William Baumol is the dean of cultural economists. His seminal book Performing Arts: the Economic Dilemma (1966), written with William Bowen first introduced the concept of 'cost-disease' in the performing arts. 'Cost disease' is caused by the inability of the performing arts to achieve the productivity gains of other industries. This results in ever-increasing relative costs of producing performances as arts institutions must compete with wages paid in the more productive sectors of the economy, yet cannot benefit from similar productivity gains themselves. Many studies of orchestra and performing arts finances draw on his concepts as they seek a solution to the seemingly overwhelming rates at which expenses increase.
In 1996, Dr. Baumol revisited his early work in an article for the Journal of Cultural Economics, entitled "Children of Performing Arts, The Economic Dilemma: The Climbing Costs of Health Care and Education". At the conclusion of this article, he finds hope for the performing arts, health care and education in the ever-increasing wealth of the economy as it benefits as a whole from productivity gains. He proposes the thesis that as the country grows wealthier, the relative cost of many things decreases, making the so-called "stagnant industries" relatively more affordable, despite the increasing costs.
We are not able to reproduce the entire article, but it is fascinating and well worth a database search at your local library, as much for its discussion of health care and education as for its treatment of the performing arts. A few key excerpts follow:
Children of Performing Arts, The Economic Dilemma:
The Climbing Costs of Health Care and Education
WILLIAM J. BAUMOL_
C.V. Starr Center for Applied Economics, New York University and Princeton University
It is thirty years since William Bowen and I (1966) first reported our analysis of what
we called the “cost disease of the performing arts,” with its profoundly disturbing
presage of the future. Soon after, we realized in general terms that it applied to a
variety of other services (we now call them “stagnant services” including health
care, education, restaurant services, library and legal services, police protection,
auto repair and a number of other services). The cost predictions that flowed from
the analysis have been amply confirmed by subsequent developments.
-----
The issue is surely complex, and no single explanatory hypothesis can pretend
to account for a set of problems whose roots are undoubtedly sociological and
psychological as well as economic. Still, there is one influence that goes far in
accounting for the difficulties that have just been described, and that at the same
time at least suggests the general directions one must pursue if a way out is to be
found. It should be kept in mind that the following discussion focuses throughout,
not upon the level of costs, but upon their real rates of increase.
------
We traced the cost disease, whose empirical history in health care and education
have just been described, to the differences in the productivity growth rates of the
various parts of the typical developed economy. It is productivity growth that, on
this view, creates both private affluence and public squalor, and it is not by mere
happenstance that at the same time, in the words of the poet, “wealth accumulates
and men decay.” For it is inherent in the technological structure of the economic
growth process that a particular set of economic activities, many of them the very
activities that are generally considered most critical for the health of the society,
are condemned to a pattern of spiraling increases in their real prices that appears
to put them beyond the reach of both the individual and the state.
It is hardly surprising that, while overall productivity in the industrial world has
been growing rapidly, the pace of growth in different industries has varied substantially.
What is more unexpected is the persistence of the pattern of differences
in productivity growth between economic sectors. A given sector of the economy
does not usually fluctuate haphazardly between periods of relatively slow and relatively
rapid advance in productivity. Rather, the industries in which productivity
has expanding slowly a century ago are, by and large, the very ones that are still
the laggards today. And the endurance of productivity stagnancy in those industries
was imposed upon them a distinctive price history that is the fundamental symptom
of the cost disease of the personal services. This cost disease phenomenon
occurs when the services, in a class that will presently be described, are plagued
by cumulative and persistent rises in their costs, increases that normally exceed to
a significant degree the corresponding rate of increase for commodities generally,
i.e., almost always outstrip the economy’s rate of inflation.
The common element that characterizes all the stagnant services is the handicraft
attribute of their supply processes. None of them has, at least so far, been fully
automated and liberated from the requirement of a substantial residue of personal
attention by their producers. That is, they have resisted reduction in the amount
of labor expended per unit of their output. Not that the growth rate of their labor
productivity has always been zero. On the contrary, in almost every case there has
been some rise in the productivity of these personal services with the passage of
time; but over longer periods it has been far slower than the rate of productivity
increase characteristic of the economy as a whole. That is why we call them the
stagnant services.6
There are at least two reasons why rapid and persistent productivity growth
has eluded the stagnant services. First, some of them are inherently resistant to
standardization. Before one can undertake to cure a patient or to repair a broken
piece of machinery it is necessary to determine, case by case, just what is wrong
and the treatment must then be tailored to the individual case. The manufacture
of thousands of identical automobiles can be carried out on an assembly line and
much of the work done by industrial robots, but the repair of a car just hauled to
a garage from the site of an accident can not be entrusted completely to automated
processes. A second reason why it has been difficult to reduce the labor content of
these services is the fact that in many of them quality is, or is at least believed to
be, inescapably correlated with the amount of labor expended on their production.
Teachers who cut down the time they spend on their classes or who increase class
size, doctors who speed up the examination of their patients, or a police force that
spends less time on the beat are all held to be shortchanging those whom they serve.
This, then, is why the stagnant services have consistently proved unamenable to
steady and substantial productivity growth, that is, to reduced labor content. To see
the implications for costs and prices, let me return to a (slightly edited) quotation
from our earliest description of the relationship (Baumol and Bowen, 1966, pp.
167–171).
Let us imagine an economy divided into two sectors: one, the progressive
sector, in which productivity is rising, and another, the stagnant sector, in which
productivity is constant. Suppose the first economic sector produces automobiles,
and the second, performances of Mozart quartets. Let us assume that in automobile
production, where technological improvements are possible, output per work-hour
is increasing at an annual rate of 4 per cent, while the productivity of quartet players
remains unchanged year after year. If wages in the automobile industry match the
rise in productivity, then the one effect on cost is exactly offset by the other – total
cost and output both rise by the same percentage. As a consequence, labor cost
per unit (the ratio between total labor cost and total output) remains absolutely
unchanged. This process can continue indefinitely, with auto workers earning more
and more each year, but cost per car remaining stationary. Matters are very different
in quartet performance. Suppose that the quartet players’ wages, though below that
of the auto workers, maintains its relative position. What does this situation imply
for the costs of quartet performance? If the earnings of string players increase, say,
by 4 per cent per year while their productivity remains unchanged, it follows that
the direct labor cost per unit of their output must also rise at 4 per cent. Moreover,
there is nothing in the nature of this situation to prevent the cost of performance
from rising indefinitely and at a compounded rate.
Ordinary price inflation plays no role in the logic of our analysis. So long as the
wages of musicians in this two-sector economy continue to increase at all, the cost
of a live performance will rise, cumulatively and persistently, relative to the cost of
an automobile, whether or not the general price level in the economy is changing;
the extent of the increase in the relative cost of the performance will depend directly
on the relative rate of growth of productivity in the automobile industry. Moreover,
though it is always tempting to seek some villain to explain such a cumulative run
of real price increases, there is no guilty party here. Neither wasteful expenditure
nor greed plays any role. It is the relatively stagnant technology of live musical
performance – its inherent resistance to productivity improvements – that accounts
for the compounding rise in the cost of performance of quartets.
It will be evident that the foregoing analysis is applicable to many other personal
services, and that is the burden of my story here. In particular, the services that
have been labelled “stagnant” all appear to have difficulties persistently impeding
growth in their productivity very similar to those that beset the musicians in our
parable. Clearly, health care has taken giant steps in quality improvement over the
decades, but while the amount of physician time spent per patient-visit or per illness
may have declined somewhat, it has done so only marginally; in education there
has been no marked change in class size, and therefore no large variation in number
of students served per teacher-hour, and it is widely judged that there has been little
if any improvement in quality. The output of an hour of police protection, or an
hour of postal delivery time, or an hour of street cleaning time has probably been
enhanced by the use of motor vehicles in terms of territory covered, but the increase
has probably been modest (criminal activity has also been “enhanced” by the use
of motor vehicles), and certainly has not been continuous and cumulative. The
productivity of trial lawyers and actors or musicians engaged in live performances
has risen to a minuscule degree at most, and while automotive repair services
have done somewhat better, the increase in their productivity has still been well
below that of manufacturing, as we will see presently. The circumstances of the
insurance industry follow directly from what has just been said, for the purchaser
of an insurance policy simply is acquiring a bundle of several stagnant services
– health care, auto repair, legal services, and so on, and as we have just noted,
productivity growth in the supply of this bundle has surely lagged. A final class of
stagnant activities to be noted here is particularly significant in terms of the state of
society. The care of the indigent, government welfare, and related programs seem
to benefit from no significant source of productivity growth – they appear to remain
fundamentally unchanged, handicraft activities.
The upshot is that all of these services suffer from a rise in their costs that is
terrifyingly rapid and frighteningly persistent.7 They threaten the strained budgets
of the individual families, the municipalities and the central governments of the
entire industrialized world. And, as financial stringency becomes more pressing, it
is understandable that spending on these services is cut back or, at most, increased
by amounts barely sufficient to stay abreast of the overall price inflation in the
economy. But since the costs of the stagnant services are condemned to rise,
persistently and cumulatively, with greater rapidity than the rate of inflation of
the economy, the consequence is that the supply of these services tends to fall in
quantity and quality. This undoubtedly is not the only source of increasing public
squalor, but it must surely have made a significant contribution.
-----
In A Connecticut Yankee, Mark Twain devotes an entire chapter to Sir Boss’s
unsuccessful attempt to explain the concept of real wages to his primitive hosts. He
argues with some passion that the monetary magnitudes of wages are irrelevant;
that, regardless of their value as expressed in terms of money, wages are really
higher only when it takes fewer hours of labor to earn the wages needed to purchase
a given set of goods. Yet, as I will shown now, precisely that is true of the stagnant
services. Their money prices are indeed rising ever higher, and their exchange rate
against manufactured goods is constantly increasing, just as I have shown here.
But in terms of the number of labor hours it takes to acquire them, over the longer
run, their cost is decreasing steadily, albeit relatively slowly. If so, it is immediately
obvious that the claim that we cannot afford them is simply a manifestation of what
economists call “money illusion.”
But how can that be? The answer is that even the most stagnant of services is
undergoing some productivity growth – slowly, and not very steadily, but some
growth nevertheless. The cost disease analysis does not claim otherwise; its workings
merely require productivity in the stagnant services to grow substantially more
slowly than that of the economy overall. To illustrate the point, let me return to my
favorite example. If, in the earlier parable, the hypothetical Mozart string quartet
had been scored for a half-hour performance, then its performance in 1990 required
two person-hours of labor, just as it did in 1790, when it might have been written.
Thus, there is apparently no scope for the slightest increase in labor productivity.8
Yet that is only an illusion. To see why, consider a recent performance by a Viennese
group of musicians played in Frankfurt am Main. A 1990 trip from their Austrian
home base to the German auditorium surely would normally have taken no more
than several hours. But when Mozart made the trip in 1790 it required six days of
extreme discomfort (and, at that, Mozart wrote that he was surprised at the speed
of the journey).9 Certainly, technical progress has reduced the number of hours of
labor required to provide a unit of the output in question, thus raising the labor
productivity of every itinerant performer, even in live performance (and we know
that performers are virtually all itinerant).
This example clearly suggests that there is no service whose productivity is
untouched by technical progress to some degree. The consequent rise in labor
productivity means, by definition, that it requires ever less labor time to produce
a unit of such a service. And every resulting reduction in labor time spent in
producing the service means that those purchasing the service must expend that
much less labor-time to acquire the wherewithal needed to purchase it. That is the
sense in which even education and medical care have really grown steadily cheaper
(albeit at a snail’s pace, compared to other outputs), even as they appear to become
steadily more unaffordable. Productivity growth in the stagnant services means
that their real costs are steadily, if slowly, declining despite the dramatic inflation
in their money prices. However, this, at best, can only make a minor contribution
toward solution of the politico-budgetary problems that stem from the cost disease.
More powerful aid must come from a second source:
PRODUCTIVITY GROWTH IN THE ENTIRE ECONOMY MEANS WE CAN AFFORD
MORE OF EVERYTHING
There is a good deal more to the sanguine side of the cost disease story. Even if
it were true that productivity in the stagnant services was not increasing one iota,
their rising prices could still not put them beyond the reach of the community; on
the contrary, it would remain true that society could afford ever more of them, just
as it has in fact been getting ever more of the health care and education that seem
steadily to become to an ever greater degree too expensive to afford.
As was pointed out some time ago by David Bradford (1969), in an economy
in which productivity is growing in almost every sector and declining in none,
it is a tautology that consumers can have more of every good and service. To
achieve this goal, some limited quantity of the inputs used to produce goods whose
productivity is growing relatively quickly (the “progressive” outputs) need merely
be transferred into the production of the stagnant services. Then productivity growth
will still permit expansion of the progressive output quantities, despite the limited
decline in their inputs, while the outputs of the stagnant service will grow because
more input is devoted to their production. To achieve such a goal – ever greater
abundance of everything – society must change the proportions of its income that
it devotes to the different products. In these circumstances, it is a fiscal illusion that
underlies the view that consumers as a group cannot afford to pay the rising costs
of education, health care, and other such services.
We can suggest the magnitudes that may be involved by using current U.S. data
on price trends and expenditures on health care and education to illustrate the point.
We will now see what would be entailed if the real prices of education and health
care continue to grow at their current rate for 50 years, if overall U.S. productivity
rises for that period at its historic rate of (approximately) two per cent, and if real
educational and health care outputs maintain an unchanged share of GDP – that
is, if the economy were to produce more of education, health care and everything
else, keeping their relative outputs completely unchanged. It should be emphasized
that the resulting numbers do not pretend to constitute anything like a forecast –
they are intended to be no more than a suggestive extrapolation. Or rather, they are
intended as an indication of what the economy will be capable of achieving for the
public, if historic price and productivity trends continue.
In 1990 actual shares of GDP constituted by health care and education constituted
less than 20 per cent of the total. We can calculate directly what will happen to
outputs if the number of hours of labor performed in the U.S. remains constant but
productivity in the economy grows at its historic average rate and each industry’s
output level is adjusted to retain the same share of total output. We find that in that
period the output of every good and service, including education and health care,
can increase to more than 3.5 times its 1990 magnitude.
We can also calculate the modification of the apportionment of real expenditures
required to achieve this result, in effect, measuring spending on each product by
the relative number of hours of labor it requires to earn enough to purchase it. Our
assumption that the total hours of labor does not change means that total spending
on GDP, measured in terms of labor hours used, must also be constant. However,
expenditure proportions will have changed drastically. Medical outlays, instead of
constituting 12 per cent of the total, as they did in 1990, must rise to more than 35
per cent of the total in 2040. And the share of expenditure devoted to education will
have risen from under 7 per cent of the total at the beginning of the period to nearly
30 per cent at the end. In other words, if current relative price trends and output
proportions continue as they are now, by the time four decades of the next century
have passed, education and health care alone will absorb well over half of GDP!
But that will not prevent consumers from nearly quadrupling their consumption of
each and every good and service, including manufactured necessities and luxuries
of every variety.
An analogy can perhaps make the sanguine character of the basic conclusion
clearer. Suppose we think of the public’s consumption of goods and services as the
purchase of a bundle containing many components, just as the purchase of a car
includes the acquisition of seats, tires, steering wheel, etc. Imagine that the price of
steering wheels is increasing at an impressive rate, but that because of the decline in
the costs of the other components, cars (equipped with steering wheels) grow less
expensive every year. Would one really conclude that steering wheels are growing
unaffordable, even when their price grows to 65 per cent of the price of the car?
------
Not the least of the remaining problems is the difficult educational task of getting
the public to recognize the difference between the reality and the illusion in the
behavior of costs. It will not be easy to convince the intelligent nonspecialist that,
even though prices of personal services appear to be rising at a rate that is out of
control, in fact the costs of those services (in terms of their labor-time equivalent)
are really gradually declining, because of increases in their labor productivity.
One can hardly blame such persons for their reluctance to be taken in by what
appears to be sleight of hand or mere theoretical gobbledygook. The difficulty of
convincing those who have not thought about the matter analytically that an item
really becomes cheaper when its price doubles if, at the same time, wages rise by
a factor of 2.1 is, after all, the main moral that Mark Twain seems to want the
reader to draw from Sir Boss’s failure in his attempt to make the matter clear to
his medieval hosts. Yet the task of explanation to the public should not be beyond
the most skilled of journalists and others who specialize in the art of effective
communication.
------
The popular answer of the hour to dangers such as the huge expansion of the
public sector just described is greater reliance on privatization. But privatization,
too, is no panacea. The public’s opposition to any threat to the survival of the
public school system can hardly be dismissed as groundless, and similar remarks
apply to a number of other services, such as police protection, currently supplied
by government. In addition, any industry beset by the cost disease that is in private
hands is sure to be suspected of greed and malfeasance. It is hard to believe
that calls for price controls to limit their cumulatively rising prices will not become
irresistible politically. But if the rising costs are caused by unavoidably slow growth
in productivity, price controls can confidently be expected to lead to deterioration
in the quality of those services or to their partial or total disappearance.
We must add to these considerations the likelihood that the cost disease affects
many services besides health care and education vital for quality of life. For the
same arguments apply to the live performing arts, to libraries, to police protection,
to restaurants, to welfare support for the impoverished and to many other critical
services. The implication is that if we do not think through the complex problems
just described, or fail to do so in short order, we face a society increasingly
characterized, in the words of Kenneth Galbraith, by private affluence and public
squalor. Already, unmistakable and disquieting signs are available for all to see.
Back to "The Facts about Orchestra Salaries"