missing. Well, the whole thing about an opinion poll is lots and lots of people are missing because the pollsters will email a whole bunch of people or they'll phone a whole bunch of people, and the vast majority of those people will not respond. And so you've polled, and I'm putting the numbers off the top of my head, but Maybe you've polled 50,000 people and 2,000 of them have actually got back to you. And so not only are you trying to predict how A very close election is going to break. You're doing so on the supposition that the 2,000 people who would actually talk to you are perfectly representative of the 48,000 people who wouldn't. And we know they're different. We know they're different in at least one way, which is that the 48,000 wouldn't talk to you. And pollsters know this. But it's one thing to know about it, it's another thing to do something about it.
- [05:45]
B
And you had that great example of a poll, I think it was in the 1930s, right, when Franklin Roosevelt was running for reelection. And it was like a massive poll, huge numbers, and it looked like he would lose in a landslide. And it was just like they were asking a certain cohort of people. And it was hugely embarrassing for a magazine, right?
- [06:04]
A
Was it, Tim?
- [06:05]
C
Yeah, it was the Literary Digest, which shortly afterwards disappeared. It's an amazing story. It's well known in statistics, but I think it's worth retelling. Alfred Landon was gonna win that election by a landslide. And the poll, they polled 10 million people, 10 million people in 1936. It's incredible. I think it's about 25% of the entire electorate, and about two and a half million people got back to them. But there was a differential response. So the Alf Landon Stans got back to them enthusiastically, and the Roosevelt fans didn't. But also, how did they find their 10 million people? Well, it was subscribers to Literary Digest magazine who were Republicans, and it was people they pulled out of the phone book who had a phone in the 1930s. And it was car registrations. There was a database of people who owned a car in the 1930s, in the middle of the Great Depression. Well, it turns out these people are not, in fact, representative of the rest of the population. So you really need to ask, well, who is included and who is not included in the statistics that we're gathering?
- [07:13]
B
And this still goes on today. You see these online polls all the time. National Review readers say that Trump is great or whatever. This still goes on.
- [07:23]
C
No, it still goes on. It still goes on. But the lesson I draw, one of the lessons I draw is actually all this excitement about big data. We have to understand that all these data sets that people get so excited about that are very powerful and can deliver all kinds of insights, but they don't necessarily include everybody who you might want to include. There was one going around on Twitter just this past few days. Cathy o', Neill, I think, had a field day with this, of course, as she would where you could put a photograph in and it would turn your photograph into an old school oil painting. And it turns out if you put a photograph of President Obama in, he comes out a white dude. And in fact, I think Oprah Winfrey comes out a white lady because all of the oil paintings they put in are of white people. And so even if you put a photograph of a black person in, you get a white person out. That's just an example of how you think you've got all the data. And in fact, all you've got is all the old oil paintings, which is not the same thing as all the data that you might want.
- [08:28]
A
It strikes me that you're saying how to truth with statistics. But a lot of this comes back to there are problems. And on some level I feel like what you're saying is, well, you can look at a poll, but then you need to look at the number of people polled and whether they're representative and get into the nitty gritty of the methodology. And is that realistic?
- [08:50]
C
I would boil it down to three pieces of advice. First is calm. Be calm about the number you're looking at. The second is get the context. Where did this come from? Is it big? Is it small? Is it going up? Is it going down? And the third thing is curiosity. I want to use this information to understand the world rather than to win some stupid argument. Because the moment you're trying to win an argument, you're making yourself dumber. So the three Cs, calm, context, curiosity. And you're absolutely right, Felix, you can't do that for every single fact that you happen to observe. But what you can do is to ask yourself, is the source that I'm looking at the social media post, the journalist, whatever, are they doing it? Good journalists will give you the context, will explain the problems. We'll say where this came from, we'll say what this number looked like 10 years ago. And a stupid JPEG circulating on Facebook will not give you the context. So that's your basic way.
- [09:46]
A
Can I come out with a dirty secret of journalism here?
- [09:49]
C
Please do.
- [09:50]
B
Don't tell him.
- [09:51]
A
Don't tell anyone. But you sleep money listeners, you get to find this out. But just don't tell anyone because it would do a horrible service to the noble calling of journalism, which is that it's becoming increasingly difficult to trust news outlets. There are 100% great, statistically responsible journalists at every single news outlet. And if you read a piece by Tim Harvard in the Financial Times, you can trust that. That and he's reporting on some statistic, you can trust that that statistic is, is worthy of reporting. But the velocity of news at every single outlet in the world is now so high that you can't just trust the outlet anymore that there will every single outlet. There is someone who doesn't really understand this stuff and a harried editor who doesn't really understand this stuff, and a bunch of largely innumerable people who are like, yeah, that looks okay, let's feed the content beast. And it's becoming harder and harder to just say, oh, well, I read it in the Wall Street Journal, so it must be true.
- [10:56]
B
Ouch. That brings me to something I would like to talk about, which is the past year of vaccine and coronavirus statistics and whether we think broadly that we've done a good job measuring the pandemic. I mean, I remember at the beginning there was a lot of talk about there were all these models projecting certain number of deaths that didn't turn out to be true. And there was a lot of noise around that. Then it was hard to track the cases. Like, Tim, where do you think we kind of stand now? Do we have the good picture using statistics of this pandemic?
- [11:32]
C
Well, it is, of course, a mixed bag. There are some sources that have been fantastic. I mean, epidemiology, Twitter is amazing and there are some sources that have been terrible, as you would expect. The models, actually, you say the models didn't pan out. I went back to one of the most famous modeling exercises just this week because we're coming up to the first anniversary of it. It was report nine, I think, by the Imperial College Covid Response Group. And it came out in mid March 2020. And it said that if nobody does anything about this, there's no government regulation and there's no voluntary response to the virus, then about 80% of the American public are going to get it and about two, two and a half million people are going to die. Was that wrong? I don't think that was wrong because half a million people did die and counting and that's despite a fairly substantial response. So I think they got the big thing right. They got the infection fatality rate about right. They predicted this is not just flu again, and they predicted hundreds of thousands of deaths. And they said, look, there will be a response. It's not going to be two, two and a half million people because there will be a response, but it's in that ballpark. So I think even the early modeling, when you go back and look at it, I think it Stands up surprisingly well.
- [12:58]
A
Well, as we were talking to Charles Kenny last week, that modeling clearly failed in India and large chunks of the rest of the world.
- [13:08]
C
Yeah, well, the Imperial College Modeling only looked at the UK and the US but, yeah, I mean, of course we made a lot of mistakes. I think the modellers have done okay. I think the epidemiologists have done okay. I have been surprised at how well some of the pretty early stuff stood up. There have been some big mistakes, like advising people that masks probably wouldn't help. That wasn't great, but I think it has reinforced the message I've been trying to get across for a long time, which is that the statistics matter. The statistics are showing us something important. They're not just about winning some stupid argument. They are telling us about this deadly threat that's coming in and how to direct our resources. And the other lesson I draw is that there were holes in the statistical data that were perfectly predictable and you could point to them in advance, and we just didn't bother to fix them. So in the US for example, Alexis Madrikal from the COVID Tracking Project, which is just an amazing project which recently shutted its doors, he told me that at the beginning of the pandemic, the United States of America did not know how many hospitals there were in the United States of America. So just a sort of piece of information, and we just take the data for granted. We assume it's in some spreadsheet somewhere and all you need to do is download it. But very often we don't have the numbers if we haven't made them a priority to collect. And it takes effort, it takes skill, it takes money.
- [14:37]
B
From that perspective, maybe the US should get a better grade, considering we're starting from zero tracking something we'd not heard of until December. And now we're up to speed and the COVID Tracking project was created. I can look on my New York Times app every day and see the number of cases, the number of deaths, and a bunch of other statistics. It seems like there's been a robust statistical response to the crisis the last nine months.
- [15:04]
C
The statistical response has been good. The crucial first few months were not great. And not just in the United States. I'd love to blame Trump for that, but it's not just Trump. Amy Maxman was writing in Nature over the summer about the US having a Covid data crisis because just the infrastructure was built to deal with little outbreaks of salmonella, and people were just faxing each other information. Just wasn't buil deal with hundreds of thousands of cases every day. And given the resources that Silicon Valley has, given the amount of money there is in the United States and the technological sophistication, you would hope that the US could have done better. And my own country, the uk, we should have done a lot better too.
- [15:47]
A
Although the one thing which, of course, the most important use of statistics, it's important to know what's going on. But the most important use of statistics is the vaccine trials, where you have really rigorous double blind trials. And we had a large number of vaccines from Sanofi and Merck and places like that, which just weren't. They didn't pass muster, according to these trials. And everyone said, yep, that one isn't working, we're going to discard that and go with the ones which are working. And that worked out astonishingly well.
- [16:24]
C
Yeah. And something else that worked out well along the same lines that gets less attention is something called the recovery trial, which was organized in the UK and that was a very rapidly arranged trial of a whole bunch of candidate treatments for Covid, so hydroxychloroquine, the one beloved of Trump, but others, and they found that, for example, dexamethasone, which is a very cheap steroid, is really good, it's very effective. And they found a load of others and they realized that a load of people would be doing all of these silly little ad hoc trials all over the world and not really finding good evidence either way because the trials weren't big enough. And so they just got the whole thing sorted and this trial infrastructure for coordinating what people were doing and in just a few weeks produced results as to what effective treatments were and were not available.
- [17:19]
B
I wonder too, when talking about the COVID statistics, part of the problem is just public perception of this data. Like, we talked about the models being right, but I think people at the time interpreted those models as predictions and not thinking about like, well, if we take X action, then millions of people won't actually die. And looking at how people talk about the vaccines too, I'm confused because it's like, well, one vaccine is 95% good and the other vaccine is 79% good. So which vaccine are you going, oh.
- [17:50]
A
This one drives me up the wall. The percentage effectiveness statistic is. I mean, if we just never had that statistic, we would be in a much better place than we are right now. I hate that statistic more than words can say. And it has really created this genuine feeling in the minds of millions of Americans that, like, I don't want The Johnson and Johnson vaccine, because the effectiveness number is lower and, you know, and. And I think it's truly damaging, but it's. I have a real problem with the way that one has been communicated.
- [18:29]
B
But can one of you then explain why. Why is. Why should we ignore those numbers? Like, I've heard the advice, just get whatever vaccine is available. The Johnson and Johnson vaccine is really good. Da, da, da. But no one said to me, what is 79% or whatever it is versus 95%. What is the difference? And why is that a terrible thing to know and judge the vaccines on?
- [18:52]
C
So, in a nutshell, the trials have been focused on severe illness. You can very quickly say, look, did this vaccine prevent severe illness? Did it prevent any infection that created symptoms? But there's all kinds of other things that you might have wanted to focus on. For example, did this vaccine prevent death? It would be really nice to know. But fortunately, not enough people die. Even in the trials, which are tens of thousands of people, not enough people die in either arm of the trial to let you compare. Because, I mean, if nobody died in the vaccine arm of the trial, but only one person died in the placebo arm of the trial, I'm making these numbers up. We don't really know whether that was just chance or whether the vaccine helped. So it seems as though all of these vaccines are highly effective at preventing hospitalizations and deaths, but the trials themselves were focused on something more common, which is, did you get fluid symptoms? Did you get shorter breath? Actually, we're not really worried about a whole bunch of people getting fluid symptoms. We were worried about people dying or going to hospitality, and that's not what the trials measured. And there are other things you might want to know, like does the vaccine prevent you from transmitting the disease? And it looks like these vaccines probably do. We speculate that they do, but we haven't really got great data on that yet. So one of the challenges we've got is the clinical trials. As Felix says, they're incredibly important. We're in a hurry, so they're not going to measure everything we want to measure. They'll measure whatever's convenient to demonstrate that the vaccine is useful and then just push. Push it out and get into people's arms.
- [20:27]
B
Yeah.
- [20:28]
A
The idea is that if. If you are exposed to Covid and you've been vaccinated with absolutely any one of these vaccines, then there is an unbelievably high chance, very close to 100%. Even if you do wind up with fluid symptoms, as Tim says, it's that's about as bad as it's going to get. And that's the thing that really matters. Now obviously it's intuitively true that if you wind up with slowy symptoms, then you're more likely to be infectious than if you don't wind up with fluy symptoms. But that's where we start getting into the realm of supposition and speculation. So we don't want to spend too much time there. But the idea that the vaccine, if it's 75% effective, people have this idea that 25% of the time that means it does no good at all. And that's just totally, totally not true. Basically, 100% of the time it protects you from getting anything worse than those slowy symptoms. And then 25% of the time you might get those fluid symptoms. That's a fantastic vaccine. And the other thing about something like Johnson and Johnson is, you know, we are human. We are not always great at doing things like turning up for our second shot if there's, if we have a one shot vaccine, people are more likely to get fully vaccinated once you add in the probability that with a two shot vaccine they might not get the second shot. There's a bunch of moving variables here and the big takeaway really is that everyone should try and get vaccinated as soon as they can.
- [22:00]
B
Well then, yeah, this then goes back to your point about not being able to trust the media no matter where it's coming from lately because everyone reported the 95% number and everyone reported the, the 75% number and no one explained what it meant all that well and perpetuated.
- [22:21]
A
It was really terrible, terrible, terrible job of explaining it.
- [22:24]
B
And I think a lot of the messaging around the vaccine and the data has been pretty, pretty bad. It's kind of like this downer message, like, oh, we don't know.
- [22:34]
A
It's almost as though journalists are bad at communicating the meaning of statistics.
- [22:40]
C
Yes, it's tricky. I mean, it's not just the journalists. I think public health officials and epidemiologists as well, because I mean, Zeynep Tufechi, I probably mispronounced her name. She's writing amazing stuff about this in the Atlantic at the moment. Her point is that in an effort to be cautious and not to oversell the vaccine and to make promises we can't keep, we don't know, for example, does it prevent you being infectious? We don't know if it's absolutely, if there are no side effects. There are all kinds of stuff we don't really know. And because we don't know absolutely for sure, the messaging from public health officials as well as from journalists has been very cautious. Whereas in fact, I think the big picture is, well, the vaccines are awesome, they work really well, they're super safe, there are loads of them, Everybody get one right now. Maybe we should be sort of adopting that a bit more.
- [23:33]
A
And the criterion shouldn't be, do you know that it works, it's criterion for wearing masks. It's not, you know, do we have a rock solid double blind placebo, you know, trial proving that masks are effective? No, it's like we understand how coronaviruses work and it's pretty obvious that masks are going to help. Like, similarly for whether or not you're infectious once you've been vaccinated. Like, in general, we understand how these viruses work. They get into your lungs, they multiply, they get worse, and then once they do that, you become infectious. If the virus prevents that from happening, then you're not going to be infectious. Like, this is basic common sense and you don't necessarily need some incredibly rock solid trial to understand intuitively that if you're vaccinated, you're much less likely to infect someone else.
- [24:22]
B
It's so weird to think about because on so many other things, people are so gullible to what are obviously ridiculous statistics and messaging like CBT will help you relax. Or like, Tim tells a story of smoking and how people were so resistant to. I mean, I guess it goes back to emotions, but people were so resistant to believe smoking was bad for you when, like, come on, like they're all waking up with their smoker's coughs. They didn't figure that out before. I mean, but then on the flip side, it's like, we need to make sure the vaccines are. We need 100% on those vaccines, otherwise we totally don't believe it. It's so obvious that there's some other thing at play. Emotions and not science.
- [25:02]
C
I mean, emotions really matter. You're absolutely right, Emily. And I guess the other point that I make in the data detective is that it's important to just understand what is being said. So I have a chapter on what I call premature enumeration, which is when you start slicing and dicing the numbers, comparing numbers and doing all this fancy stuff and when you actually aren't sure what the numbers mean. And we had this problem back in the financial crisis with people measuring risk, but the numbers that they were doing, very sophisticated analysis of that were tangentially connected to risk. They didn't really know what risk really meant. And in the case of the vaccine, as you rightly say, we've got these numbers that measure effectiveness or efficacy, but efficacy in what way? Lots of different ways to measure it. And if we don't ask ourselves what is actually being measured or what is actually being counted, even if we get the numbers right, we can misunderstand what's happening.
- [25:59]
A
Okay, so let's move on to the absolutely massive $1.9 trillion news of the week, which I feel has been kind of weirdly undersold. Like, you come out of the Trump crazy and everything feels normal in relation to Trump. But this is just an astonishingly huge stimulus bill coming on top of a series of astonishingly huge stimulus bills. And here's a statistic for you, Tim, that I did some back of the envelope math. It's just division. It's very easy.
- [26:35]
C
My favorite kind of math.
- [26:37]
A
But if you take the five, if you add up all the stimulus bills, it comes to $5.3 trillion over the past year. If you divide that by the number of households in America, it comes to $43,000 per household, which is like, people are like, I got a $2,000 check. It's like, yay, that's awesome. But do you realize how big this stimulus is? The stimulus in total comes to $43,000 per household. Now, it hasn't all been spent yet, but that's an insane amount of money.
- [27:11]
C
Yeah.
- [27:12]
B
It's such great news. I'm so excited about it.
- [27:16]
C
Thank you, Felix, by the way, for following the data detective advice and making it into a comparison that actually makes sense, like per household. And you can get your head around just how big the number is. But, Emily, tell us about why you think it's such a great plan.
- [27:28]
B
Because it takes care of the people and the sectors who really need help. It's a rescue package that doesn't rely on the old paradigm of, like, trickle down. Basically, this package delivers money to the people who need it. It contains. The thing I'm most excited about probably is this thing called the child tax credit, which they've expanded so that the parents who need it most are now going to get for a year, monthly payment, depending on how many children you have. Could be it's $300 per kid, which is just wild. The US was always standalone country when it came to supporting parents. And the tax credit as it used to exist mostly benefited people who made a lot of money already. It was people who paid taxes, people who paid taxes. So the people who needed it most weren't able to get it. So this package with this child care tax credit, it expands healthcare coverage to middle class people. It gives them more support to buy into Obamacare. It expands the earned income tax credit, which helps very poor people. It makes it easier to tap that tax credit now if you don't have children. And of course it has the checks and it really helps the people like the middle class and the low income earners. It cuts child poverty in half. It's just, I think it's revolutionary in how redistributive it is and how it's looking at the economy and saying we can really help the economy by helping the people who need the most money and who will spend it, as opposed to what we saw in the Trump years, which was like, we can help the economy by giving rich people tax cuts, which we all know is dumb.
- [29:05]
C
Right.
- [29:06]
A
So there's two things I want to say about that. The first thing which just to make this clear, is that this is what is known as a refundable tax credit, the child tax credit, which means that you get the money even if you didn't pay any taxes. So that, that's the really big thing about the child tax credit is it becomes very much like a basic income that you can pay no income tax or no tax at all because you just don't earn enough. And the government will give you money. It's called a tax credit, but it's not really a tax credit. It's just the government giving you money. And then the other thing they're doing is while rich people, if you give them, you know, $3,000 to help support their childcare costs for a year, they'll be like, that's awesome. I get $3,000 to the year. Poor people, you know, living paycheck to paycheck, they actually want a monthly income. This is not something that is best done once a year when you file your taxes. This is something that is best done by giving people money every month. And there is a lot of work going into making this child tax credit a monthly check rather than something that you need to file taxes to get. And that is also just fantastic. The one minor quibble I will have with what you said is when you said that this is redistributive, like it does give to the poor. There's no doubt that what the stimulus is doing is giving to the poor. I'm just not sure that it takes from the rich. There's no pay for here.
- [30:34]
B
Yeah, you know what you're redistributing from the rich. Yeah, no, that's true. The focus is just giving to the poor and the middle class. Like. And the other part that's revolutionary, which you mentioned in your newsletter, is that it's money from nowhere.
- [30:47]
A
Right?
- [30:48]
B
I mean, it's just money machine goes, brr. Like, no one's really whining or saying anything about that because we've been doing it for a year now, just printing the money, giving it out.
- [31:02]
C
So I have a question for the two of you. So, Felix, you describe this as a stimulus, and Emily, you described it as a rescue package. Now, of course, it can be both, but I find the distinction interesting. And I think it's not just about political framing. I mean, which is the more appropriate phrase to. To use to describe it, do you think?
- [31:26]
A
Well, Tim, let me ask you as the economist first. What's the difference?
- [31:30]
C
Okay, that's a good question. So a stimulus, classic stimulus, is there is something called an output gap in the economy. People, for whatever reason, generally fear are not going out and spending because they're afraid for their future jobs. They're afraid they don't have the money. So people aren't spending. There's productive capacity in the economy, and the gap between what people are willing to spend, between demand and what the economy could supply, is the output gap. And the government just dumps a load of cash in there, either gives people money or it purchases goods directly. And that is the stimulus. And it gets demand back up to where output could be and the recession's over. That's your classic stimulus rescue package. Basically says there's people out there who are in desperate need. They've lost their jobs because they worked in restaurants, for example, which have closed or maybe much more indirectly have been affected by the pandemic. They need help, and we're going to send them the help. Now, these two things are conceptually completely distinct. So, for example, you could have a rescue package that wasn't in any way a stimulus. You just tax a bunch of rich people, or you cut spending elsewhere in the economy and you direct it at the working parents, the people who lost their jobs. And you can also have a stimulus that isn't really a rescue package. You just say, well, we'll just give everybody some money and maybe a little bit.
- [33:03]
A
Or you just say, let's increase defense spending. That would be the stimulus. Without any rescue.
- [33:09]
C
Yeah, pure stimulus. No rescue. Now, it seems to me that there's quite a lot of rescue in this. There's also a fair amount of stimulus. And it's fair to ask whether there's. Even if the rescue package is appropriate, is there too much stimulus? Should they have raised taxes on the rich right now because they're going to overdo it? I don't know the answer, but I think it's an interesting question to ask.
- [33:30]
B
I think politically, I mean, it's not just a pure economic play here. Obviously, there's political considerations. And the thinking, even if it made sense to raise taxes on the rich to offset some kind of fear of overstimulation, that it would have been much harder to message the package and get it through the way they did. That's my sense.
- [33:52]
A
I think the other part of this is that economists love to talk about this concept of too much stimulus without necessarily being entirely clear about what they mean. And we saw this when Larry Summers came out and said that $1.9 trillion was too much. He kept on using the word overheating or overheated. And you would look at him and go like, what are you talking about? Is the economy suddenly become an oven? Can you be a bit more specific? And eventually, if you asked him like five or six times, he would come out and say, well, something, something, something, something, inflation. That really, ultimately, when people worry about this concept of too much stimulus, the thing they're worried about is consumer prices will rise so fast as to be in some way damaging. And once you articulate that, that the fear is that the stimulus is so big that it's going to cause not just inflation, but too much inflation, then I think it's easier to sort of step back and say, well, we've had too little inflation for over a decade. The Federal Reserve has been desperately trying to create inflation for years and years without any success. The idea that the federal government can come in with a single bill and solve the inflation problem, that up until now has very clearly been that there isn't enough of it. It's hard for me to see as a really big problem and not. And a reason not to do this, as opposed to like a side benefit and a feature, not a bug.
- [35:38]
C
Yeah, I can think of one thing that has inflated from approximately zero to approximately $70 million. But I guess we'll come to that after the break.
- [35:47]
A
Tim, I have to say, you are a podcaster extraordinaire. We need to bring you back every week just for your segues. Okay, Tim, what or who is a Beeple?
- [36:00]
C
Beeple is an artist. Felix, surely you know that. And I think his real name is Mike Winkelman, and he's just some regular guy who was a pretty successful designer until Covid hit. And then he increasingly started focusing on his digital art, which he's been producing for a very, very long time. And a few months ago he started getting interested in NFTs, non fungible tokens, which has accidentally led to one of his pieces of digital art selling for 69 point something million dollars. So it's very exciting because it's this rags to riches thing, like he's just sort of doodling on the Internet and suddenly one of his pieces, which you can have for free, it costs $70 million. So this gets us into all kinds of interesting stuff. So I mean, I want Emily to explain to me what a non fungible token is first and then we can talk about the economics of this.
- [37:06]
A
As the former editor of IP Law & Business, I feel like you're the right person to explain.
- [37:10]
B
This. Felix, not one episode has gone by over the past six months where you have not mentioned my experience at IP Lawn business. But anyway, so.
- [37:18]
A
Non. Which is. It's.
- [37:19]
B
Important. No, it's very important. So non fungible token, as far as I understand it, is. It's a unique item. You can't replicate it. It's its own unique.
- [37:28]
C
Thing. So the token is the thing that's unique. I mean, the artwork can be replicated, but the token that says this is. I mean, it's kind of like, you know, my wife's a photographer, Fran Monks, she's a portrait photographer. She's very good. She's done all this zoo photography photographs over zoom during the pandemic. And the National Science Museum in London said, we want to acquire copies of your zoom photography for our permanent collection. And she said, okay, that's great. I mean, they're zoom photographs. So they're like, they're digital. And they're like, yeah, yeah, but we want prints, that's fine, okay, I can make you a print, no problem. And they said, yeah, we want you to tell us how many prints you're going to make. It's going to be a limited edition, but just tell us, is it a limited edition of. Well, they didn't even set a number. So she's like, right, okay, so I have to create scarcity to sell prints of my digital work to the National Science Museum. And I don't know, should it be a limited edition of five or a limited edition of 100? I mean, she's got no idea because this is new territory. So in a way, any photographer who can create limited edition prints and just sign and Say, well, this is the only one, or this is one of a hundred or this is one of ten thousand is in a similar situation. Just the NFT technology lets you do that with a purely digital good. And to say there's an infinite number of copies of this thing available, but there's only one copy stamped with the NFT that says, this is the.
- [39:00]
A
One, this is the original, except for it's a little bit more complicated than.
- [39:04]
C
That. Say it ain't.
- [39:05]
A
So. There is no one copy that is stamped with the nft. The copy of this artwork that I downloaded this morning onto my computer, it's about 0.3 gigabytes. It's big, but it's not that big. And it's fun to zoom into. And I can recommend that you do it. That copy is absolutely identical to the copy that the owner of the work, the person who bought the NFT at Christie's, has. There is literally no difference between them. It's not like one of them has a digital stamp and the other one doesn't. It's not like one of them has a certificate of authenticity in the form of an NFT and the other one doesn't. They're identical. And in no way is it true, and this is a misconception that a lot of people have. In no way is it true that if I exhibited this work on a screen in my home that it would be inauthentic or fake or anything like that. It is the genuine, authentic digital file that exists in multiple places on the Internet and is just as real as the authentic digital file as the one that the owner of the NFT owns. What the.
- [40:13]
B
Difference? I thought. I thought the digital thing that you buy, the nft, it was different because it had some kind of tokeny data y encrypty thing in it, like how you have a picture and has the copyright information embedded in it or.
- [40:29]
A
Something. So this is the crazy thing, right? So one of the. I put this thing up on Twitter on Friday morning that the more you understand what's going on here, the more difficult it becomes to answer the very simple question of did Christie sell an artwork on Thursday? Because they didn't. On some level they didn't. On some level they did. As I say, it's a very difficult question, but the thing that was bought was not the artwork. If you look at the non fungible token, there's no art embedded in that artwork. There is no JPEG there. It just points to a URL@makerspace.com where you can download or anyone can download this artwork. The thing that was bought in the non fungible token is the ownership of the artwork. It's the title to the artwork. But the artwork itself is distributed and lives on the Internet and can be exhibited by.
- [41:25]
C
Anybody. It's so much fun this, it's so interesting. So I've been trying to think of parallels. So we've already talked about the. The idea of a photographer who can of course make multiple prints of her work, but can create artificial scarcity and say, well, yeah, but this is a limited edition print or this is the only.
- [41:44]
A
Original. Well, artificial scarcity is still real scarcity. If your wife only makes five prints of that work, there are only five prints of that work in the world. So it might like. It's artificial in the sense that if she wanted to, she could make six, but it's also real in the fact in the sense that there are only five. And whereas in this case the, the scarcity is slightly different because the number of versions of this, number of copies of this artwork is potentially infinite. And no one is stopping anyone from downloading it. And of course, whenever you visit a webpage or download a file on the Internet, you're copying it. No one like the copying the file is, is lossless. There's no degradation of quality when you, when you copy it or download it. And it increases the number of versions of this file that exist in the world. But that is not the scarcity that we're worried.
- [42:37]
B
About. So you're just selling a feeling, which is really what you're.
- [42:42]
A
Selling. Title. You're selling.
- [42:44]
C
Ownership.
- [42:44]
B
Yeah. So I mean, it's just, in a way, it's not that different from. From anything that you're selling a title, an emotion. You can. So the guy who buys it, the lady can just say, I paid $69 million for this thing.
- [43:00]
C
Right. So here's another parallel action comics number one. It's published in 1938. It's got this classic photograph of, well, not photograph, this classic print of Superman raising a 1930s auto above his head and smashing it down. And the bad guys are stunned by this Superman. It's the first comic in which Superman appeared. You can buy a replica. You can buy a nice hardback book on Amazon with a whole bunch of the original comic strips in. Or you can buy the original and mint condition original will cost you $3 million, which is, I mean it's no artwork by Beeple, but it's a chunk of change. So I mean, you can say there is scarcity. There are only so many of the originals out there, but the art itself can be reproduced. So is that similar or is that very.
- [44:01]
A
Different? I think that's exactly the same as your wife's.
- [44:05]
C
Photo. I'll tell her she should be charging.
- [44:07]
A
More. Ever since, you know, Walter Benjamin wrote about the work of art in the age of mechanical reproduction, in what, like the 19 teens, 20s, whenever that was, we've understood that there's the, the, you know, the idea of additioning works and, and real world scarcity. Like the reason why Action Comics number one is valuable is because it is scarce and there aren't that many of it. And the thing that you need to know about the art world that like it has astonished me how many people I've been talking to over the past couple of weeks who don't understand this about the art world. If you buy an artwork, any artwork, it could be a photo or an oil painting or anything you like, a digital video. If you buy that artwork at auction or from a gallery or directly from the artist, 99.9% of the time, you do not buy the copyright. You do not own the copyright. You have no control. You aren't allowed to, to license it, to appear in a book. You aren't allowed to make postcards of it and sell it or anything like that. You don't own the image. The copyright nearly always remains with the artist. And a lot of people just assume that if you own the artwork, you can do what you like with it, you can reproduce it. That's not true. The thing that you're buying, the thing that all art buyers have always bought for well over a century, is the object, not the image. And so in that sense, there's nothing new here. Like the image is out there floating in the ether and anyone can download it. And the object that was bought was this token, this nft, and not the artwork. So I think that's the easy and the, the not new thing about what happened here. The new thing that happened here is that the image is weirdly kind of owned by a person. Now when I download that image to my computer, which I did, the image that I am looking at on my computer is owned by some person who, I don't even know who it is, the anonymous person who won the Christie's auction. And that I think is new and.
- [46:27]
C
Different. So can I ask a question about where this goes next? Because in a way, the fact that an artist sells a work or whatever was sold, sells a thing for $70 million is not that remarkable. I think the fact that clips of Basketball action are selling for thousands of dollars with these. Well, again, the clips are not selling. It's the NFT that's selling. So people are trading them like they were trading baseball cards. I think that's interesting. And what I find interesting is that you've got the NBA producing clips. I mean, there's almost not an infinite supply, but there's a very large supply of clips of basketball players doing cool things. And so then the question is, how many of these do you actually want to issue and certify? And of course other sports organizations will want to get in on the action, and everyone's going to want to get in on the action. And the economist who I think gave us some insight into the difficulties here was Ronald Coase, Nobel Memorial prize winning economist. In 1972, he published something called the Durable Goods Conjecture. And here's how it goes. So you're a monopolist. You produce, I don't know, a photocopier or an iPhone or whatever. You produce some cool thing that no one else can produce. You can charge $1,000 for it to a bunch of people who are early adopters. So you sell it for $1,000. And then having done that, you think, okay, well, now let's cut the price and we'll sell it for $900 to the next bunch of people, and then we'll produce some more, and then we'll sell the next bunch for $800 to the next group of people. And the problem is that all of the people who are willing to pay $1,000 go to themselves. Hang on a minute. This monopolist is going to keep churning this stuff out. And it's not just a theoretical concern. Steve Jobs had this problem with the early iPhones. They were sold for $600, then they cut the price, and everyone who paid $600 wanted a refund. And in fact they got a refund because they felt they'd been screwed by Apple. So given that there's almost unlimited capacity to manufacture more of these NFTs, you just attach them to anything you like. At what point does the market get swamped? And at what point do people get dissuaded from spending money on them because they anticipate the market will be swamped, or maybe it'll never happen. Maybe the people issuing them will get smart. I don't.
- [48:52]
A
Know. So the most interesting analysis of this that I've seen came from Chris Dixon, who is a partner at Andreessen Horowitz, and he was talking about what he calls filling the area under the demand curve, which is basically what you're talking about here, selling as many of these things at $1,000 as you can, and then selling as many more at $900 as you can, and selling as many more at $800 as you can. And the way that NFTs solve this problem, which is kind of clever, is that you can change the edition size. So you sell one, you know, unique NFT for $6 million, and then you create an addition of 10 for the next one. It might look the same, but it's like. Or slightly different, but it's an addition of 10. So then that one you can only sell for maybe $100,000, and then you create an addition of a thousand for the next one, and that one you sell for, you know, $500. And then basically, wherever you are in terms of how much money you want to spend, there is an NFT for you. And because the NFTs are different, because they come in different edition sizes, they are distinguishable. Now, the question you're asking is like, what.
- [50:08]
C
If? And flood the market because they could get.
- [50:11]
A
Greedy. Well, yeah, the question you're asking is like a meta question. It's like, I can buy a super valuable LeBron James dunk, which comes in an edition of three, and it's super valuable because it's an edition of three. But what happens if Dapper Labs, who runs NBA Top Shots, ends up producing 20,000 editions of 3? Then at that point, the rarity of my LeBron James junk, while that particular dunk is just as scarce as it always was. The idea that it's special, that this thing is only an addition of three, it becomes much less special because anyone can buy something with that degree of scarcity. And so, yeah, that I think, is the interesting question, which we have yet to really grapple with, which is, well, I mean, I guess actually we've seen it to some extent, right? Is that with cryptokitties and cryptopunks, right, that they came out at roughly the same time. They both were sort of collectibles. They both had a speculative frenzy surrounding them. And cyberpunks were capped at 10,000. And we just saw two of them this week selling for $8 million or seven and a half million dollars, depending on the price of ether, 4200 ether, whereas cryptokitties are basically worthless these days. And what is it that made the cyberpunks keep their value and keep on going up in value while the cryptokitties went down in value was just that the cyberpunks there have been no more created since 2017 or whenever they first got minted. Whereas the cryptokitties, being kitties, they like, have sex with each other and mate and create new kitties, and then there are eventually so many that they're not worth money.
- [51:52]
C
Anymore. Which, by the way, was the problem with tulips.
- [51:54]
B
Right? How is this not all tulips? How is this not Beanie Babies? It's like things have the value that people think they have for a while and then they don't anymore. Like, how is this not just like an overheating like Larry warned us all about taking place in the digital art world? And it's obviously there's a craze going on and it feels like it can't possibly be sustainable because, I mean, it's all based on people's perceptions of value more than anything.
- [52:23]
A
Else. Well, isn't that the art world.
- [52:26]
B
That has been around for hundreds of years? Yes, it is, but this seems like an overheating of a sector of the art.
- [52:32]
A
World.
- [52:33]
C
Right? I think. I mean, LeBron James is obviously an artist, but I wonder whether clips of LeBron James dunks really qualify as art. I need to do. Loyal listeners will know I have a podcast called Cautionary Tales. I think I need to do a Cautionary Tales about the Tulip mania because it's way more interesting and way more weird than a lot of people give it credit for. So I'll do that.
- [52:57]
B
Sometime. What about Beanie Babies? Is that interesting? I don't.
- [53:00]
A
Know. All I know is they went.
- [53:02]
B
Down. Yeah, that's.
- [53:02]
C
It.
- [53:03]
B
Right? That's.
- [53:03]
A
All. A lot of these things, they go up and they go down. And I think going back to our earlier discussion about the stimulus and overheating, there is definitely a part of this NFT craze that is a function of there just being an enormous amount of money sloshing around the economy and looking for a place to go. And I kind of think there's a silver lining here as well, which is that if what that money goes into turns out to be basically asset price inflation and people's bank account and non fungible tokens, that's good, right, Tim? In terms of inflation, like it doesn't show up in consumer prices, People can still buy a gallon of milk for the same price that they've always been able to buy a gallon of milk for. And so long as that's the worst thing that happens in terms of inflation, that gives us less to worry about in terms of.
- [53:59]
C
Overheating. Yeah, I Think the dot com bubble showed us that you can have a big old bubble in the price of certain assets and it doesn't necessarily wreck the economy or ruin people's lives. But Yeah, I think 2008 credit bubble showed us that you should take that for granted so we might get away with it for.
- [54:16]
A
Sure. Okay, I think it's time for a numbers round. Emily, you said you had a dumb number. Oh, yeah. Do you have a dumb.
- [54:23]
B
Number? It's dumb. My number is 75%. That is or was the increase in consumer interest in psychics from March 2020 to March 2021 on Yelp. People were nervous this year and they were looking to understand the future. And they didn't look to statistics? No, no. They looked for information on psychics. Yelp has a lot of data like this that they just posted this week where they look at various trends. Interest in takeout was up, interest in movie drive ins was up, and so.
- [54:56]
A
On. Tim, do you trust that.
- [54:57]
C
Statistic? I'm sure it's correct. I think it's correct because it fits with my priors, and my priors are shaped by Maria Konnikova's wonderful book the Confidence Game, which is all about con artists and why people fall for cons. And there's a story in there about somebody being absolutely. It's a really tragic story, completely ripped off by a psychic. My fingers are doing the air quotes thing. And it's partly that people feel desperate and they turn to, I think, implausible sources of help when they're emotionally vulnerable. And this seems to be an example of.
- [55:43]
A
That. My number is 2,400, which is the value in dollars of baby eels per pound. One pound of baby eels, little glassy things will fetch you $2,400. They are incredibly valuable because no one has worked out how to breed eels in capt, and so you need to catch them in the wild. And they are so valuable that there has been an enormous, basically crime network built up around the world trading in baby eels. But one of the interesting things I learned about eels this week, and I wasn't just learning about non fungible tokens, I was also learning about eels is the eels have been used as currency for a very long time that people in the UK used to pay taxes in eels in the 12th.
- [56:38]
C
Century. I did not know that. Yes, that's a wonderful.
- [56:43]
B
Fact. I'm rendered speechless by this number.
- [56:45]
C
Felix. So. My number this week is 30 million, which is what the New York Times has reported is the number of doses of the astrazeneca vaccine currently sitting in a warehouse on the east coast of the United States. It cannot be used in the United States because the FDA has not approved the AstraZeneca vaccine. I have to say, we're using it here in the uk. It seems to be working very well. Deaths are down 90% since mid January, and that is partly locked down, but partly also the vaccine. So these 30 million doses can't be used in the United States because they've not been approved. The European Union, which has only given out 45 million doses in total so far, which is not a lot for such a big place, they've asked if they can have some doses. I think Canada could really use the doses, Brazil could really use the doses, but they're sitting in a warehouse in the United States, and I don't understand why, at this point we can't think of something better to do with 30 million doses of a perfectly serviceable.
- [57:58]
A
Vaccine. Is this vaccine nationalism as it's.
- [58:02]
C
Known? Well, vaccine nationalism would be refusing to export doses because you want to use them yourself. Refusing to export doses that you don't want to use yourself, I think is just odd. And I don't understand. I mean, presumably the US wants to keep the doses because they think that at some stage they will maybe be approved and just in case. But I mean, these things do have a shelf life. But more importantly, people are dying in large numbers in Brazil now, and things are not great in the eu. They're desperately needed. There are lots of places where these doses would save a lot of lives quite quickly. And I think if the US isn't using them, then there must be a better place for them to.
- [58:49]
A
Go. On which note, I think that's it for us this week. Thank you so much, Tim Harford, for coming on the show to have a Slate plus segment with you about Florence Nightingale and Helena Bonham Carter, which is going to be awesome. Thank you to Jessamine and Molly for producing at Seaplane Armada in Brooklyn, New York, where I am. And this is why I sound good this week, because I am in the studio and it feels.
- [59:19]
C
Great. You sound good every week.
- [59:21]
A
Felix. Thank you very much, Jim. Thank you all for listening and for keeping the emails coming on slatemoneylate.com and we will be back next week on Sleep Money. But first, on Tuesday, we will be talking to Paul Ford about the Social.