A host of studies and academic theories that apply psychological insights to economic behavior have been grouped under the label "behavioral." Is this growing field changing how the economy is studied — and how it functions?
Q: What does "behavioral" mean in your field?
Nicholas C. Barberis: When I hear the word "behavioral," what I think of is trying to understand behavior that might be the result of less than fully rational thinking. So when we talk about behavioral finance, which is my own area of interest, we're trying to understand whether things that happen in financial markets might be the result of less than fully rational thinking.
Three areas of behavioral finance have received a lot of attention. One is what I might call applications to asset pricing — trying to understand the pricing of stocks, the pricing of houses, bubbles in financial markets, and so on. A second area is investor behavior. For example, we know that many people invest heavily in the stock of their own company, and we think that perhaps that's not wise from the perspective of diversification, and it might be the result of less than fully rational thinking. The third area is corporate finance, where we're trying to understand whether managers of firms do things that might be the result of less than fully rational thinking. We know, for example, that there is a lot of merger and acquisition activity that goes on in the corporate sector, but the evidence suggests that there's not a lot of value creation for the acquiring firm. So then why do we see so many M&A transactions?
Ulrike M. Malmendier: One way to conceptualize deviations from a standard economic model is to say that there are three types of deviations. The first one is in terms of belief. We don't always form rational beliefs. We don't form the right expectations, given the information we have.
The second deviation is nonstandard preferences. What that means is the following: While we have been working with a certain set of preferences and utility functions in traditional economics — typically they involve the maximization of our own wellbeing — in reality, people have different preferences. A famous example is self-control problems. People join the gym and plan to go regularly. That's why they get a monthly membership, which only pays off if you are actually going regularly. They end up not doing it. There is an inconsistency between their preferences about future actions and their preferences when they actually are choosing the action. The same phenomenon can be observed in finance — people choose a certain means of financing, like a credit card, which is the right instrument if you pay everything back at the end of the billing cycle, but then they end up not doing so. There is this discrepancy between how I want to act from a long-term perspective, and how I do act, which is one example of nonstandard preferences. Others are social preferences, such as altruism, shame, guilt, conformity.
Then, as the third deviation, there is what people call "failure to optimize" or "failure to maximize." We don't have a good term for that third deviation, but what that means is, even if an economic agent had the standard preferences we assume in traditional economics and even if he formed rational beliefs, the whole maximization process doesn't always work out perfectly. We sometimes just don't get it right, not because we are irrational or because we have funny preferences, but because we have limited processing capacity or limited attention. It's Friday, and we are planning our activities for the long weekend.
Sometimes, even if you're fully rational and have standard preferences, you may not act according to standard economic textbooks.
Shane Frederick: In my view, at least, a person or a study is behavioral to the extent that there is an explicit recognition of the distinction between a formal characterization of some choice object and the mental representation of that object — the idea that people may think about the same thing in different ways. The same object can yield systematically different evaluations by focusing attention on a subset of features or by describing the features differently. Everybody who is behavioral accepts that, and people who are not behavioral generally don't. They want to define the object in terms of its formal properties. It has this probability and it has this amount of winning, for example.
In marketing, I think there is really no resistance to the idea that people are behavioral. We know that all light beers are more or less indistinguishable. And yet many marketing efforts go toward changing people's mental representations of those objects. Marketing — at least a big part of marketing — is behavioral at its core, and has been historically.
Andrew J. Redleaf: Financial markets may be the last area that isn't cognizant that the economy is made up of interactions between human beings rather than automatons. In business school, when you're taking the class in negotiation, they will not tell you to start with the assumption that the person across the table from you is an automaton with a predetermined set of instructions, which either you can know or anticipate, in trying to figure out what's important to him, how he values different trade-offs, and where the possibilities for a win-win outcome are. On the other hand, if you're a banker or investor talking to a CFO about the company's capital structure, the presumption is that here is an automaton running an optimization program. In a lot of ways, financial markets are either paradigmatic or completely exceptional.
Q: Do you take behavioral insights and find advantage in them in practice?
Redleaf: Yes and no. Yes in the sense that if you're confronted with something that seems anomalous or seems to not make economic sense, that should create dissonance and you should approach with some caution. If you can find explanations that make the anomaly coherent or understandable, then you can be more comfortable. You can have some sense that what you see is a deviation from pure rationality, and a deviation from where an asset should be priced. In my experience, I've done much better on things where I can figure out what I thought people were thinking, and have a story or an explanation for why it was wrong, rather than just having the argument that it's wrong without understanding the genesis of the anomaly.
Malmendier: This is parallel to why I find behavioral economics so important. What I really love about behavioral economics is the big emphasis on trying to get it right when explaining why humans make decisions. We are trying to be humble as economists, and to learn from evidence in psychology about why people think the way they think and why they are choosing the actions they are choosing, rather than saying, "Our standard economic model may not get the psychology right, but it predicts what the correct action is." No, we really would like to understand the truth. We would like to go inside the black box and be able to ask why a person is making a certain choice. The benefit is, of course, that we will have better predictions for new situations.
Q: Nick, what do you think are the exciting areas in your field right now?
Barberis: There are a couple of things that have been in the news quite a lot, and I think those are indeed some of the exciting things. The first is associated with this well-known book, Nudge, by Richard Thaler and Cass Sunstein. The rough idea there is that, look, people make incorrect decisions sometimes, so can we help them? Can we perhaps nudge them towards making the right decision, but without actually taking any options away from them? Thaler and Sunstein call their philosophy "libertarian paternalism." One example is that a lot of people don't save as much as they'd like to for retirement, so one specific policy change that was made recently was to switch 401(k) plans from opt-in plans, where the default is that you are not enrolled and you have to fill out a bunch of forms to become enrolled, to opt-out, where the default is that you are enrolled and you have to fill out a bunch of forms to get out of the plan. Now, in a world where everyone is fully rational, it shouldn't make a difference what the default is. You can just choose whatever you want. But we know that in the real world, where people have inertia and they procrastinate, the default really matters. And so we might predict that the people who are automatically enrolled will end up saving more than people who are automatically not enrolled. Studies have demonstrated exactly that. That simple policy change could have a real impact on how much people are able to save for retirement.
And the second thing I'd mention that has also been in the news a lot, obviously, is the financial crisis, the housing bubble. I don't know if I can call the work on this exciting, because it's still at an early stage, but we have an amazing event in history to work on. How did this housing bubble form?
Frederick: There's a really interesting study about organ donations by Eric Johnson published in Science, where they showed that whether or not people elect to have their organs donated if they should die in an auto accident goes pretty much from 0% to 100%, depending on whether you are defaulted in as an organ donor or defaulted as a nondonor. This default setting has as big an effect as you could imagine for one of the more important decisions you could imagine.
I don't think there's anything in marketing that's as exciting or as important as that. What I do find exciting is the work on the role of metacognitive judgments — for instance, that people form a sense of whether they are an expert or not based on cues as subtle as the font in which something is printed. If the font is unclear or if it's an atypical or unfamiliar font, you get the sense that this is hard to process and feel insecure. And from that, your choice strategy actually differs. You couldn't imagine, ex ante, that these subtle changes could have these significant effects. These things are real discoveries, so I find that pretty exciting.
Q: Thaler says that he thinks marketers and advertisers have been nudging forever. Does that make sense to you?
Frederick: I think that's certainly true. Along with two of my colleagues at Yale, we just published a paper which featured an experience I had once had in a store. I was buying a stereo, and wasn't sure which one to buy — the cheaper one for $600 or the more expensive one for $900. And then the salesperson said, "Think of it this way — if you buy the $600 stereo, you can have $300 worth of CDs." Suddenly, the choice, which had been very tough, was quite clear. I would buy the cheaper one.
For a neoclassical economist, this makes no sense whatsoever, because the decision to buy should go through the concept of opportunity costs. Is the opportunity cost to have the nicer stereo worth it or not? But here's somebody who studied decision making for a decade, and I just didn't think of the choice in that way. Have people been nudging for decades? Yeah, this guy did.
Malmendier: Given that we've shown that being serious about behavioral insights can make a big difference to important economic decisions, both firms and policymakers are becoming more open to our data demands and our suggestions. For example, in 25 minutes I'm going to be running a behavioral lunch. One of the projects we'll discuss is a weatherization program, which is supposed to help more low-income households with insulation and other energy-saving remodeling. Having in mind that people who want to apply for the funding might procrastinate, the Department of Energy is open to allowing some randomization in how the procedure is implemented, which will help researchers gather good data on the relevance of this behavioral bias and ultimately to improve the process.
Redleaf: I have to say, listening to all of this, I find it interesting, but I also feel some trepidation. I want the world to be materially and economically better off, and I think the right use of all of this information is to better allocate resources. And it seems to have great potential there. But I also worry that the insights are apt to be taken in policy directions that I don't like and that have the risk of making people worse off. I mean, most of the time, market imperfection is used as an excuse for government expansion, which is anathema to people on my side. So I'm made nervous by people saying you get different results if you have an opt-in savings plan versus an opt-out savings plan, because I'm nervous about who decides which is actually optimal. I'm for heightening awareness of any insights that come out of this line of inquiry, as opposed to explicit use by government.
Q: Why do you think behavioral approaches have become more prominent recently?
Barberis: By the time we reached the 1990s it just became clear that lots of things in finance were hard to explain in a model where everyone in the economy was rational. I think when the tech-stock bubble came along, people were more convinced that we had to find an alternative to the rational paradigm, and that the behavioral finance paradigm seemed like a very plausible alternative. Now that we've been through the housing bubble, I think people are all the more convinced that this is a very promising direction to explore. The field has also gotten an extra push in the past year because of the new administration in power in Washington. A lot of the economic advisors to President Obama have actually worked on behavioral economics and behavioral finance during their academic careers.
I do want to say one other thing, which is about the power of a conceptual advance. If you talk to people on the street and say, "If people are irrationally exuberant, they're going to push the stock market up, and if they are irrationally pessimistic, they're going to push the stock market down," that sounds very plausible. But when you think about it, it's not so clear. If irrationally pessimistic people push the stock market down, then there is a very strong incentive for smart people, for hedge fund managers, for example, to come in and buy the stock at those bargain prices and thereby push the stock market back up to its correct level. This is sometimes called the arbitrage argument, and it suggests that irrational people cannot affect the prices of financial assets for a very long period of time. That argument was very convincing to academics for decades, and it's the reason that the rational paradigm was so dominant in universities. What happened in the early 1990s was we realized that this argument didn't make sense. Irrational people can affect prices for a significant amount of time, because it's actually not so easy for hedge fund managers to correct mispricing. Hedge fund managers face a lot of risks in correcting mispricing and so they may allow mispricings to survive for quite a long period of time. It was really that conceptual advance that opened the gates to more behavioral finance research.
Malmendier: What I've observed over my research life — the last five or ten years — is a change in the presumption that if you go outside of the standard models and assumptions with which we work, you're giving up all rigor and discipline. This was, of course, always a very weird argument, if you think about it, for scientists. We want to understand what truly drives economic decision making and how we can better predict how people will act in the future, not how to squeeze it into an existing but possibly wrong model, just to prove our "rigor."
I would say that ultimately the truth won. We economists realized that we could only get so far with the assumption of rationality. And we realized that we don't let hell break loose if we allow for some novel assumption. If we dare to allow for unconventional but realistic assumptions about human behavior, we find that they have a really big impact on investment decision making, consumer decision making, or firm decision making and help us to explain some long-standing puzzles.
For instance, in development economics, researchers would look at why farmers don't use fertilizers that would increase their income. They thought there must be standard economic mechanisms at work, such as asymmetric information about the usefulness of fertilizers, so let's do an experiment which reduces asymmetric information and see whether that will have an effect. They found very little effect. Once they opened up to the idea of a time-inconsistency problem — if I give the farmers the money for fertilizers now, they may waste it on going out for drinks or other spending that doesn't help in the long term, but if you offer them a commitment device, maybe that makes a difference and they do actually go for the fertilizer — they found huge effects. And so people in development have started to be serious about modeling behavioral traits and incorporating them into the design of their field experiments.
Redleaf: It's an interesting question to what extent the academic belief in efficient markets filtered down to the masses, as it were, to investors — and to what extent that has facilitated bubble behavior and the bubbles that we've had in the subprime area. At the cultural level, people know to be wary of used car salesmen and to take what they say with the appropriate shakers of salt. On the other hand, even for relatively financially unsophisticated people, I think they have a sense that financial markets are different somehow. It's not just that the practitioners are professionals with professional codes different than used car salesmen. I think there's some sense that financial markets get things right. If I can get this loan, I must have a reasonable prospect of being able to pay it back, even if it doesn't seem to add up to me. Because it's touching finance, it must be okay. They gave me the loan, after all. So apart from explicit fraud on either the lender's or the borrower's part, there was a susceptibility on both sides, given a cultural understanding of the efficiency of the financial markets. I don't think behavioral finance has changed that yet.
Q: Are there ways that behavioral research is filtering out to have an impact on the spheres of business and policy?
Barberis: One set of people that has been interested in behavioral finance is the investment management community. And there are a number of large investment firms out there that, I would say, fall under the umbrella of behavioral finance in the sense that they see themselves as systematically exploiting mistakes made by less sophisticated investors.
An interesting question I have been getting in recent years from investment managers is, okay, we know the ways in which people are less than fully rational. Can you now help us design an investment process that will get rid of those errors in our own thinking? I'll just give you one example that I heard from a hedge fund manager in New York. He said, "When my analysts pick a stock, they often set a target price, as well. So I'm buying a stock at $50 and I expect to sell it when it reaches $65." What he was finding was, often, when the stock reached $65, his analysts wanted to put on a second-round bet that the stock would rise even higher. And he often found that that second-round bet didn't work out. He put it down to some kind of overconfidence — having seen their initial hunch confirmed, these analysts became too confident. To try to solve this problem, he instituted a policy where every time an analyst in his firm wanted to raise the target price, they had to send an email to everyone in the firm explaining why. This removed the problem. Indeed, he was just using a well-known technique in the psychology literature for de-biasing overconfidence, which is forcing people to give a public rationale for their actions.
Redleaf: In financial markets, writ large, though, I don't think that behavioral economics has had any effect to date. I would guess that that would be at least 10 years off.
Frederick: I agree that not a lot of ideas have made it into business. One thing that I've recognized recently in our field is that there's a lot of attention to one-off behaviors, like whether you can get people to choose option A rather than option B by setting up the choice set in just such a way in the stylized micro-world of the experiment. The assumption is that if you can do so, that translates directly to business. But what I think is often ignored is that there are downstream effects. Suppose, for example, you trick me into buying something too expensive, say, a dinner, by the way you set up the menu. I end up spending $120 on dinner and wine. Afterward, I might feel bad about that. And so I have this negative affective tag associated with the restaurant experience, and I don't go back there again. While you have succeeded in a narrow sense, the bigger message is lost: Was this good for business? It's not entirely clear.
I do think there are some ideas that people outside the academy are considering more now. The idea that complexity is bad is catching on. We did a field study where we had prices for juice that were either all high or a mixture of high and low. Strict economics would say it's a pretty clear knife-edge prediction what's going to happen: You should sell more in the second case. At least for awhile, we sold less. Psychologically you can sort of see why. The decision becomes very complex with multiple prices. It's not just a question of which is my favorite fruit juice, it's which is my favorite and is the intensity of my preference sufficient to justify the extra expenditure? It becomes a much more complicated decision and people have a certain distaste for that.
I love Brian Wansink's work on anchoring. He has taken this basic insight that you can specify any number and people will anchor on that, and he put up signs saying "limit of 12 per customer" for cans of soup. People weren't buying 12, they weren't buying close to 12. But the faux limit put this notion in people's heads that 12 is the reasonable number of cans of soup that they should buy. And he showed that you sold significantly more soup in that case. Some things like that are catching on.
Malmendier: When I hear that question, I immediately reverse it. I get my ideas from business. For example, I have worked on how firms design the pricing of goods that challenge consumers' self-control. Namely, if firms know that their good is tempting and people will tend to over-consume, firms charge a low flat fee and then charge above-market costs each time you consume — for instance, each time you use your credit card to borrow money. If there's a good that you will underuse, because it's something like going to the gym, which you hope you will do a lot but you're not going to end up doing it, then firms charge you a whole lot up front and less per usage. That study was inspired by talking to health clubs, originally, and later the credit card industry. I wrote another paper comparing how large investors and small investors react to analysts' recommendations. People often under-appreciate incentive misalignment, so, small investors don't understand that analysts tend to exaggerate recommendations — a hold recommendation really means sell, buy means hold, et cetera. That idea was already out there. Very often, as an academic, I learn from what's done in the business world and then analyze it from a behavioral perspective.
As I said in the beginning, what's important to me is to really try to capture the truth of what's driving human thinking and actions, whether it's rational or irrational, standard preferences or nonstandard.
Interview conducted and edited by Jonathan T.F. Weisberg.