The Economics of Superstars
This Quantum course on the "Economics of Superstars" is based on the following detailed article. It is provided here for those of you who wish to read further about Sherwin Rosen's theories.
Reprinted from The American Scholar, Volume 52, Number 4, Autumn 1983.
Copyright © 1983 by the author
The Economics of Superstars
WHO IN RECENT YEARS has not felt his gorge rise upon learning the staggeringly high salary of a shortstop, a movie star, an opera singer? A basketball player on a losing team earns $1.2 million, an author sells the paperback rights to his book for $800,000; a television interviewer switches networks and signs a contract calling, for her to receive an annual income of just under $2 million. And the gorge continues to rise. The spectacle of people doing work that doesn't always seem overweighted with significance for annual (and, in the case of rock singers, sometimes nightly) sums of money that figure to exceed what you and I may earn in our lifetimes-this, as they say nowadays, does not give off good vibes. What's going on here?
What we are talking about, of course, is the phenomenon of superstars, wherein relatively small numbers of people earn enormous amounts of money and seem to dominate the fields in which they are engaged. This phenomenon appears to be increasingly important in the modern world-certainly, with the breakdown of economic privacy, it is an increasingly visible phenomenon. The very word superstar implies inflation in our most precious currency, language; to be a star would have been sufficient in my youth. Yet we appear to be stuck with the term. As for the phenomenon itself, viewed from the standpoint of an economist, it may not be as puzzling as it at first glimpse seems.
The first thing to be said in this connection is that certain economic activities admit extreme concentration of both personal reward and market size among a handful of participants. Every economic activity supports considerable diversity of talent and significant inequality in the personal distribution of rewards. Activities where superstars are found differ from those in which most of us make our livings by supporting much less diversity and much more inequality, in the distribution of earnings. The bulk of earnings goes to relatively small numbers of practitioners-typically, the few regarded as among the best in their fields. Similar distributions of earnings in the industrial sector would ultimately come to the attention of the Federal Trade Commission or the Justice Department, but so far as I know, proceedings in restraint of trade were never brought against Caruso, Babe Ruth, Picasso, or the Beatles.
It is difficult for students of income inequality to
obtain systematic data on these
matters, because most royalty arrangements and salary negotiations are
private and confidentiality laws prohibit public disclosure. Still, the
phenomenon is obvious to everyone and is nowhere more apparent these
in professional sports. We are bombarded by stories of fabulous amounts
money earned in salaries, sign-up bonuses, pensions, and incentive
the contracts of professional athletes. A typical player in the
Basketball Association is reported to earn somewhere in the vicinity of
$250,000 per year in salary equivalent alone, and this figure will soon
$300,000. We hear of genuine superstars, towering figures as it were,
contracts guaranteeing numbers in the seven digits per year, and of a
significant fraction of not-so-esteemed players who earn in the high
baseball, winter now brings a different form of hot-stove league than
in the old
days. The revolution in the free-agency contract (which, by the way,
features in common with the breakdown of the old star/stock company
system in Hollywood of the 1930s and 1940s) raises speculations about
player might even be located next year, never mind the unbelievable
contracts negotiated by many players who have become free agents. How
be that many a mediocre free agent earns far more than the secretary of
and the president of the United States combined?
Detailed official earnings figures are more readily available among professional golfers and tennis players, since the nature of competition in these sports puts prize money on a different status than in most other sports. As every reader of the sports page knows, earnings in the top ranks among golfers and tennis pros have grown from a pittance twenty years ago to enormous sums today. And the numbers quoted in the press represent only the tip of the iceberg because they ignore unofficial winnings, endorsements, and other payments that are often two, three, or more times as large as official earnings. True, some of these figures simply are the result of price inflation: a $60,000 salary in 1960 had the equivalent purchasing power of $200,000 in 1982. But can anyone doubt that earnings growth in these activities has not far surpassed the rate of inflation?
There is another more important and more interesting sense in which the publicized contracts of athletes are the tip of the iceberg. They refer only to the successes. In sports, we seldom if ever hear about what happens to the unsuccessful. A boxer of championship caliber, if sufficiently careful about his contract and the finaglings of his entourage, might retire after a relatively brief career with wealth on the order of $20 million or $30 million. Sugar Ray Leonard, from all reports, has done so. Sums of that magnitude are, however, skewed to a minute fraction of boxers. The median boxer cannot even make a living at the game and has difficulty generating more than a few thousand per year. Only the hopefuls, including both those with genuine prospects and those who have not yet perceived how dim their prospects really are, can sustain interest and commitment in boxing at those earnings. Ranking contenders earn enough to work full time at the trade but hardly make fortunes, especially when physical risks and the brief period of their productivity a reconsidered.
Every job has its risks and rewards, and one can never tell in advance precisely how things will turn out in the end. Fortunately for the majority of us, most occupations admit far less concentration of rewards and allow a person to find an acceptable niche relative to personal abilities and motivation than do superstar activities. There only a select few take home most of the marbles. The National Basketball Association has about 250 players; there is probably a lesser number of tournament quality golfers and tennis players; and there are at most a few dozen highly successful boxers. The number of people attempting to break into the top echelons is larger by many orders of magnitude.
These examples point to another characteristic of the activities where superstars are found. Rewards and the probability of success appear to rise more than proportionately with talent and ability. In this we are on a little dangerous ground because in many instances it is difficult to end objective measures of personal productivity. If in football a running back is half a step quicker than the defense, that might have enormous effect on his productivity. Then there are various intangibles to consider Intangibles may even define the star and superstar. Nevertheless, consider the following: The top five money winners on the pro golf tour have annual stroke averages that are less than 5 percent lower than the fiftieth or sixtieth ranking players, yet they earn four or five times as much money. A twenty-game winning pitcher in baseball earns far more than the sum of two ten-game winners. In other words, the reward structure in most sports is highly nonlinear-that is, it rises in great disproportion to personal ability. This is just the other side of the observation that, in the fields where superstars exist, most income accrues to a very small number of persons.
An analogy will perhaps nail the point. A salesman' s productivity is easily measured by the value of goods he sells relative to their cost. Payment on commission basis guarantees a roughly proportional relationship between personal productivity and pay (roughly, because most commission systems are not strictly linear). If the nature of com petition was such that the person who sold the most in the firm received, say, 80 percent of the firm's total compensation to salespersons, the distribution of reward would be much more concentrated and skewed to the to p ranks than it actually is. But, then, this is precisely what defines a superstar.
In a competitive market economy, of which the United States is a tolerable approximation for the purposes of this discussion, competition ensures that workers are paid in proportion to their persona l contribution to national output. Were someone paid less than that contribution, a competing firm would bid more for his services. A person perceived as twice as productive receives twice as much. By the standards o f the day, this kind of social arrangement is generally thought to be reasonably equitable. In addition, a central theorem in economics proves that payment by appropriate contribution is the efficient outcome of a decentralized competitive market mechanism under ordinary circumstances. It is efficient in the sense of making the best out of resources available. To be sure, most of us perceive our own talent with a bit more acuity than the way others see it, but misperceptions on that score are, with a few exceptions, ones we can live with. Superstar phenomena appear on the surface to be rather different. There a person with a slight edge in talent receives significantly larger rewards. The puzzle is confounded by the fact that the activities in which superstars engage are characterized by an extreme form of competition. Does this suggest that the principle o f payment by contribution has been abandoned?
Before attempting to answer that question, it may be useful to delimit further the scope of our discussion and place it in a broader context. Social policy of the last two decades has focused the attention both of the public and of professional economists on inequality in the distribution of income and particularly on the plight of the poor and the problem of poverty. However, a much longer tradition in social science has concerned itself with the nature and sources of inequality in economic life generally. Of particular interest here is an observation, first studied systematically by the great Italian economist Vilfredo Pare to in the late nineteenth century, that the distribution of income contains an unusually large proportion of top earners: that is, among the rich rather than the poor. A visual image will perhaps clarify what is meant by "unusual" in this connection. Imagine a graph plotting IQ scores on t he horizontal and the frequency of scores on the vertical. The result is a familiar bellshaped curve. The peak of the bell occurs at a score arbitrarily scaled at 100 and the curve falls symmetrically on either side o f 100. Now picture a similar graph, except with earnings on the horizontal. The resulting curve is unbalanced and nonsymmetrical-a bell that is definitely out of whack. To the left of the modal (peak) value it appears much like the IQ frequency curve. However, to the right of the mode it does not fall as fast as it does to the left. It looks as if someone ha d stood at the right end of the curve, placed it over his back like a rope, and dragged and stretched it out a very long distance. The upper or right-hand tail of the distribution of income is much thicker than the lower, left-hand tail. The extra weight on the right lends a certain skewness to the distribution of income. What this comes down to is that the distribution of earnings is far from proportionate to the distribution of ability. Amazingly, Pareto's observations have been qualitatively duplicated in virtually every era of every society for which data on income distributions can be found.
For the phenomenon of superstar income to exist, certain conditions must exist alongside it. The attention of the media to t he activities in which the superstars engage is one such condition. This becomes evident in the world of show business, of which professional sports might be considered a subset; but it is also evident in arts and letters, two other fields that produce superstars. Show business first. Plausibly informed opinion has it that the number of full-time comedians in the United States does not exceed a few hundred. This is probably a smaller number than was employed in the days of vaudeville. Among contemporary comedians, the most popular are reported to earn extraordinary sums-an d none earn more than those who appear regularly on television. Again, t he capacity of television to produce large incomes is manifest in the enormous salaries paid to news broadcasters, especially those who work for t he networks and for stations located in large local markets such as New York and Chicago.
Popular music, which is also dependent upon promotion in the media, abounds with superstars. Classical music plays to a smaller market, but it has its fair share of superstars. The market for classical music has never been larger than it is now, yet full-time soloists on any given instrument are said to number in the few hundreds (and to be much fewer for instruments other than violin, piano, and voice). Performers of the first rank comprise small handfuls out of these already quite small numbers. Interestingly, income differences between first-rank and second-rank performers are substantial, even though, in a blind hearing, an infinitesimal portion of the audience could detect more than minor differences among them.
Writers seeking superstar status increasingly appear on television or lend themselves to newspaper and magazine interviews. A popular author whose work is seldom if ever discussed in this journal has sold more than twenty million copies of his novels in the past twenty years. But superstars of a sort - minor superstars - even appear in the less well known field of textbook writing. For example, undergraduate Course enrollments in economics have taken an unexpected leap forward in recent years. Enrollments have become so large that a leading elementary text can sell upwards of a hundred thousand copies in its first year or two. At a royalty rate of three or four dollars per volume, this is not a bad piece of change. Little wonder that a plethora of new titles appears year after year with hardly more than the most superficial distinctions among them.
Wherever superstars are to be found, I believe at least one of two elements will also be found-elements that are necessary to support and sustain both stars and superstars. One element is that the technology of consumption or use of the services provided by the activity must be such that poor talent is an inadequate substitute for superior talent. More about this presently. The other element has to do with certain peculiarities in the technology of the production of services through the use of audiences. These activities must admit duplication of a kind so that a person-the superstar-can deliver services to many buyers simultaneously. Once again, here the use of media is instrumental.
That greater and lesser talents are imperfect substitutes for each other is a notion that applies to a wide variety of work, but it applies in particular to the service sector of the economy. Substitution between persons of varying talent is often easier when the products of their labor are homogeneous. Farmers pay harvest workers by the pound or bushel and are willing to take on all comers at a set price because it is the total amount harvested, not its distribution over workers, that is most relevant to the profitability of their enterprise. Two persons who harvest half as much are about as good as one whose productivity is twice as large. Among the practically infinite varieties of work required by modern economies, there are fine gradations in the relative importance of the composition or distribution of productivity among many persons as compared with its sum. At one pole it is the sum that matters most; at the other pole the distribution is paramount, and substitution may be very difficult indeed.
Sometimes these differences are inherent in the valuations put upon services by buyers. If one surgeon is 10 percent more successful in saving lives than another, who among us would not be willing to pay much more than a 10 percent premium to have the more skillful person perform the operation? A company engaged in a $30 million treble-damages lawsuit is rash to scrimp on the legal talent it engages. Stockholders and directors would look askance at hiring mediocre talents under those circumstances. Other times differences are inherent in the nature of services demanded. Hearing a succession of second-rate singers does not measure up to hearing one outstanding performance by Placido Domingo. Contracting for a legal defense with two lawyers, each of whom would be likely to lose the case half the time, would not elevate the probability of winning much above one-half and may actually decrease it.
Here it must be added that the positive effect of talent for insuring successful outcomes is not necessarily independent of the nature of services to be rendered. Some tasks are so routine and so circumscribed by existing practice that nearly any competent person achieves about the same outcome. Others are more difficult, more uncertain, and, this being so, allow greater possibilities for alternative courses of action and decision. Such tasks offer greater scope for superior talent to stand out and make its mark. More capable physicians spend smaller fractions of their time on routine cases and larger fractions on difficult ones than do physicians of more modest ability, and it is socially desirable that they should do so. Untested apprentice jockeys never ride the favorites in big-stakes races.
Such considerations are important for understanding the market for executive officers in large firms. Unusually good information on executive compensation is available from public proxy statements circulated to stockholders by requirement of the Securities and Exchange Commission. Examination of these statements is instructive. They reveal Superstar-scale rewards that are highly concentrated among the top half-dozen executives in these firms. More detailed study indicates that the top incomes vary systematically with the size of the organization. Large firms pay executives more than smaller firms do. Even the occasional, well-publicized dollar-a-year man falls in line once stock options, pensions, and other forms of deferred compensation are properly accounted. The value to the organization of good top-level decisions and avoidance of bad decisions is abundantly clear once the nature of control of resources on such a vast scale is considered.
Common use of the term Officer for corporate executives Suggests certain parallels with the military. A good or bad decision by a platoon leader does not have much effect on the overall fortunes of war, but the same cannot be said of decisions made by the chief strategists. The value of extra talent is much larger at the top of the organizational hierarchy than at the bottom because those decisions percolate through the enterprise, and they have much further to travel in a larger enterprise than in a smaller one. Competition in the market for executive talent of this high order is what sustains salary and other payments of the superstar size reported in the proxy statements.
That a little talent often goes a long way takes us a fair distance in understanding Pareto's observation on the concentration of density in the right-hand tail of the distribution of income. However, it has substantial difficulty in coming to terms with the other aspect of superstars-the marked concentration of reward and Output among relatively small numbers of people in certain lines of work. Here the elusive quality of box-office appeal, or the ability to attract an audience and generate a large volume of transactions, must be confronted. Current and prospective impresarios will find no guidance from economists on what makes for box-office appeal. One might as well consult psychiatrists on how to raise children. But that doesn't mean people can't recognize it when they see it, or that where and when superstars will appear, and to what extent, might not be predictable, even though it is impossible to tell in advance who the lucky ones will be. The general importance of box-office appeal in the creation of superstars should not be underestimated. The jockeys who obtain mounts in the big races need the credential of a winning record. Aspiring, executives cater to a small clientele but still need to attract sufficient attention by past performances to be in the running for the top positions.
The superstar is someone whose audience is enormous relative to the scale on which most of us operate. Personal markets of that magnitude are almost exclusively sustained by use of media as a cooperating resource. These markets represent technologies that, in effect, allow a person to clone himself at little cost. More precisely, costs do not increase nearly in proportion to market size; and if costs are the same, the more tickets that can be sold is, as they say, all gravy. Once an author delivers a manuscript to a publisher, it can be duplicated at small expense practically indefinitely. A television or radio program is communicated virtually costlessly and identically to whomever happens to tune in. The performer or author puts out more or less the same effort whether one thousand or one million people show up to listen to the concert or buy the book.
Most economic activities are far more constrained in this respect. In the generality of such activities, costs increase more nearly in proportion, or more than in proportion, with output. When this is the case, it is not necessarily advantageous to work on the grand scale. The ultimate constraint here is the limitation of time. Services rendered by most professionals and other workers require significant one-to-one contact with buyers, and each of these contacts takes time. Time can be conserved, but only at significant cost. For example, a fancy and nimble dentist might manage to keep himself fully occupied by shifting waiting time to patients, by keeping the waiting room full of patients, and by working three chairs sequentially with several assistants. Many patients remain willing to pay the time and money costs if the services provided are sufficiently good, but imagine what would happen to the concentration of supply of dental services if a practitioner could serve a thousand patients simultaneously.
Here it becomes clear that technologies that enable sellers to cater to mass audiences account for the small number of successful practitioners in the fields we commonly associate with superstars. It just doesn't take many people to supply the entire market demand for these services when each one can effectively duplicate himself through the media. This, combined with a little market competition, also accounts for why the successful few are among the talent elite and why their incomes are so large. In such economic activities, a person of lesser talent is dominated by a person of greater talent who charges the same price. The greater talent captures all the business, and it is worthwhile to get as much business as possible because costs don't increase by very much. But the more talented person can do even better. His extra margin of talent allows him to raise prices above what the less talented can charge without losing significant audience and market share. Once again a little extra ability goes a long way. The return on each unit sold may be very small, but total reward is enormous because unit reward is multiplied by a large number. The fundamental limitation on the superstar's reward is t he potential size of the market out there to be attracted and the relative edge of the superstar's talent over those of others waiting in the wings - those who are willing to supply services to the market should the occasion arise and who keep trying to do so.
Several factors determine whether the relevant market is shared by a larger or smaller number of performers. First, the many facets of talent and box-office appeal, indeed the intangibles that make it so difficult to predict who will emerge as superstars, give rise to substantial disagreement, differences of opinion, and perception about who best caters to the individual buyer's preferences and tastes. There seems to be less room for these kinds of differences in professional sports than in many other areas. All that is required in sports is enough people to provide the competition and uncertainty about outcomes that maintain fan interest. The book trade appears to offer greater scope for variety and therefore to support large numbers of authors. Readers' preferences span a broad spectrum of education and interests, and each demands a certain amount of variety in and of itself. An author's skills are much more specialized than that.
Second,in those areas where creativity and product differentiation are crucial to the trade, limitations on the creative process often impose sharp constraints on individual supply. It is the rare book whose life on the best-seller list is as long as it took the writer to craft it. There is also a danger for the artist or entertainer of overexposure and competition with his own past performances. On the other hand, it often happens that an innovator attracts many successful imitators simply because the crank can't be turned fast enough by any single person. Finally, in some areas practitioners must contend with legacies: competition is especially tough because the current generation competes with the giants of the past. Cursory examination of subscription programs of major orchestras reveals that in this competition contemporary composers seldom win, and it is well known that nowadays remarkably few new musical compositions receive even one official performance. Depreciation of the past occurs at a much greater pace in other fields.
Personal market expansion, which ultimately requires using inferior production techniques that dilute and degrade the quality of performance transmitted to the audience, is another powerful force that often works toward spreading the business around. A clarinet concerto in Yankee Stadium might as well not be performed. Recordings are a superior way of reaching large audiences but are inferior in most dimensions to live performances before smaller groups. Another way of putting this is to say that costs do in fact increase with overall audience size. However, if buyers are willing to put up with greater dilution for the pleasure of watching or listening to the more talented performers, marketing costs of superstars do not rise as much as for other purveyors of goods, and superstars do retain a competitive edge in using large audience production methods. Slim pickings are left for the less talented. The point is related to the high-quality physician who chooses to spend less time with each patient, charging lower prices than otherwise, and spreading his skills over a much larger number of patients. Dilution of a large talent maintains competitive advantage over smaller talent if the price is right.
One hears complaints from time to time about the necessity for debasing the inherent quality of one's product to appeal to mass audiences and coarser tastes. And so it may be. Yet there are too many counter examples not to warrant certain skepticism about this argument. A good soldier of social science, I prefer to think about it in less value laden terms. The potential size of audiences to be tapped varies across different fields. Why this is so is territory where angels and most market researchers fear to tread. Many of these differences cannot be easily explained. Why, for instance, is soccer more popular abroad than in the United States? Economics does, however, provide a plausible speculation of modest generality. We are not born with refined tastes, by definition. Rather, they must be developed by education, effort, and investment. A serious writer imposes great demands on an audience, and only those who have undertaken prior investments in literary culture gain commensurate rewards. It is, furthermore, not altogether obvious that serious artists have either the capacity or the desire to produce the good trash that captures larger markets and satisfies less refined tastes.
Whatever the sources of these differences, economics of the scale afforded by modern media imply that it is monetarily more advantageous to operate in a larger overall market, and it is increasingly advantageous the more talented one is. Here lies the attraction of New York and not Cincinnati for actors, writers, artists. Similarly, the best doctors, lawyers, and architects are more frequently found in the large r cities. Whether we like it or not, today it is more lucrative to be a violinist than an accordionist, a heavyweight than a flyweight, a rock musician than a folk singer, a tennis player than a gymnast, and a writer of elementary texts than of monographs.
Changes in the technology of communication and control of distribution have decreased the cost of cloning of talent in many areas and contributed substantially to turning mere stars into superstars. Motion pictures, radio, television, phono-reproduction equipment, and other changes in communications not only have generally decreased the real price of entertainment services but also have increased the possible size of each performer's audience. The effect of radio and recordings on pop singers' incomes and the influence of television on the incomes of news reporters and professional athletes are good cases in point. There are finer gradations within these categories. Television is a more effective medium for American football than for bowling, and incomes reflect it. Television nevertheless has had an enormous influence oil the fortunes of top bowlers, golfers, and tennis players because it has enabled their markets to become much larger. Nor are these changes confined to the entertainment sector. Reductions in the costs of communication and transportation have expanded potential markets for all kinds of professional services and have allowed many of the top practitioners in the arts, journalism, and elsewhere to work on national and international scales.
Is all this fair? Probably not, Few people grow to be seven feet tall, never mind with the agility of a cat. Fair or not, it is the necessary and natural outcome of the unusual technology with which we now live. The distribution of rewards would look much different if modern technology did not admit such large economies of scale, but it is by no means obvious that society as a whole would be better off without it. The sums earned by first- and second-rank stars today are sources of envy and disgust in some quarters and give rise to mumblings about crass commercialism and the evils of cutthroat competition. In my view, a more balanced perspective is possible once one understands how technologies that sustain such sums have at the same time reduced the relevant real pr ice and cost of these services to consumers to remarkably small proportions compared with earlier days. Bringing back the good old days of restrictive reserve clauses and stock-company movie star contract systems surely would reduce the incomes of those stars. It just as surely would simply transfer the gains to club owners and producers because it would do nothing to eliminate the fundamental sources that support them. Because of the technology and the demand, the money is there; the only question is how is it to be divided up.
The great singer Elizabeth Billington earned somewhere between ?0,000 and ?5,000 in the 1801 season of the London Opera. (Neither Grove's nor the Encyclopedia Britannica, 11th edition, gives information on endorsements.) I imagine after adjusting for 1981 prices, Mrs. Billington would remain a pale shadow beside Pavarotti. What would her income have been had phonograph records existed in 1801? What changes in the future will be wrought by cable, video cassettes, home computers, and hard copy? Stick around. As the vaudeville comedians used to say, "You ain't seen nothin' yet!"
SHERWIN ROSEN is professor of economics at the University of Chicago and Research Associate at the National Bureau of Economic Research and at the National Opinion Center.
Reprinted from The American Scholar, Volume 52, Number 4, Autumn 1983. Copyright 1983 by the author. This permission is for nonexclusive world rights. It does not cover future editions and revisions of the above work.