Do people know what they want to know?

Lene, one of my cobloggers at A Better Earth, points to an interesting paper called “Sentiments and Acts Towards Genetically Modified Foods.” The study examines consumer behavior in the Netherlands after the government required foods with genetically modified ingredients to be labeled as such:

Because the Netherlands was one of the first EU countries to enact mandatory labelling, it provides an excellent case study for examining European consumer response…

Based on the results of the various attitude surveys carried out during the period of interest, one would expect a sizeable market segment (anywhere from 44% to 85% of Dutch consumers) to avoid GMFs if confronted with the choice. Likewise, one would expect Dutch consumers to be armed with information about GMFs based on their high level of awareness, knowledge, and intense media coverage. As indicated earlier, Dutch consumers were indeed confronted with such choice for a meaningful number of processed foods. The question then is how they actually chose?

What do you think they did?

Netherlands consumers did not significantly change their purchasing behaviour towards foods that received labels indicating the presence of GM ingredients. Nor did they alter their purchasing behaviour towards such foods after the labels were removed nearly three years later. There are no abrupt adjustments or gradual shifts away from GMFs.

That’s fascinating. The implication is that consumers responding in surveys are not good at predicting what kind of labeling and information they will actually demand in the marketplace. Yet those kinds of surveys are what justify a wide variety of regulations. Origin labels are being considered for meat because consumers supposedly want to know where their meat is coming from. Diageo says alcoholic beverage makers should be forced to put nutritional information on their bottles because customers demand it. Smoking ban supporters defend their policies on the grounds that bars and restaurants refuse to go smokefree in the face of widespread desire for such establishments. Are those expressions of consumer sentiment reliable?

I’ve argued in the past that they are not (meat labels here, alcohol here, smoking bans here). I believe that these cases of “market failure” are actually cases of “introspection failure.” Responding to survey questions removed from any actual trade offs, people overestimate their desire for information and their aversion to bad things like smoke or genetically modified foods. Their real, revealed preferences in the market are actually much more moderate. In the paper above, even when the costs of avoiding GMFs are virtually zero, consumers didn’t do so. If that’s a typical outcome, regulations based on publicly expressed opinions are bound to exceed what consumers actually demand, needlessly increasing costs or reducing consumers’ options.

This paper only applies to GMFs, but the authors provide a thorough discussion of the many methodological problems that undermine consumer attitude surveys. I urge you to read the whole thing and decide if they can plausibly be extended to the cases above. I think that they can.

So in summary of many blog posts from the past few years, here are three reasons the nanny state grows beyond what consumers really want:

1) People are ignorant of their actual preferences. (If you’re not facing real trade offs, it’s hard to imagine what you’ll actually demand. Introspection is a poor and easily biased guide.)*

2) People want to hold “good” preferences. (If supporting smoking bans, nutritional info, or GMF labels is the politically correct thing to do, you’ll respond that way in a survey, but not when those beliefs carry personal costs.)**

3) People will gladly externalize the costs of their preferences onto others, exaggerating them if they’re weakly held. (If you don’t like smoking, you may not be willing to bear the cost of going to the bars that don’t have it. But if by passing a ban you can impose the costs of your preference onto other people, you’ll exaggerate your aversion to smoke.)

What do you think?

* I’m treating revealed preferences as true. Alternatively, one could say that introspective preferences are true and consumers inexplicably ignore them at the grocery store, even when the monetary and informational costs are very low. That doesn’t strike me as a very sensible or useful interpretation.

** See Bryan Caplan’s excellent book The Myth of the Rational Voter for a lot more of this kind of analysis of democracy.

Update 8/31/07: See also Frank Newport of the Gallup Poll expressing similar doubts about whether consumers will really avoid Chinese products. Hat tip: Caleb Brown.


7 thoughts on “Do people know what they want to know?”

  1. I think you might be missing one reason. This might be nothing more than an idiosyncrasy of the human race, but I think we generally get greater happiness and a more “full” enjoyment of our consumption if we are more informed about our consumption. We may not plan on making purchasing decisions based on information we receive about the products, but we nevertheless want the information because it will make us happier with our eventual use of the product. Also, perhaps we feel better knowing that we’ve made an informed decision. If I’ve carefully researched the available products in a particular market and ultimately decided with certainty that one is the best for me to purchase, I am likely to be more pleased with my purchase than if I buy the same product but am never sure whether there is something better out there.

    Perhaps those aren’t valid concerns to an economist–they probably have little to do with maximizing society’s efficiency or wealth or whatever it is economists ultimately care about, but they do have the potential to increase happiness. For me, that’s a better goal.

  2. I agree, and I think that’s why detailed labeling has become such a notable trend for lots of foods. But there’s still an optimal amount of information that people desire and I’m not convinced that government regulators are somehow more in tune to what that is than markets are.

    Mandating information might even be detrimental. Right now, I can use the fact that some companies offer detailed information about the source of their products as an indication of high quality. If even the cheap-o products had to do that, that useful signal would be diluted.

  3. I’m not sure there is an optimal level. As long as the information is provided in a well-designed way (something that government is certainly unlikely to effect), I don’t think there is any ceiling to the greater happiness provided by more information about the products.

    I’m personally okay with simply mandating that a minimal level of information be available in proximity to a product-on-sale (i.e., ingredients), as well as slightly more detailed information being available upon request (nutritional information, origin), or on the company’s website, of any American food company. But I’d really like it if companies provided a whole lot more. But I do agree that the market is the best way to ensure that the information provided is provided in a nice way, from an information design perspective.

  4. Honestly, how costly is the labeling process anyway? Do labeling requirements (especially labels that don’t require a detailed scientific analysis of the product) really represent that much of an added cost? I’m curious about this…

  5. I can’t give you a detailed answer, but from what I understand the cost would be in the supply chain. If you’re using lots of uniform grade ingredients from lots of different sources, keeping everything separate and organized through shipping, storage, processing, and packaging could be a major imposition.

  6. I find your first asterik especially intriguing. I’m slightly more inclined to say introspective preferences are more accurate, but that there is some unknown barrier that keeps them from becoming revealed (money, information, other?). But I’ve got nothing to back that up.

    Even more likely, I’d say, is that preferences are generally crap, and that people are terrible at predicting/knowing what they really want/what will make them happy. Rational actors my foot, I say.

  7. Sorry to say. Caplan’s book is full of illogical and contradictory arguments, mangled terms, cultural prejudice, and a whole lot of other weaknesses. It’s also pretty scary when you really think about what he is arguing for. Like a lot of cloistered academics, he’s hermetically sealed inside his own thinking and theories, and totally unhinged from the real world… past and present. I won’t recap the whole list of objections here… but it’s on my site. (

Comments are closed.