Almost twenty years ago - in 1983 to be exact - Glenn Withers and I undertook what we understand to be the first ever application of contingent valuation methods (CVM) to the arts. We carried out a randomsample survey of the adult inhabitants of Sydney which sought to measure the community’s willingness to pay (WTP) for the perceived public-good benefits of the arts. Around 825 respondents were questioned about the nature and extent of the nonmarket benefits they enjoyed from the existence of the subsidised arts in Australia - literature, visual arts, music, theatre, dance etc. - and they were asked to nominate the dollar amounts they would be willing to pay out of their taxes to support the arts, under conditions of both liability and nonliability for actual payment. With appropriate caveats, we concluded from our research that aggregate WTP for the public-good benefits of the arts in Australia at that time exceeded the thenprevailing tax-price of cultural subsidy.
What did we think we were measuring in this study and what did we actually measure? As far as the arts were concerned, our work was predicated on two principal motivations, one theoretical and one practical. The theoretical drive came from a desire to test the longstanding proposition that the arts were a case of market failure. This hypothesis, first articulated in the 1960s (Baumol and Bowen, 1966; Peacock, 1969) and elaborated at length in our own book of 1979 (Throsby and Withers, 1979), remains to this day a cornerstone of efficiency-based arguments for public support for the arts, yet at the time of our early 1980s study it had remained empirically unexplored. The practical motivation for our work sprang from the political and economic trends affecting Australian cultural policy at the time: a sense that the arts needed to demonstrate their economic importance, and the fact that public expenditure programs were coming under sharper scrutiny in times of increased budgetary stringency.
We used the rhetorical question "What Price Culture?" as the title for the report on our study that was written for a popular audience (Throsby and Withers, 1984), reflecting our belief that we had indeed been able to place an economic value on the nonmarket output of the arts. We argued that art has its price: for those producing it, for those consuming it for their private enjoyment, for those making voluntary donations to support it, and for those required to contribute to it by way of compulsory taxation. The prices received or paid by the first three groups could be readily observed; our work, we suggested, rounded out the picture by placing a value on those benefits not captured in market transactions.
Nevertheless, despite any satisfaction we might have felt at having brought art so decisively into the economic calculus for the purposes of formulating cultural policy, some doubts have lingered. True, our work did seem to help make some sense of arts support programs in Australia in the years that followed, and our study was replicated in at least one other country. But the view has continued to be expressed by politicians, artists, cultural theorists and others, whenever economists try to estimate the worth of cultural goods, that ultimately the value of art cannot be expressed in monetary terms. Are these doubters simply ignorant of the power of economic analysis, are they trying to impose their own preferences on the rest of us, or are they expressing something that we as economists would rather not contemplate? These are fundamental questions which I have raised here in reference to one particular study, but which could equally be asked in connection with any of the multitude of CVM applications to culture which have appeared over the last decade or so.
In this paper, I want to explore some of these issues. What does CVM tell us about the value of cultural goods and services, and what does it leave out?