By Brent Skorup and Michael Kotrous

In 1999, the FCC completed one of its last spectrum “beauty contests.” A sizable segment of spectrum was set aside for free for the US Department of Transportation (DOT) and DOT-selected device companies to develop DSRC, a communications standard for wireless automotive communications, like vehicle-to-vehicle (V2V) and vehicle-to-infrastructure (V2I). The government’s grand plans for DSRC never materialized and in the intervening 20 years, new tech—like lidar, radar, and cellular systems—advanced and now does most of what regulators planned for DSRC.

Too often, however, government technology plans linger, kept alive by interest groups that rely on the new regulatory privilege, even when the market moves on. At the eleventh hour of the Obama administration, NHTSA proposed mandating DSRC devices in all new vehicles, an unprecedented move that Brent and other free-market groups opposed in public interest comment filings.  As Brent wrote last year ,

In the fast-moving connected car marketplace, there is no reason to force products with reliability problems [like DSRC] on consumers. Any government-designed technology that is “so good it must be mandated” warrants extreme skepticism….

Further,

Rather than compel automakers to add costly DSRC systems to cars, NHTSA should consider a certification or emblem system for vehicle-to-vehicle safety technologies, similar to its five-star crash safety ratings. Light-touch regulatory treatment would empower consumer choice and allow time for connected car innovations to develop.

Fortunately, the Trump administration put the brakes on the mandate , which would have added cost and complexity to cars for uncertain and unlikely benefits.

However, some regulators and companies are trying to revive the DSRC device industry while NHTSA’s proposed DSRC mandate is on life support. Marc Scribner at CEI uncovered a sneaky attempt to create DSRC technology sales via an EPA proceeding. The stalking horse DSRC boosters have chosen is the Corporate Average Fuel Economy (CAFE) regulations—specifically the EPA’s off-cycle program. EPA and NHTSA jointly manage these regulations. That program rewards manufacturers who adopt new technologies that reduce a vehicle’s emissions in ways not captured by conventional measures like highway fuel economy.

Under the proposed rules , auto makers that install V2V or V2I capabilities can receive credit for having reduced emissions. The EPA proposal doesn’t say “DSRC” but it singles out only one technology standard that would be favored in this scheme: a standard underlying DSRC . 

This proposal comes as a bit of surprise for those who have followed auto technology; we’re aware of no studies showing DSRC improves emissions. (DSRC’s primary use-case today is collision warnings to the driver.) But the EPA proposes a helpful end-around that problem: simply waiving the requirement that manufacturers provide data showing a reduction in harmful emissions. Instead of requiring emissions data, the EPA proposes a much lower bar, that auto makers show that these devices merely “have some connection to overall environmental benefits.” Unless the agency applies credits in a tech-neutral way and requires more rigor in the final rules, which is highly unlikely, this looks like a backdoor subsidy to DSRC via gaming of emission reduction regulations.

Hopefully EPA regulators will discover the ruse and drop the proposal. It was a pleasant surprise last week when a DOT spokesman committed that the agency favored a tech-neutral approach for this “talking car” band. But after 20 years,  this 75 MHz of spectrum gifted to DSRC device makers should be repurposed by the FCC for flexible-use. Fortunately, the FCC has started thinking about alternative uses for the DSRC spectrum. In 2015 Commissioners O’Rielly and Rosenworcel said the agency should consider flexible-use alternatives to this DSRC-only band.

The FCC would be wise to follow through and push even farther. Until the gifted spectrum that powers DSRC is reallocated to flexible use, interest groups will continue to pull any regulatory lever it has to subsidize or mandate adoption of talking-car technology. If DSRC is the best V2V technology available, device makers should win market share by convincing auto companies, not by convincing regulators.

Those of us with deep reservations about the push for ever more unlicensed spectrum are having many of our fears realized with the new resistance to novel technologies using unlicensed spectrum. By law unlicensed spectrum users have no rights to their spectrum; unlicensed spectrum is a managed commons. In practice, however, existing users frequently act as if they own their spectrum and they can exclude others. By entertaining these complaints, the FCC simply encourages NIMBYism in unlicensed spectrum.

The general idea behind unlicensed spectrum is that by providing a free spectrum commons to any device maker who complies with certain simple rules (namely, Part 15’s low power operation requirement), device makers will develop wireless services that would never have developed if the device makers had to shell out millions for licensed spectrum. For decades, unlicensed spectrum has stimulated development and sale of millions of consumer devices, including cordless phones, Bluetooth devices, wifi access points, RC cars, and microwave ovens.

Now, however, many device makers are getting nervous about new entrants. For instance, Globalstar is developing a technology, TLPS, based on wifi standards that will use some unlicensed spectrum at 2.4 GHz and mobile carriers would like to market an unlicensed spectrum technology, LTE-U, based on 4G LTE standards that will use spectrum at 5 GHz.

This resistance from various groups and spectrum incumbents, who fear interference in “their” spectrum if these new technologies catch on, was foreseeable, which makes these intractable conflicts even more regrettable. As Prof. Tom Hazlett wrote in a 2001 essay, long before today’s conflicts, when it comes to unlicensed devices, “economic success spells its own demise.” Hazlett noted, “Where an unlicensed firm successfully innovates, open access guarantees imitation. This not only results in competition…but may degrade wireless emissions — perhaps severely.”

On the other hand, the many technical filings about potential interference to existing unlicensed devices are red herrings. Prospective device makers in these unlicensed bands have no duty to protect existing users. Part 15 rules say that unlicensed users like wifi and Bluetooth “shall not be deemed to have any vested or recognizable right to continued use of any given frequency by virtue of prior registration or certification of equipment” and that “interference must be accepted.” These rules, however, put the FCC in a self-created double bind: the agency provides no interference protection to existing users but its open access policy makes interference conflicts likely.

There is a concerted effort, then, by some wireless industry associations, tech journalists, and tech-focused nonprofits to ignore the Part 15 rules and suggest that open access no longer applies. In particular, there are suggestions that LTE-U must or should comply with wifi-like listen-before-talk mechanisms before using the unlicensed commons. Chris Lewis at Public Knowledge insinuated as much in a blog post on the issue. He states the correct but legally irrelevant fact that early versions of LTE-U don’t use listen-before-talk protocols and then adds a confusing non sequitur, “This is in violation of basic Wi-Fi standards.”

The notion that LTE-U or any other new technology must employ the wifi industry’s preference, listen-before-talk, is wrong. There are tens of millions of Part 15 devices that don’t use listen-before-talk, including cordless phones, garage door openers, Bluetooth devices, and RC toys. There are different sharing etiquettes and the FCC has generally been hands-off regarding what etiquette device makers should use since, first, the strict Part 15 power limits mitigate most problems and second, interference is typically reciprocal and parties have an incentive to coordinate.

Interestingly, the FCC has required some unlicensed devices to employ listen-before-talk protocols in the unlicensed PCS band. Never heard of it? The band is a wireless graveyard. Aside from a few cordless telephones, it’s had very little use, in part because the FCC required a complex listen-before-talk etiquette that raises the cost of producing equipment. In light of this failed experiment, the FCC probably has little appetite (or aptitude) for predicting via technology mandates which sharing etiquette will most benefit consumers.

Further, unlicensed spectrum incumbents show a selective sensitivity to interference considering their unlicensed devices face interference daily. It’s impossible to approximate the severity and regularity of everyday interference but focusing on potential interference from new services like LTE-U or TLPS, which use spectrum sharing etiquettes, and ignoring the effects of, say, poorly configured or legacy wifi access points or microwave ovens in the 2.4 GHz band is akin to complaining about hearing your next-door neighbor’s TV volume when there’s a rock concert playing in your front yard. Microwave ovens are powerful emitters, typically around 400 to 800 watts compared to a 1-watt wifi device. While microwave ovens are built to shield most emissions from escaping, none are perfect and they are a frequent source of wireless interference in households and offices around the country. Relatedly, in apartments, condos, or dormitories with unmanaged wifi systems, interference occurs regularly.

The FCC sends very mixed signals regarding unlicensed policy. It formally provides no interference protection to unlicensed users but frequently solicits comment about possible harms to these existing users. No wonder, then, that some Wall Street investors have strenuously opposed Globalstar’s multi-year attempt to get approval for its TLPS technology to provide wifi-like Internet access. Why would a hedge fund take an interest in the intricacies of Part 15 rules? Recent tech reporting is suggestive.

Bloomberg BNA reported that one intervenor who has filed comments against Globalstar’s TLPS application “runs a hedge fund [and has] said he is short-selling Globalstar’s stock, so he has been very active in the Globalstar TLPS FCC proceeding.” The New York Times similarly reports on another frequent filer in the TLPS proceeding, “a little-known activist investor [who] has declared war on the multibillion-dollar satellite communications company Globalstar, contending that it is worthless.” Existing device makers likewise may see a competitive threat from new devices that provide similar services, as Hazlett notes, and pile on in these proceedings.

Singling out a company with important business before a regulatory agency is not unheard of but the FCC only encourages financial gamesmanship by requesting that parties weigh in on interference potential for users that formally aren’t entitled to interference protection. Is this how the spectrum commons dies?

The most effective tactic to use when the FCC is likely to do something you dislike is to induce regulatory delays. The public interest groups can see much of this and their responses have been relatively muted relative to the commercial interests. I suspect many are deeply uncomfortable with what is occurring because it undermines the idea of a commons and the intent of the Part 15 rules. Nevertheless, they favor the status quo because wifi works pretty well and consumers have reliance interests. Knowing that the Part 15 rules don’t help them, they typically resort to asking for more studies about interference potential. It sounds like an innocuous request but anyone following telecom policy knows that “more study” from the FCC is the kiss of death because it simply gives time for opponents to agitate for reinforcements (like powerful members of Congress) and to scare off investment.

Congress, by the way, foresaw this risk–pressure groups compelling the FCC to kill entrants with delay–and in 1983 added the little-known Section 7 of the Communications Act, which requires the FCC to approve new technologies within a year. By requesting parties weigh in on interference potential and delaying indefinitely Part 15 approvals for TLPS and LTE-U (assuming they show they comply with Part 15) the FCC violates the spirit of the law. The agency has a statutory duty to companies with new technologies to make a decision quickly, but these lengthy unlicensed proceedings send a chilling message to the tech industry (so much so the IEEE asked then-Chairman Genachowski for Section 7 guidance in 2011).

The FCC knows spectrum NIMBYism is a big, developing problem. The unlicensed incumbents are agitating more and more as new technologies encroach on “their” spectrum. It should be enough for the FCC to respond that these unlicensed device makers knew the tradeoff going in–you can avoid expensive licensure and use spectrum freely but you cannot object when interfered with. Firms that want interference protection and higher QoS are free to spend millions or billions of dollars on licensed spectrum. Increasingly, however, by largely remaining silent and delaying approvals, the FCC gets bogged down in proceedings and undermines the purpose of unlicensed spectrum–encourage innovators to experiment with new wireless technologies. If the delays in approving TLPS and foreseeable delays for LTE-U are any indication, the FCC is quietly slipping towards de facto beauty contests, the infamous practice of picking technology winners and losers.

The most pressing challenge in wireless telecommunications policy is transferring spectrum from inefficient legacy operators like federal agencies to the commercial sector for consumer use.

Reflecting high consumer demand for more wireless services, in early 2015 the FCC completed an auction for a small slice of prime spectrum–currently occupied by federal agencies and other non-federal incumbents–that grossed over $40 billion for the US Treasury. Increasing demand for mobile services such as Web browsing, streaming video, the Internet of Things, and gaming requires even more spectrum. Inaction means higher smartphone bills, more dropped calls, and stuttering downloads.

My latest research for the Mercatus Center, “Sweeten the Deal: Transfer of Federal Spectrum through Overlay Licenses,” was published recently and recommends the use of overlay licenses to transfer federal spectrum into commercial use. Purchasing an overlay license is like acquiring real property that contains a few tenants with unexpired leases. While those tenants have a superior possessory right to use the property, a high enough cash payment or trade will persuade them to vacate the property. The same dynamic applies for spectrum.

Overlay licenses have been used to reassign non-federal spectrum but never federal spectrum. The paper presents new evidence from a 2006 spectrum auction (AWS-1) that suggests that billions of dollars of underused federal spectrum could be deployed more quickly than other policy alternatives. Crucially, overlay licenses allow agencies to receive payment for spectrum sales and this reordering of spectrum rights would benefit taxpayers and wireless broadband users.

Policymakers are interested in spectrum policy because spectrum availability improves broadband access and generates substantial government revenues. Further, conservative estimates place the consumer surplus losses from misallocation of spectrum at hundreds of billions of dollars per year. Therefore, policymakers should favor reform proposals, like overlay licenses, that show promise in repurposing federal spectrum relatively quickly. The paper compares two policy proposals for spectrum reform: regulation-intensive dynamic spectrum sharing and market-oriented overlay licenses.

Regulation-Intensive Approach. A 2012 President’s Council of Advisors on Science and Technology (PCAST) report promotes complex spectrum-sharing technologies to enable consumer use of fallow federal spectrum in order to avoid clearing agencies from their spectrum.

• According to the PCAST report, widespread dynamic spectrum sharing would take decades to implement. The proposal relies on precise government planning and complex device requirements to enable intensive use of federal spectrum. However, the sharing technologies contemplated are in early development and will not be in routine deployment for many years. Social welfare losses mount quickly in the interim.
• Despite recognizing that agencies have no incentive to improve efficient use of their spectrum, this proposal does little to encourage efficient government use of spectrum. Dynamic spectrum sharing techniques allow wasteful legacy systems to operate indefinitely, and PCAST recommends against clearing inefficient federal users.
• Implementation of the PCAST proposal would likely degenerate into regulatory failure. Previous attempts at spectrum sharing between different wireless systems, like the TV White Spaces allocation that PCAST lauds, frequently resulted in rent seeking, severe deployment delays, and relatively few consumer benefits.

Market-Based Approach. A superior reform proposal is to auction off overlay licenses to certain federal spectrum bands. These winning overlay licensees can put unused federal spectrum into service rapidly. For the remaining spectrum that agencies are using, the winning licensee can pay the agency to vacate the bands or upgrade to more efficient systems. Agency resistance may be mitigated because agencies can negotiate compensation for selling rights to their spectrum.

• The FCC has conducted overlay auctions in the past and they represent an “off-the-shelf” tool to reorder spectrum rights. In previous overlay auctions, the process was effective and winning bidders compensated existing users like state public safety agencies and public utilities to vacate their valuable spectrum.
• Overlay license auctions and clearing deadlines transfer spectrum into the market and to its highest-valued uses. For example, in as few as two years after the 2006 AWS-1 auction, existing users and federal agencies vacated their spectrum, allowing carriers to invest billions of dollars into networks and deploy mobile broadband in cities like San Francisco and New York.
• A combination of clearing federal agencies from their spectrum and using overlays to clear nonfederal users has freed about 210 MHz of prime spectrum for mobile broadband use, supplying over one third of spectrum held by mobile carriers today.

Government agencies sit on wireless spectrum worth hundreds of billions of dollars rent-free. This federal spectrum is often unused or underutilized and the misallocation of this valuable resource is socially costly. My paper proposes that Congress permit agencies to sell some of their spectrum to private parties after an overlay auction. No other reform proposal has enabled widespread consumer use and economic investment as rapidly as have overlay auctions combined with clearing deadlines. Overlays and clearing deadlines in the recent past have permitted commercial deployment of cutting-edge wireless technologies in encumbered spectrum within a few years.

Related Research Reclaiming Federal Spectrum: Proposals and Recommendations

The FCC is being dragged–reluctantly, it appears–into disputes that resemble the infamous beauty contests of bygone years, where the agency takes on the impossible task of deciding which wireless services deliver more benefits to the public. Two novel technologies used for wireless broadband–TLPS and LTE-U–reveal the growing tensions in unlicensed spectrum. The two technologies are different and pose slightly different regulatory issues but each is an attempt to bring wireless Internet to consumers. Their advocates believe these technologies will provide better service than existing wifi technology and will also improve wifi performance. Their major similarity is that others, namely wifi advocates, object that the unlicensed bands are already too crowded and these new technologies will cause interference to existing users.

The LTE-U issue is new and developing. The TLPS proceeding, on the other hand, has been pending for a few years and there are warning signs the FCC may enter into beauty contests–choosing which technologies are entitled to free spectrum–once again.

What are FCC beauty contests and why does the FCC want to avoid them? From the 1930s to the 1990s (aside from a few short-lived spectrum lotteries), the FCC handed out valuable spectrum licenses for free to applicants who showed they would benefit the public with their planned services. TV broadcasters, taxicab dispatchers, satellite communications companies, medical facilities, and others lobbied to claim their stake when new spectrum became available.

These time-consuming proceedings became known as beauty contests, reflecting the subjective nature of giving away an input often worth tens or hundreds of millions of dollars to “deserving” applicants. The inefficiency, delay, and predictable corruption of beauty contests were widely criticized, but it wasn’t until the 1990s that Congress permitted auctioning spectrum. Allowing markets to allocate spectrum greatly improved the chances spectrum would go to the firms that had financial incentives to put it to good use, rather than the firms that had the most persuasive insiders.

But not all spectrum is auctioned today. Decades ago the FCC realized that short-range, innovative new services could be deployed without expensive and time-consuming licensing. The agency decided to authorize low-power devices in certain bands of spectrum. Essentially, any device maker could freely deploy technologies in these bands as long as they complied with a few basic FCC rules, the Part 15 rules. The FCC left technology choices to the device makers, who share the spectrum with other–sometimes interference-prone–device makers and users. While wifi technology is the most popular and most economically significant user of unlicensed spectrum, there are many other technologies coexisting in unlicensed bands. Today, hardware companies make dozens of short-range technologies like toy RC cars, wireless speakers, Bluetooth earpieces, baby monitors, garage door openers, cordless phones, and wifi routers.

Unlicensed spectrum has downsides for device makers, however. As the FCC said in a recent proceeding, “As a general condition of operation, Part 15 devices … must accept any interference that may be received from [licensed users] or other Part 15 devices.” Operators like AT&T, Sprint, and Dish pay millions or billions of dollars for their licensed spectrum at auction. In return, however, they can exclude other wireless operators from using their spectrum assignments. In contrast, using free unlicensed spectrum means you have no protection from interference from other unlicensed and licensed users. This is intended to create an environment of permissionless innovation, where wireless entrants can be free to try new services.

In theory, this means unlicensed users cannot object when other unlicensed users deploy new technologies. In practice, however, now that unlicensed spectrum is occupied by services like Bluetooth and wifi-delivered Internet, new entrants often modify their technology to be “good neighbors.” The potential for interference also motivates established players to prevent entrants like TLPS and LTE-U from using the bands.

Richard Bennett has a good explanation of the LTE-U engineering issues before the FCC. TLPS has slightly different issues. After a few years of testing, TLPS may be approved soon, but not without a fight. TLPS is a novel wireless technology that uses a channel of spectrum that straddles unlicensed spectrum and licensed spectrum. The licensed portion is currently used by Globalstar for satellite communications but the FCC generally wishes to get away from mandating certain services–like satellite communications–and to allow licensees to use their spectrum for whatever service is demanded by consumers. For that reason, the FCC has sought, since releasing the 2010 National Broadband Plan, to make this relatively unproductive “satellite spectrum” available for land-based wireless broadband use. Knowing that the FCC is willing to be flexible to meet growing consumer broadband needs, Globalstar saw an opportunity to merge its licensed spectrum with a portion of the free, adjacent unlicensed spectrum. With this wider channel, some of it shared with existing unlicensed users, wireless broadband delivered via TLPS technology became feasible. As TLPS approval nears the finish line, however, some unlicensed users are objecting that TLPS will interfere with their services.

The FCC proceedings reveal a technical debate about interference measurements. These claims distract from the larger issue: Either the Part 15 rules mean what they say–unlicensed users have no interference protection–or the FCC is increasingly back in the business of beauty contests and deciding which services are entitled to free spectrum.

Henry Goldberg, a communications lawyer who represented Apple years ago in getting more unlicensed spectrum allocated, predicted these fights at a 2008 Information Economy Project conference.

[I]f you are a company or a municipality or a port authority or a university who has invested in unlicensed spectrum to provide a WiFi services for a fee, you’re not so sure you want someone using unlicensed spectrum to compete with you. Such players may try to use contractual rights, lawsuits, etc. to seek to limit additional entry to what has become “their” spectrum. If a “not-in-my-back-yard” dynamic takes over, the very essence of Part 15 is compromised. Vigilance is needed to fight Part 15 NIMBY.

It’s this growing Part 15 NIMBYism that concerns many spectrum policy watchers. No one wants the return of beauty contests and the FCC picking winners among different technologies.

But Goldberg has a discouraging addendum to his prescient warning against NIMBYism in unlicensed bands:

Supporter of unfettered grazing rights that I am, it doesn’t offend me to have the town permit grazing by sheep and cows, but forbid elephants.

Herein lies the problem. The FCC is being pressured to declare that TLPS is an elephant that should not be allowed in the commons filled with wifi sheep and Bluetooth cows. LTE-U will be the next target.

If the FCC encourages these kinds of complaints, the result will be customary law that is destructive to innovation in unlicensed bands. Firms will sink investments in technologies and business plans that comply with the rules, and only later learn they are violating unwritten rules.

The bigger problem is that the FCC is entering beauty contest territory once again. Even if the FCC someday prohibits “elephants” in unlicensed–current Part 15 rules say unlicensed users have no protection against others–the agency has to determine what that means. The FCC does not want to go back to the bad old days of beauty contests, specifying, in the face of intense lobbying, that only certain technologies were allowed on certain frequencies in certain places.

As firms find ways to intensely use free unlicensed spectrum, more conflicts like these may arise. Unfortunately these fights politicize FCC decisionmaking and could stymie new wireless innovations.

It may be that NIMBYism in unlicensed is inevitable. If interference in unlicensed is a regular problem and the FCC finds itself picking winners, the FCC needs to be much more cautious about allocating unlicensed spectrum. It’s worth noting that auctioning spectrum removes the temptation to engage in the ad hoc dispensations of spectrum that plagued the agency for decades. In any case, the results of the TLPS and LTE-U proceedings will have ramifications beyond the approval or denial of those technologies.

Related Reading: Super Wifi and Unlicensed Spectrum: “Spectrum Condos” How the FCC Killed a Nationwide Wireless Broadband Network

There is renewed interest in unlicensed spectrum as the FCC approaches the TV white space issue (again). Tim B. Lee reports on some of the unlicensed supporters,

Activists at the South by Southwest Interactive festival in Austin, TX, built a free wireless network to help publicize the power of unlicensed “white spaces” technology. The project is part of a broader campaign to persuade the FCC not to auction off this spectrum for the exclusive use of wireless carriers.

Unlicensed spectrum for high-powered devices has been called Super Wifi (“wifi” in this context is used loosely; Super Wifi is a PR term and has nothing to do with the wifi technical standard). Frankly, there are many reasons to be cautious about assigning more unlicensed spectrum, especially given the confusing information out there about the technology. (For instance, despite a popular rumor, Super Wifi would not provide free Internet access to everyone with a device, as Matt Yglesias and Jon Brodkin point out.)

The unlicensed/licensed debate is several years old and often technical. I won’t rehash the old issues here, but there is a point I’d like to highlight about the nature of unlicensed spectrum: In spectrum assignments, you generally want to create “apartments, not condos.” Like most, I favor unlicensed spectrum under certain circumstances. However, we should be aware of the rigidity unlicensed spectrum imposes on future reassignments.

If you’re a property developer in a city and you want to raze and build on property occupied by a residential high-rise, you want that high-rise to be an apartment complex, not a condominium building. With apartments, you can bargain with the property management company and, with time, all tenants can be cleared out. Not so with condos, many urban developers are finding. Even if most condo owners in a building are contacted and compensated for leaving, the remaining owners have an effective veto over the new development.

Similarly, unlicensed device users can veto the future reassignment or transfer of the spectrum they occupy. Smartphone and satellite radio users, for example, have no veto ability–they are “apartments,” essentially leasing space from a spectrum “owner.” Like real property, you really need small-numbers bargaining to transfer and lease spectrum for its highest-valued use. Many unlicensed “owners” in a band creates a tragedy of the anticommons. Control over devices drives most unlicensed spectrum advocates mad, but it is also what permits technology upgrades and relatively fast spectrum transfers. (Mobile phones with 1G (analog) are long gone. Not so with old baby monitors, cordless phones, and garage door openers, which are all unlicensed. There’s no spectrum manager to clear these old devices out.) Once unlicensed devices populate a band, the spectrum almost certainly cannot be transferred and used for other technologies.

The time will come when–not if–a brand new social need arises that requires substantial amounts of spectrum as an input. If the FCC wanted to reassign spectrum in the future for, say, driverless car technology, Super Wifi bands are out of the question. It’s simply impractical to locate all the (mobile and transient) high-powered Super Wifi devices that will be using the band, install a new radio, and move them to another band. Even if you could identify most of them, people who buy or sell devices–many of whom will be powerful institutions like public safety, transportation, and tech companies–will have built business models based on the unlicensed spectrum. Entrenched users will not relinquish their spectrum easily after making substantial investments in the technology.

Ideally, you want a spectrum manager that can be compensated to discontinue services or move their users to another band when better uses come along. This is not to say we should not have “Super Wifi” or other unlicensed bands. But we should hesitate before creating these spectrum condos, particularly in the valuable bandwidth under 1 GHz. By permitting unlicensed operators, future spectrum reassignment of unlicensed bands moves from the marketplace to lengthy administrative resolution* by the FCC and NTIA because of the fragmented and numerous users–which is what the Congress and the FCC have tried to avoid for the past 20 years with auctions and secondary markets. Instead of negotiation and compensation, the reassignment becomes a shouting match between interested parties and their lobbyists. In the end, consumers typically lose.

* Recent history is illuminating. Just look at LightSquared’s dealings with Inmarsat (apartments) versus GPS users (condos). Conflicts with GPS users killed LightSquared’s new nationwide LTE network because there were too many GPS parties to bargain with. For another example, observe how NextNav is running into interference problems with WISPs (condos).

Over at his always-informative Spectrum Blog, wireless guru Michael Marcus brings to my attention a new report that will definitely be of interest to everyone here about “The Economic Value Generated by Current and Future Allocations of Unlicensed Spectrum.”  It was written by Rich Thanki of Perspective Associates, a UK consulting firm. I haven’t had time to finish the whole thing yet, but it basically lays out the argument for opening up more spectrum, especially “white spaces,” to unlicensed use.

Anyway, Mike Marcus has an much better write-up of the report than I could ever do, so head over there to check out his discussion.  One important thing that Mike stresses is the importance of technical flexibility:

But the key issue here is not the presence or absence of a license, the key issue is deregulation. A major reason why unlicensed networks have been so innovative is that the descendants of the FCC Docket 81-413 rulemaking, e.g. Wi-Fi, Bluetooth, and Zigbee have been in spectrum bands with great technical flexibility… If you overregulate unlicensed systems, they can stagnate just as much as licensed one often do.

I think that is an important insight and essential lesson that we should always keep in mind when it comes to spectrum policy, regardless of whether we talking about licensed or unlicensed spectrum.  Although I’ve always been a bit torn about how much spectrum should be allocated on an unlicensed (or “commons”) basis versus auctioned (property rights model), as Marcus suggests, flexibility is crucial in either case.   In all the heated catfights over licensed and unlicensed spectrum, that point sometimes gets overlooked.

The white space debate has been the subject of much attention lately, with Microsoft, Dell, and Google pitted against the CTIA on the question of how to allocate white spaces between UHF channels. The two competing proposals are 1) auction off white spaces, similar to the 700mhz auction, or 2) leave them unlicensed and managed (like 2.4Ghz) but allow devices which don’t cause interference.

This controversy again raises the issue of the desirability of unlicensed spectrum. I’ve been reading about the merits of unlicensed spectrum, inspired by a 2006 exchange between Jerry Brito and Mike Masnick on TLF and TechDirt. Jerry makes a compelling argument that command-and-control commons rules might hinder the emergence of superior networks operating with devices emitting greater than 4w EIRP.

The public interest is to allocate the spectrum in the most economically efficient manner, so if unlicensed spectrum uses do not make the best use of scarce airwaves, unlicensed bands should be auctioned off. Tim envisions privately managed commons that would provide for much the same openness now offered by unlicensed spectrum, but without a monolithic regulator imposing centralized rules.

Privately managed commons are naturally appealing to libertarians, as they accomplish the virtue of openness and participation without the need for government intervention. But there’s the question of whether a spectrum commons would actually emerge. With the immense profit potential that comes with owning even a tiny chunk of spectrum and offering restricted communications services, it’s not easy to imagine firms ponying up billions simply to manage spectrum for open, unrestricted use—unless each user must pay to become a commons member. Supporters of licensed spectrum suggest that wireless device manufacturers could confer spectrum usage rights to the end user for a small fee, rolled up in the cost of the product. But with myriad small wireless devices now on the market made by upstart overseas companies, abolishing unlicensed spectrum would necessitate that each device have some privately granted right to broadcast, possibly adding new entry barriers. And what about the geek who wants to build a low-power, home wireless device with off-the-shelf parts from Radio Shack? Does she have to buy spectrum rights from a third-party in order to broadcast in her own home without facing “spectrum infringement” litigation? If so, perhaps Faraday Cages will one day become a nerd luxury.

Currently, mobile phone services run fairly well using licensed frequencies. People spend hundreds (if not thousands) of dollars a year for wireless service plans. Perhaps the ability of unlicensed spectrum to greatly reduce transaction costs explains its appeal to consumers and producers alike. According to the Coase theorem, unhindered market allocation of property rights should generate a Pareto optimal outcome. But that outcome might be skewed if transaction costs eat up a huge portion of consumer and producer surplus. The ability to broadcast a Bluetooth signal ten or fifteen feet, for example, is far less valuable than mobile phone service.

The amazing success of the unlicensed 2.4Ghz band, even compared to widely deployed 3G networks like EVDO, suggests there’s something to the argument for unlicensed, lightly regulated spectrum. Who would have envisioned a mere 73mhz chunk of spectrum playing such a major role in technology?

Unfortunately for unlicensed spectrum management, the economic calculation problem is unavoidable. The FCC cannot possibly ascertain the socially optimal rule-making scheme for unlicensed spectrum. Who knows if 4W EIRP is too much, or too little? Despite this problem, 2.4 Ghz has still arguably resulted in a hotbed of innovation rivaling the achievements of any other band on a megahertz-for-megahertz basis.

To be sure, spectrum rights should be sold on the free market, with the federal government acting as a registrar of spectrum deeds. The real question is, should every last bit of the spectrum be licensed, or is there a valid case for setting aside a small portion of the airwaves for open, unlicensed, government-regulated use?