In my last post I made passing reference to David Throsby’s “Determining the Value of Cultural Goods: How Much (or How Little) Does Contingent Valuation Tell Us?”, from the Journal of Cultural Economics, 2003. But it is a really important short paper, and so I’d like to take a deeper look at it.
If you take a course in cost-benefit analysis, one of the first things you learn is that for ordinary goods traded in competitive markets, your best guide to their value is the going market price. The cost of a construction project is what you will have to pay to build it. The benefit is what people have shown is their willingness to pay in the market for the goods and services it will provide. Hayek’s great insight is that prices are a source of information - a very useful thing for the policy-maker.
Sometimes prices do not work so well, and a correction is needed. For example, if one of the contractors on your project is a monopolist, who has the market power to charge you more than their actual costs at the margin, then some of what you are paying them is the true cost of resources you are using up, but the rest of what you are paying them is in effect a pure transfer payment they are able to gain as a monopolist. So, in social terms, the cost of what they are supplying is something less than what they are charging. But that will come up in your cost-benefit textbook as well, maybe in chapter two.
But what if your project involves some good that simply has no available market price? What is the value to the public of preserving the forests and rocky shores of the Schoodic peninsula in Maine from industrial development? What is the value to the people of Bloomington to be able to attend Shakespeare in the Park every summer, for which no tickets are sold?
One solution is to just ask people, through a poll: what would you be willing to pay to ensure that Schoodic stays free from development? What would you be willing to pay to ensure Shakespeare in the Park can continue? This method is known in the trade as “stated preference”, or “contingent valuation”, the latter name coming from the idea of what people would pay if there were in fact a market.
What could go wrong? Two things.
The first is that the answers people give will, in general, not be very good, even if they are trying their best. I can tell you what I am willing to pay for ordinary goods that I have bought in the past, reasonably accurately. I go to the grocery store a few times a week, and will pick and choose amongst the produce according to “that is a good price”, or “that’s not really worth it this week.” Even for infrequent items, say looking to buy a Camry or Accord of an age and mileage, I have some idea of what is worth it and what is not. But contingent valuation is asking me to put a number on something which I haven’t really thought about before. I might give a number that just sounds sort of right - ten dollars? - without considering the magnitude of what is being proposed. My answer might change according to how the poll-taker has framed the question: did they mention other possible uses of my money? Did they make clear exactly what was going to be financed? Diamond and Hausman spell it out: you are simply not going to get very accurate numbers. Not because people are dumb. But because you are asking a really difficult question, although it might mask itself as a simple one.
But Throsby has a different take. He know contingent valuation models well, and has done some himself for arts funding in his native Australia. His issue is: what if, contra Diamond and Hausman, you could get very accurate takes on what the public is collectively willing to pay to fund the arts, or to preserve a cultural or environmental monument. Should that number guide the “benefit” side of your cost-benefit analysis? He says, no:
… economic value – which is not synonymous with financial or commercial value, although it is ultimately expressible in terms of either a numeraire good or (preferably) money – comprises any direct use values of the cultural good or service in question, plus whatever nonmarket values it may give rise to (which may perhaps be amenable to evaluation by CVM). Cultural value on the other hand is multi-dimensional, unstable, contested, lacks a common unit of account, and may contain elements that cannot be easily expressed according to any quantitative or qualitative scale.
The characteristics of cultural goods which give rise to their cultural value might include their aesthetic properties, their spiritual significance, their role as purveyors of symbolic meaning, their historic importance, their significance in influencing artistic trends, their authenticity, their integrity, their uniqueness, and so on …
and from his conclusion:
… citizens may be asked in political referenda how much they would hypothetically be prepared to pay for this or that budgetary measure, and in responding they may well account for a range of intangible effects in expressing their trade-offs. My purpose has not been to suggest that such measures are a waste of time, or that well-conducted CVM studies of cultural goods tell us nothing about consumer preferences. Rather it has been to argue that if we go beyond the day-to-day world of practical politics or empirical number-crunching, there are fundamental issues at stake concerning the true value of cultural goods and how that value should be constructed.
Even so, there is a legitimate question to ask at the end of all this, namely: So what? Suppose there are other sources of value that are not captured by CVM or any other methods in the economist’s assessment of the value of cultural goods. Do they matter for economic decision making? Since both public and private decisions in the cultural sphere ultimately come down to questions of resource allocation, where the resources have opportunity costs, isn’t a realistic assessment of the economic value of cultural goods all that counts?
This question could be addressed by putting it in a more concrete form. Suppose a cultural policy-maker has a choice between two projects involving the renovation of two different heritage sites, each project having a capital cost of $10 million. Only one of the projects can be undertaken because the budget constraint is exactly this amount and neither project is divisible. The benefit-cost ratio of project A, counting in all market and nonmarket effects correctly measured, is 1.1, that of project B is 0.9. By some means an independent assessment of the cultural value of the projects is obtained, which shows unambiguously that the cultural value of project A is low and that of project B is high. Accepting the validity of these measurements, which project is the policy-maker to choose? An economic criterion would suggest project A, a cultural one would indicate B. The tradeoff between the two sources of value can be framed either as indicating the economic price that would have to be paid to achieve a culturally desired outcome, or conversely the cultural price that would have to be paid to achieve an economically desired outcome, in choosing either project over the other.
The point of this illustration is not just to make the fairly trivial point that tradeoffs are inevitable in this life, but rather to advance the more significant argument that cultural value, for all its ephemeral, shifting, incoherent and even irrational properties, is likely to influence peoples’ decision-making in regard to cultural goods and might therefore affect desirable patterns of resource allocation in this area in ways that cannot be fully captured by standard economic analysis. If this is so, there is a challenging task ahead, namely to work out whether methods such as CVM and other approaches can be extended to account for these wider dimensions in their application to art and culture, or whether entirely new techniques of measurement, perhaps adapted from other fields, need to be developed.
In a nutshell: market prices, even hypothetical ones developed by the very best contingent valuation methods, ought not to be our sole guide to evaluating the true and complete value of current cultural offerings or cultural preservation.
This is a big deal. For it takes the whole field of cultural economics (and it is in the house journal for the field that this article is published) and reminds it: economics is a limited, partial perspective on the value of culture, and the ordinary economic methods used in policy analysis are a limited guide to cultural policy. It is a challenge to the utilitarian foundations of economics.
I remember when he first presented this paper at a conference, and at the time I wasn’t really persuaded: if we were not going to use the public’s stated values to guide us in policy decisions in cultural spending and investment, what were we going to use? If the art history scholar says “the public has got it wrong”, should that affect the policy-maker’s decision? But I’ve come around to it - this is a paper that changed how I see things, but slowly, to the point that my own book concludes that the arts policy-maker cannot simply rely on measurements of economic value.
I suppose the best commentary to this article would be W. B. Yeats To a Wealthy Man Who Promised a Second Subscription to the Dublin Municipal Gallery if it Were Proved that People Wanted Pictures