The Radio Act of 1927, which bore the signature of President Calvin Coolidge and was the handiwork of his Secretary of Commerce, Herbert Hoover, forced wireless innovators to play a game of “Mother May I?” Each imaginative product launch or service twist, whether in FM radio or space satellite systems or cable TV or mobile phone networks, had to first win the hearts of regulators divining “public interest, convenience or necessity.” Stern policymakers have proven risk averse. Given their fondness for existing service suppliers, and their credulity when presented with evidence supporting the status quo (as reliably packaged for them by such incumbents), it isn’t surprising that the administrative creation of 1927 reliably resisted the forces of change for generations afterward.
Yet just as water pressure gradually erodes rock and leaves a raging river, policy restrictions were ground down. The failure of administrative planning for the radio spectrum, generally followed by the tragicomic ritual of “holding their feet to the fire,” has been shown to yield “public interest” payoffs that invariably benefit interests far more than the public.
As feared, incumbents have often been swamped. But predictions that the “public interest” would be completely neglected have proven wildly off target. Instead, the march of progress has stormed new playing fields in mass media and telecommunications. Vast new networks have been imagined, and built, under more liberal spectrum rules. Markets have coordinated what was asserted to be chaotic. Innovation and rivalry have been encouraged by means of de facto frequency ownership rights.
While liberated radio bands have produced these impressive results, Hoover’s administrative controls still stand. The great majority of useful bandwidth is allocated now as it was when the Radio Act was passed—even though we know the enormous opportunity costs of locking up the best wireless resources the planet has to offer.
Unfinished Business
Today, ambitious wireless entrepreneurs imagine the impossible. One outlandish idea is to bring cheap, ubiquitous high-speed internet access to the entire planet, from the tiniest island in Indonesia to the remotest village in Namibia to the highest outpost in the Himalayas. Indeed, there’s even competition to fulfill this fantasy. A venture headed by serial visionary Elon Musk, SpaceX, is attempting to deploy hundreds of low-earth-orbit satellites. Google X’s Project Loon launches stratospheric air balloons. Meanwhile, Facebook is planning a worldwide network supported by drones. These wild aspirations are economically straightforward: There may be money to be made in extending service to the “bottom billions.” Some ventures will surely fall short, while others—perhaps including those not yet envisioned—may soar.
But these wireless dreams will surely be toast without sufficient radio spectrum. In some regions, flexible-use spectrum rights allow these ambitious internet entrants to bargain for access to bandwidth controlled by existing licensees. Project Loon fortifies emerging LTE (Long Term Evolution) networks, sharing mobile operators’ flexible spectrum rights. Yet elsewhere, bandwidth requests must join the bureaucratic queue. That’s where the dreams go to die.
There will be no pure solution to the problem of spectrum allocation. Models are clear, but the world is a mess. Nor will there be complete “deregulation.” As economist Ronald Coase noted in 1959,
How far this delimitation of rights should come about as a result of a strict regulation and how far as a result of transactions on the market is a question that can be answered only on the basis of practical experience. But there is good reason to believe the present system, which relies exclusively on regulation and in which private property and the pricing system play no part, is not the best solution.
Scientific expeditions to the frontiers of wireless are giving us wonderful new communications options, but their implications for regulation are often misunderstood. Technology has not put an end to spectrum scarcity. Just as breakthroughs continue to provide fatter conduits, entrepreneurs ruthlessly compete to stuff them full of more communications. Supply creates its own demand, and then some. Access to prime radio spectrum has generally become more contentious, not less.
We have long heard the counterclaims. This “end of spectrum scarcity” refrain has been sung since at least March 1941. That was when Radio News broke the exciting story that as “FM becomes universal, there will be no physical limit on the number of stations in one town. The interference problem is solved.” Well, not quite then. And not quite yet.
Innovation should challenge incumbents; may the most efficient ecosystem win. But that victory is compromised when regulators seek to divert usage with new, unpriced allocations that effectively become locked in whatever configuration regulators create. The requests for additional unlicensed bands, of the sort set aside to support Wi-Fi or Bluetooth services, are submitted to the authorities and approved by them—not won or lost by bidders playing with their own stakes. This is Hoover’s top-down allocation system. The FCC favors one set of demands and excludes others, on the basis of what the Commission has itself deemed the “Wise Men Theory” of allocation.
This approach produces a consensus of dissatisfaction. To improve matters, the FCC’s top spectrum-policy experts, Evan Kwerel and John Williams, proposed in 2002 that contests over rules and bandwidth set-asides should be decided via competitive bidding. In 2008, the FCC tasked economists Mark Bykowski, Mark Olson, and William Sharkey with devising an auction format that would include offers from both licensed and unlicensed partisans. They wrote:
The allocation between licensed and unlicensed use…is based on the FCC’s judgment, which in turn relies on information provided by interested parties who seek to use the spectrum. One method of reducing the incentive that parties have to exaggerate the value they place on a given regime involves creating a market for such rules.
The authors designed a process to assign liberal licenses authorizing whatever services the grantees might like to devise. Mobile carriers, for example, could then expand the capacity of their networks by bidding for spectrum against companies, such as software firms or cable TV operators, which desire more unlicensed bandwidth. Under this scheme, a license-exempt band would be created out of the spectrum allocated to a license if the sum of the unlicensed spectrum bids (for that license) exceeded the top bid made by a rival seeking exclusive rights.
This approach was promising. Yet the FCC need not design special licenses or auctions to accommodate coalition bids. Flexible-use permits allow high bidders to determine business models on their own. And, under existing FCC auction formats, industry consortia have already bid for, and won, wireless licenses. Moreover, the assignment of frequency-use rights to third parties, including radio manufacturers—the parties who directly shape the use of unlicensed spectrum—is routinely a part of how liberally licensed frequencies are used. Nor is the practice unique; patents are widely shared in an analogous manner.
This type of commons, however, would benefit from an important reform. Under current rules, licenses typically come with build-out requirements. These rules mandate that license buyers create networks within a given time frame—say, five years. If a buyer does not make appropriate progress, the rights revert to the government. Many competing business models, including unlicensed “parks” that supply bandwidth for low-power wireless devices (Bring Your Own Network) might be excluded. Such requirements should be relaxed to permit them.
Some of the most contentious allocation contests are those among interests seeking different flavors of unlicensed bandwidth. One controversy features a satellite phone operator who aims to create “a private Wi-Fi channel and charge for access to it,” opposed by corporations that want to freeze the initiative to protect adjacent 5.8 GHz unlicensed frequencies. The latter group fears that enhanced traffic will diminish the performance of existing Wi-Fi service. Meanwhile, another battle involves General Electric and aircraft maker Boeing: GE proposes that its medical devices be permitted to transmit data (including patients’ vital signs) over 2.4 GHz unlicensed frequencies (a band also hosting popular Wi-Fi usage), while Boeing claims that this activity will interfere with its use of the bandwidth “to test the safety of planes.” Life or death in a hospital bed, versus life or death in the sky—which does the “public interest” favor? In another faceoff, oil companies want to use airwaves dedicated for educational institutions to reach deep-water oil rigs because “the only schools in the Gulf of Mexico are schools of fish.” The educators counter with a science lecture of their own: Sharing their offshore channels will create harmful interference because “wireless signals tend to travel farther and faster over the warm Gulf water, causing greater interference on shore.”
And then there is the brisk controversy over LTE-U (long-term evolution—unlicensed), a 4G mobile technology allowing carriers to improve internet access (and data downloads) by more tightly integrating the use of local Wi-Fi services. The Wi-Fi Alliance, which represents cable operators and tech companies, has made strong protests, claiming that the innovation tends to hog unlicensed frequencies for mobile subscribers. Some see the conflict as simply technical, but scientific research seems to follow economic self-interest. “Recent tests to see whether LTE-U technology interferes with Wi-Fi signals prove conclusively that LTE-U poses no problems whatsoever for Wi-Fi networks,” reports a trade journal, “and also that LTE-U will drown out Wi-Fi, depending on which party is to be believed.” It would be stunning were it otherwise.
The FCC does not know the optimal solution to these conflicts. Nor do I. But the agency will nonetheless, after lengthy deliberation, impose its guesses. And it will not permit market transactions to undo its decisions, as the fragmented, nonexclusive access rights it distributes in unlicensed bands cannot be easily reconfigured. An auction enlisting bids from the opposing parties, however, would help to drag the relevant costs and benefits out of the dark.
Rivalry in a Pastoral Setting
One objection raised against this competitive solution is that the value created by the “spectrum commons” cannot be captured in auctions. It is “akin to asking users of public parks to bid against developers to decide how land is to be allocated,” as economists Paul Milgrom, Jonathan Levin, and Assaf Eilet wrote in a 2011 paper. But the chosen analogy actually proves the opposite. First, when a government agency sets aside unlicensed bandwidth, it is bidding against other “developers” who would seek to use the airwaves differently. It simply does so in a nontransparent and monopolistic manner that suppresses competing valuations. Second, when open auctions are used for wireless licenses, the parties making offers—including mobile networks—shoulder the task of aggregating the disparate demands of millions of future subscribers (many of whom are not yet even born). The process encourages extensive research and careful calculations, given the risks involved for bidders, and puts prices on public display. This information improves decision-making for all parties, including governments, by exposing values and opportunity costs.
Finally, the idea that public parks are analogous to unlicensed band allocations perpetuates a common misunderstanding. As the FCC has written, “A mechanism based on markets…will be most efficient in most cases. However, government may also wish to promote the important efficiency and innovation benefits of a spectrum commons…much as it allocates land to public parks.”
The “however” crisply defines regulatory confusion. Public parks sit on land allocated through a system of private property rights. A market where basic resource rights are preempted in favor of “public interest” determinations is something quite distinct. That is what led Ronald Coase to characterize the spectrum allocation system as equivalent to a Federal Land Commission (FLC). The analogy was extended by New York University economist Lawrence J. White, in a thought experiment about how a Federal Land Use Commission (FLUC) might hinder the productive use of real estate. (Acronyms are also a leading output of the spectrum allocation system.)
Land, a key input into parks, is not held in abeyance, with rights dribbled out on a case-by-case basis as government administrators plan where all the public facilities (now and in the future) should be located. Instead, ownership rights are defined, and the assets made widely available. They are largely distributed, transferred, combined, and subdivided by marketplace transactions. Bids are registered not just by private developers but also by nonprofit organizations pursuing social objectives. These include governments providing green spaces, parkland, and other amenities. The latter institutions have multiple ways to bid—eminent domain, for instance, requires compensation while mediating holdout problems—and zoning regulations. But the market for real estate does not funnel each choice through a narrow administrative spigot; rather, it broadly cedes property rights to decentralized actors.
This yields a wide variety of land-use modes. Real-estate markets enabled the City of New York to acquire the necessary rights to create Central Park, and to understand the relevant trade-offs in expanding (or reducing) it. While White argued for property rights in frequencies, he declined to call this “privatization,” because he wanted to avoid the mistaken impression that “public ownership of spectrum is not part of the concept.”
Reforms
There is a droll, probably apocryphal story about a graduate student who fell asleep during Milton Friedman’s macroeconomics course at the University of Chicago. Friedman hurled an eraser at the snoozer and bonked his target. The young scholar awoke with a start. “I apologize, Professor, for falling asleep during your lecture,” he said. “But the answer is—reduce the money supply.”
The spectrum-policy answer here—if you’ve dozed off—is auction overlay rights. The method strategically introduces new spectrum-use rights, unleashing competitive forces by dispensing with centralized micromanagement. It empowers entrepreneurs who then achieve the cooperation needed for progress. It protects legacy systems but not needlessly; existing users face market incentives to accommodate the future.
The reform paves the way for advanced “spectrum sharing” by allowing gains from trade to flow to those parties whose active assistance—say, by upgrading technology, revamping networks, adopting bandwidth-saving applications, switching to less-contested bands, or exiting the market—contributes to consumer-pleasing outcomes. Government studies may assume that these forms of sharing are too complicated to arrange, but the white flag is hoisted due to a dubious choice of tactics. The nub of an economic agenda for reform would include:
- Auctioning the overlay rights to bands allocated to traditional licenses. These liberal authorizations would vest existing wireless users; TV licensees, for example, would be grandfathered in and maintain their rights to broadcast. But new, fully flexible rights would be created and sold at auction. These “overlays” would allow owners to deploy all the idle channels immediately. Moreover, overlay licensees would enjoy secondary rights to use the spaces occupied by incumbents. This enables spectrum-sharing based on the bargains struck. Entrants would compensate legacy interests for their cooperation in unleashing amazing new stuff.
- Auctioning the overlay rights to bands assigned to public agencies. Massive problems attend the reservation of approximately half of the most valuable radio waves for military and government use. Bureaucratic incentives are widely known to block efficiencies, with agencies effectively hoarding bandwidth while resisting technological change. Overlay licenses—“hunting permits,” awarded by auction, that grant private owners secondary rights to access public bands—can provide a slick solution here, too, although they require special measures. In particular, government agencies must be able to make deals and acquire money and other resources (including new and improved wireless systems) when overlay holders extend them advantageous offers. In these trades, entrepreneurs can drive spectrum economies, deploying advanced technologies to upgrade services while sharing radio waves more effectively. Policymakers have experience with such arrangements, having already steered successful transitions in public bands. Incumbent microwave users were cleared out of the 1.9 GHz band to make way for Personal Communications Services (PCS, or “2G”) in the 1990s; Advanced Wireless Services frequencies (AWS-1) supplanted dozens of federal agencies’ wireless networks in the late 2000s. Regulatory templates exist.
- Holding incentive auctions. The FCC’s National Broadband Plan called for considering overlay auctions if the approach adopted—a two-sided “incentive auction” with TV stations (offering to sell) and wireless carriers (stating prices at which they would buy) bidding in succession—should prove disappointing. Given the delays and limitations in the latter, it is time to dust off Plan B. But the demoted Plan A might still prove useful. We should take lessons from the FCC’s policy experimentation with how capacious spectrum reallocations can be achieved more expeditiously.
- Blanket liberalization of new and existing wireless licenses. It is not necessary to reinvent the wheel by devising more complex rules for transmitters or receivers (with extra dimensions defined by regulators). Liberalization can occur by tweaking the existing flexible-use templates that govern most mobile licenses. Where restrictions still block innovative services, technologies, or business models, deregulation can supply extra breathing space. One option is to invite licensees to request the relaxation of use rules, with presumptive approval for any noninterfering activity. Complaints lodged by protesting parties would be limited to border disputes, and adjudicated under efficient arbitration rules with strict time limits.
- Requiring government entities to purchase wireless services via competitive bids. Public agencies, including first responders, should buy wireless services from efficient suppliers, not build their own networks. The latter is a model constructed by political wheeling-and-dealing and bureaucratic turf protection, and it has achieved widespread failure in practice. Police and fire departments, in concert (coordinated by state authorities, perhaps) should issue Requests for Proposals. Competitive bids should then be received from commercial wireless service providers. These will inevitably share spectrum with mass-market civilian customers, outperforming boutique networks dedicated to special tasks but lacking the scale economies that make better technologies affordable and functional. “So as not to distort spectrum usage decisions,” writes former Pentagon official (and economist) Dorothy Robyn, “the government should subsidize the desired social good (i.e., public safety) directly and then let the relevant group acquire spectrum or spectrum-based services in the market.”
- Allowing markets to create unlicensed allocations. Interests favoring set-asides for non-exclusive spectrum access, such as those that support Wi-Fi or Bluetooth, should bid for liberal licenses. Where a compelling determination is made that bids for such business models will suffer from public good (“free rider”) problems and be inefficiently under-provided, explicit subsidies should provide the remedy. Prices are then transparent, not hidden, and competing interests have the opportunity to demonstrate rival valuations in arms-length transactions.
- Implementing complementary policies. Many reform proposals have outlined policies to support spectrum liberalization. I will mention just one: U.S. policy could do more to overcome the NIMBY (not-in-my-backyard) problem encountered in building cellular towers and base stations. Metropolitan jurisdictions routinely block new radio facilities, sometimes out of concern that the radiation from such transmitters is harmful to human health. To the extent that a radiation threat exists, it is associated with mobile phones held close to the user’s brain. People who are nervous about cell phone radiation are advised to use the speaker function, which “drastically reduces RF [radio frequency] exposure,” according to a 2010 Time article. The power emitted by phones increases when there are fewer base stations; radios amp up to reach the more remote tower. Hence blocking nearby towers actually exacerbates exposure to radio emissions.
With coming 5G technologies, mobile carriers will attempt to make their networks significantly denser, adding vast numbers of “small cells” to ramp up capacity and reduce latency (the lags in between interactive communications). But they must first surmount the roadblocks erected by local governments. Holdups here are endemic, and Federal efforts in the United States to impose a “shot clock” on municipal approvals (enacted in 2011) recall some official descriptions, circa the 1970s, of the results of the Vietnam War: an “incomplete success.”
Non-Judgmental Liberation
A commonplace in the spectrum policy debates is that the regulator’s task is to carefully select the best outcomes for society: if more broadcast TV or wireless broadband would be welcomed by the public, then regulators should allocate more spectrum to support them. That idea masks a subtle but powerful policy error. The unique correspondence between government rules and market services holds only if we’re doing public policy wrong. Flexible spectrum rights unleash competitive forces that discover tomorrow’s “killer apps” in unforeseen ways. Planning for more of this or less of that by administrative fiat—not so much.
When Ronald Coase broached the idea of auctions for spectrum rights, it was received as a “big joke” whose odds of adoption—even decades later—were equal to those of “the Easter Bunny in the Preakness,” as two members of the FCC scoffed in a 1977 opinion. These dismissals were accompanied by a phalanx of regulatory defenses. The FCC’s chief economist in the 1950s, Dallas Smythe, declared it impossible to define wireless rights and sell them. That wisdom went unchallenged for decades.
Coase might have been naive about the politics, but his economics were spot on. Not only could rights be defined and sold, but competitive bids could be used to speed up innovation and dramatically improve services for consumers. Auctions of wireless assignments, authorizing firms to offer whatever services or technologies fit within their designated frequency spaces, eliminated obstacles and flooded markets with innovations previously blocked. It would have been impossible for smartphone ecosystems, with their hundreds of devices and millions of applications, to emerge under traditional regulation. Liberalization even enabled elegant reforms in other markets. From electricity to water to pollution allowances to fishing rights, newly constructed markets have fashioned superior alternatives to command-and-control regulation. Today, economists and systems engineers are at work designing ever more ambitious bidding mechanisms, revealing hidden values, improving resource use, and saving the planet. Many of these inventions—according to Caltech’s Charles Plott, a central figure in this revolution—emerge from “spectrum auctions.”
The foundational idea was dismissed in 1960s Washington not only because incumbent licensees and bureaucrats had their own agendas, but because, as policy disruptor Clay Whitehead noted, the President of the United States had his. This was perhaps the only time the matter of spectrum allocation got such high-level attention. As Bobby Baker, Lyndon Johnson’s top Senate staffer (later imprisoned on corruption charges) wrote of his boss:
It was no accident that Austin, Texas, was for years the only city of its size with only one television station. Johnson had friends in high places…. LBJ demanded, and received, the opportunity to pick and choose programs for his monopoly station from among those offered by all three of the major networks.
Ironically, the regime that accommodated this sordid record is trumpeted, even decades later, as a noble enterprise protecting the public. In 2013, FCC Chairman Tom Wheeler bemoaned broadcast video content and suggested that “maybe the industry was in need of another Newton Minow ‘Vast Wasteland’ moment…to call them to the angels of their better programming natures when it came to violence or indecency.”
With deregulation, unlicensed cable networks created exponentially greater variety in video choice. Network information sources beaming bland “News from Nowhere” in three dull shades were replaced by a raucous rivalry of 24/7 services ranging from CNN to Fox News to Comedy Central to Vice. Diversity further increased as the unregulated internet came to sit atop emerging broadband networks. Better angels did make it to television—Touched by an Angel was a hit series—but it was a very earthly set of reforms that set aside “public interest” rule-making, removing power from the hands of the corruptible few.
Spectrum Detox
Spectrum allocation enjoys a long history of paradox. The best tool for understanding it is not the physics of radio waves but the economics of public choice, which explains how special interests craft political coalitions and ally with regulators to distribute favors that bless the anointed while shorting entrepreneurial risk-takers.
The “deregulation wave” of the 1970s changed history, and many of its positive externalities were political blessings. The Open Skies reform broke the government-backed monopoly in satellite communications, Comsat, in about 1975. Anti-cable rules were relaxed in the late 1970s. By the late 1980s, liberalized cellular licenses allowed competitive firms to decide what services, phones, technologies, business models, or content to offer. In the 1990s, mobile networks eclipsed broadcasting as the preeminent wireless sector. The cozy spectrum allocation club became overrun. In 1997, a confident Peter Huber wrote:
It appears that old-style broadcasters will carry the regulatory baggage of the 1934 Act for another decade or so. Early in the next century, however, this dismal regulatory era will finally come to an end. Broadcast spectrum will be dezoned. Roseanne will have to compete for airtime with the more civil, uplifting, and profitable expressions of ordinary people talking on wireless phones.
By the 2000s, global wireless networks had brought modern information services to billions, helping lift millions up from poverty. Carriers competed to offer text, data, and video services atop voice networks; to generate daring new platforms populated by smartphones, tablets, netbooks, and dongles; and to provide connectivity for millions of applications. These emerging machine-to-machine services are disrupting markets, altering social intercourse, toppling governments, and even extending human life. Crime rates fell with the introduction of cellphones, and mHealth apps push patients to take their meds while monitoring their vitals.
Spectrum policies gave markets room to roam, limiting the “controllers.” Broadcast TV, tightly licensed and subject to the Fairness Doctrine and Equal Time and the Public Interest Standard, led no revolutions. Under traditional authorizations, services were preordained and innovations lost. With liberal rules, up popped green shoots.
The Internet of Things is just revealing its shape and scope. Tech writer Vivek Wadhwa predicts that, using “sensors and the apps that tech companies will build, our smartphone will become a medical device akin to the Star Trek tricorder.” These coming advances in science, culture, and economics stretch far beyond radio spectrum policy. But they are related.
As the dreams of visionaries grow, the drag imposed by anti-competitive spectrum regulation becomes all the more damaging. Wireless is a key component of the drive for a better world, so it becomes increasingly curious that society would slow its progress. Wireless technologies have opened up new vistas; we can see the future from here. The political spectrum ought to stop blocking the splendid view.