tom lenard is a senior fellow and president emeritus of the Washington-based Technology Policy Institute. Larry White is a professor of economics at NYU's Stern School of Business.
Published January 21, 2019
To the applause of economists — not to mention stakeholders in the vast wireless communications industry — the FCC has driven the transformation of America’s deeply inefficient system from one that regularly assigned spectrum use by government fiat to one based on market allocation. A substantial portion of the “airwaves” is now flexibly licensed under a structure in which the government retains formal ownership, but licensees have considerable latitude to use or transfer their “rights” to others.
Still, much remains to be done to employ markets to allocate spectrum to uses where it is most valued as the country moves into the next generation of wireless technologies — often lumped together as “5G.” In particular, realizing the potential of 5G for gee-whiz applications will require that large amounts of additional spectrum be diverted from other applications.
That’s more easily said than done. Much of the spectrum needed for next-generation uses is currently controlled by federal agencies, ranging from the National Weather Service to the Defense Department, which pay nothing for it and therefore have no incentive to give up their stranglehold.
The good news here is that a big chunk of government-occupied spectrum could almost certainly be reallocated to private use without degrading government services. The goal, then, ought to be to give these agencies the appropriate incentives to relinquish enough of that spectrum to balance the value of public and private spectrum services at the margin.
The Long and Winding Road
Judging from past experience, getting there won’t be easy or rapid. The rules by which the FCC allocated blocks of spectrum for spezcific uses (namely, public safety, radio and television broadcasting) and assigned licenses to specific users were adopted in 1927. Nobel-laureate-in-the-making Ronald Coase drew a map for reform in 1959, explaining how giving away or selling flexible, transferrable licenses for spectrum would lead to more efficient use. Yet Congress waited until 1993 to give the FCC authority to auction such licenses to the highest bidders — and, even then, Congress’s motives were more about generating revenues for the federal budget and less about efficiency.
Meanwhile, the compelling need to open another big swath of spectrum for wireless communications has been clear for some time. The FCC and the National Telecommunications and Information Administration, the agency in the Department of Commerce that is responsible for managing federal spectrum, have been studying a number of government-occupied spectrum bands to determine the feasibility of repurposing them for commercial use or for creating ways for private users to share them.
The logjam shows signs of breaking. On its own, the FCC is completing a proceeding to determine auction and sharing rules for the 3.5-GHz (3,550-3,700 MHz) band, a portion of the spectrum currently in the hands of the military. And, more recently, the NTIA announced it was eyeing an additional 100-MHz (3,450-3,550 MHz) band also being used by government agencies, as mandated by the recently passed Mobile Now Act. But if the interest-group tussle over spectrum follows precedent, only a fraction of the 100 MHz will be freed for 5G anytime soon.
Indeed, “slow” has always been the name of the game in spectrum reallocation. For example, back in 2010, the Obama administration’s National Broadband Plan proposed that the spectrum set aside long ago for over-the-air TV be transferred to uses in mobile applications. The mechanism the FCC adopted to do this — voluntary auctions — targeted 120 MHz, of which only 70 MHz was successfully transferred. This spectrum is now scheduled to become available in 2020.
The most recent examination of the broad policy issues, a 2012 report by the President’s Council of Advisors on Science and Technology (PCAST), concurred that outright transfer of agency-controlled spectrum was problematic. But it did see a plausible alternative in sharing bands between public and private users.
This approach, we believe, has serious drawbacks. Even if government agencies could somehow be induced to share spectrum they use only intermittently, the difficulty of coordinating multiple uses would sharply reduce the value of the spectrum to private users and thus the investments they will be willing to make — a reality shown by the generally lower prices at auction in sales of rights to encumbered spectrum.
Indeed, the PCAST report did not conclude that sharing would be a desirable way to break the impasse. Rather, it reflected the view that agency resistance to simply relinquishing spectrum made that approach a losing proposition — and consequently, at least by comparison, made sharing worth pursuing. From the agencies’ perspective, spectrum rights are a freebie. Even if they have no current use for the spectrum, the option of using it in the future has value.
In this same vein, the Spectrum Now Act, introduced in both the House and the Senate, provides more funds for government agencies to study the feasibility of relocating or sharing spectrum with private users. But it’s unclear why agencies would be interested in conducting these studies in the first place.
A Way Out
At the root of the impasse, of course, is the difficulty of giving public agencies appropriate market incentives. Even if an agency had the option of selling the rights to spectrum for which it had little use, it would surely be reluctant to do so because any windfall it gained could easily be taken away in the give-and-take of federal budget reviews.
Britain has taken a shot at solving the problem. It instituted a system of “administered incentive pricing” to provide direct incentives for government agencies to use spectrum more efficiently — but, for the aforementioned reason, with ambiguous results.
There is, however, another route toward incentive pricing for government-held spectrum that we think holds more promise.
Consider first a parallel issue: the way that the federal government manages its vast real estate holdings. The Government Services Administration owns or leases much of the office space that is used by federal agencies. (The Defense Department is the major exception.) In turn, the GSA leases that space to agencies at market-comparable rates. The requirement that they pay rent creates an incentive for the agencies to economize on space and/or location. The GSA, incidentally, can use the rent to acquire additional property if needed without approval from Congress.
The GSA model offers a template for how government-held spectrum could be handled. Suppose that all government-used spectrum were “owned” by a central government agency — we call it the Government Spectrum Ownership Corporation, or GSOC — and leased to government users.
Under our proposal, the existing users would retain assured access to the spectrum they currently control, but would have to pay for it at rates reflecting its value in private uses. In the first year, the Office of Management and Budget would add a sum to each agency’s budget equal to the rental payment, so the first year’s financial transactions would be a wash.
In subsequent years, agency budgets would start from the base that included the initial allocations and rental charges, but the GSOC would adjust the rental rates in light of updated information about the market value of the spectrum. The agencies and OMB would then negotiate budgets (as they do now). But there need not (and should not) be a one-to-one budget adjustment in which agencies automatically get enough cash to cover the fluctuating rent for spectrum. If an agency economized on spectrum usage, it could keep the windfall — perhaps using the money to purchase more modern telecom equipment that uses less spectrum, thereby creating a virtuous circle of sorts.
As the agencies economized on spectrum and handed back the surplus, the GSOC would sell or lease the newly created surplus to the private sector. Note that if this surplus were auctioned, the GSOC would receive a bonus in the form of updated information about the private value of spectrum, which it could use to set future charges for government agencies. The same auction platform might also be used by the private sector as a secondary market to buy and sell spectrum.
From the agencies’ perspective, spectrum rights are a freebie. Even if they have no current use for the spectrum, the option of using it in the future has value.
The GSOC could also help implement a variant on the same theme of creating incentives for efficient use by public agencies. Here, the incumbent agency would retain primary rights to the spectrum it uses, taking into account the required rental payments that reflect private opportunity costs. But there would be one important difference: secondary rights to unused spectrum in the band could be auctioned to commercial users through “overlay” licenses issued by the federal government.
These licenses, proposed by the economist Tom Hazlett and fleshed out by Brent Skorup of the Mercatus Center at George Mason University, would give the winning bidder the exclusive right to bargain for a transfer from the government incumbent — much the way a buyer of real estate can bargain with a long-term leaseholder to vacate before the lease has expired. For example, the overlay licensee might agree to pay for equipment upgrades for the government user in return for freeing up some of the now-surplus spectrum.
In this sense, although overlay licenses have somewhat of the flavor of spectrum sharing, they would give the interested parties — both the government incumbent and the private newcomer — incentives to work out creative ways to induce more efficient use of government-held spectrum.
Local politics is not the impediment to spectrum reform. Rather, the problem is a lack of appropriate incentives for the agencies themselves.
Several objections have been raised to the GSOC approach. First, estimating the appropriate spectrum rents to be charged to federal agencies would be difficult — more difficult than for GSA’s building rents. Market rental rates for spectrum, after all, would depend on a host of factors running from frequency assignments to locality.
Yet, while admittedly no walk in the park, obtaining reasonably accurate estimates does not seem an insurmountable problem. The FCC currently pays consultants tens of millions of dollars to design and implement spectrum auctions. Developing and updating a reliable model to estimate spectrum value would seem, if anything, to be a less difficult task.
Second, critics point out that many big government agencies — notably, the Defense Department — administer their own real estate. Indeed, all told, GSA administers only about 12 percent of the federal building inventory. This has a direct implication for our GSOC proposal: if big agencies can keep their real estate out of the hands of the GSA, presumably they’d have a good shot at retaining control of their spectrum, too.
Perhaps — though it would be worth trying to wrestle it away. And in any case, enough spectrum would still end up with the GSOC to free up billions of dollars’ worth of surplus for private use.
The Brac Digression
Skeptics about the GSOC idea point to another approach for freeing spectrum. The fact that the Defense Department had (and no doubt continues to have) a surplus of military bases may be partially attributable to the agency’s ability to avoid GSA control of its real estate. But, from time to time, Congress has managed to stand up to the DoD on this issue.
In the case of surplus military bases, the fix is the Building Realignment and Closure process, which has been used four times. Under BRAC, an independent commission is tasked with proposing a set of base closings, which the Congress must accept or reject in its entirety. The objective is to give individual members of Congress political cover for closures that would inflict economic pain on their constituents.
In our view, the BRAC approach is virtually irrelevant to the spectrum impasse. It misses the point because local politics (as represented in Congress) is not the impediment to spectrum reform. Rather, as discussed above, the problem is a lack of appropriate incentives for the agencies themselves.
Of course, no government agency would be happy to give up its current de facto ownership of “free” spectrum, but agencies aren’t happy about having to pay rent to the GSA for office space, either. The change would have to be sold as a means of achieving better overall management of a multibillion-dollar resource while offering sweeteners for agencies that figured out how to meet their responsibilities with less spectrum.
Yes, It Really Matters
It’s hard to get the public excited about spectrum allocation. For most of us, the very idea of spectrum is (to borrow from Winston Churchill’s wordplay on another subject) a riddle wrapped in a mystery inside an enigma — one more elusive concept from physics we were grateful to forget after the final exam. But spectrum allocation is way too important to be left to the experts — or, more to the point, to lobbyists and bureaucrats.
Much of the technological change of the past few decades has depended on spectrum access. And it’s not just about fun and games on those ubiquitous screens. Cheap wireless communication has been a key to economic development in much of the world, and promises to revolutionize everything from education to medicine in our own world. The failure to make the most of what is an increasingly scarce resource would constitute a self-inflicted wound of the first order.