Bryan Caplan  

Geoffrey Miller Briefly Makes Me Embarassed to Be An Economist

Econlog Makes WSJ's Top 25... Massachusetts Health Reform, V...
One of the best passages in Geoffrey Miller's Spent shows economists at our worst:
My crisis point came at a 1999 conference that I organized in London on the origin of people's economic preferences.  We psychologists thought that economists would enjoy hearing about our preference experiments, so that they could develop more accurate and sophisticated models of human economic behavior.  How wrong we were.  It became clear that economists still followed a "revealed preferences" doctrine, which holds that consumer preferences are psychological abstractions - hidden, hypothetical states that cannot be measured or explained apart from the purchases that they cause.  If preferences are revealed only through purchases, and not through questionnaires, interviews, or focus groups, then it is redundant to study preferences apart from actual consumer spending patterns, to speculate about the origin of preferences, or to conduct market research on preferences for hypothetical products... So the economists gradually drifted away from the conference, leaving the psychologists to nurse our bruised egos, in the company of some strange-looking folks we hadn't seen before.

These folks weren't like the academics at the conference... They were the marketers, and they were hot for psychology.  They actually cared about people's preferences - where they come from, how they worked, and how to profit from them.  I talked for hours with them, and a new world opened up.
This anecdote rings true.  Many - perhaps most - economists would scoff at Miller's search for the origin of preferences.  Why?  Because it's not economics.  And why isn't it economics?  Because we don't do it.

To any psychologists who have encountered this mentality, all I can say is: We're not all like that.  I found Miller's questions immediately fascinating.  I furrowed my brow when I realized that economists usually leave these important questions to a field I know only by name - "marketing."  There's no excuse for it.

After I read this passage, I was eagerly expecting Miller to give economists a primer in marketing.  But the chapter on "what economists should learn from marketing" never comes.  Instead, Miller quickly starts giving marketers a primer in intelligence and personality research!  It's a confusing rhetorical strategy:

Step 1: Tell economists they need to learn about marketing. 

Step 2: Tell us very little about mainstream marketing research. 

Step 3: Tell marketing people that they need to learn more psychology.

Frankly, Miller barely seems more interested in actually listening to marketers than the typical economist.  Still, this would be easy to forgive if Spent showed that intelligence and personality research made mainstream marketing research obsolete.  As I'll soon argue, though, Miller dismisses mainstream marketing far too quickly.  Intelligence and personality research may be able to supplement standard approaches, but they can't credibly replace them.

Comments and Sharing

COMMENTS (13 to date)
ionides writes:

I don't see why you're puzzled. Miller figured out that economists don't want to hear about marketing. So why should he waste his time telling them about it?

Les writes:

I remember a faculty meeting at the Stanford Graduate School of Business. A professor of marketing exclaimed: "Why do we wait until the second semester of the MBA program before we teach marketing? Let's hit them with marketing on the very first day!"

A finance professor replied: "Great idea! Now, what should we do on the next day?"

Of course marketing is an important function in business. But that does not make it a science.

Tyler Cowen writes:

Marketing insights are very important, but they're notoriously slippery, context-specific, and also hard to communicate in print.

fundamentalist writes:

Ohe primary markets for BA economists is marketing, even if it doesn't appeal to academics. But there are two types of marketeers--communications marketeers and research marketeers. Most of what passes for marketing int he country is marketing communications, which focuses on writing ad copy, designing graphics, video production, and some accounting.

Marketing research is applied economics. I would guess that Miller encountered marketing communications people, not marketing researchers because his group's findings would have been old hat to the researchers.

Marketing communications people work by rules of thumb, most of which marketing research has proven wrong. The two disciplines rarely get together, but when they do buy stock in that company.

Marketing researchers have spent decades and billions of dollars trying to figure out what makes people tick, but they'll be the first to tell you they still don't know. They know a lot, but not the most important things, like what new products will customers buy. Market research on new products is almost worthless because people think they want something but when they have to put down cash, their wants seem to change. Around 50% of all new product launches, after extensive market research, fail.

Miller, and Caplan assume that academic economists aren't interested in new fields of research, but my guess is that they aren't interested in micro-economics. Micro makes macro look really, really silly.

John Jenkins writes:

" Market research on new products is almost worthless because people think they want something but when they have to put down cash, their wants seem to change."

That sounds interesting. Maybe we could give it a catchy name. We could call it "revealed preferences."

Psychology is bunk.

SWH writes:

"Because it's not economics. And why isn't it economics? Because we don't do it."

Macro guys dont do it because it is easier to make models reach conclusions with presumptive preferences. If macro had to support all the preferences used, it would grind to a stop. No fun there.

twv writes:

Economists generally treat the fantasies of people that are not revealed in action as just that, something for psychologists. The preferences that are revealed in action are generally treated as "a given," and it is up to social psychology to explore how tastes change, etc.

It's just part of model building, right? You choose what you want to explain, and prescind out the parts of the complexity that don't immediately effect what you study.

Of course, good science brings back in the complexities after the simplified models have run their course. So, economists should be interested in psychological research for that reason. They may be most interested in what happens after a preference is revealed in choice, but . . .

Do I have this wrong?

The ghost of Ludwig von Mises seemed to be nudging me as I wrote the above, up until the point where I called his method a model....

Lance Cahill writes:

George Stigler (1951) had an interesting perspective on this:

"[E]ach decade, for the past nine or ten decades , economists have read widely in the then-current psychological literature. These explorers have published their findings, and others in the field have found them wanting—wanting in useful hypotheses about economic behavior."

Penske File writes:

Anyone who has ever actually worked with Marketing folks will understand the scenario described. My experience is that they are essentially witch-doctors who are looking for some scrap of science on which to latch and use to justify their work (and budgets)

MattW writes:

There is some validity behind what he called "revealed preferences." What people say that they want or think doesn't match up 1:1 with what they do. But it's pretty arrogant to reject a whole set of ideas and information because of imperfect links. There is a lot to be gleaned from psychology and marketing.

Tracy W writes:

What's the point, from the economics point of view, of learning about why people have the preferencers that they do? What questions would economics answer better with a set of preferences rather than a general statement that people maximise utility, however they define it? I can just imagine spending ages arguing over what people prefer for no particular benefit in terms of policy-making.

Lauren writes:

Tracy asked an excellent question:

What's the point, from the economics point of view, of learning about why people have the preferencers that they do? What questions would economics answer better with a set of preferences rather than a general statement that people maximise utility, however they define it?
As a grad student, I initially scoffed at trying to explain preferences. What possible difference could it make? Preferences are what they are; economists take preferences and endowments as given and move forward with the optimization problem. I finally figured out why it matters in the second semester of Gary Becker's price theory course. What goes into the utility function matters for our understanding and ability to make predictions. For example, if people learn over time--a kind of feedback--or change what they value in a way that is not random, we as economists will be better able to explain what is going on.

As a simple example, suppose people simply have a taste for prejudice against those who do not look or behave like them. We could take that preference--that people simply get pleasure out of disliking others, say blacks or people who look Middle Eastern or women--as given; and we'd actually do a pretty decent job of explaining racial, cultural, and gender discrimination in the workplace.

However, we could probably do a better job as economists by noticing that over time, people become more familiar with and more tolerant of those who are different from the primary surrounding culture.

You can think of this as shrugging your shoulders and saying that tastes simply change over time, and we as economists have nothing to say about how or why. Preferences are simply data we take as given.

Alternatively, you can work on a way of modeling it--say, by noticing that the degree to which any particular prejudice is weighted in the utility function depends on the relative population of the outside group to the whole.

And as a deeper alternative, you can try to figure out what variables really are inside the utility function in the first place. Maybe it's not that people have a taste for discrimination at all, but that people have a taste for sameness. Modeling or thinking about a taste for sameness results in a much richer economic theory, suggesting that people also enjoy habits like the same toothpaste when they wake up each morning, or that they have a taste for avoiding risk, etc.

In summary, it depends on what you are trying to explain. For many questions, it's fine to take tastes as given. For questions of changing behavior over time, it's rather interesting and helpful to think about any observed regularities in how people's tastes are formed.

Tracy W writes:

How does modelling or thinking about a taste for sameness result in a much richer economic theory? What does that tell me about the allocation of resources that just postulating learning costs doesn't?

And, stopping being an economist for a moment, I'd say that the postulated theory that people become more familiar and tolerant of those who who are different from the surrounding culture is wrong, or at least incomplete. For example, the Israelis and Palestinians have had over 60 years to become tolerant of each other, but it does not appear that they actually have become more tolerant of each other.
I am a NZer, and can travel the world in a relaxed way, I think in part because people don't know much at all about NZ.
In Frederick Douglas's autobiography (he was a black man born into slavery in the Southern USA who became an abolitionist and escaped and became a famous abolitionist), he talks about how wonderful it was to travel in England and be treated as an equal - when immigration by blacks to the UK rose in the 1950s said blacks reported the opposite of wonderful treatment.

I find psychology and sociology interesting in their own right, I have no objection to understanding where people's tastes come from, but I think part of the value of economics is that you can have an understanding of the world that is generic - eg you can think about the best way to save fish species from overfishing separately from why we would want to save fish species from overfishing.

Comments for this entry have been closed
Return to top