I Get to Pretend That I Am a Scientist for a Day…

… because today the Science journal published a short commentary [subscription required] written by myself and John List, on the topic of behavioral economics. Our piece begins like this:

The discipline of economics is built on the shoulders of the mythical species Homo economicus. Unlike his uncle, Homo sapiens, H. economicus is unswervingly rational, completely selfish, and can effortlessly solve even the most difficult optimization problems. This rational paradigm has served economics well, providing a coherent framework for modeling human behavior. However, a small but vocal movement in economics has sought to dethrone H. economicus, replacing him with someone who acts “more human.” This insurgent branch, commonly referred to as behavioral economics, argues that actual human behavior deviates from the rational model in predictable ways. Incorporating these features into economic models, proponents argue, should improve our ability to explain observed behavior.

But we remain somewhat skeptical:

Perhaps the greatest challenge facing behavioral economics is demonstrating its applicability in the real world. In nearly every instance, the strongest empirical evidence in favor of behavioral anomalies emerges from the lab. Yet, there are many reasons to suspect that these laboratory findings might fail to generalize to real markets. We have recently discussed several factors, ranging from the properties of the situation — such as the nature and extent of scrutiny — to individual expectations and the type of actor involved. For example, the competitive nature of markets encourages individualistic behavior and selects for participants with those tendencies. Compared to lab behavior, therefore, the combination of market forces and experience might lessen the importance of these qualities in everyday markets.


Dan

such.ire, Jacob B:

You could argue that we should readjust the utility curve, but I think that doing so has the potential to lead you to seemingly absurd arguments.

Jacob, you hypothesize, for instance, that the opportunity cost of buying the "wrong" jam is higher than the cost of not buying any jam. That strikes me as one such absurd argument; surely in both cases the opportunity cost is defined solely by the value the buyer would have gained from buying the "right" jam, no? In other words, in both the case where we buy jam and the one we don't, we run the risk of failing to buy the better jam--the opportunity cost. The only difference I see, rationally, between the two cases is whether there was a marginal gain by buying the jam we did buy. This shouldn't change based on how many other possible jams there were.

So this brings us to the subjective experience of the buyer. Yes, there's a utility cost, in some sense, for feeling bad (overwhelmed by how many choices, perhaps). But isn't it simpler, and more reasonable, to simply say that the buyer estimates, inaccurately, the potential opportunity cost (as Jacob described) by using the number of jams available? In other words, the buyer uses this as a rough heuristic and, seeing how many jams there are, decides (irrationally) that the risk of making the "wrong" choice has increased (it hasn't, since he is guaranteed to miss the better jam if he never buys jam at all).

So it doesn't strike me that this example is so easily reconciled. Rationally, the possible opportunity cost hasn't increased. But the best we can say about the buyer is that he estimated that the opportunity cost went up.

(As a side note, my personal view was always that the buyer feels a compulsion--nonrationally--to inspect every possible jam before buying. If he's presented with enough jams as to make that costlier than the value of the jam he wants, he won't buy. This, I should guess, is why we aren't put off by the number of possible cars, computers, or houses--these are all expensive items, so the cost of looking at all the options is comparatively negligable--but we are put off by doing so for something worth a lot less, like jam.)

Read more...

Dave

The blurb seems to imply that economics is a sort of self-fulfilling prophecy:

behavioural psychology may tell us many things about human motivation, but economics guides us to construct markets in such a way as to discourage people from behaving in these more complex modes and instead to behave according to the market rules.

Thus, economics is quite good at predicting market behaviour but tells us almost nothing useful about human motivation or satisfaction - these studies belong in other disciplines, and our responses need to be guided by those disciplines as well as economics.

Jack

Responding to Rob:
The homo economicus camp generally allows that observed individual behavior will deviate from the predicted rational decision. However, they believe that those deviations will be non-biased, thus cancelling out on average as you suggest. (E.g. some firms at an auction for oil rights bid more than they rationally should while others bid less.)

The behavioralists look for deviations that are systematically biased in one direction. In the auction context, several studies show that participants tend to bid higher than they "should" because they fail to account for a phenomenon sometimes called the "winner's curse." Thus, the deviations don't average out to zero.

To my knowledge, this phenomenon occurs in many real-world auctions for common-value goods (such as oil rights), not just in lab experiments.

Jacob B

I tend to agree with such.ire that even the novice jam buyers' behavior could probably be modeled as a utility maximization which makes me curious about why, exactly, they demonstrate the seemingly irrational behavior they do. I haven't studied economics beyond micro 101 but my first thought is that the opportunity cost of buying a given subset of jam varieties increases with the number of available varieties (because also increasing is the number of varieties rejected) -- and that at some point as the number of varieties increases, because the buyer might get his/her own preferences 'wrong' (which is more and more likely the finer grained choice he/she is forced to make) and regret, in particular, the *rejections* he/she made, leaving the buyer with the option to still purchase any given variety by NOT spending the money allocated for the jam budget becomes the most preferred option.

Dave

How about the work that Barry Schwartz cites in his book, that more choice makes people less happy? A particularly salient example is the work by Sheena Iyengar (Best name in decision research) demonstrating (in actual grocery stores) that shoppers are less likely to buy jam when given more (21, I think) choices than when given fewer (5-6). Any rational model I can think of would argue that more choices leads to likelier match to stable preferences. These data show otherwise.

AaronS

Perhaps it is time for homo economicus to evolve....

Very simply, while the purely rational doings of homo economicus are a benchmark of sorts, he may not be telling us very much about the real world. For while he is Spock...the rest of us are the more emotional Kirk or Bones.

And as homo economicus evolves, he begans to take on the average nuances, histories, traditions, religious beliefs, values, biases, and so forth that we humans have--and that, taken together, cause us to make decisions differently than originally supposed. Once we have this "average" model that takes into account the many internal influences of a man, we might actually be able to "predict" with some precision how endeavors and tests will turn out.

Kevin L

As a non-economist, I like to give my take on the jam issue. More isn't necessarily better. Given 21 different jams, what exactly are the variables that change...flavor, jar size, brand name? Or can those 21 really be categorized into fewer sets? Is the variety of jam really just commoditized and the average consumer really can't tell the difference?

As for the lab vs real world discussion, isn't by definition lab experiments only focus on a finite set of factors? It's not feasible to examine all of the factors and as such, people can always look back to the real world and throw another factor in. If lab experiments could incorporate all factors, then why use labs...just look at the real world.

Speaking of irrationality vs. rationality, maybe we should look at systems more closely. If you create a system (in the lab or market), and people consistently respond in an "rational" or "irrational" way, then wouldn't it just all be rational? Then would we say that irrational is just very unpredictable. So we're saying there are systems where people sometimes respond rationally and sometimes irrationally? Or is it systems where a percentage of people respond rationally and the rest otherwise? Are there systems out there where everybody behaves rationally? Going back to the jam experiment, if we can establish that when people are given a relatively higher choice, that they are less likely to buy, then we can now predict how consumers will behave and that's not so irrational of them.

On a lighter side of things, homo economicus and psychologist try to understand the behavior of homo sapiens but is it ever the other way around. Despite years of research, the experts still haven't gotten it down pat yet. How useful is the "rationality" model then? I think you would have to convert all homo sapiens into homo economicus and then we won't even have to discuss this topic of behavioral economics. Can you imagine a world where everybody acts rationally? Is there such a thing as a rationality index? Has anybody tried to measure the world's rationality?

Read more...

Dan

Kevin L:

I wanted to just address one thing you said--having already rambled on about the jam thing myself for far too long. ;)

"Going back to the jam experiment, if we can establish that when people are given a relatively higher choice, that they are less likely to buy, then we can now predict how consumers will behave and that's not so irrational of them."

I don't think you can conflate rationality with consistency. A group of people might be consistently irrational--in the sense that their actions do not maximize their utility. They might be consistent, but wrong.

(For that matter, there might be situations where rationality demands unpredictability, so that the rational actors are the ones who appear to act inconsistently. For example, if the stock market were perfectly efficient, a rational actor would do well to pick stocks purely by flipping a coin. He'd be unpredictable, but rational.)

Read more...

Jacob B

Dan wrote: "Jacob, you hypothesize, for instance, that the opportunity cost of buying the "wrong" jam is higher than the cost of not buying any jam. That strikes me as one such absurd argument; surely in both cases the opportunity cost is defined solely by the value the buyer would have gained from buying the "right" jam, no?"
and Dan also wrote: "In other words, the buyer uses this as a rough heuristic and, seeing how many jams there are, decides (irrationally) that the risk of making the "wrong" choice has increased (it hasn't, since he is guaranteed to miss the better jam if he never buys jam at all)."

The buyer is committed to having missed the better jam once he makes a purchase, but until then he leaves himself the opportunity of nabbing it later. The opportunity cost is not the value of the best jam in the case where he doesn't buy any, since he has not given up that opportunity, only delayed it. The non-purchasing shopper only has to pay the opportunity cost of not having the best jam immediately (but a friend's recommendation could clarify things and allow him to return to the store tomorrow not worried about making the wrong choice) whereas the purchaser who has expended his jam budget has to wait for an entire budget/jam cycle and is stuck with a lesser jam.

Read more...

Meds

I offered my own take on the "Ultimatum Game" earlier this month.

http://justenjoytheshow.blogspot.com/2008/02/ultimatum-game.html

mike roddy

Choosing jam is never a rational decision. If it were, the buyer would not respond to colorful labels, corn syrup laden product, and chemical flavorings. Home economicus will always be a flawed model, expressed as a curve showing those who pay attention on one end and those who respond to other stimuli on the other.

If people behaved rationally, Burson Marsteller would not be in business. Nor would we continue to build 4,000 foot houses with things like double offices and pantries, all built with two by fours and toxic formaldehyde laced carpet, cabinets, and sheathing.

S. Heaton

First, the claim that Levitt is making with respect to individual behavior (not mediated by markets) has been widely disproven. This has been shown time and time again both in the lab and in the field.

The point that Levitt tries to be making is that lab results may not generalize to the field. This may be true, in some cases, but not all. Not to mention, the economic models have been spectacularly poor at predicting just when these cases might be.

Moreover, Levitt and List's "Big idea" about generalization has been widely studied in psychology at least as early as the 1930s. List even cites some of these papers in his own work. The difference is that List and Levitt have no model for how the mind works -- and provide no concrete mechanism for how individuals get "de-biased" by market forces.

That said, Levitt and List's research is potentially very important, but it has nothing to do with a critique of "Behavioral economics." That's a bet that Levitt is likely to lose, big time.

Read more...

Browning

And some vote for George W. Bush - twice.

Dan

Jacob B:

That's a good point, but wouldn't it assume--or at least, wouldn't it assume that the buyer assumes--that he later does more jam research and comes back to make a more informed jam purchase?

Dan

mike roddy:

"Home economicus will always be a flawed model, expressed as a curve showing those who pay attention on one end and those who respond to other stimuli on the other."

Doesn't this describe the issue not as one of rational behavior, but as one of imperfect information? The buyer may not know about the potential harm of corn syrup--he may use the label as a (poor) heuristic for jam quality. He may not know about formaldehyde-laden 2x4s (I didn't). But imperfect knowledge is not the same as irrational behavior.

misterb

Taking the behaviorists' side in the jam debate, I think that the numbers of choices are critical. Consistent psychological research shows that we can only distinguish somewhere between 7-10 similar items as individuals. A gathering of more items will be perceived as a crowd rather than as a group. Some jam buyers may be intimidated by a crowd, and decide that they don't want to spend the mental energy to disambiguate it. Good vs bad jam doesn't enter into it because these potential buyers never get to the point of distinguishing one jam from another.

All told, there is no reason other than mental laziness for homo economicus; if economists didn't persist in this ungainly myth they would have to start revealing unpleasant truths to the powerful who fund them.

angryman

I'm baffled by commenters such as #18 and 19 that describe homo economicus as "flawed" or "clearly flawed". Can you not think of circumstances where it is valuable to know what a rational person would do?

Sure, it might also be interesting to look at why some people make non-rational decisions - but how could you even identify a non-rational decision without first looking at our good buddy homo economicus.

G.V.varma

Does ther exist "behavioural macroeconomics"? If yes, could you please tell me some references?

Doug Wolkon - Author of The New Game

Free Time (or lack there of) to choose is the economic variable that is missing outside of the lab. How much time does the individual have? Do they think and act based on a balance of short and long-term goals, or are they in a hurry to decide and therefore forced to think and act irrationally. If we can afford the ultimate luxury of taking our time to choose, our ability to think and act rationally increases.

Sue Bristol

I'm a high school English teacher of 16 years, now a department chair, at a VERY good school in San Diego, wielding a little power now, which sometimes surprises me, given my rebellious nature. But, anyway, I just finished your book, for my book club of "super liberal powerful women" (tongue in cheek, but hey, I guess some are). Anyway, I LOVED it! So, one of my sweet innocent girl students who is struggling in my Honors American Literature class gave me a book today that she wanted me to read. I knew as soon as saw it that it was one of those books designed as "high interest" for young minds. It's titled "Give a Boy a Gun," by Todd Strasser. I'm horrified by the statistics and fear tactics displayed here! My kids have to use MLA to site all sources, and this guy gives stats on every page without any sources! The story line is cool for kids, i.e. don't bully, but although he states it is fiction, he puts stats on almost every page (usually without sources). I find it really disturbing. Any thoughts? Have you heard of this book? I'm thinking of just giving her a copy of "Bowling for Columbine" so she can hear the other side of the story. Help.

Read more...