ABSTRACT

Failures to understand the constraints and incentives facing decisionmakers have resulted in the creation of the myth that property rights and spectrum markets would have been superior to the regulatory system of the Radio Act of 1927. Discussions of hypothetical spectrum property rights in the 1920s fail to take account of (1) the vast differences between the radio propagation conditions in the radio spectrum in use then and propagation in the bulk of the radio spectrum today and (2) the technical limitations of equipment at that time. The author concludes that spectrum property rights would have resulted in more radio service in urban areas, a substantial loss of rural service, and diminished consumer welfare.

A revisionist myth asserts that a system of spectrum property rights would have been superior to regulation by the Federal Radio Commission (FRC) under the Radio Act of 1927.1 Institutions reflect the constraints and incentives that lawmakers, industry, and consumers face. Property rights could not have functioned efficiently in the 1920s because of the nature of the interference externalities in broadcasting at that time. Efficient property rights could not be defined because of (1) the nature of radio wave propagation in the broadcast band, (2) radio technology in 1927, and (3) the need for an efficient industry that reflected the preferences of the listening audience.

Proponents of the myth do not point to any contemporaneous description of a workable spectrum property-rights system, nor do they describe a property rights system that reflects radio propagation in the AM band. Proponents of property rights also fail to address the fact that, in the early days of radio, most radio frequencies were like waterways or rights-of-way—used briefly by one party and then returned to a common pool and often involved two unrelated parties communicating. Ostrom is a better guide to the management of the radio spectrum in 1910 or 1920 than is Coase. Policy considerations beyond efficiency are also important in this context. Moss and Fein showed that Congress feared concentrated control of mass communications and that the turn away from property rights in radio reflected that fear.2

I first provide some background on the evolution of broadcasting technology and the broadcasting industry in the early 20th century, followed by a discussion of the early years of radio regulation and the passage of the Radio Act of 1927. I then present a summary of Coase’s approach to spectrum property rights and demonstrate the impracticability of applying that approach to the allocation of broadcast stations in the 1920s.3

Broadcasting in the 1920s

Radio broadcasting evolved swiftly between 1920 and 1929. Technology improved, the structure of the industry changed, and consumer usage exploded. In 1920, only a few thousand households had radio receivers; in 1929, about 12 million households—40 percent of all households—had radios.4 Broadcasting expanded from a handful of stations operating a few hours each day and sharing a single frequency to hundreds of stations operating on ninety different frequencies.

In early 1927—just before the passage of the Radio Act of 1927—there were about 700 licensed broadcasters. Some broadcasters transmitted more than twelve hours per day; others transmitted as little as one hour per month. One “station” did not even have a transmitter.5

In the 1920s, most investment in the broadcasting industry was for purchasing and maintaining home receivers. Radio receivers were expensive—in 1927, the average cost of a receiver was $1246—equivalent to $2,120 in 2023 dollars.7

Table 1 shows expenditures on home radio apparatus (radios and radio parts) and radio advertising during the 1920s. Radio advertising began in about 1926, and by 1930, advertising revenues were less than 10 percent of spending on receivers. Many broadcasters were manufacturers or retailers of radios who provided programming to stimulate receiver sales.

TABLE 1

Radio Broadcasting Revenues in the 1920s

YearSales of Home Radio Apparatus ($ millions)Radio Advertising Expenditures ($ millions)Radio Advertising Expenditures as a Percentage of Home Radio Apparatus Sales
1922 60   
1923 136   
1924 358   
1925 430   
1926 506   
1927 426 4.8 
1928 691 14.1 
1929 843 26.8 
1930 501 40.5 
1931 309 56.0 18 
YearSales of Home Radio Apparatus ($ millions)Radio Advertising Expenditures ($ millions)Radio Advertising Expenditures as a Percentage of Home Radio Apparatus Sales
1922 60   
1923 136   
1924 358   
1925 430   
1926 506   
1927 426 4.8 
1928 691 14.1 
1929 843 26.8 
1930 501 40.5 
1931 309 56.0 18 

Sources: Sales: Scott, Peter, and James T. Walker. Producer-Driven Supply Chains for Inter-War Entertainment Radio: Were Dealers “Over-Sold” on Marketing? Reading: Centre for International Business History, University of Reading, 2014, 20; expenditures: Sterling, Christopher, and John Michael Kittross. Stay Tuned: A Concise History of American Broadcasting. Belmont: Wadsworth Publishing Company, 1978, Table 4, author’s calculations.

Between 1925 and the end of 1927, consumers invested about $1.4 billion in home receiving apparatus. Audience investment in receivers was about 40 times larger than broadcasters’ investment in transmitters.8

Terminological Confusion

The meaning of terms in radio technology has changed. In the 1920s, station referred to an enterprise quite different from a modern broadcasting station. Many stations operated only a few hours a week—Hilliker9 reported that station KFPR typically transmitted one time per month in 1925. Per Slotten,10 the University of Vermont station WCAX broadcast only one hour per week in the early 1930s.

The transmitters and antenna systems of many stations were primitive and relatively inexpensive. Stations were often quite low power. In mid-1927, 28 percent of stations operated at 50 watts or less, whereas only 95 stations (14 percent) operated at 1,000 watts or higher.11 In the 1920s, transmitter modulation was inefficient, which reduced the coverage of a station but left its ability to create interference unimpaired. Some stations were portable—in the sense that they were built into a truck or railroad car and could be moved quickly from city to city.12

Another word that can mislead is interference. Interference between broadcast station signals was far more severe in the 1920s than today. Two phenomena—heterodyne interference and sky wave propagation—and less “pollution” of the airwaves from electrical and electronic devices than is the case today made interference in AM broadcasting substantially different from today’s interference.

In 1927, AM transmitter frequency control was imprecise—stations operated within only several hundred hertz of their assigned frequency.13 Consequently, an interfering signal would create a loud tone at the audio frequency equal to the difference between the carrier frequency of the desired station and the carrier frequency of the interfering signal, and a much weaker signal would cause interference than would be the case today.14 This effect was called heterodyne interference.

At night, AM signals reflect off the ionosphere, a layer of ionized air about 200 miles above ground, and return to earth hundreds of miles from the transmitter site—a phenomenon called skywave propagation. See Figure 1. Because many farms and rural locations lacked electrical power, on most farms there was no background static caused by electrical transmission lines, electrical motors, or other electrical equipment. In such areas, a skywave signal often provided excellent service in the evening. However, in an area where two skywave signals on the same frequency overlapped—even if one was much weaker than the other—there would be objectionable heterodyne interference. Contemporary analysts concluded that at night during times of good skywave propagation, operating only one station at a time in the United States on each channel would permit service to rural and remote points.15 If multiple stations operated on a channel, each station would serve only a small area near the transmitter.16

FIGURE 1

Skywave Signal Transmission

FIGURE 1

Skywave Signal Transmission

Close modal

In the AM band, Skywave is a nighttime phenomenon. Skywave occurs both day and night for signals between about 3 and 30 MHZ. For historical reasons, these frequencies are called shortwave (they are longer than the waves in 99 percent of the useful radio spectrum). An example of skywave propagation is given by Voice of America transmitting shortwave broadcasts to Africa from a transmitter in North Carolina.

The interference issues that Congress and the FRC considered in the late 1920s were vastly different from most such issues considered today because of the combination of (1) the long-range propagation possible on essentially all available spectrum and (2) the technical limitations of transmitters and receivers in 1927. Understanding the choices made by Congress and the FRC requires knowledge of the problems they faced and how they viewed those problems.

The Birth of Radio Regulation

In 1910, Congress enacted the Radio Ship Act, which required larger ships to have radio facilities and to keep watch on the emergency calling frequency. In 1912, Congress ratified the Berlin Convention of 1906, which required that (1) radio systems exchange communications with radio systems obtained from any manufacturer and (2) all ships be able to operate on 300 meters. That same year, Congress enacted the Radio Act of 1912, which required licensing of both transmitters and operators and implemented many elements of the Berlin Convention.17 Although technology changed and radio use expanded greatly, the 1912 Act governed the use of radio until the passage of the Radio Act of 1927.

The Early Years of Radio Regulation

Radio broadcasting began around 1920 and grew rapidly. In late 1921, Westinghouse negotiated with the Department of Commerce for the assignment of a specific frequency to them for broadcasting. On September 15, 1921, the Department of Commerce assigned the wavelength of 360 m (830 kHz) to Westinghouse’s station KDKA18 and immediately afterwards licensed other broadcast stations on this same frequency. In December 1921, it made 485 meters (620 kHz) available for broadcasting. This second frequency was limited to “broadcasting crop reports and weather services,” and broadcasters were required to not interfere with maritime communications that took place on the same frequency. In late 1922, the Department of Commerce made 400 meters, or 750 kHz, available for broadcasting. Stations that operated on 400 meters were denoted class B and were required to meet substantially higher technical standards than the typical station broadcasting on 360 meters. For example, the minimum power allowed was 500 watts.

By March 1923, 524 stations were authorized to transmit on 360 meters. In April 1923, the Department of Commerce created the AM band by dedicating a contiguous block of spectrum—81 frequencies or channels—to radio broadcasting.19 Stations were divided into three classes—A, B, and C. Separate channels were set aside for class A and class B stations. Class A stations were low power and could be located relatively close together—say one station in New York City and another in Washington, DC. Class B stations were high power—corresponding to the higher power of the earlier class B stations on 400 meters. Most class B frequencies had only a single station operating on them at night, and frequencies used on the East Coast were reused on the West Coast. To obtain a class B license, a broadcaster had to promise to build a transmitter with at least 500 watts of power and meet several other technical requirements that improved the quality of the station’s signal and the reliability of the station’s operation.20 Stations continuing to operate on 360 meters were denoted as class C stations. They were permitted to remain on 360 meters but were encouraged to switch to the new class A and class B frequencies.

The Bureau of Navigation in the Department of Commerce regulated radio during the 1920s. However, a pair of court decisions weakened the Department of Commerce’s authority. In November 1921, in a proceeding that did not involve a broadcasting station, the Supreme Court of the District of Columbia ruled that the Secretary of Commerce had no discretion when acting on applications for radio licenses—a proper application must be granted. In April 1926, in the Zenith case, a federal court ruled that the Secretary of Commerce lacked authority to specify the time and frequency to be used by broadcasters.21 The court, reviewing the 1912 statute designed to manage a shared resource—similar to navigational waters—declined to find that the statute permitted the government to restrict stations to specific frequencies or to assign exclusive rights to individual broadcast stations. In mid-July, the acting Attorney General issued an opinion that the court was right—the Secretary of Commerce lacked the authority to limit the power, frequencies, and time of day of broadcast operations.22 The period referred to as “chaos in the airwaves” then ensued.23 Broadcasters changed channels and broadcast times, creating widespread interference.

The Radio Act of 1927

In early 1927, Congress passed the Radio Act of 1927,24 which created the FRC and gave it the power to regulate radio. The FRC could assign powers and frequencies, set technical standards, and otherwise exercise powers that the Department of Commerce had been exercising before the Zenith decision. The Department of Commerce supplied the FRC with office space and technical support.25 Department of Commerce staff members who had been involved in radio regulation earlier performed much of the day-to-day work of the FRC in its first years. In 1932, the Radio Division of the Department of Commerce was transferred to the FRC.26 The first Chairman of the FRC was retired Admiral William H. G. Bullard—a well-respected electrical engineer who had played key roles in the development of radio in the U.S. Navy, had served as head of the Department of Electrical Engineering at the U.S. Naval Academy, and had been instrumental in the creation of RCA after World War I.

The FRC and Technocratic Values

In his excellent and enjoyable examination of the evolution of broadcast regulation in the United States, Slotten repeatedly speaks of “the technocratic policies of the commission.”27 Slotten characterizes technocratic as referring to the belief that (1) technical experts should make political decisions involving complex technologies and systems and (2) abstract values of rationality and efficiency should guide such decisions.28

My review of contemporaneous engineering literature and FRC reports and orders indicates that the following were key technocratic values associated with broadcast regulation:

  • Radio should deliver relatively high quality audio signal.

  • The listening audience should be maximized.

  • Everyone in the country should receive the signals of at least a few stations.

  • The right to transmit must be used “efficiently.”

The following important values were sometimes at odds with this technocratic view:

  • Media diversity is a worthy objective.

  • A competitive industry structure is important.

  • Local programming and local voices are valuable.

Below, I explain how I reached my conclusions regarding these values.

Audio Quality

Multiple sources show that the technocratic vision included a broadcast service that delivered—or that would be able to deliver when receivers improved—a high-quality audio signal. Explaining General Order 116, the FRC’s Chief Engineer stated that a signal-to-interference ratio of 26 dB should be considered acceptable in most rural areas where “the standard of reception is not of the highest.” But such a standard would not be regarded as acceptable in metropolitan areas.29 Similarly, Hogan stated that a 40-dB signal-to-interference ratio marked “the limit of the useful service zone” in the non-heterodyne context.30

The FRC kept in place the 10-kHz separation of broadcast channels implemented by the Commerce Department.31 In contrast, the FRC felt that police radio systems—systems that must deliver intelligible voice but not music—could operate on 6 kHz channels.32

Audience Maximization

The FRC’s Chief Engineer’s description of the recommendations of the FRC’s Engineering Committee’s channel allocation plan provides an example of the technocratic view that maximizing audience was an appropriate policy goal:

The number of channels (50) indicated for class C stations is the minimum that should be provided, in view of the far greater service, both distant and local, that will be rendered by such channels, owing to the absence of heterodyne interference and the consequent possibility of the use of greater power.33

He also provided his analysis of counterproposals made by the National Association of Broadcasters, the Radio Manufacturers Association, and others. He denoted these proposals as The Broadcasters’ Plan. In contrast to the Engineering Committee’s proposals, The Broadcasters’ Plan lacked clear channels. Only the engineers’ recommendations provided the maximum possibility of avoiding interference and providing service to rural areas.34 Similarly, the Chief Engineer wrote that following the engineers’ recommendations would result in “an astonishing improvement in the broadcast service available to the listeners.”35 Hogan’s study of heterodyne interference also noted that one station to a channel would deliver the best service to the largest number of listeners at the expense of not delivering local and regional service.36

The FRC’s General Order 40 created 40 clear-channel stations—frequencies with a single station transmitting at night. The order reflected the technocratic desires to maximize the audience served and to provide service to all Americans.

Efficient Use of the Right to Transmit

Consistent with technocratic values, three FRC orders—General Orders 97/111, General Order 105, and General Order 116—required broadcasters to operate more “efficiently.” The easiest case to understand is General Order 105, which required a broadcaster with a full-time license to operate at least twelve hours per day, with three of those hours being between 6 p.m. and midnight.37 The efficiency benefits of such a rule are obvious. A station that is not transmitting is not providing a signal for the listeners—idling consumer investment in receivers.

General Order 97, quickly replaced by General Order 111, required stations to use transmitting equipment that could provide modulation of 75 percent or better.38 Increasing the modulation factor from 40 percent to 80 percent increased coverage as much as would a fourfold increase in transmitter power. But, unlike increased power, increased modulation efficiency did not create additional interference. More listeners received the station; existing listeners got a better signal. If 500 stations had to spend $10,000 on average to upgrade their modulators, that $5 million investment would upgrade the service delivered to the more than $1 billion consumer receivers.

In 1931, General Order 116 required broadcasters to maintain their transmitter to within 50 Hz of its nominal frequency.39 If all stations transmitted with this accuracy, heterodyne interference would be substantially reduced or effectively eliminated, and every station’s coverage would expand. The FRC estimated that requiring stations to accurately control their carrier frequency would expand the area covered by each station by a factor of 4.6—thereby extending coverage.40

Some authors have noted an anticompetitive element to these actions. For example, Thierer and Skorup wrote

Further, the FRC defined “public interest” in ways that systematically excluded smaller competitors and entrants through capital requirements, advanced technology requirements, and requirements to broadcast continuously. These regulations were devastating for nonprofit, educational, and small radio operators; many folded.41

Slotten provided examples of two stations that needed to spend $3,900 and $3,000 to comply with such “rigorous” technical requirements.42 He also described a third station, one that broadcast only one hour per week, that gave up its license rather than spend $6,000 to $10,000 to comply.

General Orders 111 and 116 imposed costs on broadcasters but substantially expanded industry output—as measured by the number of useful signals available to the average household. For perspective, $3,000 was roughly the investment required to put receivers in 30 households.

These rules may have arisen from a desire to stifle marginal stations. But it is simpler and easier to accept that the technocratic thinkers noticed that a few investments, tiny compared with the public’s investment in radio receivers, would yield a substantial increase in industry output—as measured by the number of signals available to a consumer and by the quality of those signals. The technocrats promoted the necessary rule changes and got those changes adopted.

Assertions that these improvements reflected anticompetitive motives conflict with the fact that they increased competition. Many listeners saw an increase in useful signals; essentially none received fewer signals.

Media Diversity, Competition Policy, and Localism

A preference for maximizing the audience served appears to have led the engineering community—including the FRC staff—to prefer alternatives that provided more signals to the average listener at the expense of diversity and localism. In contrast, Moss and Fein convincingly show that Congress considered regulation to be the best tool for preventing industry concentration:

The bulk of the evidence strongly suggests that the fear of concentrated control over mass communication mattered a great deal in the making of American radio regulation. The record also suggests that this concern about concentrated political power provided lawmakers with a perfectly reasonable basis upon which to conclude that a property-rights solution would not have been socially optimal.43

The FRC did act to enhance media diversity. For example, it consolidated nearby stations that were under common ownership—freeing up opportunities for others to broadcast. For example, on September 1, 1928, the FRC consolidated the Chicago-area stations WGN, WLIB, and WTAS into the single station WGN.44

Efficiency of FRC Regulation

The efficiency of FRC regulation is illustrated by (1) rural radio service, (2) elimination of blanketing interference, (3) repacking, and (4) output-enhancing technical regulation.

A 1928 study of radio reception in Minnesota, North Dakota, and South Dakota illustrates the benefits of rural skywave coverage. Jansky gathered data on both daytime radio coverage (local stations) and the ownership of radio receivers in those three states.45 He noted that the probability that a household had a radio receiver was essentially independent of whether the household was served by a local radio station. Jansky calculated that only 40 percent of the radio receivers in those states could receive a useful local signal and inferred that many listeners valued distant signals received at night highly enough to justify investing in a radio. He concluded that radio can provide greater economic and social benefits to rural areas than to cities.

A decade later, in 1938, the Federal Communications Commission (FCC) estimated that the only nighttime radio service for 21 million people, or 17 percent of the nation’s population, was the secondary service provided by the clear-channel skywave signals.46 The high-power clear-channel stations also provided the only nighttime service to another 23 percent of the population with their interference-free groundwave signal. That is, 40 percent of the population relied on clear-channel signals for nighttime service.47 The FRC had moved high-power transmitters away from built-up areas where they created substantial interference to nearby receivers.48

In a series of orders culminating in General Order 40, the FCC rearranged the stations on the AM band, thereby reducing interference by increasing the geographical separation of stations operating simultaneously on the same frequency.

Concluding Thoughts on Technical Regulation

As broadcasting technology improved, the FRC imposed regulations that required broadcasters to adopt such improved technology. These regulations increased the quality and quantity of broadcast signals available to the average listener.

An Alternate Approach—Spectrum Property Rights

In his famous article “The Federal Communications Commission,” Ronald Coase suggested that property rights in the radio spectrum would serve efficiency and illustrated this concept with broadcasting examples.49 The various property rights systems for the broadcasting industry that Coase describes all have the feature that the property is owned by the broadcaster and is “used” when the broadcaster transmits. No rights are assigned to the receivers owned by consumers. Coase described interference negotiations that take place between broadcasters—the entities that create interference. Consumers—who are denied service because of interference—have no role in the bargaining regarding acceptable interference. Both broadcasters and their audiences are users of the frequency, but Coase envisioned a world in which all property rights reside with the broadcaster.50

Now, Coase was a smart guy, and he understood that the interests of the audience should be considered in determining the optimal level of interference. He offered a weak, nonquantitative explanation for why vesting rights with the audience should not be considered:

In general, as the distance from a radio station increases, it becomes more and more difficult to receive its signals. At some point, people will decide that it is not worthwhile to incur costs involved in receiving the station’s signals. A local station operating on the same frequency might be easily received by these same people. But if this station operated simultaneously with the first one, people living in some region intermediate between the stations may be unable to receive signals from either station. These people would be better off if either station stopped operating and there was no interference; but then those living in the neighborhood of one of these other stations would suffer. It is not clear that the solution in which there is no interference is necessarily preferable.51

The interference described by Coase matches that in TV or FM broadcasting but does not match nighttime interference in the AM band. Figure 2 illustrates the differences between the interference models. The left portion of Figure 2 shows the interference model that Coase considered. An added station creates a region of interference between the two stations. The right portion of Figure 2 shows nighttime interference in the AM band as it was in the 1920s (the heterodyne era). Interference destroys reception over the vast majority of the nighttime coverage.

FIGURE 2

Skywave interference compared with more conventional interference

FIGURE 2

Skywave interference compared with more conventional interference

Close modal

Coase’s Broadcasting Property Right—First Try

Coase considered alternate approaches to defining spectrum property rights for broadcasting. Perhaps the simplest but most restrictive approach regarding subsequent market transactions would be to auction off the current licenses. Coase describes this option saying

If the right to use a frequency is to be sold, the nature of that right would have to be precisely defined. A simple answer would be to leave the situation essentially as it is now: the broadcaster would buy the right to use, for a certain period, an assigned frequency to transmit signals at a given power for certain hours from a transmitter located in a particular place.52

Such a property right could work well—but, as Coase noted, the “owner” could not modify spectrum usage to reflect changing technology and demand.53 Moreover, it does not allow the regulatory authority to repack the radio band when changes in law or technology require such repacking. If the property right were modified to be the right to operate a broadcast station with specified coverage at a given location—but on a frequency to be determined by the FRC—then the right would be quite similar (in practice, if not in theory) to the rights granted to broadcasters by the FRC when it licensed them. For example, the FRC stated

The first important general principle in the validity of which the commission believes is that, as between two broadcasting stations with otherwise equal claims for privileges, the station which has the longest record of continuous service has the superior right. This is not a doctrine of vested rights or an extension of the property law to the use of the ether; it applies only as between private individuals or corporations operating stations and not as between either of them and the plenary power of the United States to regulate interstate commerce.54

Coase’s Broadcasting Property Right—Second Try

Coase suggested that a more general property right could be created by auctioning off the nationwide rights to use a radio channel (say, the AM channel with carrier frequency 700 kHz) to a single bidder and allowing that purchaser to subdivide that property. Coase described this option thus:

The simplest way of doing this would undoubtedly be to dispose of the use of a frequency to the highest bidder, thus leaving the subdivision of the use of the frequency to subsequent market transactions.55

In Coase’s scenario, an additional cochannel station creates a small amount of interference in the region midway between the two stations. That is more or less true for FM or TV broadcasting, but as explained above, that is not at all the case for AM coverage at night—skywave interference substantially shrinks a station’s nighttime coverage—in Hogan’s example, the area covered shrinks by 99 percent.56

Coase recognized that one needed to consider the loss of service to “people living in some region intermediate between the stations” when evaluating the costs of added interference. But he concluded that the benefits of a new broadcast service to those living near the new station might outweigh the costs of the added interference. That is, he recognized that in broadcasting, it is reception by the consumer that is harmed by interference. He was able to drop the audience from the analysis by asserting that the benefits to one part of the audience from the interfering signal might exceed the costs to another part—although he made no attempt to quantify the costs of interference or the benefits of added service. He made it clear that, as a general proposition, the benefits to the audience would not always exceed the costs.

The broadcast audience most closely corresponds to the doctor in Coase’s recounting of Sturges v. Bridgman—a case in which noise and vibration from machinery disturbed a doctor at work. Of course, interference to reception of a broadcaster’s signals would probably affect the profits of the broadcaster. For some reason, Coase saw no need to assign property rights to the listening audience or incorporate their preferences into the spectrum market that he described for broadcast frequencies.57 Much of the analysis offered in Coase’s rightly famous article is sound. Coase considers many radio systems in which the same entity owns both the transmitter and the receiver. In such cases, bargaining between owners of transmitters and receivers is far simpler than it would be in broadcasting.

Perhaps Coase chose to use the example of broadcasting to illustrate his general insight because doing so allowed for easy and understandable exposition. Readers were familiar with the basics of broadcast coverage—such as a radio signal weakening and finally fading out as one drives away from a city. Coase could have considered a radio service such as land mobile radio, in which the user both transmits and receives. The use of an example that was less familiar to readers would have made exposition more difficult. Coase offers arguments that apply to services such as land mobile but do not apply to broadcasting. For example, discussing adjacent-channel interference, he stated

And the operator whose signals were interfered with would be willing to pay to stop this interference an amount up to the decrease in the value of his service which it causes or the costs he has to incur to offset the interference.58

To prevent adjacent-channel interference, an operator of a land-mobile system can replace the existing stock of receivers with receivers that resist adjacent-channel interference better. This is possible in the land-mobile context because the operator owns the receivers, and there are relatively few of them.59 A broadcaster cannot require consumers to buy new receivers, and a single broadcaster would not be able to afford to buy a city’s worth of improved receivers.

A second possible explanation for Coase’s focus on broadcasters is that he was troubled by some of the consequences of licensing broadcasters, such as (1) the tension between First Amendment values and broadcast regulation and (2) the corrupting incentives that occur when bureaucrats give away rights worth millions of dollars.60

A Hypothetical Coasian Property Rights Market in 1927

Consider what the outcome would have been if the FRC had auctioned off the 90 frequencies in the AM band that were used by stations in the United States.61 I assume that different firms purchase each frequency and that, contrary to fact, there are no adjacent-channel constraints.62

Far Fewer Clear Channels Signals

I know what I would have done if I had purchased the frequency of 700 kHz from the FRC. Rather than using it at night for a clear-channel station in Ohio—as did the entity licensed by the FRC to use 700 kHz and as the successor firm still does—I would operate nighttime stations in New York, Chicago, Los Angeles, and so on—adding relatively high-power stations in major cities until adding the next station would not be profitable. Strong local stations in a few big cities would be more valuable than a clear-channel station in Cincinnati. I would follow the logic of Coase’s example—and build stations that served the cities but ignored the regions between the cities. I suspect that other owners of frequencies would follow similar strategies—calculating that offering local advertising and local programming on a strong signal in a handful of large cities would be far more profitable than operating a clear-channel station that provided local advertising in one city and national advertising to many other cities and rural America on a weak signal that was subject to fading and was susceptible to static and electrical noise. This view is supported by the fact that, when the FRC was studying repacking the broadcast band, the broadcast industry presented a plan to the FRC with many local stations and no clear-channel stations.63

The National Electrical Manufacturers Association—representing, among other interests, radio receiver manufacturers—saw more benefit in clear channels. But even they recommended only twenty-eight such channels.64

Addressing station location and advertising in an earlier publication, Coase concluded

As an abstract proposition, it is incontrovertible that, in these circumstances, stations would tend to be established where the revenue was highest. . . . Perhaps it was assumed that the broadcasting service would be financed by revenue obtained from advertisements. But all this argument tells us that if there were not control of the location of broadcasting stations and the service were financed by means of revenue from advertisements, broadcasting stations would not necessarily be set up in the right places.65

Three cases indicate that market forces would have operated as I predict—sacrificing rural skywave coverage for more urban coverage.

WGY/KGO

The FRC’s General Order 40 reduced General Electric’s station WGY in New York (790 kHz) from full time to daytime-only operation to protect the clear-channel signal of KGO. Station KGO in San Francisco, also owned by General Electric, was the only station allowed to transmit at night on 790 kHz. General Electric successfully appealed the FRC order and gained the right to transmit in both New York and California at night. General Electric regarded nighttime service in New York as preferable to higher-quality skywave service in the rural West.

Breaking Down the Clears

Foust stated that, by the end of 1934, only thirty-two of the forty clear channels of General Order 40 remained clear.66 All stations authorized to share a clear-channel frequency at night had the consent of the original clear-channel station to do so. Many original stations had been paid to give this consent. Foust stated that many stations resisted accepting such buyouts because they hoped that the FCC would permit them to operate at much higher power in the future. Foust indicated that the possibility of being able to operate at 10 times higher power, with the consequent substantial improvement of a station’s local and regional signal coverage, was more valuable than the buyout offers from cochannel stations, but that the buyout offers were more valuable than any increased advertising revenue made possible by a station’s skywave signal.67

Ray Livesay, founder of the Daytime Broadcasters Association, recounted that he had approached FCC Chairman Rosel Hyde regarding the possibility of daytime broadcasters buying one or more clear channel stations and allowing daytime broadcasters to operate on that channel at night. Livesay also stated that Chairman Hyde was strongly negative and indicated that there was no chance that the FCC would approve such a transaction.68

Rate Cards and Revenue Statistics

Prices charged for advertising on clear channel stations and regional stations also indicate that operating multiple stations on a frequency might have been more profitable than operating as a clear channel station. An article in Broadcasting in 1938 reporting on FCC hearings on clear channel policy noted that the advertising rate for clear channel stations was about $1,000 per hour, with WABC in New York receiving $1,250.69 WLW, then operating at 500 kW, received $1,080 per hour. The markedly superior coverage of WLW’s signal, transmitted at 10 times the power of other clear channel stations, does not appear to have resulted in exceptional advertising revenues. In contrast, the highest reported hourly ad rates for regional stations were $475 and $440 per hour. Similarly, the highest rates for local stations were $150 and $125 per hour. In that same hearing, Joseph Maland, testifying for the Clear Channel Group, stated that the revenues of regional stations in large markets were comparable to those of clear channel stations in similar markets. These values are only suggestive, not definitive. A more complete analysis of the costs of operating multiple stations must be considered. However, note that the revenue from the daytime operation of a station would cover many of its fixed costs. Thus, the comparison should be to the marginal cost of the hours of additional operation each day.

FRC General Order 40 fit, on average, 4.25 regional stations on a channel and thirty-one local stations on a local channel. The FRC ensured that regional stations were sufficiently far apart that interference did not substantially limit each station’s local coverage.70 Assuming that the owner of a frequency would sacrifice some coverage at the edge of each station’s service area in exchange for being able to operate more stations, it is reasonable to deduce that a frequency owner could operate six stations with coverage similar to that of regional stations or forty stations with coverage similar to that of local stations. If the average hourly revenue of a regional station was $350, then a frequency with six regional stations would generate $2,100 per hour. Similarly, if the local station’s hourly revenue was $75 per hour, a frequency with forty local stations would generate $3,000 per hour. These calculations, based on a simple model of station economics, are suggestive, not definitive.71

Radio Manufacturers

In addition to advertising revenues, radio equipment manufacturers and retailers were motivated by strong incentives to provide programming to spur the sale of radio receiving equipment. In 1927, sales of receivers in rural areas motivated radio manufacturers to provide skywave service. However, the structure of the radio manufacturing industry changed rapidly in the 1930s.

In 1932, to settle an antitrust suit, General Electric and Westinghouse divested their ownership interests in RCA and withdrew from radio manufacturing for the next thirty months. Many key patents on radio receivers expired in the 1920s and early 1930s. Consequently, competition reduced the rents that the patent owners (principally RCA and its stockholders) received from equipment sales and patent royalties and retail prices of radios fell.72 These events weakened manufacturers’ incentives to provide skywave service and strengthened incentives to operate multiple stations on a frequency.

Consumer Welfare Implications of the Lack of Skywave Service

Assessing the economic efficiency of the FRC’s regulation in a Kaldor–Hicks sense in which the benefits to the rural audience of the skywave signal (measured by the rural audience’s willingness to pay for the option of additional useful signals) are traded off against the benefits of increased signal availability in other areas of the country (measured by willingness to pay) is difficult because (1) good data on the willingness to pay of both groups are unavailable and (2) the difference in radio broadcast service between the outcome under FRC regulation and the outcome under a market in frequency rights is unknown.73

However, given reasonable assumptions, one can examine the tradeoffs involved. Recall that, as noted above, about 17 percent of the population relied on the skywave coverage of clear channel stations for nighttime service, and another 23 percent relied on the interference-free groundwave signal of the clear channel stations. Assume that, in the frequency market world, skywave coverage is essentially eliminated and groundwave coverage shrinks. In such a scenario, the fraction of the population that would lose nighttime coverage would range from 20 percent to 40 percent. Of course, in this scenario, a substantial fraction of the population—maybe 50 percent or so—would see an increase in the number of available signals. Assume that (1) 50 percent of the population sees an increase from five useful broadcast signals to eight, (2) 20 percent of the population sees a decrease from three useful broadcast signals to zero,74 and (3) willingness to pay for an additional broadcast station when one already has n useful signals is proportional to 1/(n+1)—the marginal value of additional signals begins at a high value but falls off fairly quickly. In this scenario, the benefits to the population receiving improved coverage under the “market” solution are smaller than the costs to the population losing coverage.75

If the marginal utility of an additional broadcast signal declines sufficiently fast, the outcome under FRC regulation was more efficient than the outcome would have been in a Coasian market of nationwide frequency rights. There is also an issue of social equity. Should a substantial fraction of the nation be denied access to the improved flow of news and entertainment made possible by broadcasting even though such access could easily be made available? Would such denial be sustainable in Congress?

Adjacent-Channel Constraints

Auctioning off single frequencies would have been complicated by the existence of adjacent-channel interference. The broadcast receivers of the late 1920s were imperfect. A strong signal would block the reception of weak signals on nearby frequencies. The FRC concluded that such adjacent-channel interference extended four channels on either side of an AM station’s frequency. If a station were assigned to operate on 1,000 kHz in Seattle, the FRC would not authorize stations on 960, 970, 980, 990, 1,010, 1,020, 1,030, and 1,040 kHz in Seattle. Table 2 shows the separation distances that the FRC believed in 1931 were appropriate for 1,000-watt cochannel and adjacent-channel stations. The use of the frequency 1,000 kHz in a city blocks the use of eight other frequencies in that city.76

TABLE 2

FRC’s Nighttime Separation Requirements for 1,000-Watt Regional Stations

Frequency Offset (kHz)Required Separation (miles)
1,060 
10 230 
20 105 
30 67 
40 54 
50 
Frequency Offset (kHz)Required Separation (miles)
1,060 
10 230 
20 105 
30 67 
40 54 
50 

Source: Federal Radio Commission. Fifth Annual Report of the Federal Radio Commission to Congress for the Fiscal Year 1931. U.S. Government Printing Office, 1931, 35.

Coase considered adjacent-channel interference and concluded the following:

What this analysis demonstrates, so far as the radio industry is concerned, is that there is no analytical difference between the problem of interference between operators on a single frequency and that of interference between operators on adjacent frequencies.77

That is, the frequency owners can negotiate to determine which owners will operate in New York. I saw nothing in Coase’s article on the FCC that indicated that interference could impair not only reception on adjacent frequencies but also reception on second-, third-, and fourth-adjacent frequencies. There is an analytical difference if the frequency user must reach agreement with eight parties rather than with one.

Return now to our hypothetical auction of ninety frequencies. The most valuable broadcast market in the United States was New York. Which frequency owners would get the right to operate there? There would be a difficult bargaining problem in New York City alone. It is even harder to imagine ninety frequency owners sitting down in a room and reaching agreement on a channel plan that works not only in New York but all up and down the East Coast.78 Coase’s discussion of bargaining regarding adjacent channel interference concludes as follows:

Whether the number of parties normally involved in transactions involving users of adjacent frequencies would be unduly large and call for special regulation, only experience could show.79

Well, perhaps. But, if we accept the FRC’s adjacent channel rules as representing the underlying physics and state-of-the-art technology, the owner of 1,000 kHz would need to reach agreements with eight other frequency owners in order to gain the right to transmit in New York. Moreover, each of those eight frequency owners has, in addition to its adjacency relationship with 1,000 kHz, an adjacency relationship with seven other frequency owners. All ninety AM frequencies are tied together by a chain of adjacency relationships. Because Boston and New York are less than 200 miles apart, the 230-mile first-adjacent channel constraint also precludes the use of 990 and 1,010 kHz in Boston if 1,000 kHz is used in New York.

Other approaches might solve the adjacent channel interference problem. Perhaps the courts could impose a first-in-time doctrine whereby, if there were conflicts over the use of adjacent channels, a frequency owner that put a station in operation would be protected against adjacent-channel interference from stations beginning operation afterwards. Such a doctrine would give all frequency owners a strong incentive to immediately put stations in operation in order to stake claims for such protection.

A second solution would allow one firm to buy up adjacent channels. If a single firm owned all 90 channels, then all adjacent channel interference would be intrafirm, and these bargaining problems would be eliminated.80 Of course, the single firm solution results in a monopoly—and concerns about efficiency and media diversity.

Coase analogized interfering signals from adjacent channels to smoke:

Of course, if there were only one source of smoke and only one person were harmed, no new complication would be involved; it would not differ from the vibration case discussed earlier. But if many people are harmed and there are several sources of pollution, it is more difficult to reach a satisfactory solution through the market.81

Coase does not tell us how big many and several must be for regulation to be likely to outperform the market. It seems clear that a “new complication” existed in radio broadcasting in 1927 because transmitters generated adjacent-channel interference to reception on eight other frequencies.

Elimination of Heterodyne and Blanketing Interference

Recall that the FRC imposed regulations that eliminated heterodyne interference, thereby increasing each station’s coverage area by more than a factor of four, thereby benefiting consumers and probably benefiting most broadcasters. Could a property-rights system have achieved the same outcome and, if so, how? Each station benefits if all other stations act to eliminate heterodyne interference. But, the last station on a frequency to act generates far larger benefits to the others than it does for itself. The incentives for holding out are strong and, unlike the situation in real estate, broadcasters cannot build around a holdout. Given the magnitude of the benefits, broadcasters would likely eventually negotiate an agreement. But given the diversity of institutions and individuals holding broadcast licenses, reaching agreement would be slow—denying the benefits for months or years.

The FRC also moved high-powered transmitters out of built-up areas. High-power transmissions block the reception of other stations in a region near the transmitter—a form of adjacent-channel interference that extends over more than four channels. A broadcaster had little incentive to move its transmitter out of town—doing so would cost money and would improve reception of its competitors’ signals.

Applying Coase’s Qualifying Criteria to the Literature

A recent book, Hazlett’s The Political Spectrum, consistently repeats this founding myth.82 It would be an interesting exercise to compare Coase’s careful qualifying language with Hazlett’s exposition of the specifics of a property-rights system. Unfortunately, Hazlett does not describe the property rights system he envisions. He describes what he calls common-law property rules that existed from 1920 to 1926 but were abolished with the passage of the Radio Act. More correctly, they vanished when the court concluded that the existing statute did not give the Commerce Department the authority to enforce them. The property rules that Hazlett describes—priority in time gives priority in right—were endorsed and supported by the FRC when it stated

The first important general principle in the validity of which the commission believes is that, as between two broadcasting stations with otherwise equal claims for privileges, the station which has the longest record of continuous service has the superior right.83

Hazlett cites one court case for the proposition that property rights in the AM band could be defined and enforced. Yet, the decision in that case was limited. It did not consider the effects of its order on the sky-wave service delivered in other states or on other stations in the Chicago area. It is a weak foundation for a complete theory of property rights.

Below, I rephrase Coase’s qualifiers in a form that facilitates analysis. Coase’s qualifiers include the following:

  • Efficient use of the broadcast spectrum requires consideration of audiences’ preferences.84

  • The communications and computation tasks associated with a property rights scheme must be considered.85

Hazlett does not address how a property-rights system in which the property right was held by the broadcasters would incorporate the preferences of the audience. A simple example of the necessity of considering audience preferences is given by a simple world with two cities—City A and City B and two frequencies, F1 and F2. Each frequency can be used in either city, but use in one city blocks use in the other city. City A has a population of 300,000, and City B has a population of 100,000. Broadcasting is advertising supported, advertising revenues are proportional to the number of listeners, and listeners who have a choice between two stations choose randomly between the two. Broadcaster Alice is the first to enter the market. She buys the worldwide rights to F1 and chooses to operate in City A. Broadcaster Bill buys the rights to F2. Bill also chooses to operate in City A. Having a potential audience of 150,000 (half of the 300,000 population of City A) is better than having a potential audience of 100,000 (the population of City B). However, if consumers greatly value a first broadcast service and mildly value a second broadcast service, consumer benefits will be greater if broadcaster Bill operates in City B. Hazlett describes a property-rights system for broadcast spectrum in which property rights are held exclusively by broadcasters. Recall Coase’s statement, quoted above, that an advertiser-supported broadcasting system could not be guaranteed to locate stations in the right places.

Similarly, Hazlett never discusses the computational problems of market clearing with the interference effects that existed for radio in the 1920s. At night, AM signals create interference at distances of thousands of miles. Adjacent-channel interference effects link all channels in a region into a daisy chain of interlocking interference problems.

Conclusions

Many Americans received better radio service under the regulatory system of the FRC than they would have under a property-rights regime in which ninety entities bought the right to operate on one of the ninety AM channels available to the United States. The FRC was a reasonable surrogate for the listening audience, which property-rights advocates leave out of the market. If one were to interpret public interest, convenience, and necessity to mean represent the consumer’s interest in good broadcast service, one could explain a large fraction of the FRC’s regulation of AM broadcasting.

In setting forth its understanding of public interest, convenience, and necessity, the FRC used technocratic language to describe such a consumer surrogate role:

The commission is convinced that the interest of the broadcast listener is of superior importance to that of the broadcaster and that it is better that there should be a few less broadcasters than that the listening public should suffer from undue interference. It is unfortunate that in the past the most vociferous public expression has been made by broadcasters or by persons speaking in their behalf and the real voice of the listening public has not sufficiently been heard.86

Recall that Coase explicitly, albeit not very prominently, caveated his analysis with regard to broadcasting. He identified the costs and benefits to the audience as an important factor in assessing the efficiency of outcomes. Unfortunately, many readers of Coase’s article have failed to note and think through the implications of Coase’s observation.

The logic of Coase’s analysis of the problem of radio interference is that bargaining between the interferer and the the party receiving interference can reach an efficient outcome. In situations in which only two parties are involved—say a T-Mobile microwave transmitter is creating interference to reception of a Verizon microwave signal—such bargaining can occur and remedy the problem. But, the interference problems in AM—indeed in all radio frequencies in use in 1927—are far more complex and often tie together dozens or hundreds of broadcasters and thousands or millions of listeners. Moreover, a system of property rights in AM broadcasting that maximizes advertising profits will not necessarily maximize consumer welfare. As Coase recognized, “if there were not control of the location of broadcasting stations,” relying on broadcaster maximization of profits would provide no assurance of an efficient outcome.

Radio regulation in the United States was born in 1912, when safety-of-life-at-sea and national defense were the primary uses of radio and most uses of radio frequencies were as a form of common property—often cooperative use of common property. Regulation was expanded in 1927. At that time, (1) a single broadcast signal could interfere with the reception of dozens or hundreds of other stations, (2) only a tiny fraction of the radio spectrum could be exploited, and (3) Congress was concerned about monopolization and concentration of media control. The structure and powers of the FRC, and later the FCC, reflect this history. The thought that a property rights system would have been either practical or politically acceptable in 1927 is a myth—a myth that can impair one’s understanding of our institutions.

ACKNOWLEDGMENTS

Evan Kwerel, Paige Jackson, David Moss, Ann Jackson, Thomas White, William Scott (WE7W), and two anonymous reviewers all provided useful insights, valuable data, and research support. I thank all of them. Any remaining errors are mine.

FOOTNOTES

1.

Epstein. In 1927, about 700 broadcasters were operating in the AM band. It seems likely that few of them would have supported the auction that Epstein describes. Epstein’s hand-waving invocation of the common law and first possession fails to account for the possession of radio spectrum by consumers who have purchased radios; Hazlett.

2.

Moss and Fein, 389. Additional issues make broadcast policy difficult. First, broadcasting meets the economic definition of a public good—in the sense that when a listener tunes into an existing broadcast signal, the broadcaster incurs no additional cost. But some mechanism is needed to pay for the creation of broadcasts. Alternative mechanisms for covering the costs include (1) voluntary donations to broadcasters from the listening audience, (2) finding broadcasters willing to transmit for free, (3) charging listeners, (4) government support, or (5) advertising. Each of these alternatives has advantages and disadvantages, and all have been used to support broadcasting, although the technology available prior to 1970 made charging infeasible. Second, radio broadcasting can carry political speech. These issues exist independent of the nature of interference in broadcasting, but they complicate the policy choices and the political process around those choices.

3.

This is not a general argument against property rights in the spectrum. I believe that, in many contexts, spectrum property rights systems can serve efficiency. My PhD thesis addressed technological definitions of spectrum property rights. I was part of the team of consultants that designed the spectrum property rights system of New Zealand, and I participated, in a minor way, in the drafting of New Zealand’s Radiocommunications Act 1989—the first statute to incorporate spectrum property rights.

4.

See U.S. Bureau of the Census, Historical Statistics of the United States Colonial Times to 1970. That table reports that 60,000 households had radios in 1922. Series A 288-319 reports that there were 29.9 million households with radios in 1930.

5.

Federal Radio Commission, Second Annual Report.

6.

Caldwell, author’s calculations. For context, the average household income in 1929 was $2,335. See Historical Statistics in Series G308.

7.

Inflation adjustment made using the Bureau of Labor Statistics Inflation Calculator at https://data.bls.gov/cgi-bin/cpicalc.pl. The inflation multiplier is 17.1.

8.

A 1937 survey by the FCC of 618 commercial broadcasting stations showed a total investment of $46.2 million and an investment in technical equipment of $28.3 million. The stations surveyed included a 500-kW station (WLW) and thirty-one 50-kW stations (Federal Communications Commission, Fourth Annual Report, 220). In 1927, no stations were yet operating at 50 kW, and some stations were homemade.

9.

Hilliker.

10.

Slotten, Radio’s Hidden Voice, 143.

11.

Federal Radio Commission, Second Annual Report, 55–64, author’s calculations.

12.

Schneider, 11.

13.

The recommendations of the Second National Radio Conference of March 20–24, 1923, called for broadcasting stations to keep within 2 kHz of their assigned frequency. Department of Commerce, Bureau of Navigation, Amendments to Regulations.

14.

Hogan stated that heterodyne interference occurred with an interfering signal that was 14 dB lower (a factor of 25) than would be regarded as causing interference today (Hogan, 1359).

15.

Ibid., 1363.

16.

Ibid., 1361.

17.

Radio Act of 1912. Howeth provides a thorough description of the development and passage of this legislation. (Howeth, 153–65.) Epstein’s description of this legislation is amusing. He states, “In 1912, federal legislation gave the United States control over the spectrum, much of which was turned over to the navy for maritime operations. With one stroke the government bypassed the traditional common law rules of first possession.” Epstein, 280, footnote omitted. It is not clear how “first possession” would apply to a shared resource such as the emergency and calling channel 300 m. The only significant division of spectrum between naval and other uses is a time-division requirement that, in some circumstances, restricts the Navy to transmit only during the first fifteen minutes of the hour and reserves the other forty-five minutes for stations other than naval and military stations. Of course, because radio was the only way to communicate with ships beyond the horizon, the Navy operated many radios.

18.

White.

19.

Department of Commerce, Bureau of Navigation, Amendments to Regulations.

20.

Department of Commerce, Bureau of Navigation, Recommendations of the National Radio Conference.

21.

United States v. Zenith Radio Corporation.

22.

Report of the Attorney General, July 8, 1926. In fact, the opinion did allow limits on broadcast frequencies—but only those limits specified in the 1912 statute (188–500 kHz).

23.

Moss and Decker.

24.

The Radio Act had passed both houses and gone to conference on July 3, 1926, before the acting Attorney General issued his opinion.

25.

Federal Radio Commission, Annual Report, 2; Federal Radio Commission, Second Annual Report, 2; Federal Radio Commission, Third Annual Report, 3.

26.

Federal Radio Commission, Seventh Annual Report, 1.

27.

Slotten, Radio and Television Regulation, 43.

28.

Ibid., xiv.

29.

Federal Radio Commission, Fifth Annual Report, 33.

30.

Hogan, 1326.

31.

Federal Radio Commission, Third Annual Report, 6.

32.

Federal Radio Commission, Sixth Annual Report. Note that, to accommodate transmitters with poor frequency stability, the FRC later modified the police channel plan to use 8 kHz channels.

33.

Federal Radio Commission, Second Annual Report, 132, emphasis added.

34.

Ibid., 143.

35.

Ibid., 144.

36.

Hogan.

37.

Federal Radio Commission, Fifth Annual Report, 94.

38.

Federal Radio Commission, Sixth Annual Report, 23.

39.

Federal Radio Commission, Fifth Annual Report, 23.

40.

Ibid., 24.

41.

Thierer and Skorup, footnotes omitted.

42.

Slotten, Radio’s Hidden Voice, 143.

43.

Moss and Fein, 391, 409.

44.

Federal Radio Commission, Second Annual Report, 162.

45.

Jansky, 1362.

46.

Federal Communications Commission, Fourth Annual Report, 198–9.

47.

“FCC Winding Up Hearing on Allocations,” 41. Author’s calculations.

48.

Federal Radio Commission, Second Annual Report, 169.

49.

Coase, “The Federal Communications Commission”; Coase is more measured and less dogmatic than Epstein and Hazlett about the feasibility of a market in spectrum, and he frequently caveats his general observations with careful qualifications.

50.

If one accepted the common-law rule of first possession as defining property rights, it would seem reasonable to say that a consumer who bought a radio to listen to signals on 700 kHz had taken possession of 700 kHz in the neighborhood of his or her home and should have rights against any new station that interfered with listening to programs on 700 kHz. Of course, defining property rights this way would lead to a maze of overlapping, nonexclusive rights. But, if first possession were the rule as Epstein suggests, then it would be reasonable to say that a receiver owner possesses a block of spectrum.

51.

Coase, “The Federal Communications Commission,” 27, emphasis added.

52.

Ibid., 25.

53.

In fact, many of the traditional, non-property-rights licenses issued by the FCC did accommodate technological change. For example, point-to-point microwave links evolved from FM to AM to digital. In these systems, the same entity operates both the transmitter and receiver and can upgrade them together. Similarly, Equatorial Communications was able to deploy a digital, spread-spectrum system using satellites that had originally carried analog TV signals.

54.

Federal Radio Commission, Third Annual Report, 32.

55.

Coase, “The Federal Communications Commission.”

56.

Hogan; Attempting to apply Coase’s frequency rights system to analog TV and FM broadcasting would not encounter the problems created by skywave interference. There would still be issues associated with interference on second and third adjacent channels that might make Coase’s discussion of adjacent channel interference inapplicable. In contrast, a Coasian approach to property rights would work well for modern wireless systems such as those operated by AT&T, Verizon, and T-Mobile. In fact, given that the FCC permits subdivision of such licenses and permits licensees to negotiate interference disputes, one could argue that it does work well.

57.

If consumer surplus were to correlate perfectly with broadcaster profits, then it would be reasonable to ignore the listening audience and focus on the broadcasters.

58.

Coase, “The Federal Communications Commission,” emphasis added.

59.

These observations are appropriate for land-mobile radio at the time Coase was writing.

60.

Ibid., 39, 36.

61.

This is easier said than done. The idea of a frequency needs to be defined. A reasonable definition would include limits on (1) maximum power and (2) out-of-band emissions. It might also incorporate adjacent-channel protection rules based on in-band emissions.

62.

The FRC engineering rules required that stations in the same city operate on frequencies separated by at least 50 kHz. This rule reflected the fact that existing receivers could not reject strong signals on frequencies close to the desired signal. That is, if there was a strong signal on 1,000 kHz, reception would be impossible on 990 and 1,010 kHz.

63.

Federal Radio Commission, Second Annual Report, 135–40.

64.

Ibid., 142.

65.

Coase, British Broadcasting, 181–2. Coase was considering a model in which (1) there were about 10 channels available and (2) skywave service was not an issue. In these circumstances, there would be broadcasters in the largest cities and none in the medium-size cities.

66.

Foust, 45.

67.

The issue of superpower stations—stations operating at powers as high as 500 kW—illustrates how difficult it is to define spectrum property rights. Would a common-law property right include a limit on the maximum power transmitted on the frequency? If so, what would that limit have been? How would that limit have been determined? If technology changed, could the limit be changed?

68.

Ray Livesay, personal communication. It may be relevant that Chairman Hyde had grown up in rural Idaho. He would have known many people who valued skywave service.

69.

“FCC Winding Up Hearing,” 41.

70.

Federal Radio Commission, Fifth Annual Report, 32.

71.

One reviewer questioned whether the equilibrium outcome would eliminate all nighttime skywave coverage. That reviewer pointed out that if, roughly speaking, the clear channel audience was 20 percent of the population, then serving it might be more valuable than sharing the market in a dozen or so cities. This analysis is complicated by the facts that (1) a single station’s skywave signal covers only a portion of the nation—so a single station’s skywave signal could probably serve 5 percent to 10 percent of the national audience; (2) many advertisers, such as a store promoting a sale, are only interested in the local audience—thus the value of advertising in programming delivered to the skywave audience should be expected to be less than the value of local advertising; and (3) operating a station confined to daytime hours makes network affiliation less attractive and makes consumers less likely to tune to that station or to leave the radio tuned to that station when they turn the radio off in the evening. I agree that an equilibrium with no clear-channel stations would not have been inevitable, but it seems unavoidable not to conclude that there would have been very few, if any, clear-channel stations.

72.

Recall that, as noted above, the average cost of a radio receiver in 1927 was $124 (1927 dollars). By 1938, the average cost had dropped to $35 (Caldwell, 13, author’s calculations, 1938 dollar).

73.

See Spence and Owen, and Steiner for earlier discussions of market failures in broadcasting that could result in economically inefficient underserving of small separate populations (in terms of tastes as opposed to geographic areas) and overserving of large populations.

74.

Although there were 40 clear channel frequencies, the signals on only some of those frequencies provided useful signals at any specific location on a given night. As noted above, a skywave signal does not cover the entire nation. Many factors, such as time-of-day and the extent of sunspot activity, cause skywave coverage to vary. Moreover, many of the clear-channel stations were affiliated with major networks. Three clear channel signals, all carrying NBC, did not provide three listening choices. Thus, I conservatively assume that skywave service provided only three useful signals to the average rural listener.

75.

The relevant total benefits under regulation are proportional to 50%*2.28 + 20%*1.83 = 1.51; benefits under the market solution are proportional to 50%*2.72 + 20%*0.0 = 1.36. Notice that this argument assumes that rural and urban consumers value radio equally. Recall that Jansky thought that radio broadcasting was of “greater economic and social service” to rural audiences (Jansky, 1362).

76.

A reviewer pointed out that it might have been possible for broadcast stations to operate on closely spaced or adjacent frequencies if they operated at similar powers and were collocated. Hull discusses the performance of radio receivers in 1928 and 1929. He concludes, “Modern receivers of what might be termed medium quality and cost are quite generally able to suppress crosstalk from an unwanted signal in the adjacent channel which has the same field intensity as the wanted signal” (Hull, 1335). But he goes on to describe the testing of twenty receivers sold in 1927 and 1928 and twenty-four receivers sold during 1928 and 1929. Of the group of twenty receivers, only five were able to adequately reject an adjacent channel signal when tuned to 1,000 kHz. Of the group of twenty-four receivers sold one year later, eleven adequately rejected an adjacent channel signal at 1,000 kHz. Given that many households had receivers sold earlier, it seems unlikely that the use of collocated, adjacent-channel transmitters would have been acceptable in 1927.

77.

Coase, “The Federal Communications Commission.”

78.

In the general case, the problem of clearing such a market is NP-complete and requires simultaneous consideration of the value placed by every frequency owner for operation in each city. For a discussion of the difficulties of performing such a computation, see Newman et al. They describe a program, SATFC, that can solve an instance of a problem in this class (considering about forty frequencies) in a second or two on a modern computer. SATFC was used to compute market-clearing packings in the recent FCC incentive auction.

79.

Coase, “The Federal Communications Commission.”

80.

The natural name for this monopoly is the Big Broadcasting Company (BBC).

81.

Coase, “The Federal Communications Commission.”

82.

That book is unreliable. For example, after describing the Oak Leaves litigation, Hazlett states

Now Congress acted. An emergency measure, introduced and passed just days after the ruling, mandated that all wireless operators immediately waive any vested rights in frequencies; failure to do so would result in license termination. This provision was then included in the Radio Act, a compromise being cobbled together from separate House and Senate bills. (Hazlett, 42, emphasis added)

Hazlett states that the date of the ruling is November 17, 1926. The relevant measure, S. J. Res. 125, was introduced in the Senate on July 3, 1926, it passed the Senate and was transmitted to the House, where it also passed that day—more than four months before the ruling. The passage of S. J. Res. 125 is described in Coase with the proper date. As for the provision then being included in the Radio Act, causation went the other way. The Radio Act went to conference on July 2, 1926, and the conferees were unable to reach agreement. They asked that the Congress enact S. J. Res. 125 in order to maintain the status quo until the legislation could be enacted. The primary issue dividing the House and Senate conferees was whether the regulatory authority was to be an independent commission or the Secretary of Commerce. See Congressional Record—Senate 1926 at p. 12959.

83.

Federal Radio Commission, Third Annual Report, 32.

84.

Coase, “The Federal Communications Commission,” 28.

85.

Ibid., 29.

86.

Federal Radio Commission, Third Annual Report, 167.

BIBLIOGRAPHY

35 Op. Atty. Gen. 126
(
July
8
,
1926
).
Caldwell, O. H.
“Retail Radio Sales in the United States, 1922–1938.”
Broadcasting
,
January
1
,
1939
,
13
.
Coase, Ronald H.
British Broadcasting: A Study in Monopoly
.
London
:
Longmans, Green and the London School of Economics
,
1950
.
_______.
“The Federal Communications Commission.”
Journal of Law and Economics
2
,
no. 1
(
1959
):
1
23
.
Department of Commerce, Bureau of Navigation
.
“Amendments to Regulations, Broadcasting Stations.”
Radio Service Bulletin
73
,
no. 11
(
1923
).
_______.
“Recommendations of the National Radio Conference.”
Radio Service Bulletin
72
,
no. 9
(
1923
).
Epstein, Richard A.
The Classical Liberal Constitution: The Uncertain Quest for Limited Government
.
Cambridge, MA
:
Harvard University Press
,
2014
.
“FCC Winding Up Hearing on Allocations.”
Broadcasting
,
July
1
,
1938
:
41
51
.
Federal Communications Commission
.
Fourth Annual Report of the Federal Communications Commission
.
Washington, DC
:
U.S. Government Printing Office
,
1938
.
Federal Radio Commission
.
Annual Report of the Federal Radio Commission to the Congress of the United States for the Fiscal Year Ended June 30, 1927
.
Washington, DC
:
U.S. Government Printing Office
,
1927
.
_______.
Fifth Annual Report of the Federal Radio Commission to Congress for the Fiscal Year 1931
.
Washington, DC
:
U.S. Government Printing Office
,
1931
.
_______.
Second Annual Report of the Federal Radio Commission to Congress for the Year Ended June 30, 1928
.
Washington, DC
:
U.S. Government Printing Office
1928
.
_______.
Seventh Annual Report of the Federal Radio Commission to Congress for the Fiscal Year 1933
.
Washington, DC
:
U.S. Government Printing Office
,
1933
.
_______.
Sixth Annual Report of the Federal Radio Commission to Congress for the Fiscal Year 1932
.
Washington, DC
:
U.S. Government Printing Office
,
1932
.
_______.
Third Annual Report of the Federal Radio Commission to Congress Covering the Period October 1, 1928 to November 1, 1929
.
Washington, DC
:
U.S. Government Printing Office
,
1929
.
Foust, James C.
Big Voices of the Air: The Battle over Clear Channel Radio
.
Ames
:
Iowa State Press
,
2000
.
Goldsmith, Alfred N., and Austin C. Lescarboura.
This Thing Called Broadcasting
.
New York
:
Henry Holt
,
1930
.
Hazlett, Thomas Winslow.
The Political Spectrum: The Tumultuous Liberation of Wireless Technology, from Herbert Hoover to the Smartphone
.
New Haven, CT
:
Yale University Press
,
2017
.
Hilliker, Jim.
Time to Debunk a Myth from Radio’s Infancy: Aimee Semple McPherson’s Telegram to Herbert Hoover
.
2011
,
updated 2022
, http://www.theradiohistorian.org/lax/hoover-telegram.html
Hogan, John V. L.
“A Study of Heterodyne Interference.”
Proceedings of the Institute of Radio Engineering
17
(
1929
):
1354
64
.
Howeth, Linwood S.
History of Communications-Electronics in the United States Navy
.
Washington, DC
:
U.S. Government Printing Office
,
1963
.
Hull, Lewis M.
“Some Characteristics of Modern Radio Receivers and Their Relation to Broadcast Regulation.”
Proceedings of the Institute of Radio Engineering
17
(
1929
):
1334
41
.
Jansky, C. M.
“Some Studies of Radio Broadcast Coverage in the Middle West.”
Proceedings of the Institute of Radio Engineering
15
(
1928
):
1356
67
.
Moss, David, Jonathan Lacklow, J. B. L. Decker. “Capturing History: The Case of the Federal Radio Commission in 1927.” Chapter 8 in
Preventing Regulatory Capture: Special Interest Influence and How to Limit It
, edited by Daniel Carpenter and David Moss.
Cambridge
:
Cambridge University Press
,
2013
.
Moss, David A., and Michael R. Fein.
“Radio Regulation Revisited: Coase, the FCC, and the Public Interest.”
Journal of Policy History
15
(
2003
):
389
416
. https://muse.jhu.edu/article/48407/pdf.
Newman, Neil, Alexandre Fréchette, and Kevin Leyton-Brown.
“Deep Optimization for Spectrum Repacking.”
Communications of the ACM
61
(
2018
):
97
104
.
Radio Act of 1912
, 47 U.S.C. §51 et seq. (
1912
).
Schneider, John.
“Portable Stations—the Radio Rovers of the 1920s.”
Spectrum Monitor
1
(
2014
):
2
3
.
Slotten, Hugh Richard.
Radio’s Hidden Voice: The Origins of Public Broadcasting in the United States
.
Champaign, IL
:
University of Illinois Press
,
2009
.
_______.
Radio and Television Regulation: Broadcast Technology in the United States
, 1920–1960.
Baltimore, MD
:
Johns Hopkins University Press
,
2000
.
Spence, Michael, and Bruce Owen.
“Television Programming, Monopolistic Competition, and Welfare.”
The Quarterly Journal of Economics
91
(
1977
):
103
126
.
Steiner, Peter.
“Program Patterns and Preferences and the Workability of Competition in Radio Broadcasting.”
Quarterly Journal of Economics
62
(
1954
):
194
223
.
Thierer, Adam D., and Brent Skorup.
“A History of Cronyism and Capture in the Information Technology Sector.”
Journal of Technology Law & Policy
18
(
2013
):
Article 2
.
United States v. Zenith Radio Corporation
, 12 F. 2d 614 (
N.D. Illinois
1926
).
U.S. Bureau of the Census
.
Historical Statistics of the United States Colonial Times to 1970
,
bicentennial ed., pt. 2 at Series R 93-105
.
Washington, DC
:
U.S. Government Printing Office
,
1975
.
White, T. H.
“Building the Broadcast Band.”
2008
,
updated 2017
. http://earlyradiohistory.us/buildbcb.htm.
This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.