Based on Bradford DeLong, “Robber Barons” (1998), and Elijah Wald, How the Beatles Destroyed Rock ’n’ Roll (Oxford University Press, 2009)
Economies of scale, also known as increasing returns to scale, cause the average cost of production to fall as the volume of production rises. There are several kinds of economies of scale. The simplest one is the division of labour, which Adam Smith claimed was responsible for “the greatest improvement in the productive powers of labour.” He gave the pin factory as an example of its operation:
One man draws out the wire, another straights it, a third cuts it, a fourth points it, a fifth grinds it at the top for receiving the head; to make the head requires two or three distinct operations; to put it on, is a peculiar business, to whiten the pins is another; it is even a trade by itself to put them into the paper; and the important business of making a pin is, in this manner, divided into about eighteen distinct operations, which, in some manufactories, are all performed by distinct hands, though in others the same man will sometimes perform two or three of them.1
Smith visited a pin factory that was too small for complete division of labour. Ten men were able to produce more than 48,000 pins in a day, but if they had worked independently, “they certainly could not each of them have made twenty, perhaps not one pin in a day.”
Division of labour was the key to Henry Ford’s moving assembly line. The earliest cars were assembled by teams of workers who carried parts from storage areas to their work areas. Ford’s first attempt to improve on this method involved laying out piles of parts in rows. A work team placed a chassis on skids and dragged it from one pile to another, adding parts to the car at each stop. His next attempt, the moving assembly line, did away with the teams. The assembly of a Model T was broken into 84 steps, each of which could be carried out by one or two people. As the chassis moved along the line, the car was assembled by 140 stationary workers, each of whom carried out just one of the steps. Teams using the “work area” approach assembled a car in 12 hours; assembly line workers did it in 93 minutes.
The modern rise of Chinese manufacturing was likewise based upon the division of labour. The opening scene of Edward Burtynsky’s Manufactured Lanscapes (2008) shows the interior of a Chinese factory. Although the factory is huge, there is surprisingly little machinery in it. Very few people move about and everyone attends to just one task. Its organization reflects the principal advantages of the division of labour: no time is lost in switching from task to task, and each worker gains great proficiency in a single task through long repetition. Another scene from Manufactured Landscapes shows one of the factory’s workers assembling a circuit breaker in 45 seconds. This assembly is her only task, and her work day is ten hours long.
The division of labour is not, however, the only way to achieve economies of scale. From the late nineteenth century onwards, technological change created new opportunities for economies of scale, both for firms and for individuals.
Production with Large Fixed Costs
Industrial production involves both fixed and marginal costs. The fixed cost includes the costs that are independent of the level of output, such as investment in plant and equipment. The marginal cost is the cost of producing one more unit of output, and includes the costs of raw materials and labour. Marginal cost is often imagined to be constant until the plant’s design capacity is reached. For any level of output that does not exceed the design capacity, the average cost of producing units of output is:
Here, is average cost, is the fixed cost, is the marginal cost, and is the number of units produced. and are determined by the design of the plant. Average cost falls as output rises, which is the defining characteristic of increasing returns to scale. Average cost is smallest at the plant’s design capacity, when the fixed cost is spread over the greatest possible number of units.
A firm that engages in head-to-head competition with other firms wants its average cost to be as small as possible, so that it remains profitable even at low prices. There are two reasons why a higher fixed cost might be associated with a lower minimum average cost. First, capacity could rise faster than the fixed cost. For example, industrial chemicals are often produced in large vats. If the cost of a vat rises more slowly than its volume, chemicals can be produced more cheaply in large vats than in small ones. A similar logic applies to the steel industry’s blast furnaces and the mining industry’s ore crushers. Second, for plants with the same design capacity, the marginal cost could fall as the fixed cost rises, because the additional machinery and equipment make workers more productive. The moving assembly line, for example, reduced the number of workers needed to produce each unit of output. For the past century or so, these two factors have often worked together: bigger plants filled with more machinery have had a larger design capacity and required less labour (per unit of output) to operate.
Of course, a firm’s realized average cost depends upon its actual level of output, not its capacity. In a small market, choosing a high fixed cost / low marginal cost design is likely to backfire. The firm’s average cost would be low at design capacity, but with limited demand, it would produce much less than that and have a higher average cost. Firms with smaller fixed costs would be able to produce at a lower average cost, undercut its price and drive it out of business.
Nevertheless, in the late nineteenth century and early twentieth century, “go big or go home” was often the right strategy. Henry Ford developed the moving assembly line in order to drive down the average cost of his cars, so that he could sell them at the lowest possible price. His technology required a high rate of production. In 1914, the first full year of operation for the moving assembly line, Ford produced more than 300,000 cars — more than all other American manufacturers combined. Andrew Carnegie recognized that the most efficient steel mills were larger than current demand conditions warranted. He built them anyway, and then put his energies into selling the steel.
Networks have the cost structure described above. To operate a railroad between two places, for example, tracks must be laid and stations built at the terminal points, and a locomotive and a few passenger cars must be purchased. These are all fixed costs. The service can then be extended to more and more customers by operating the train more frequently, which involves relatively small additional costs. The greater the number of passengers, the lower the average cost. Other networks with similar cost structures include telephone service (with a “land line”), electrical distribution, the provision of clean water, and sewage disposal. The last two services were provided by municipal governments from the beginning. The others were open to competition, but the competition was often short lived. The firm with the largest customer base had the lowest average cost. It could set its prices low enough to drive its competitors out of the market, or it could earn large enough profits to buy them out. In either case, the market would ultimately be dominated by a few large firms, and perhaps by just one firm.
The tendency for oligopoly or monopoly to prevail when an industry’s production is characterized by increasing returns has an interesting implication: very large fortunes can be made in these industries. This was certainly the case in the United States.
In 1900 there were 22 fortunes that were as great as those of a modern-day billionaire. Nine of these fortunes were made by building and operating railroads, a new technology that displayed increasing returns to scale. Five more were made in finance, which chiefly meant railroad finance. One fortune belonged to Andrew Carnegie, who exploited the increasing returns to scale inherent in the new steelmaking technologies.
There were more “billionaires” by 1918, and they were involved in a wider variety of businesses: photography, retailing, chemicals, tobacco, farm machinery, automobiles, food processing, municipal railroads, oil, steel, finance. The common factor among these businesses was increasing returns to scale:
The billionaires of 1918 or so come as close as we will ever find to being examples of situations in which enormous wealth comes from being in the right place at the right time — able to build large organizations to take advantage of hitherto unexploited economies of scale, and retaining large enough ownership stakes and access to the capital market to then transform expected future profits into present wealth.2
Fortunes were made in finance through another venture that displayed increasing returns to scale: the conversion of massive private companies into joint stock companies.
In case after case — and this is where the financiers of 1918 got their billion dollar fortunes — the financier’s job was to allow the founding entrepreneur to retire, to bring in a professional management to keep the business going, and to reassure those who are going to purchase the founding entrepreneur’s ownership share that this is a prudent and worthwhile investment.3
The digital revolution of the late twentieth century opened up new increasing-returns-to-scale industries, whose exploitation led to more great fortunes. The major cost of computer software is the cost of creating it — distribution of the completed software costs almost nothing. In other words, computer software is an industry with large fixed costs and very small marginal costs, so average costs fall dramatically as output rises. The same cost structure applies to firms that operate over the internet. Although all aspects of the computing industry were once open to competition, a small number of firms now dominate it — Microsoft, Google, Amazon, Alibaba — and the founders of these firms have amassed huge fortunes. Jeff Bezos of Amazon, Bill Gates of Microsoft, Larry Ellison of Oracle, and Mark Zuckerberg of Facebook are among the ten richest people in the world. Four more (Larry Page and Sergey Brin of Google, Jack Ma of Alibaba, and Ma Huateng of Tencent Holdings) are among the twenty richest people in the world.
Some markets are dominated by a relatively small number of people. These people — the superstars — generally have extremely high incomes. Anyone who follows popular culture can identify musicians, actors, authors, and athletes who merit this designation, but there are also superstars among lawyers, surgeons, professors, architects, and coders. Among the more talented people in these professions, a small increase in ability is associated with a large increase in market share and a large increase in income. American professional baseball exhibits both of these traits. The players on AAA teams are already an elite: the odds that a high school senior playing interscholastic baseball will be drafted by the major leagues are about 1 in 200. The drafted players almost always start in the minor leagues, with AAA being the highest level of play. Despite their skills, AAA players earn very low salaries (US$12,000 for a five month season is a frequently cited figure). They play for crowds of a few thousand, without television coverage. The odds that a AAA player will reach the major leagues are about 1 in 10, but if he does reach the major leagues, his minimum salary will be half a million dollars. The rewards for small increases in ability get even larger for players in the major leagues. The average major league salary in 2018 was over four million dollars, but the best-paid player earned 34 million dollars. The major league player is also performing in front of a much bigger audience: tens of thousands in the stadium, and millions more via television. In the United States “baseball fan” and “major league baseball fan” are almost synonymous.
The income distribution for professional baseball players is heavily skewed. Minor league players outnumber major league players by about 5 to 1, so the vast majority of players are earning very low salaries. At the same time, the bulk of the income is earned by a small number of very highly paid players, who could be earning fifty or sixty times as much as some of their teammates. This radically skewed income distribution is characteristic of markets dominated by superstars: actors and musicians are little different from athletes in this regard. One reason for the skewness of the distribution is that moderate talent is often a poor substitute for great talent.
Hearing a succession of mediocre singers does not add up to a single outstanding performance. If a surgeon is 10 percent more successful in saving lives than his fellows, most people would be willing to pay more than a 10 percent premium for his services. A company involved in a $30 million law suit is rash to scrimp on the legal talent it engages.4
This factor alone explains why the fees for the superstars’ work rises faster than their talents, but superstars add to their income advantage by exploiting economies of scale. A professional baseball player moves up to the major leagues where the stadiums are bigger and the games are televised; a musician plays in bigger venues; a lawyer farms out routine work so that he can take more clients.
This strategy is more successful for the musician than the lawyer because the musician’s services are less rivalrous than the lawyer’s. In economics a good or service is “rivalrous” if each unit is consumed by just one person. A ham sandwich, for example, is rivalrous. A good is “non-rivalrous” if one person’s consumption does not diminish another person’s consumption. A GPS signal is non-rivalrous: millions of people use it simultaneously, and none of them are bothered by the others’ use of it. Between these two extremes lie other goods that are, to varying degrees, “congestible” or “partially rivalrous.” Many people can use partially rivalrous goods or services at the same time, but the quality of use declines as the number of users rises. Swimming pools and highways are both partially rivalrous. The degree of rivalry strongly influences a superstar’s ability to exploit his talents. The lawyer’s services are very nearly rivalrous. His clients expect a face-to-face consultation, and they expect him to devise a legal strategy. These tasks require a certain amount of time, so no matter how hard he is willing to work, the number of clients that he can take on is limited. The musician’s services, on the other hand, are largely non-rivalrous. Since the effort needed to perform in large venues isn’t much greater than the effort needed to perform in small ones, musicians can readily extend their markets by playing in larger venues. The degree of rivalry varies even across musicians. Rock is amplified, so rock music in a football stadium sounds much the same as rock music in an auditorium. Opera is not amplified. There are smaller and larger opera houses, but the best opera singers perform before much smaller audiences than the best rock musicians, and earn much smaller incomes.
The modern rise of superstars has much to do with the appearance of new technologies that allowed them to expand their audience. The shift from stage to film created superstar actors. Television created superstar athletes. Radio and the phonograph gave rise to superstar musicians, and also entirely transformed the musical experience.
Technology and the American Music Industry
Before the invention of the phonograph, and even for two or three decades after its invention, all music was social. There were occasional concerts given by local bands, but for the most part, people participated in their own entertainment. Most people felt comfortable singing in public, and many people played instruments. The ability to play an instrument was particularly common among women, for whom it was part of their requisite social skills. Dancing was also a popular diversion, and most bands played to accompany dancers. The bands rarely specialized in a particular type of music. They understood their role to be getting people onto the dance floor, so they played what the dancers wanted. The music industry was built around sheet music. Everyone wanted copies of the latest songs, so that they could sing them or play them.
The phonograph only gradually gained popularity, in part because recordings were difficult to produce before the development of the master recording:
As with the hand-copied manuscripts that preceded printing, every early phonograph cylinder was unique: Musicians played or sang into a bank of acoustic horns, each of which was connected to a needle that cut an impression of the sound waves in the groove of a wax cylinder. The number of cylinders that could be made at one time was limited by the number of machines that could be crowded closely enough together to capture the music clearly, and because some horns were necessarily closer to the musicians on one side of the room and some closer to the musicians on the other, each cylinder was slightly different. A recording session consisted of playing the same piece over and over, with the recording engineer putting new cylinders on the machines as each was finished. To make a thousand records of a song, assuming that you could fit ten horns into the room and that every cylinder recorded properly—which was rarely the case—you would have to perform a song a hundred times.5
Thomas Edison had intended the phonograph to be a device that recorded and played back brief messages. Its sound quality was poor, and there was no way to duplicate the recordings. The development of musical recording was carried forward by others, most notably by Emil Berliner, who invented the gramophone, and by Eldridge Johnson, who developed the electroplating of wax masters.
The gramophone, like the phonograph, used an acoustic horn to focus sound on a diaphragm whose vibrations were transmitted to a stylus. The gramophone’s stylus vibrated horizontally, following a long, inward-spiralling path along the surface of a zinc disc covered in lampblack. The disc was then placed in an acid bath, which etched a groove in the disc where the stylus had exposed the zinc. This process produced a playable “positive” image, but Berliner was interested in mass production. He electroplated the “positive” image to produce a copper “negative” image, which was then used to stamp out the final “positive” copies of the recording. The zinc image was destroyed in the production of the copper image,6 and the copper image could be used to stamp out (in shellac) about a thousand records before it wore out.
This process was soon simplified: the stylus etched a groove in a wax master, which was then coated in fine graphite (to make it electrically conductive) and electroplated to produce the copper “negative” image. This image was the new master. A mixture of wax and plaster of Paris was poured over it to create a “positive” image, which was then electroplated to produce a copper “negative” stamper. Since this process did not damage the copper master, dozens of stampers could be made from it. This procedure was used from 1901 until about 1920, after which the copper master was coated in silver (which acted as a separator) and then electroplated to produce a copper stamper. The last step was repeatable, allowing true mass production of a recording.
The next major innovation was electro-mechanical mastering, developed by the Bell Telephone Company in the mid-1920s. Here, microphones converted sound into an electrical signal, which then drove the stylus that etched the wax master. Electro-mechanical mastering produced records whose sound quality was as good as radio. It immediately replaced acoustic mastering.
Records gained widespread acceptance in the 1920s. Sales of the sheet music of the biggest “hit” songs fell from about a million to about half that; sales of their recordings averaged about 860,000 copies.7 Most customers were buying particular songs, not particular performances. Every record company would produce a version of the most popular songs, and the customers were largely indifferent between them. As Elijah Wald explains, a “hit” song was just part of the popular culture:
My father, born in Brooklyn, New York, in 1906, had a terrific memory for the hits of his youth and I grew up hearing “The Sheik of Araby,” “When Frances Dances with Me,” “Yes! We Have No Bananas,” over and over. I would sometimes ask, “Who sang that song?” — a normal question for any pop listener born after 1950. But it made no sense to him. Everybody sang those songs.8
Even as records rose in popularity, the incomes of most professional musicians continued to come from their work in dance bands, and the bands continued to play whatever music the dancers wanted. Even Louis Armstrong spent most of his time playing in dance venues — the “Hot 5” and “Hot 7” recordings, now considered to be jazz classics, were a minor sideline.
Paul Whiteman was the most popular band leader of the 1920s. He extended the scope of his talents by forming additional bands that replicated the sound of his original band. Other successful band leaders did the same. Vincent Lopez, for example, had twenty-two “cloned” bands. The people who hired a Whiteman or Lopez band weren’t expecting a unique musical experience:
Virtually all mainstream dance bands relied largely on “stock” arrangements—instrumental charts bought from the publisher rather than arranged by or for a particular orchestra—and hiring a group from a big-name front man such as [Meyer] Davis or Whiteman was considered a guarantee of quality rather than implying that it would play in any distinctive, personal style. Nor did dancers feel shortchanged because they were not getting something original or unique. After all, they didn’t go to dances to listen to the band; they went to dance.9
The very ordinariness of a musician’s work meant that even journeymen could do it. It was a way of earning a living, liking delivering ice or reporting the news.
Many of these journeymen lost their livelihoods during Prohibition (1920-1933) and the Great Depression (1929-1939). Prohibition was a national ban on the production, importation, transportation, and sale of alcohol. Hotels and restaurants had once hired bands to attract customers: the cost of the band was covered by additional profits from the sale of food and especially alcohol. This practice often ceased to be profitable when alcohol sales were taken out of the calculation. The Great Depression, which forced Americans to pare their spending, made hiring bands even less profitable. Two-fifths of the members of the American Federation of Musicians lost their jobs during the Prohibition era.10
While Prohibition and the Great Depression were extinguishing the jobs of many moderately skilled musicians, two technological innovations were extending the scope of the more talented ones. These innovations were radio and sound movies.
The economies of scale in radio were potentially very large, and they were quickly realized. The first commercial radio station opened in 1920. By 1928, NBC had 69 affiliated stations reaching 80% of the radio audience. A single show could entertain all of these listeners. Most shows were live, either recorded in the studio or relayed over telephone lines from local venues. Jazz was the dominant form of music: a sampling of local stations in 1928 showed that 64% of their airtime was devoted to it.11 But it was the jazz of Paul Whiteman, not of Louis Armstrong.
The scarcity of airtime meant that radio stations could choose to broadcast only the best bands. These bands were financially well-rewarded.
As radio became a national industry with major sponsors, it also began to provide solid incomes for its stars. By 1934, Fred Waring and his Pennsylvanians had beaten Whiteman’s live salary record by earning $10,000 for a week in a Broadway theater, but that was just a pleasant addition to the band’s regular weekly stipend of $12,500 for presenting the hour-long Ford Dealers Program. 12
As well, radio prominence translated into bigger record sales and more profitable tours. The bands competed for these rewards. They recognized that radio stations had to have a reason for choosing them over their competitors, and that the public had to have a reason for wanting their records. For the first time, bands began to develop unique personalities. Count Basie, Duke Ellington, Benny Goodman, Tommy Dorsey, and Glenn Miller are all remembered for their distinctive musical styles, but few people remember bandleaders from the time before radio.
The first superstar (and perhaps the only one of his generation) was Bing Crosby. He started out as a vocalist with Paul Whiteman’s band, and his early performances reflected the public’s focus on the song. Pitch, timing, phrasing and intonation were what mattered to Crosby as he strove for the perfect realization of the written music. But Crosby had an affable and appealing personality, and he was flexible: he could be whatever the public wanted. If it wanted swing, he would swing. Crosby became a star of all the new media — records, radio, movies and television — but it seems that there was as yet little space for superstars. Frank Sinatra, one of the leading male vocalists of the next generation, left Tommy Dorsey’s band in 1942 to attempt the same kind of solo career as Crosby, but his popularity waned within a few years. It took a role in From Here to Eternity (1953) to bring him back to the public’s attention and restart his career.
Radio, like Prohibition and the Great Depression, harmed most musicians. Radio became a popular form of entertainment, steering people away from live events. Record sales fell sharply, and since the best musicians were selling more records, the less-than-best were selling much fewer.
The sound movie was another technology that allowed the most talented musicians to extend their scope, while the less talented musicians were pushed to the sidelines. The Hollywood musical added extravagant visuals to the musical experience. The segment below features Frank Sinatra, a hundred extras, elaborate costumes, and a stunning trompe l’oeil set. The glamour of it all clashes with the song’s narrative, but that’s Hollywood.
The distribution of musicians’ incomes had been skewed by Prohibition and the Great Depression, during which “playing for a living was a victory.” Radio and sound movies increased its skewness by offering great rewards to a few and trimming the incomes of the many.
The next innovations — disc jockeys (c. 1935) and jukeboxes (c. 1933) — created more economies of scale: an announcer and a dozen records replaced a live performance; and a jukebox allowed a single record to be “rented” by hundreds of listeners. However, the economic rewards went to the radio networks and the jukebox operators, not the musicians. The royalties earned by composers and publishers on records were about one-tenth of their royalties on sheet music. Moreover, once the record was sold, they earned no further royalties. A record sold to NBC for nation-wide broadcasting earned them no more than a record sold to a private person, and a record played a thousand times from a jukebox earned them no more than a record played in a teenager’s bedroom. The composers and publishers were a necessary part of the music industry, however, and they were eventually able to negotiate royalties. The musicians were in a more difficult situation. The top performers could have royalties written into their contracts, but royalties were immaterial for most musicians because they did not make records. Many of them, like saddlers and blacksmiths, were simply displaced by the new technologies. The fraction of Americans who earned their living as musicians was cut in half between 1910 and 1940.
The big band musicians were among the hardest hit. From the swing era onwards, more and more Americans became listeners rather than dancers. The foxtrot and the waltz no longer matched the music, and only young and energetic dancers could manage the lindy hop and the charleston. Americans would not return to the dance floor until Chubby Checkers invited them to “do the twist.” But if Americans were not dancing, the dance bands were obsolete. There was a shift toward smaller bands — modern jazz came into its own — and an even sharper shift toward vocalists.
The dancer’s priority had been the song, but the listener’s priority was the performance. The record companies responded to the change in their market. In 1955 the record company Victor announced that it would no longer issue covers of popular songs, and that it would demand full control over the songs that it chose to record. Victor more or less held its course, and other companies followed its lead. This change is reflected in the way that we think about the music of the past:
We continue to think of the hits of the 1930s and 1940s as “standards,” performed by jazz and pop artists around the world, but of the hits of the 1950s as “oldies,” usually heard on recordings by the original artists. 13
Listeners cared about the uniqueness or creativity of a performance, so record companies began to emphasize these attributes when scouting new talent. The period from the mid-1950s to the mid-1970s yielded everything from doo-wop to acid rock. A good illustration of how far the pendulum had swung is Janis Joplin’s rendition of Summertime, a song that had been sung by Bing Crosby a half-century earlier. Bing Crosby thought of himself as a conduit for the song; Janis Joplin thought of the song as a conduit for her own turbulent emotions.
- Adam Smith, The Wealth of Nations (1776), Book 1, Chapter 1. ↩
- Bradford DeLong, “Robber Barons” (1998). ↩
- Bradford DeLong, “Robber Barons” (1998). ↩
- Sherwin Rosen, “The Economics of Superstars,” American Economic Review (1981), p. 846. ↩
- Elijah Wald, How the Beatles Destroyed Rock ’n’ Roll (Oxford University Press, 2009), pp. 86-7. ↩
- The copper image was essentially welded to the zinc image, so sulphuric acid was used to eat away the zinc. The key to later metal-to-metal mastering (c. 1920) was a layer of silver between the master and the electroplated stamper, which allowed the master and stamper to be separated without damage. ↩
- Elijah Wald, How the Beatles Destroyed Rock ’n’ Roll (Oxford University Press, 2009), p. 85. ↩
- Elijah Wald, “Mitch Miller’s Part in Pop History” ↩
- Elijah Wald, How the Beatles Destroyed Rock ’n’ Roll (Oxford University Press, 2009), p. 63. ↩
- Elijah Wald, How the Beatles Destroyed Rock ’n’ Roll (Oxford University Press, 2009), p. 65. ↩
- Elijah Wald, How the Beatles Destroyed Rock ’n’ Roll (Oxford University Press, 2009), p. 92. ↩
- Elijah Wald, How the Beatles Destroyed Rock ’n’ Roll (Oxford University Press, 2009), p. 94. ↩
- Elijah Wald, How the Beatles Destroyed Rock ’n’ Roll (Oxford University Press, 2009), p. 177. ↩