The Winner's Curse

A discussion with Nobel Laureate Richard Thaler

Richard Thaler is one of the founders of behavioural economics, collaborating with Daniel Kahneman, Amos Tversky and others in defining the field. In 2017 he was awarded the Nobel Prize for Economics for his contributions. He has won numerous other awards and honours and is a past president of the American Economic Association.

Thaler has written several bestselling books - most famously “Nudge” which he co-authored with Cass Sunstein and led to the creation of “nudge units” to inform policy in multiple countries, including the UK. The aim of nudging is to use “choice architecture” to change the way people make decisions, without mandating or forcing them to do anything. He has a new book out - a fully updated version of his first book “The Winner’s Curse” - which looks at behavioural anomolies that can’t be explained by traditional economics. This new edition is co-authored with Alex Imas.

In this interview we discussed the impact of “nudging” on policy, frustrations with the current political environment, and how Thaler’s work has, and has not, changed economics.

---

Sam: I want to start by asking about your experience with the UK government. There was already interest in behavioural economics but when you published “Nudge” it really took off. What were the successes and frustrations of that experience?

Richard: Certainly the impact of “Nudge” could not have been anticipated. No trade publisher wanted the book. It was published with a University Press. Amusingly, the publisher of “The Winner’s Curse” didn’t bid on “Nudge”, and I asked him why. He said, well, I read the section in “The Winner’s Curse” on how publisher’s overbid.

In the UK, policy-makers were very interested. Rohan Silva [one of David Cameron’s SPADs] read “Nudge” and gave it to Cameron and Osborne, which led to the Behaviour Insights Team being set up. Cameron and Osborne were keen. And I had talked to them a fair amount, even before the election. So then they get elected and they want to create this Unit. David Halpern, was an ideal first director. He knew his way around government and was a true expert in behavioural science so the team had real potential.

As far as I was concerned, I always say that although I had I coined the term “choice architecture” I was just putting two words together. It didn’t mean I knew how to do choice archecture. So we had to figure it out as we went along.

Sam: How involved were you with the with the team at the start?

Richard: At the start, I was a lot involved. There were only five people on the team and there had not been an organization like this before so it was trial by fire. However, in addition to David Halperin we had the benefit of Gus O’Donnell serving as Cabinet Secretary. Gus has a PhD in Economics and he knew everyone in government and how everything worked. Over the years, Gus and David have become good friends. I would typically come just for a week at a time. For a while we were located in some offices they found in Admiralty Arch. There was a rare London snowstorm, perhaps a full inch on the ground that we would not even notice in Chicago, but traffic was at a stand-still and George Osborne had turned off the heat in an austerity measure so I put on every article of clothing I had with me.

We would talk to people - ministers or their deputies - and see whether there was some way we could help. The primary lesson I took away from this experience was to always start by asking them to describe the problems they were facing rather than present a list of tricks we had up our sleeve that we could try out.

Frankly, I think the biggest success of behavioural science during this period does not belong to the Nudge Unit, but to Adair Turner who designed the augmentation of the UK’s private pension system. The changes, which were influenced by our research but predated the creation of the Nudge Unit, said that if an employer is not offering a retirement plan, they must do so, and they must automatically enroll their workers. If the employer didn’t want to choose a provider, then there was a national stand-in [NEST – National Employment Savings Trust].

The implementation was very wise. They started just with large firms and gradually added smaller ones. And at the beginning, the initial savings rate was just 1 percent, what Robert Cialdini would call “a foot in the door”. At that low rate they got little resistance. And everybody said fine so then they went up to 3% 5% 8%. but they they’ve kept 90% enrollments. Of course one can quibble with various aspects of it, but the basic design is completely consistent with behavioural economics research.

Sam: What were the barriers the team found in getting departments to engage with the ideas?

Richard: Surprisingly, there was no problem finding teams that were keen to talk, I would say the biggest problem was and is our ability to fundamentally change systems.

Let’s distinguish between choice architecture and persuasion. The pension changes we talked about were pure choice architecture. They set up a structure. Firms must offer a plan and make joining the plan the default. Having done that there was little need to spend a lot of time and effort to convince people to join. You just set it up and make it easy: that is the biggest tool we have.

The problem is that nudge units rarely had the ability to change the way systems work. They can message and that’s it. That is not to say there were no successes from persuasion. One win from back in the early days was getting people who owed money on their taxes to pay more quickly through changing their wording of the letters. There was a Minister whose job it was to collect that money and he did so by sending letters to those who owed money. The team asked to see the letters, and we were able to come up with better letters. And that worked. That is an example of just persuasion using different language. It’s free. You’re going to send a letter. You might as well write a good one.

The effect size are not huge. It’s not like you get 90% enrollment in the pension. Maybe you increase the number of people who pay within the next month by 10% or 15%. Now that’s not nothing. In fact in brought in millions of pounds earlier than before. There were other innovations along those lines, sending people utility bills that tell them how they’re spending compared to their neighbors. Those are also free and they work, but the effect sizes are relatively modest, say, 3% or 4%. But I don’t want to belittle such changes. The benefit/cost ratio is gigantic!

Sam: But you were rarely able to implement choice architecture in the design of policies?

Richard: So here’s my big disappointment. There are very few examples, like with pensions, where we got to design the entire system. I’ll give you a counter example of where I’d have loved to be more involved: Obamacare. The economists were confident that it had to be mandatory. They were worried that if it wasn’t mandatory healthy people wouldn’t join and costs would spiral out of control.

So there was a mandate. The mandate became a big political fight. A big part of the hatred of Obamacare at the time was the mandate, because it made it compulsory. If it had been up to me I wouldn’t have made it a mandate. I would have just automatically enrolled everyone – use the power of default. Then what happened? Well, the Supreme Court, a couple years later, said the mandate was illegal. They got rid of it. The predicted death spiral of healthy people dropping out did not occur because most people like having health insurance!

Then there were little design issues. For some reason they grouped plans into categories and decided to label the categories by metals: platinum, gold, silver, bronze, and I remember asking at the time: this is choice architecture, right? What do we think we get from these labels? I was visiting a friend in the White House one day and I asked to talk to the people designing the interface, and they show me screen shots of the signup page with all these metals. And then I see there’s another category of high deductible plans that are called “catastrophic”.

I said, “wait, no you’re not having a category called that are you?” You know, laundry detergents have brands like Tide and Era and so forth but no one would think of naming a laundry detergent: Catastrophic! So the guy says: “well, that’s what economists call high deductible insurance plans”. I said “fine but it’s not a good brand name”. A couple colleagues and I ran a quick survey that showed that changing the name of the catastrophic category to “economy” or “value” would have reduced the number of uninsured by a significant amount. And now there’s been lots of research since, some of which we mentioned in the book, that high deductible plans are often the best bet, and in many cases, a high deductible plan dominates other plans, meaning you spend less no matter how much health care you consume. So they were giving what might have been the best plan for some people the worst brand name.

Sam: On the subject of system changes and autoenrollment I was interested to read that you didn’t support the idea of going to opt-out for organ donation, which was considered under Cameron and then, has more recently been introduced [under Boris Johnson]. It’s often cited as a positive example of the use of behavioural science in policy-making.

Richard: So I’m glad you brought that up. Defaults work. If you make something the default it will work, for sure. I don’t know of an exception. You’d have to be defaulting people into something they hate for them not to take it. We did a stunt once at a conference at the White House where we sent people a letter describing the agenda for this conference, and it included something about a lunch on the first day, and the default option is some dreadful sounding vegan sandwich, but no one opted out. I was asked to give the the first talk and I mentioned this fact, I could see people frantically wondering where the nearest sandwich shop was.

So on organ donations there’s a paper by two friends of mine that shows that in countries that have opt-outs on organ donation, hardly anyone opts out. Again, defaults do “work”. The problem is the donors are not the decision makers. The donors are deceased. The decision makers are the donor’s family. And there is no country in the world that has the policy “routine removal” in which useful organs are simply taken for transpant use. That might be a policy that would save a lot of lives, but it has understandably not been adopted. But if the policy is just presumed consent, family members are still consulted and they can only be told “well, your loved one did not opt out” which is not much of a signal as to the donor’s actual wishes.

So you get resistance. I much prefer the system we have in the US, where people are asked when they get their driver’s license if they want to be a donor, and you’re asked every time you renew. There is a nudge there. It’s a different type of nudge. But that’s combined with “first person consent”, a law which states that if the donor has expressed a preference, those wishes should be followed. So family members are told “your loved one wanted to be a donor, and by law, we’re required to honor that request”. The family can put up a fuss, but it’s rare. They’re less likely to if they know that was a real decision. It also gives you a second chance, because, if the donor hadn’t signed up, well, the donor didn’t actively say no. So now you can go to the family and say “your loved one didn’t express a preference but those organs can save several lives”.

I would not call this out as Boris Johnson’s worst mistake, but that is a high bar.

Sam: We’ve been talking about the Cameron era and the Obama era. It feels like politics has changed a lot since then. Now we’ve got Trump and populists and those people don’t have much interest in scientific approaches to policy. Do you feel like we’ve moved to a more existential place in politics and things like choice architecture are less relevant?

Richard: Well, the problem is quite general. Certainly in the United States the Trump administration has completely abandoned any pretense of policy analysis. They may have killed “woke” but they also killed “wonk”! I don’t know a single economist who thinks that it is a good idea to have tariffs which jump up and down on a whim or fancy gift.

In the US people complain rightfully about our health care system, that it costs twice as much as elsewhere but with only mediocre results. Neither political party has anything like a plan to fix that.

Then there’s lots of discussion of the high cost of building housing , and high speed rail construction is a disaster in both the US and UK. This is what in the final edition of “Nudge” we call “sludge”, the details of bureaucracy that slow things down and make it costly to do anything. It’s clear lots of well-meaning regulations are imposing a lot of costs. The book “Abundance” rightly raises many of these issues but stops short of saying which environmental and safety regulations we should abandon.

Sam: Yes that’s what I said in my review - here’s no theory of regulation in there. It’s just examples of where regulations create problems, which I think we could all find examples of.

Richard: The other book that I think has a similar problem is “Breakneck” by Dan Wang. The book has a simple message: China is run by engineers so they can build things. The US is run by lawyers so they can’t. But Wang also stops short of saying how American can build better or how China can improve its policies on human rights. It is certainly cheaper to build a gigantic dam if you can simply order millions of farmers who lived along the river to move into high rise apartments.

A good current example are the fires in LA and the many homes that must now be rebuilt which will take years. [Governor Gavin] Newsom has talked about cutting red tape, but you probably don’t want to cut all of it, right? This is a fire zone, and maybe we need new rules to make another fire less likely? Certainly there are too many regulations and required inspections, but how exactly can we do this more efficiently? I’m sure there are lots of ways one could speed up the permitting process. It is a very important problem, but one that is difficult to get people very excited about.

Everyone’s opposed to red tape. But unless you know what you actually want to regulate you don’t know what to take away.

Sam: I wanted to ask you about “sludge”. You’ve talked about examples of well-meaning regulation that accretes over time because people think it’s a good idea. But sludge can also be deliberately used to make people’s life worse. You use the example of newspapers that make you ring up to cancel a subscription. But in government, I’m seeing more and more examples of barriers being put in the way of claiming entitlements to save money. So sort of reverse nudging, where you say we’re going to make the forms longer and more painful, and you’re going to have to wait five weeks for this to happen, because they don’t want people to claim what they’re rightfully entitled to. Choice architecture in reverse.

Richard: Certainly there’s lots of intentional sludge in the private sector. At the very end of the Biden administration they passed a rule requiring “one click cancelation”, a policy that we proposed in the final edition of “Nudge”. Some friends of mine were instrumental in writing that regulation. I don’t know whether Trump has noticed it because I haven’t heard that it’s been revoked, but who would oppose such a policy?

And yes there are examples in the public sector too. There are new requirements to get SNAP benefits [food benefits in the US] and most of that burden is having to prove you’re working. So it’s exactly sludge, deliberately trying to make it harder to do something.

There are other negative examples of choice architecture in the private sector. Take the online investing apps like Robin Hood. It’s not an accident that that looks like a casino. It encourages behaviour that is the opposite of Warren Buffet’s buy and hold philosophy. Buying one day and selling two days later is not really anybody’s idea of sensible investing advice and is why the vast majority of amateur investors lose money.

When the first version of The Winner’s Curse came out, there was a chapter on the equity premium puzzle, which is about the surprisingly large difference in returns over time in stocks versus bonds. Around that time I was telling my psychology colleagues, put your money into diversified porfolios of stocks, and then only read the sports section! That was my investment advice. And there are many psychologists around the world that thank me when they see me.

Sam: One of the new sections in the book is about meme stocks like GameStop and the cinema chain, AMC, which is a result of this gamification of trading, or partly a result of this gamification.

Richaed: Well, two modern inventions contribute to meme stock mania: on-line trading apps with zero commissions and gamification plus social media. For GameStop there was a guy who used the name “Roaring Kitty”, on line trying to drum up support for GameStop. Social media gave him a platform. It’s as if you gave Mr. Ponzi a megaphone. We’ve built a world where bubbles are easier to create.

Sam: This is one of the frustrations with understanding irrationalities, is that even when you understand them and can see them, there isn’t necessarily any way you can make many money out of them, because you don’t know if people are going to continue to behave in the same way.

Richard: Yes. When I talk about the efficient market hypothesis I always say there are two components, “the price is right”, and “there’s no free lunch”. The “no free lunch”, you can’t beat the market, it’s approximately true. Most professional asset managers do not beat their benchmarks.

The other “The price is right” – the idea that it’s never possible for markets to set the wrong price - is incorrect. The chapter in the book on the “law of one price” is where we provide counter examples. There are some very clear examples where there’s no possible defense of irrational pricing. [One famous example quoted in the book is when 3Com spun off it’s then innovative “Palm Pilot” business. The new company’s share price rocketed up whereas 3Com’s fell even though it retained 95% ownership of the new business, thus valuing the rest of 3Com at a negative $22 billion].

When I hang out with professional investors at the moment they tend to be worried. They think the market’s more likely to be overvalued than under right now, but that was also true for years in the late 1990s, and yet prices kept going up.

Sam: I guess you’ve got both this question about AI as a bubble, but also Bitcoin and crypto, which is very large amount of money built on very little.

Richard: I’m a golfer, and back when Bitcoin started I would jokingly say, “all right, well, we’ll play for one bitcoin payable in a decade”. The implication being, we’re playing for nothing. I am glad no one kept track of those bets!

I asked my friend Gene Fama [an economist known for his defence of the efficient market hypothesis] what he was more embarrassed about: Bitcoin or GameStop. He said Bitcoin shouldn’t exist. Whereas he puts GameStop in the category of Thaler’s annoying examples. They’re small enough that he can just say I’m trying to annoy him. But Bitcoin is big and he thinks it really should not exist unless you’re criminal.

Sam: As we’ve moved on to economics, I wanted to ask about a really interesting reflection at the end of the book, which is that whilst obviously, behavioural economics has become a field in itself, with Nobel laureates and every university Department having somebody who does it, it hasn’t got into the textbooks. It hasn’t created this paradigm shift in economics that one might have expected it to do ten or 20 years ago. Why hasn’t that happened? And what would it take to happen?

Richard: Well economics as a field is maximizing selfish agents engaging in markets. That’s what distinguishes it from psychology, sociology, anthropology. If you write a textbook and you start to incorporate our stuff it makes it messy. Some of the textbooks are written by behavioural economists, and what they do is they’ll put in a box so there’s a chapter on the theory of consumer and then there’s a box about mental accounting [the tendency to treat money differently based on its source, rather than viewing it as all interchangeable - named by Thaler].

The way I would teach it is I would make this normative and descriptive distinction. I would say, look, if you’re a firm and you want to maximize profits, here’s the way you do it. Marginal cost equals marginal revenue is the solution to that problem. You know, life cycle saving, that’s the way a rational person would want to solve this problem. But then here’s what actually happens. I think you could write a book structured that way.

I have some friends who wrote one of these books, and their solution was to have a behavioural economics chapter at the end. And they told me it’s a very controversial chapter, because it’s the most popular chapter with the students, except they say – “hey wait this disproves everything else – that was a bait and switch”.

It’s not like we’re bitter about this. We’ve succeeded way beyond our wildest dreams. You know, I was at Oxford on Monday and gave a talk in the economics department. The chairman’s a behavioural economist who works on mental accounting, right? But when it comes to microeconomics students are still learning the same old stuff.

Sam: That’s even more true with macro-economics, right?

Richard: Yes, macro economics has been slow to incorporate behavioural features. I’ve been on a campaign recently to get macro economists to incorporate mental accounting.

Here’s a good example. There’s lots of discussion about the so called wealth effect. So, say, stock markets are up 10% is that going to keep us out of a recession? Because people are spending all that new wealth?

And my answer to that is, well, let’s put people into three categories. There’s the poor, they basically own no stocks. Then there’s the middle class which I’m going to include everyone up up to rich, and if they have stocks, it’s in their retirement account, and people don’t spend more now if their retirement account goes up. Then there’s the very rich and if the stock market goes up, well they’re already spending everything they want. How many weddings does Jeff Bezos need? Or yachts? So, I think just having a capital W for wealth in a model without disaggregating it into where it sits and how it got there means you’ll get a bad forecast.

A British friend who we’ll call Sarah, got a tax refund recently because the withholding had been done wrong or something. So £1,000. She gets a letter saying there’s a cheque on the way. She told me she spent it before the cheque arrived and then I asked her how much money she has in her retirement account - it is about £100,000 and it has gone up but she’s not sure by how much. She hasn’t looked, she certainly hasn’t spent it.

But that £1,000 tax refund isn’t new wealth, right? She always had it. But she spent it.

Sam: I think mental accounting of all the things that you’ve done is probably the thing most people would find most useful in their everyday life if they understood it. It’s something I think about a lot when I spot that I’m making obviously irrational decisions. In the book there’s the example of people keeping money in a savings account while not paying off credit card bills. They’re paying a fortune in interest unnecessarily but they keep the two things separate in their heads.

One final question - of all the economists who are still skeptical of behavioural economics and the work that you’ve done, what’s the what’s the best challenge that you’ve had from any of them that you found the hardest to answer?

Richard: The challenge I hear all the time is, well, it takes a theory to be the theory. And so give me your new theory and I don’t have an answer to that. In fact, I say there will not be a new grand theory. The same theory can’t explain the ultimatum game, and mental accounting and Palm/3Com.

Psychology has no grand theory it has hundreds of theories and humans are interesting in hundreds of ways. They’re all departures from perfect rationality and perfect selfishness. So I say if you want a single coherent theory of economics you should probably stick to the one we have, but pay a lot of attention to the anomalies. They matter.