Here is another in the string of posts inspired by my weekly class on cost-benefit analysis. Last night’s topic was stated preference methods, like contingent valuation. These are controversial because they are often used to put prices on things people don’t normally think of as having prices, like the “existence value” of whole species, pristine natural environments or the avoidance of risks to public health.
My view is that many of the confusions in economics can be traced to ambiguities in language. We often use words to mean multiple things and then try to apply what works for one meaning to a different meaning, where it doesn’t. Case in point: preferences. I prefer A to B means I want state-of-the-world A to occur rather than state-of-the-world B, whether A and B are two pairs of shoes that could sit in my closet or two destinies for wild salmon along rivers that drain the Olympic mountains in northwestern Washington State. They are similar in the sense that both pertain to my wanting something, but they are also different.
I propose two kinds of preferences based on different motivations. One I will call normative; this reflects my judgments regarding what I deem to be right or wrong. The other is experiential, what I want based on how I would personally benefit from it. Economists sometimes say that ethical judgments are essentially experiential, since you derive pleasure from seeing right triumph over wrong, but I disagree. Experiential preferences cause you to want A over B because A makes you happier or gives you more “utility”. Normative preferences give you happiness or utility if a choice process selects A, and you believe A is ethically preferred to B. These are clearly not the same thing. In the first case utility is a cause, in the second an effect.
An example of a fundamentally normative preference is the one exercised by a jury deliberating a civil or criminal dispute. It would be absurd to have verdicts determined by jurors expressing a willingness to pay to convict or acquit, and then adding up the totals to see which is greater. This is because juries are supposed to deliberate based on a conception of justice, not on what’s in it for them, personally. An example of a fundamentally experiential preference is the question of whether to publicly subsidize a sports stadium in a city. Taxpayers’ preferences will be based on the degree to which the team that plays in the stadium gives them some sort of personal excitement, satisfaction or pride. This could well be captured by a technique that measures their willingness to pay for the stadium.
Of course, preferences that are primarily normative can have a secondary experiential component, and vice versa. In the stadium example, for instance, one effect of a subsidy is to transfer public money to private investors in professional sports teams. This has an ethical aspect, which may play a role in how preferences are established. In fact, in a society with glaring shortfalls in public programs for health, education and other essential services, like Brazil, the ethical component may become primary, as we saw in the protests over the World Cup. Where public funds are not so constrained and the gaps not so severe, the decision turns on what the local population expects to derive, personally, from better facilities for professional sports, and questions of ethics are secondary.
Most existence values for environmental goods, I would argue, are essentially normative preferences. They are about what people believe to be right or meritorious, not what gives them personal satisfaction. Willingness to pay in these circumstances makes about as much sense as a decision tool as it does in jury trials. We might be misled by elements of experiential preference that enter the mix, but our well-being as members of a society that makes choices of this kind is an effect, not a cause of what we wish to see happen.
If this analysis is correct, CBA can help us put numbers on the experiential aspects of a policy choice, recognizing that some other process is needed to assess its normative merits.