Any economic examination of the business of the performing arts inevitably must deal with the phenomenon known as "cost disease." This was first described by William Baumol and William Bowen [1965, 495-502; 1966, 161-172](1) and refers to an inherent inability of performing arts firms to enjoy productivity increases from labor, resulting in continuously rising costs relative to revenues.
To be more precise, cost disease is a consequence of increased productivity and wages in the manufacturing sector in combination with an absence of higher productivity in the performing arts. Using a crude definition of product for the moment, the performance of a Haydn symphony requires the same number of musicians working for the same amount of time as it did when the piece was first written 200 years ago. Therefore, the only way that the marginal revenue product of the musicians could increase as it has for factory workers would be via increases in marginal revenue for the output.(2)
Once compensating wage differentials between employment in fine arts and manufacturing are exhausted,(3) one or more of three things will occur: the orchestra will be starved of its labor supply, it will be increasingly reliant on patronage in order to be able to continue to produce, or it will have a constantly rising output price (relative to the factory's output price) and falling employment level-the sector will shrink away.
The purpose of this paper is to examine how the problem of cost disease might be solved on the demand side.(4) Over the course of the analysis, a dichotomy between two basic approaches (which I refer to as "Veblenian" and "Marshallian" for reasons that will presently be made clear) will become apparent.
While this paper is limited to the performing arts industry, it should be noted that the issue of low or zero productivity growth has been extended by Baumol [1989, 611-615, for example] to the services sector in general. Thus, the conclusions reached here may be useful in a more general setting.
The next section of the paper describes ways in which demand can be changed, while the following two sections explore "Veblenian" and "Marshallian" cures for cost disease. The fifth section examines approaches to demand modification currently in use in the United States and Europe. Finally, the last section draws conclusions from the preceding analysis.
If demand for the performing arts is assumed constant, there will be an eventual loss of the labor force, or increasing subsidies will be required to cover a growing wage bill. This will be the case if cost curves are not allowed to rise; rising relative output prices coupled with lower employment will result if rising costs are passed on. If demand is no longer held constant, all of these results need not necessarily obtain: demand increases, if sufficiently large, could lead to increases in revenue that match or even exceed the increased costs.
The three most likely ways that demand for the performing arts could increase would involve increasing the number of consumers, increasing "taste" for the performing arts, and increasing consumer incomes. The way in which an increased number of consumers or a greater taste for the performing arts would affect demand is self-evident. The effect that increased consumer incomes would have over the demand for the performing arts depends, of course, on the income elasticity of demand for the arts.(5) A reasonable guess is that the magnitude of this coefficient is somewhere above unity, since the performing arts could hardly be construed to be a necessity by very many. However, as both Glenn Withers [1980, 735-742] and David Throsby [1994, 1-29] point out, enjoyment of the performing arts is time-intensive. This means that increasing demand stimulated by rising incomes might be diminished to some extent if the opportunity cost of leisure increases (an assumption that is itself open to doubt). Whereas a "pure" income effect might well be above unity, that which is actually seen econometrically might not be. Indeed, Withers calculated this coefficient to be just about exactly 1, while other estimates have placed it both above and below 1. For purposes of argument, let us assume that the coefficient is somewhere above unity.
Although it does not cause changes in demand, another factor that should be mentioned that could mitigate the ill effects of cost disease is relative price inelasticity. If the price elasticity of demand for the performing arts is low, then wage increases can occur and be passed on without much shrinkage of the sector.(6) While the price elasticities of demand for different parts of the performing arts sector have generally been estimated to be less than unity (Withers  found it to be about -0.65, while T.G. Moore [1966, 79-87] estimated it at -0.63), it would still be beneficial to firms that this elasticity be somehow minimized (in absolute terms). How this can be done is addressed in the next section.
A "Veblenian" Approach
One approach to the problem of cost disease involves an increase in demand via increases in consumers of the performing arts and greater expenditure through increased share of rising income spent on the arts. Specifically, the industry should find some way to portray its product as an attractive, desirable luxury for those who might not currently see it as such, which in addition to bringing in additional consumers, would also increase the income elasticity of demand for the performing arts. New demand would be stimulated, and subsequent income increases would have their greatest effect.
What tactics might be used? A reading of Thorstein Veblen's The Theory of the Leisure Class  suggests the enhancement of certain characteristics that the performing arts possess in abundance. In two famous passages, Veblen alludes to the particular qualities which a luxury good should have:
Conspicuous consumption of valuable goods is a means of reputability to the gentleman of leisure [1899, 75].
The point of material difference between machine-made goods and the hand-wrought goods which serve the same purposes is, ordinarily, that the former serve their primary purpose more adequately . . . This does not save them from disesteem and depreciation, for they fall short under the test of honorific waste. Hand labor is a more wasteful method of production; hence the goods turned out by this method are more serviceable for the purpose of pecuniary reputability . . . Commonly, if not invariably, the honorific marks of hand...