Omnivoracious

Now It's Time for the Group Hug (Author One-on-One: Tim Harford)

Like Dan, I have enjoyed this discussion very much, and I hope that's true of Omnivoracious readers. It has been hard work--in a good way--and forced me to think carefully.

Let me finish by writing a bit about a subject I've been thinking about recently--it might help us find a bit of common ground and show why this discussion is important. Let's think about the problem of climate change, which is widely believed to require us to cut back a lot on our emissions of carbon dioxide. So what should we do?

Standard economics has a solution: a carbon tax, or a tradable permit scheme, to raise the cost of emitting carbon. We'd all have an incentive to cut down by driving less and turning down the heating or air conditioning, and also by insulating our homes and finding more efficient cars and fridges. Businesses would also cut back, and would have an incentive to develop new technologies.

Behavioral economics draws attention to different considerations. For example, we're not as smart as conventional economics assumes, which means that we might fail to notice ways to reduce our energy bills. We might also need to be reminded about what our neighbors are doing: psychologist Robert Cialdini showed that we're more likely to recycle if we think everyone's doing it. That may hold true for carbon-saving too. And although we're reminded of high gas prices every time we fill up, we really need to be reminded of them when we're buying a new car or a new fridge.

So which approach is right? That's a strange question. They can both be right, and I think that they both are.

Of course, the economic incentives will work. I don't think there's any serious doubt that raising the price of carbon would persuade us to pollute less. The Logic of Life showed that we are sensitive to these simple incentives in the most unexpected situations. For instance, raising the risk of unprotected sex persuaded American teens to have more oral sex instead; and two Australian economists, Joshua Gans and Andrew Leigh, have found that tax incentives can persuade pregnant women to delay labor, and even deter people from dying too quickly. (It's really true: the death rate in Australia dropped in the week before inheritance tax was due to be abolished, and then sharply rose the week immediately after. A lot of people were clinging on to life to avoid their estates being taxed.) If people will change their date of death or the day they have a baby in response to a tax incentive, I am quite sure that they will be more environmentally friendly, too.

But I am also sure the behavioral insights will produce results. Cass Sunstein and Richard Thaler report one dramatic success: people cut back dramatically if the electricity company gives them an "Ambient Orb" that glows red when they're using a lot of power. There's no standard economic theory that explains why that should be; and that is just one example.

So we need insights from both economics and psychology to make the world a better place. And I know that Dan will agree with me when I say that whatever we try, we should be checking carefully to see whether the idea is working as predicted, because the world is full of surprises.

Thanks to all Omnivoracious readers for dropping in on our debate.

Your logical friend,
Tim Harford

Designing for the Real, Irrational World (Author One-on-One: Dan Ariely)

Dear Reader,

If you have examined the debate between Tim and myself over the past few weeks, you must have realized that, in fact, Tim and I agree on many things. Particularly, that our respective perspectives are not the only useful ones (although we each believe that our individual perspective is more useful than the other's).

However, one major disagreement that Tim and I do have is about the weight of the evidence that supports the other's position. In general, Tim doesn't see how the results from lab experiments translate to the "real world." In essence, he does not believe that these lab-based irrationalities become full-fledged irrationalities outside of the lab. I, on the other hand don't see how the results from many of the statistical analyses he and others report necessarily demonstrate that people are rational. I do see how these studies show that people react to incentives in a way that is compatible with economic theory, and that they are sensitive to the general structure of the economic environment (and psychologists and behavioral economists would say that people should react to these), but I don't see the evidence that people react to these in an optimal way--in a rational way. I hope that Tim can explain this to me in our next exchange, but in the meantime I want to answer the two questions he posed for me in his last post.

At the end of his YouTube video (and by the way Tim has one other great video on YouTube) Tim posed the following two questions:

1. How can we take the insights from behavioral economics and apply them to economics?
2. How can we the apply insights from behavioral economics to make real changes in the real world?

Let me try to give my perspective on both of these questions. In terms of applying the insights from behavioral economics to economics: I don't see this as the goal of behavioral economics, I don't expect that we will ever be able to successfully achieve this integration and in fact, I don't want to "fix" economics. I think economics is beautiful, interesting, and that it has provided us with many useful insights. So where do my objections to standard economics come from? It comes from Tim's second question about using insights to change the world. Here I think that relying blindly on standard economics is dangerous, and behavioral economics has oodles to contribute.

When it comes to making changes in the world, such as laws, policies, or even business and individual practices, standard economics assumes/claims that it can provide the correct and complete answer. The answer! After all if people are rational then what else is there to take into consideration? This is what welfare analysis and the Chicago school of economics is all about. This is also what bothers me about economics, and what I would like to change. If it were up to me, economics would remain as it is in terms of an academic discipline, but we would consider other perspectives, including behavioral economics, when making recommendations for implementing changes in the world. Moreover, I want us to take to heart one of the main lessons of behavioral economics--that our intuitions are not always correct. By doing so, instead of just implementing a policy based on our intuition, we should first experimentally test our ideas to ensure that we are getting what we expect.

To better understand the role that behavioral economics can play in the design of everyday life consider the following analogy: When we design physical products such as phones, cars, and pens, we carefully consider our physical limitations. We don't design products for Superman, and by taking our physical limitations into account we are able to design better products, and live a better life. Why not do the same for products that rely on our cognitive abilities such as mortgages, health insurance, and saving plans? Why don't we learn to recognize our cognitive limitations and by doing so, design products that better fit our actual ability? This is the promise of behavioral economics--once we recognize our cognitive limitations we can design the world in such a way that it will not demand from us the type of computations that we simply cannot do.

So Tim, if we want to live in a world with less financial crisis, better health, and higher savings, then we must apply insights from behavioral economics to make changes in the real world.

Irrationally yours,
Dan

The Laboratory and the Real World (Author One-on-One: Tim Harford)

[Ed.: And now Tim Harford has picked up his camera to reply to Dan Ariely in kind. See the full discussion here.]

The Rational Economist on the Couch (Author One-on-One: Dan Ariely)

[Ed.: Dan Ariely has chosen to make this week's response in his discussion with Tim Harford via YouTube. Enjoy.]

Teenagers Rational About Sex? (Author One-on-One: Tim Harford)

Dan, thanks for the kind comments about The Logic of Life. Anyone who read my Financial Times review of Predictably Irrational will know that the admiration is mutual.

Let me just make one small correction: I certainly don't say that high CEO pay is desirable; I think it grotesque. Half that chapter is devoted to explaining why--why can a system full of rational shareholders can deliver grotesque pay packets?

Now, let's discuss the central question: are human beings rational or irrational? Perhaps the best way to understand the different perspectives of the books is through an example, and since the sexual examples in both books have captured everybody's attention, I'll compare and contrast.

Dan wants to know whether our behavior changes when we're sexually aroused and more importantly, whether we are caught by surprise when it does. As always, his approach is to run an experiment, and he paints a picture that will forever be etched in my memory, of students masturbating over Saran-wrapped laptops, while answering multiple choice questions about whether they would use condoms, consider using date rape drugs, have sex with animals... (In case you're wondering, no, this kind of subject matter is not the only reason that Predictably Irrational has been a runaway success.)

What does he find? First, that his subjects give very different answers to these questions when in a calm frame of mind than they did when sexually excited. Second, that they did not predict the way their answers would change. Now, one could argue about this research. Maybe the Saran-wrapped laptop is not a good substitute for a real sexual encounter. Maybe the pornography did not change the subjects' preferences, but merely encouraged them to be more honest. But basically, I buy it. The research seems sound, and Dan's witty description of it is pretty unforgettable.

So, how is The Logic of Life different in the way it approaches this topic? I take a bird's eye view of the whole controversy over the so-called "teen oral sex epidemic"; I find out that, behind the hype, yes, teenagers are having more oral sex than they used to. But I also find that they are losing their virginities later and are more likely to use condoms. That isn't so much an oral sex epidemic as a safe sex epidemic. I argue, informally, that this would be a rational response of a typical teenager to increased education about pregnancy and particularly to the now near-universal knowledge of the threat of HIV/AIDS. (You can read a short article I wrote about all this in Slate. You can also see what Stephen Colbert made of it all here.)

Fine. But that's still a bit speculative. So then I look at some of the latest evidence from economics, which shows that state by state, when state legislatures introduce laws tightening up on the availability of abortion to teenagers, the rates of sexual infections in teenagers falls in those states, relative to that of adults. That shows that teens are switching to safer sex, or abstinence. This isn't some vague correlation (more conservative states = less risky sex) but a specific response from teenagers to laws that affected only teenagers.

"But that's rational!" spluttered one venerable journalist, when I told him about this. Well, yes--it seems so, doesn't it?

Okay--who's right? When you see both examples set against each other you realize that's not a smart question. Dan explains that our behavior changes when we're excited; I explain that teenagers have safer sex if a legal change increases the consequences of pregnancy. Obviously both those things could be true at the same time. I believe that both of them are true.

At the risk of making this seem like a big group hug in which everyone is right about everything, let me raise a point of disagreement. The Logic of Life leans on all kinds of evidence, everything from the kind of careful statistical work I describe above to the use of lab rats. Theory counts too, and so does the clever simulation work inspired by the likes of Thomas Schelling. Every type of evidence has its flaws. I think the weaknesses inherent in pure theory are obvious; so, too, are the weaknesses of the sort of data-detective work very common on modern economics. (I will let Dan say more about those weaknesses if he wants to.)

But the weaknesses of laboratory experiments are not always quite so evident, especially when they are described as compellingly as in Predictably Irrational. So, let me point them out. While laboratory experiments are great for creating controlled conditions, they also create artificial conditions. There are several examples of important clashes between what happens in the laboratory and what happens outside. We know, for example, that procurement managers systematically screw up when bidding in a laboratory auction, but they do much better job in the (apparently identical) real life auctions situations they face everyday.

The economist John List has tried to replicate some famous "predictably irrational" results from the laboratory; the results tend to evaporate in more realistic settings. In one example I describe in The Logic of Life, List shows that the "irrational" result (which is that people given an unexpected raise work much harder than they could get away with) only lasts for ninety minutes. A gratitude effect that lasts ninety minutes is not the basis on which to rewrite your company's human resources policy.

I'm not aware that any of Dan's experiments have been challenged in this way, and I was pleased to see that he often tried to carry out his work in realistic settings such as restaurants and bars. So this isn't an attack on his work--it's more of a question. Dan, how can we be confident that these experimental results hold up in real life? And what further work would you like to see, to make us more confident in them?

Your logical friend,

Tim

Rational or Irrational: The Logic of Life or Predictably Irrational? (Author One-on-One: Dan Ariely)

The Logic of Life is a wonderful book. In prose that is witty and entertaining, Tim Harford uses his keen economist's perspective to peel away the layers from a variety of unconventional examples, revealing the unexpected rationalities that lie at the hearts of subjects such as large executive salaries, smoking, and even racism. His exploration of these subjects enables the reader to understand how economists think, and the logic that often lies beneath the surface.

Take for example the case of executive pay. As Harford explains, there are some interesting economic principles underlying what we might otherwise consider exuberant CEO salaries (for example the $800 million that Disney paid Michael Eisner over his 13 years of service as Disney's CEO). First, Harford argues that very high salaries are desirable because they cause the executives to have more skin in the game and, as a consequence, will care more about their company. Harford also points out that these high salaries are not only intended to motivate the person who is getting the high salary, as they are also intended to motivate the people below them who might pull all-nighters for years and work much harder in hopes of one day attaining that enviable position. (Harford proposes a clever test of this idea in which the CEO doesn't do any work, making his position even more enviable.) Finally, Harford also points to the rational selfishness of the CEOs, the difficulty of tracking their actions, and the mysteries of accounting as other ways in which CEOs further increase their salaries.

These are very interesting ideas with a compelling and fascinating logic (although they are not very well tested from an empirical perspective). In my mind this is economic theory at its best. Beautiful, elegant, requiring some convoluted logic, and shedding light on what we see in the world around us.

What should we do with these insights from economics? If we trust the rational economic perspective completely, we should take steps to pay CEOs more, maybe give them more vacation time or pay them even more with stock options. But here is the catch: in order to move from a perspective that sheds light onto the question of CEO pay to an action plan that we can adopt wholeheartedly, we must not only believe in these principles, but we must also believe that nothing else matters (or at least that nothing else would make a big difference). So, should we trust that our powers of logical reasoning could provide us with a complete understanding of the role of incentives? The field of behavioral economics suggests that we should not rush to make this assumption and that, in fact, there are many forces that influence us without our understanding.

Let me give you an example. Imagine that I described to you some puzzles of different sorts that required memory, concentration, thinking, and creativity. Suppose I then asked you to predict how well people would perform on these games if they were promised a payment that depended on their performance. I would also tell you that for some people the payment was equivalent to 1.5 days of their income, for others it was equivalent to 2 weeks of their income, and for a third group it was equivalent to 6 months of their income. Who would you predict would do the best? If you are anything like the real participants in our experiments, you would predict the third group--the one with the highest pay--would have the highest level of performance.

When we ran this experiment (in India where we could pay people this much without completely depleting our research accounts) we found that participants did the worst when they were offered the highest pay. It turns out that money is a double-edged sword: it is a motivator but it is also a stressor. To get a better insight about this idea, imagine that I offered you $10,000 to come up with a creative idea in the next 5 minutes. How focused would you be on the creativity task and how stressed would you be? How much time would you dedicate to coming up with an idea and how much time would you spend thinking about the money that you are not going to get?

What is the big idea here? First, it turns out that incentives are a bit more interesting and complex than standard economics would have us believe--sometimes we can pay more and get worse performance. Second, it appears that our intuitions and logic are sometimes inaccurate in their ability to predicting human behavior. What does this say about relying on standard economics and pure logic as a tool to guide the type of incentive mechanisms we create? It tells us that when we try to understand these mechanisms, and in particular when we make recommendations for how these mechanisms should be designed (which is something economists do often), we should take all the input we can into account, including our irrational characteristics. Standard economics and logic can help us create systems that are useful for perfectly rational people, but behavioral economics will help us design a better world for the rest of us.

Irrationally yours,

Dan