Federal 
Why Revealed Preferences Beat Stated Preferences:
The case of compact fluorescent light bulbs
30 Apr 2007 in Regulatory Economics
Both "revealed" and "stated" preferences find their way into benefit-cost analysis. "Revealed" preferences are obtained from real-world data showing what people actually do. "Stated" preferences" are obtained from surveys showing what people say. It's an old maxim that what people say isn't what they actually do. That's why economists are In benefit-cost analysis, stated preferences are useful only if they adequately mimic revealed preferences.skeptical of pubic opinion polls. So surveys that measure, even accurately, what people say are only useful if they mimic what people actually do -- or would do, if there were markets.
The battle of the sexes over compact fluorescent light bulbs confirms that economists' bias in favor of revealed preference is justified.
Paraphrasing Art Linkletter, people say the darndest things when they know the whole world is watching. That's especially true when there are cultural mileposts involved. Today, it has become declasse to question global climate change, or to doubt the relative magnitude of human influence on climate, or to profess not to care about such things. For that reason, everyone is in favor of energy conservation, including the use of compact fluorescent light bulbs (CFLs).
CFL advocates, such as former vice president Al Gore, have made the case that CFLs save money, and there is little evidence that the public does not believe that these cost savings are real. Advocates also say that the light output from CFLs is as good or better than from incandescent bulbs. It's hard to find a technical review that says anything bad about CFL.
So why is it that actual CFL use in the US is so low?
Washington Post staff writer Blaine Harden reports that from Ground Zero of the battleground, CFLs don't pass "the wife test." Women didn't like early CFL models because they make skin look "pale, wrinkly and old." Harden says these problems have been overcome:
CFLs present an excellent example of how stated and revealed preferences can differ, and often by a lot. Women notice significant differences in light quality and are not willing to give up what they like just to save money, even if they say that they will when asked by pollsters. Men either don't notice the difference between incandescent and CFL lighting, or they notice but don't care. For men the tradeoff is worth it; for women it's not. (Wives win this domestic battle by default if their husbands really can't tell the difference. If their husbands do notice, wives resort to more aggressive tactics, beginning with pointing out that promises of equivalent lighting quality are provably false. Where logic fails entirely, other methods dominate, as Harden's story relates.)
This gap between stated and revealed preferences appears to be dramatic. Harden says only 6% of residential light bulbs in the US are CFLs. But according to the results of a Washington Post co-sponsored poll released last week, seven in 10 people surveyed said they had replaced incandescent bulbs with CFLs. It's plausible, but highly unlikely, that both statistics are true. It would have to mean that in the vast majority of US households CFLs are used rather sparingly. Which statistic is wrong is simple to ascertain: CFL sales statistics are observable in the market, whereas claims about having installed them are not.
The poll showed that hypothetical support for regulations mandating CFLs was decidedly weaker, with just 56% in favor and 41% opposed. It's unlikely that respondents displayed a uniquely libertarian streak when it came to CFLs. More likely, it's one thing to exaggerate your support for something widely regarded as socially responsible so long as it's purely hypothetical. But it's quite another thing if you think that legislators might take your socially responsible but hypothetical musings seriously. Actual support for a CFL mandate is surely much lower, and concentrated among those who have already installed CFLs.
Surveys systematically overstate public support for regulatory mandates. That's because they obtain opinions, not enforceable commitments. Of the five hypothetical regulations "tested" in the Washington Post poll, support was correlated with the extent to which someone else besides the respondent would bear the burden of changing behavior. 82% favored requiring supermarkets to provide paper bags, but it's unclear what respondents would be willing to pay for this change. (If compliance is free, then the only thing interesting about thus result is that 16% opposed it.)
75% favored mandatory recycling of some items, but 70% said they already did recycle, so the poll question should be understood as a mandate imposed on other people. 71% and 59%, respectively, favored requiring low-flow toilets and low-flow shower heads. Both have been federally mandated for years, though only 3% and 1%, respectively, knew this. In any case, low-flow shower heads can be easily sabotaged. The ability to evade or avoid a regulation increases support for it.
Stated preferences are generally unreliable guides to what people actually do, or would do if given the chance. The case CFLs is an obvious example. Every husband who's installed CFLs knows this very well. As does every wife who's quietly replaced them with incandescants.
Washington Post-ABC News Poll: Environment Trends
Friday, April 20, 2007
Washington Post
...
38. Do you use any compact fluorescent light bulbs in your home, or not? Those are the ones that last longer but cost more than regular light bulbs.
Yes: 70%
No: 28%
NA: 2%
39. Would you support or oppose a law in your area (READ ITEM)? How about a law (NEXT ITEM)?
...
d. requiring the use of compact fluorescent light bulbs
Support: 56%
Oppose: 41%
NA: 3%
