by Peter W. Huber, Michael K. Kellogg, and John Thorne
Low-volume, local exchange service is a natural monopoly; high-volume, long distance service is not. In 1982, the Bell System was broken into eight pieces on the strength of that assumption. But it is now clear that the assumption was wrong. In fact, it was worse than wrong: the architects of the Bell divestiture got it backwards. If there is going to be a monopoly, it will be in the long-distance market. The local exchange should be -- and soon will be -- competitive.
In 1982, the architects of divestiture observed that long-distance communication depended on microwave radio, and concluded that the economics of radio made competition economically viable. But in the long-distance network, steel towers and microwave radio have since given way to glass and infra-red light, which have radically different economics. In 1982, the theorists examined the local exchange and concluded that monopoly was economically inevitable here because the local exchange depended on underground wires. But in the local exchange, underground wires are now rapidly giving way to microwave radio.
We have AT&T's own word on it, in papers filed with the FCC in May 1992. "[T]he land, towers, and buildings used to support the microwave portion of [AT&T's long-distance] network," AT&T cheerfully notes, are now "available to support other services."(1) Local exchange services in particular. Its "ultimate goal," AT&T declares, is the provision of "affordable, nationwide," radio-based personal communications service, with "[f]eatures and quality comparable with the wireline network."(2) See Figure 1.1 [omitted].
In sum, the people who planned divestiture in 1982 fundamentally misunderstood the role of wire and radio in both local and long-distance telephony. And they confused past regulatory policy with future economic reality. It took a decade of litigation to dismantle the old Bell System; it will take a decade or more to dismantle the economic edifice erected by antitrust lawyers in its place. But the dismantling is now well under way. Within a decade at most, and more likely five years or so, the divestiture decree will have to be abandoned. [Graphic for endnote 3 omitted.]
Bell patented his telephone in 1876. The telephone was a great invention, of course, but useless if you needed to make a call from a ship. So Marconi set about searching for something better. In 1895 he had in hand the first "wireless telephone."
Initially, there were high hopes (or, at Bell, deep fears) that radio would sweep away the telephone monopoly that was then emerging. It didn't. The trouble with radio was that the airwaves seemed to get crowded very quickly, whereas wires could be deployed in almost limitless numbers. With the early technologies, radio's great advantage was not in point-to-point messaging, but in point-to-everywhere broadcast. Congress, in its wisdom, decided to turn a temporary technological limitation into a semi-permanent regulatory barrier. In the Radio Act of 1927, Congress separated the telephone and broadcast industries, and that for many years meant a separation of the wireline and wireless media. The telephone monopoly was thus secure.
But technology has a habit of surmounting regulatory barriers. Researchers at Bell Labs developed microwave communications, and Bell itself began to deploy microwave in the long-distance network. Microwave was first used in commercial long-distance service in 1950. It enjoyed rapid growth and quickly and dramatically eclipsed all other forms of long distance transmission. See Figure 1.2 [omitted].
Already accustomed to distinguishing the competitive world of radio from the "natural" monopoly of wireline telephony, the FCC concluded that this new use of radio technology did not necessarily have to be left to a phone company monopoly. The Commission accordingly began issuing microwave licenses to individual (albeit large) private users,(4) and not just to the (very much larger) Bell company. In this way, it slowly came to be understood that, however "natural" the telephone monopoly might be, radio-telephone operations could be competitive. "Microwave Communications Incorporated" (better known today as MCI) grasped that better than anyone. From the 1960s until the 1980s, during the heyday of microwave technology, MCI built up its business by attacking the Bell monopoly, first before the FCC, then in the courts, and finally in the marketplace. MCI's economic argument was straightforward enough. The basic building block in microwave transmission is a transceiver capable of handling 12 voice calls or equivalent amounts of data or video traffic.(5) Long-distance telephony typically carries a lot more traffic than that, and therefore requires multiple microwave links. Transmission costs thus increase with traffic volumes, and the unit cost of carrying traffic remains constant (or else rises) as traffic volumes increase. These radio services, in other words, were plainly not "natural monopolies." Neither were radio-satellite services, which developed not long afterward. [Graphic for endnote 6 omitted].
The antitrust lawyers at MCI and in the Department of Justice caught on to this in the 1970s, and went to work. They hired economists. The economists, who understood economic theory much better than technological reality, explained everything to the lawyers' satisfaction. Wireline networks -- like the local exchange -- are natural monopolies. A single telco can wire an entire local market more cheaply than could two or more. Radio networks -- like the microwave long-distance network -- are competitive. Mixing up competitive services (like long-distance microwave) with monopoly ones (like local wire) leads to all sorts of economic hanky-panky, so it is better to separate the two. And in 1984, that was precisely what was done.
By then, however, microwave was finished in the long-distance industry. Long-distance carriers were already beginning to dynamite their radio towers to the ground. The telephone industry didn't need radio any more. Except perhaps in a little-noticed new business at the fringe of telephony called cellular telephone.
By 1984, the techno-economic rationale for divestiture was already obsolete. Where radio was really needed now was where Marconi had intended it -- on ships. But every human, even the most committed landlubber, is a sailor of sorts, or else a driver, or a flyer, or at least a pedestrian. After almost a full century of development, the telephone still had a very fundamental shortcoming: telephone wires don't move. People do.
Marconi had solved half the problem; radio telephones worked just fine, on shore as well as on ships. But there wasn't enough spectrum available to serve very many users. A few dozen stations pretty much fill up the dial of a radio -- and radio telephone requires radio stations in pairs to sustain two-way conversation. The radio-telephone networks that operated until the 1980s could typically support a total of 25 channels, only about half of which could be used at a time. Since 30 subscribers might be licensed for every channel, calls often could not be completed.(7) The available spectrum supported only 140,000 subscribers nationwide; demand far exceeded supply. By 1976, Bell mobile service in the New York metropolitan area served 543 customers and had a waiting list of 3,700.(8) Would-be subscribers often waited six years for service.(9) Figure 1.3(a)(10) [omitted].
The solution to the problem had already been worked out by the same people who had developed microwave radio, and at almost exactly the same time. In 1947, engineers at Bell Labs had hit upon the idea of "cellular" radio networks. Radio telephones would be low-power, short-range devices. The same frequencies would then be used again and again, just as they are today with millions of cordless home telephones. A radio conversation on East 42nd Street would not interfere with another one on the same frequency on West 51st. Any area could be divided into separate cells, each one served by its own, low-power transmitter. Figure 1.3(b)(11) [omitted]. The capacity of a cellular system could then be increased almost indefinitely by shrinking cells and increasing their number. Figures 1.4(a), (b) [omitted].
The only problem was that cellular telephony required highly sophisticated transmitters and receivers, and massive coordination among cells to "hand off" calls and coordinate frequencies as the car phone on East 42nd Street moved toward West 51st. No one had the technology to perform this -- until the microelectronics revolution of the 1970s. Then coordination of this kind became perfectly feasible, and an entirely new industry was born. Figure 1.5 [omitted].
In 1977 the FCC granted Illinois Bell's application to construct a developmental cellular system in Chicago,(12) and two years later the Commission authorized Illinois Bell to offer commercial cellular service to the public.(13) Despite the disdain in which it was then held by many antitrust lawyers, the FCC correctly understood the competitive possibilities of radio, as it had since the earliest days of commercial radio telephony. The Commission first opted in favor of competitive local radio telephone services in 1949. Thereafter, the FCC either split the allocation of new radio-telephone frequencies between telcos and nonwirelines, or awarded licenses on an "open entry" basis to the best applicants. When cellular telephony finally came of age, the FCC stuck by its pro-competitive policies.
In 1984, when the Bell System was dismantled, only 32 cellular systems had been licensed; they served some 92,000 customers. In the next five years, however, the FCC completed the licensing of cellular systems in both urban and rural areas. Competition in the cellular telephone industry rapidly filled the space the FCC had allocated. The cellular industry reached the million-customer mark in 1987;(14) by the end of 1991, there were over 7.4 million cellular subscribers in the United States, almost a million of which had first subscribed in the last six months of 1991.(15) By the end of 1992, McCaw Cellular -- unaffiliated with any local telephone company -- was, by any measure, the largest cellular telephone company in the world. There are now some 10 million cellular telephones in operation worldwide. [Graphics for endnotes 16, 17, and 18 omitted.]
But this was only the beginning. With cellular technology, particularly digital cellular technology, old and inefficient uses of frequency can be converted to new and extraordinarily efficient ones. The capacity of the airwaves can be expanded indefinitely, albeit always at a cost: more users require more cells, and more cells require more radios. In economic terms, that means no "natural monopoly." Quite the opposite, in fact: the more crowded the airwaves, the more economically viable competition becomes. A single, very powerful, centrally owned and maintained transmitter is economically favored only in the land of Kim-il Song or Fidel Castro, where one voice says it all. With many voices to carry, there can be many carriers.
And many carriers there will be. Today, there are only two cellular carriers per market, allegedly because of scarce spectrum. But the cellular network is by no means saturated, and the (comparatively inexpensive) transition to digital radio will increase capacities by ten-fold or more, making the argument for additional carriers overwhelming. Moreover, as of August 1992, the FCC had issued over 150 experimental licenses to test "personal communications networks" ("PCNs") and other personal communications services (PCS). In February 1992, the FCC proposed to allocate 220 MHz of new spectrum for PCS and other "emerging telecommunications technologies." (By comparison, the current cellular carriers are each licensed to use 25 MHz of spectrum.) In a July 1992 order, the Commission asked for comments on its proposal to license as many as 5 providers of PCS -- including both current cellular licensees and local telcos. The FCC now projects 60 million PCS users in the United States within ten years.(19) Not long before, the Commission had authorized Fleet Call to convert its dispatch networks into an all-digital, cellular mobile telephone service in direct competition with established cellular carriers.
The increase in available spectrum and the outpouring of new competitors will place tremendous pressure on the local copper loop. The numbers are already strikingly clear. Copper loop averages between $1,200 and $2,000 per access line.(20) Today's cellular industry, by contrast, has invested a total of $10 billion (roughly) and already serves 10 million customers -- an average investment of about $1,000 per customer.(21) Copper technology is stagnating, but every few years, radio technologies grow more reliable and less expensive. If local telcos were to rebuild from scratch today, they would do so mostly with radio, at a cost of about $800 per subscriber.(22) The main thing that discourages them from doing so is the billions upon billions they have invested in what is now obsolete copper plant. It is difficult, after all, to embrace the technology that is going to wipe tens of billions of dollars of undepreciated assets off your balance sheets. Newcomers in the radio market don't have that problem, however, and the local telcos will either adjust or be swept aside. "Make no mistake about it," George Calhoun declares in his seminal *Wireless Access and the Local Telephone Network*, "we are witnessing the beginning of the end of the natural monopoly."(23)
Meanwhile, exactly the opposite is taking place in the long-distance network, which has replaced radio with fiber. For better or worse, the economics of fiber optic cable are as "naturally monopolistic" as any economist could ever imagine. The up-front costs of deploying fiber-optic cable are enormous. Rights of way must be secured, and deploying cable is horribly labor intensive. Costs are incurred at the front end and they are fixed; wire networks have almost zero (or even negative) salvage value; costs are irrevocably sunk before they generate a single dollar of revenue. These costs are also largely the same whether the fiber-optic cable contains one pair of optical fibers or a dozen, whether the fiber is "lit" (i.e., connected to functioning electronics) or "dark," and whether the lit fiber carries a million telephone calls or none at all. Once fiber is in place, however, its carrying capacity can be increased almost indefinitely, and at extremely low cost. Every four years since 1975, the opto-electronics at the ends have quadrupled their power to pump information down the existing strands of glass. See Figure 1.6 [omitted]. The capacities of the long-distance fiber optic networks already in place vastly exceed demand, and carrying capacities are being increased year by year at very little cost.(24) There is, in short, strong reason to believe that long-distance, fiber-optic transmission entails sharply declining average costs over the entire range of current demand. And that, in economic jargon, is a "natural monopoly."
Why then do MCI, Sprint, and a host of smaller competitors survive? They survive because antitrust lawyers and utility regulators expect them to survive, and heads would roll if they didn't. Government can maintain unnatural competition, just as it can maintain unnatural monopoly, and that is just what government is now doing. The long-distance market today contains three facilities-based carriers, additional regional carriers, and hundreds of tiny resellers. Yet AT&T could wipe them all out in very short order, and would do so quickly enough if political, regulatory, and antitrust inhibitions were ever swept aside. Competition in the long-distance market today is an illusion, a triumph of yesterday's elegant theory over today's economic fact. [Graphic for endnote 25 omitted.]....
1. Request for a Pioneer's Preference at 6-7, In re Request of AT&T for a Pioneer's Preference, No. 91-314 (FCC May 4, 1992).
2. Id. at 8.
3. Source: Request for a Pioneer's Preference, In re Request of AT&T for a Pioneer's Preference, No. 91-314 (FCC May 4, 1991).
4. Allocation of Frequencies in the Bands Above 890 MHz, 27 F.C.C. 359 (1959), recon. denied, 29 F.C.C. 825 (1960). Prior to that time, the FCC's general policy had been to license microwave systems only to common carriers and other special entities like right-of-way companies (railroads, pipelines and power companies) and state and local governmental agencies involved in police, fire, and highway maintenance activities.
5. The capacity of the channels can, like fiber optics, be increased by using multiplexing techniques.
6. Source: AT&T Bell Laboratories, A HISTORY OF ENGINEERING AND SCIENCE IN THE BELL SYSTEM: TRANSMISSION TECHNOLOGY (1925-1975), at 781 (E. F. O'Neil, ed. 1985).
7. Of 23 channels supporting 700 users in New York City, for example, only 12 could be used simultaneously.
8. George Calhoun, DIGITAL CELLULAR RADIO 31 (1988).
9. Mayer, Walkie Talkie Telephoning, WASHINGTON POST, June 6, 1982, at K1.
10. Source: Calhoun, DIGITAL CELLULAR RADIO 41.
11. Ibid.
12. Application of Illinois Bell Tel. Co., 63 F.C.C.2d 655 (1977), aff'd sub nom. Rogers Radio Communication Services, Inc. v. FCC, 593 F.2d 1225 (D.C. Cir. 1978).
13. Telocator Network of America v. Illinois Bell Tel. Co., 70 F.C.C.2d 713, 716-717 (1979).
14. Busy Signal A Welcome Sign in Cellular Telephone Industry, CHIC. TRIB., Oct. 18, 1987, at 10C.
15. Dep't of Commerce, 1992 U.S. INDUSTRIAL OUTLOOK 30-7 (1992).
16. Calhoun, DIGITAL CELLULAR RADIO 42.
17. Id. at 43.
18. Id. at 45.
19. Notice of Proposed Rulemaking and Tentative Decision Para. 26, Amendment of the Commission's Rules to Establish New Personal Communications Services, GEN Dkt. No. 90-314, ET Dkt. No. 92-100 (FCC Aug. 14, 1992). Telocator projects that there will be more than 50 million PCS users by the end of the decade, and more than 60 million users of paging, cellular, and specialized mobile radio. Telocator Study Says PCS Licensing by 1994 Could Bring 23,300,000 Customers by 1997, Telecommunications Reports, June 1, 1992, at 19.
20. George Calhoun, WIRELESS ACCESS AND THE LOCAL TELEPHONE NETWORK 383-384, 547 (1992).
21. By comparison, the typical cable TV system costs about $300 per home to build; a state-of-the-art cable system, such as Time Warner's 150-channel digital fiber system in Queens, cost only about $600 per home.
22. As Calhoun notes, "radio is much less expensive than wire for a very large number of access applications." Calhoun, WIRELESS ACCESS AND THE LOCAL TELEPHONE NETWORK 383.
23. Calhoun, WIRELESS ACCESS AND THE LOCAL TELEPHONE NETWORK 40-41.
24. While interexchange minutes of use has grown by an average of 14.6 percent per year between 1986 and 1991, the capacity of existing and new fiber has grown an average of 285 percent. Industry Analysis Div., FCC, LONG DISTANCE MARKET SHARES: FIRST QUARTER 1991, Table 1 (June 1992); J. Kraushaar, FIBER DEPLOYMENT UPDATE, Table 2 (Mar. 1992); Emmanuel DeSurvire, Lightwave Communications: The Fifth Generation, SCIENTIFIC AMERICAN, Jan. 1992, at 114, 116.
25. Transmission Capacity of optical fibers has increased 10-fold every four years since 1975. The five generations of the technology are shown. The open circle represents the results from an experiment that simulated long-distance transmissions using a loop of fiber amplifiers and optical fibers. Source: SCIENTIFIC AMERICAN, Jan. 1 992, at 116.