Averting a spectrum disaster: Now for the hard part

With the passage last week of legislation authorizing the FCC to conduct new spectrum auctions, you might think that the looming spectrum crisis has been averted.

Nothing could be farther from the truth--or more dangerous to the continued health of the mobile ecosystem.

To avoid severe service interruptions or outright collapse of mobile networks, the FCC's 2010 National Broadband Plan estimated that mobile users will need an additional 300MHz of spectrum by 2015 and an additional 500 MHz by 2020. Many industry insiders believe these estimates are actually low.

The FCC now has the authority to conduct auctions to get that capacity into the hands of mobile carriers. The problem is that we don't have anywhere near that much usable spectrum left.

The frontier is now closed
Barely a blip a few years ago, mobile broadband is growing at an astronomical pace. AT&T reports that since offering the iPhone on its networks in 2007, data volumes had increased by 8,000 percent by 2010. According to a report last week from the White House Council of Economic Advisers (PDF), mobile data traffic will increase twenty-fold between 2010 and 2015.

Existing networks simply cannot handle that increased demand without access to more bands of usable radio spectrum.

That would have been easy in the old days. Radio frequencies were plentiful, and users were few and far between. But as George Mason University economist Thomas Hazlett noted last week in Washington, after 85 years of handing out spectrum licenses, often at minimal charge to the licensee, the U.S. has run out.

We don't have 500 or even 300MHz of usable spectrum left to auction, at any price. Today's available inventory is closer to zero.

While technological innovation expands the range of usable frequencies, there's no doubt among engineers and policymakers that as things stand today, mobile users will soon hit a very unforgiving wall. The "frontier" is closed, just as historian Frederick Jackson Turner concluded about the American West in 1893. Going forward, spectrum will no longer be allocated. It can only be reallocated.

How have we come so perilously close to running out of spectrum? Part of the problem has to do with the FCC's increasingly outdated licensing system. Assignments have historically been based on transient and idiosyncratic criteria that favored once-promising new applications and technologies (e.g., UHF television, pagers, satellite radio).

This "command and control" model has resulted in a badly splintered and increasingly unmanageable allocation table of more than 50,000 localized licenses. Many of these licenses arbitrarily limit their use of spectrum to applications that have faded or disappeared, but there's no easy mechanism for reclaiming spectrum that could be put to better use. The FCC doesn't even have a working inventory of all its licenses.

(The federal government itself holds vast swaths of spectrum, much of it warehoused, but no central authority has the power to free up under- or unused bands.)

New auctions aim to dislodge underutilized frequencies
In the 1990s, the FCC finally shifted to an auction model, removing some of the whimsy from the process and, not incidentally, generating billions of dollars for the Treasury. But the agency still has a hard time resisting old temptations. Instead of picking winners and losers directly, the FCC now attaches conditions or limits auction eligibility to micromanage emerging markets and industries--or try to in any case. One result of this tinkering has been that several recent auctions failed to meet their reserve price.

The legislation enacted last week will curb some of these abuses. It will also test a novel approach to reallocating existing spectrum licenses. Over-the-air television broadcasters, who hold spectrum particularly well-suited for mobile broadband uses, will be asked to name a price to give back some or all of their current allocations.

If enough volunteers come forward, the agency will auction off that spectrum to mobile providers--or anyone else, including other broadcasters--who values the frequency more than the current licensee. The government will then share the proceeds of the auctions with the participants, reducing the deficit and redirecting spectrum to higher-valued uses.

New licenses will come with flexible use permission, making it easier for future market transactions to reallocate it again when future applications or technologies find a better use. (Existing spectrum licenses can be sold today on secondary markets with FCC approval, but use limitations and conditions still apply.)

This "incentive auction" model is promising, and Congress and the FCC are to be commended for passing this critical legislation after two years of logjams and tangential fights that kept even bipartisan proposals stalled.

But the law doesn't come close to solving the spectrum crunch--not by a long shot.

For one thing, it isn't at all clear that enough broadcasters will volunteer. Over-the-air viewership has fallen dramatically over the last two decades as over 90 percent of all households shifted to cable, satellite, and now broadband Internet alternatives. But the economics of local television stations are complicated. For example, federal law allows local broadcasters to force cable providers to carry their signal or negotiate a price for retransmitting it. Careful exploitation of this right often masks what are actually failed businesses.

And while the FCC will have the ability to "repack" nonparticipating channels to create contiguous nationwide licenses, that process will be long and contentious. In the lead-up to passage of incentive auction legislation, broadcasters lobbied intensely to limit the agency's ability to maximize auction outcomes. The lobbying will only get more aggressive as the FCC gears up to design the new system.

Time now for the short and medium-term solutions
At best, it will take upwards of 10 years before significant new spectrum for mobile broadband can be deployed from the incentive auctions. And we're already two years into the FCC's own doomsday clock toward spectrum exhaustion.

So now that the legislative battle is over, it's well past the time to think about short and medium-term plans to stave off an epic failure of the mobile revolution. The stakes are high. The mobile industry is one of the few bright spots in the otherwise sour economy. According to a recent Deloitte study, investment in 4G networks could range from $25 billion to $53 billion over the next four years, generating up to $151 billion in GDP and as many as 771,000 new jobs. And that doesn't count the revenue from app stores and the services they make possible.

So what can we do while waiting for the incentive auctions to get under way? Until recently, carriers in need of more spectrum could merge with other carriers to achieve economies of both scale and technology. In the past six years, the FCC approved nearly a dozen mobile mergers, nearly all of which were motivated by the need to make better use of limited mobile bandwidth.

But in rejecting AT&T's proposed merger with T-Mobile last year, the FCC sent an unmistakable signal that it will no longer allow market transactions as a work-around to its own plodding and sclerotic mismanagement of the nation's airwaves. The battle is heating up, for example, over Verizon's pending acquisition of AWS spectrum from a consortium of cable companies. T-Mobile, ironically, is now aping the familiar claim that allowing Verizon to purchase any additional spectrum will harm competition. The spectrum being transferred, however, is not currently being used for anything.

Besides more spectrum, the most significant way a mobile broadband carrier can enhance performance and capacity is to add more cell towers and upgrade antennae at existing sites to improve network density and site efficiency. Mobile carriers already spend billions of dollars each year to upgrade and expand their core infrastructure. They would spend even more--if only local zoning authorities would let them.

They won't. Despite a 2009 FCC rule requiring local authorities to decide on cell tower modification and construction requests within 90 and 150 days respectively, thousands of applications are languishing in political limbo. A U.S. Court of Appeals in Texas recently upheld the FCC rule, but it has rarely been enforced.

Areas with some of the most vocal complaints about existing network quality, not surprisingly, also have the worst record for approving applications, even to add equipment to existing towers. A 2009 study from wireless industry group CTIA published just before the FCC's "shot clock" was imposed found that cell tower applications in the San Francisco Bay Area were regularly stonewalled for 28 to 36 months. Nationwide, according to the FCC, "of 3,300 pending zoning applications for wireless facilities, more than 760 (nearly one quarter) had been pending for more than a year and 180 had been pending for more than three years."

The new federal law authorizing incentive auctions took some modest steps toward curbing these abuses. That's a good starting point. But coordinated state and federal action will be needed to collapse the black hole of infrastructure zoning delays.

Improving spectral efficiency will also require clear-cutting generations of other encrusted and obsolete rules at all levels of government. In 2010, for example, the FCC cleared the use of the "white spaces" between television channels for unlicensed wireless technologies. The agency didn't approve the first new device to use white space until late last year, however. Quick approval of spectrum transfers, more flexible licensing, and relief from onerous wireline regulations that limit the use of fixed networks as both support for and competition with mobile services also need to happen, and quickly.

Mobile broadband providers will also have to rely on technological solutions to improve network performance. A wide range of innovations, including smart antennae that can easily switch bands, miniature cell towers, home-based femtocels, and software that allows multiple uses of the same bands without interference are all being deployed to make better use of existing allocations. Smartphones can also be programmed to switch from cellular networks to local Wi-Fi, offloading wireless traffic to high-capacity wired networks whenever possible.

Incentives for consumers, both carrots and sticks, could likewise help stave off network failure. Providers will need to offer more incentives to quickly retire older mobile technologies. Since each new generation of cellular protocol makes more efficient use of spectrum than its predecessors, getting customers off 2G and 3G networks and onto 4G (especially 4G LTE) networks will save considerable bandwidth.

LTE, for example, can handle roughly six to eight times the capacity of a 2G network. Some of those savings would be lost to users taking advantage of video and other high-bandwidth services available on LTE, but not so much as to use up all the increased efficiencies.

Graduated or tiered bandwidth pricing, likewise, discourages excessive network use by a few extreme customers, especially at peak times.

Coordinated efforts are key
This is only the start of a much longer list of important initiatives. Short- and medium-solutions to the spectrum crisis are possible, but won't come easily. Avoiding disaster in the mobile ecosystem requires a combination of smart technology investments, innovative business practices, and policy reforms likely to offend vested interests.

Each is valuable on its own, but coordination will be crucial if we are to improve spectral efficiency enough to keep mobile users going while we wait for the incentive auctions to run their course.

Even if we get through the next few years, it's clear that staving off future crises will require radical changes to spectrum management. The patchwork quilt woven by 85 years of quixotic and often political decision-making has left U.S. airwaves dangerously inflexible and unnecessarily fractured. The accelerating pace of technological innovation is on a collision course with command-and-control assignment of spectrum. Something has to give.

For the long term, we need to rethink the entire spectrum map. Given current and future advances in radio technologies and software, in fact, we may soon find we won't even need a map--or a regulator that believes it can do a better job allocating spectrum in fits and starts than a market that runs on Internet time.

That assumes we survive the current, largely self-inflicted crisis. First things first: regulators have much more work to do to clean up the current mess.