Progress and Freedom Foundation has recently published this piece by Adam Theirer and Barbara Esbin on how encouraging a deal between broadcasters and wireless providers to reduce the spectrum used by broadcasters and auction more spectrum for wireless use would serve the public interest. The piece raises some good points. For one thing, it is happily free of the “broadcasters are obsolete and we ought to take their spectrum back” rhetoric that often accompanies these proposals (not from PFF, I should add, but from a number of others). But the paper is woefully short on specifics. It touts the value of such a deal (freeing up spectrum for wireless) and lays out some general approaches, then urges the FCC and Congress to broker a deal between the broadcasters and the wireless industry through a number of possible auction mechanisms.
And now, the FCC has issued a public notice in the National Broadband Plan soliciting input on what they should think about using broadcast spectrum as part of the national broadband plan.
This got me thinking. Is there a mechanism the FCC could use, consistent with existing law, which would allow for the sort of broadcast band clearance the FCC would like to see? And, as a bonus, could this also clear some space for white space use? After some consideration, I hatched the scheme below. It is somewhat slower than than the wireless industry would like. I expect it would take about 5 years to finish the transition. But that is not bad given that it took 4 years to manage the DTV transition and auction from the time Congress set the hard date in 2005 to the end of analog broadcasting in June 2009. Also, my plan would allow continuing gradual build out, and combines some sticks to go with the carrots.
I’ll add that I’m not convinced this is worth doing. I think the current obsession with broadcast spectrum as the solution for the upcoming spectrum crisis suffers the same myopia as focusing on offshore drilling to cure the energy crisis — it defers the crunch but doesn’t solve the underlying problem. Wireless demand is going to continue, and we need to fundamentally change how we manage spectrum access (rather than spectrum allocation) to remain on a sustainable path for growth. I also point out, as we discovered while doing the broadcast white spaces proceeding, that there are a lot of non-broadcast uses in the existing television bands that are not broadcast users. These secondary services are going to get awfully squeezed if we crunch the broadcast bands further.
All that said, a well constructed auction could free up a nice chunk of spectrum in the short term that could promote wireless services and competition — especially if it came with a spectrum cap so VZ and AT&T didn’t hog all the good stuff again.
More below . . . . .
The basic model I would use here is a variation of the one Thierer and Esbin describe as “Option 2: Auction spectrum with the right to compensate incumbents.” However, unlike Theirer and Esbin, who are handicapped by the ideological conceit that broadcasters have a property right in their spectrum (which, as I have pointed out repeatedly, the Communications Act explicitly rejects no fewer than four times; see 47 USC 301, 304, 309(h), and 309(j)(6)(B)), I am not averse to using the FCC’s license renewal process and its authority under 47 USC 316 and 47 USC 303(f) to grease the wheels. Furthermore, while I am quite happy to use market mechanisms when convenient, I also understand about setting up a dynamic to drive things in the right direction (also anathema to the free market purists as an affront to Coase).
Why Auction?
That’s a good question for several reasons. The justification for the “slash, burn, clear & auction” approach to spectrum is that current licensed-based business models need big blocks of spectrum allocated on a national basis. It also adds loot to the Treasury, a fact that the FCC is technically prohibited by law from thinking about (see 47 U.S.C. 309(j)(7)(A)) but, like telling teenage boys not to think about sex, they always end up thinking about it to the exclusion of anything else. Certainly the thought of money to the Treasury will grease the wheels for what will be very contentious if the broadcasters fight tooth and nail.
Nor do folks like the idea of relying on existing secondary markets. As I observed when I wrote about this awhile ago, broadcasters could make excess spectrum available through existing secondary market mechanisms. Problem is, they generally don’t want to. Nor do existing cellular providers (and those are the folks driving licensed spectrum policy these days, it’s all about the smart phones) like leasing. For all that spectrum auctions are hugely expensive, providers like the certainty of owning. They also have difficulty getting financing (so I’m told) for secondary-market based deployments because the business model is still uncertain. But that problem is somewhat re-enforcing. As it increasingly looks like the spectrum part of the National Broadband Plan will have a component focusing on secondary markets, this may change over time. But in the short term, no one who wants to free up broadcast spectrum wants to rely on secondary markets instead of some kind of slash-burn-clear-auction.
I’m not sold on it, but I’m not opposed either — at least not in theory. I am not a believer in perpetuating the existing exclusive license scheme, preferring to focus on technologies that increase shared spectrum access and spectrum re-use (like the unlicensed spectrum or the “licensed lite” 3.65 GHz band, which now includes a WiMAX device that is “contention based” and can function in a more open, unlicensed-like environment). On the whole, I’d rather see the FCC spend its time working on rolling out broadcast white space technology than chasing another slash-burn-clear-auction. I’m also not wild about perpetuating the “license as property” idea or that the highest best use of spectrum access is to auction off limited government monopolies to the highest bidder.
At the same time, however, I’m a pragmatist and recognize that (a) licensing is not going away any time soon), and (b) it really does meet some consumer needs. So while I’m not sold on the idea that compacting the broadcast band so we can have some new blocks of spectrum to auction is worth it, I’m not violently opposed either — as long as it doesn’t get in the way of what I consider the more promising rout of promoting cooperative sharing and secondary markets.
The Parameters
Lets assume therefore the following. We want to clear blocks within the broadcast band, without violently dislocating broadcasters or depriving them of existing must carry rights — which are based on holding a full power broadcasting license (47 USC 534 and 535). We want to protect existing white spaces allocation (or at least, I do). We want to accomplish this entirely with existing FCC authority, because getting Congress to pass a statute on this will take a long time, if it happens at all.
I’ve deliberately left off loads of complications that arise in the context of any clearance scheme — such as the fact that lots of people other than broadcasters use the “broadcast bands” on a secondary basis and the uses in question are not compatible with the sort of cellular use envisioned by existing cellular licensees for smart phones. I’ll save those for a later post.
How It Would Work: Basic Model Similar To MDS Auction (FCC Auction 6)
The underlying model I would use is similar to the one used in the MMDS band back in the 1990s, but learning from those mistakes.
Back in the 1970s, the Commission started giving out licenses for microwave closed circuit TV. The idea was to create little television equivalents. This sort of worked for the educational flavor of the service, the Instructional Fixed Television Service (ITFS), but it never caught on in the commercial side, the Multipoint Distribution Service (MDS, latter the Multichannel MDS, or MMDS). In the 1980s, some folks tried to put together enough capacity to offer “wireless cable,” a would be competitor to wireline cable that bombed for a number of reasons including the difficulty in assembling enough capacity to offer the 30+ channels needed at the time to compete with cable.
In an effort to overcome this problem, the FCC held an auction of MDS overlay rights. They divided the country into “Basic Trading Areas” (BTAs) and auctioned off the BTAs subject to the previously issued licenses. The BTA holder got any area not covered by an existing license within the BTA, and the right to try to buy out or lease the spectrum from any existing license holders.
The execution of this was crappy, with the result that, 15 years later, the FCC is still trying to clean up the MMDS Band, now the Broadband Radio Service (BRS) Band in the valuable for WiMAX 2.5 GHz band. In fact, the most recent auction to clean out the dribs and drabs finished last month. Done right, however, it could be done relatively quickly and efficiently in broadcast.
Implementing the MDS Auction Model to Clear the Broadcast Band
I don’t know how much spectrum it actually takes to do a single high-def signal. But based on what I’ve read, we could take the existing 6 MHz currently allocated to broadcasters and chop it down to 3 MHz. In other words, we could halve the number of allocated channels and require existing licensees to double up and partition their existing license into distinct 3 MHz licenses sharing a set of physical facilities (or even not sharing). ION Media and Bob Johnson have proposed such a partition and separate license under existing rules to give Johnson’s proposed new cable network a boost, and the idea is based on some early FCC licensing precedents.
The next part of this scheme relies on the fact that every 8 years, the FCC renews broadcast licenses on a market by market basis (47 USC 307(a)). Although these renewals are effectively rubber stamps, a licensee does not, as a legal matter, have a right to an expectation of renewal. Indeed, as noted above, the FCC can shift individual licensees (as it did to clear the 700 MHz band for the final band plan) or can simply eliminate a service after a notice and comment rulemaking (although generally the FCC does not like to do this and will migrate services to other bands rather than kill them outright).
So the FCC announces it will hold an MDS-type auction for licenses in the broadcast band divided geographically (I’d recommend using existing cellular areas rather than BTA’s, but I leave this sort of detail to uber-spectrum geeks who actually know economics, like fellow Wetmachiner Greg Rose). The licensee will have the ability to use any space not currently subject to a broadcast license, and have the right to pay existing licensees to migrate to another channel in the same market in the bands still allocated for broadcasting.
OK so far, but how do we avoid the MDS mess? At the same time, the FCC announces that it will reduce license size in the broadcast service from 6 MHz down to 3 MHz channels on a rolling basis along with license renewals. Furthermore, licensees in the auctioned bands will not get renewals for even a 3 MHz channel on their existing license allocation. Instead, they will be required to migrate out to another channel in the remaining “broadcast band” in the same market, which will be partitioned to create a home for the “exiled” broadcast licensee. Licensees that agree to take in a “refugee” will be entitled to compensation for any expenses (and a reasonable bribe fee for inconvenience from the auction winner migrating the refugee). Licensees that do not take in an exile get no compensation for “losing” half their spectrum, as they are not legally entitled to any compensation and don’t have to incur any new costs.
This creates an obvious incentive for licensees in the remaining auction bands to take in refugees. As an added inducement, the FCC can reserve the right to relocate any displaced refugee broadcasters without compensation on the newly created unassigned channels. All the rest of the unassigned “half channels” thus created in the remaining broadcast bands become available for white space use.
Until license renewal time (or some other date certain if we can’t get this resolved before the next renewal window begins in 2012), the auction winner and the refugee broadcaster can strike any deal they want. But once we get to auction renewal time, if the refugee and the licensee still can’t cut a deal, they go to baseball-style arbitration to resolve the question of migration fees. (In baseball-style arbitration, each side sets out what it thinks is a fair offer, and the arbiter must select which of the two packages is most reasonable. The arbitrator has no authority to craft a different package. The result is to drive parties toward the most reasonable first offer rather than to start far apart in the expectation they will be driven toward the middle during the process since if you are manifestly unreasonable, you will lose.)
The result should be that in any market, the broadcasters that are migrating and the winning licensee have an incentive to cut a deal before going to arbitration, which is always a chancy business. Broadcasters wanting to get rewards will move quickly, because they are much more likely to get better terms early in the process. The non-migrating broadcasters will have incentive to receive “refugees” to smooth the process, since if they do not take in a refugee they get no money for “losing” half the spectrum. Wireless licensees winning the licenses will also have incentive to cut deals quickly, as they want access to the spectrum as soon as possible. The FCC can permit winning licensees to pay a premium over the real market value of the “lost” spectrum by permitting the incoming auction winners to compensate broadcasters for bogus “lost business opportunities” from early migration — thus avoiding the problem of paying broadcasters for spectrum they do not legally own and creating further incentive to close the deal early.
Simple In Theory, Hard In Fact
I recognize that such processes are always messier in practice than in theory. To my mind, this is fairly elegant and meets all the goals. It frees up space for auction, creates white space to offset the lost white spaces in the auctioned channels (although it would be nice to set aside a vacant channel or two for white space use), creates a mechanism to bribe broadcasters without being too outrageous, creates incentives to clear the band with reasonable speed, and minimizes the problem of hold outs. And, best of all, it does not require any Congressional action.
I do note that this does not maximize auction revenue, as the bidders for new licenses will discount the unknown amount they must pay for clearance. But while auction revenue is the sweetener to get Congress to go along with this in the face of broadcaster protests, maximizing revenue is really not the point. The goal is to get licenses for existing cellular providers (lets not pretend we’re going to get new competitors out of this), and this does it in a way that clears the band in a window of about 5 or so years (based on this table, which places the next window to begin in 2012) (licenses can remain with renewals pending until the migration/arbitration ends).
I also don’t try to solve the question of how many bands I would clear this way. I expect there is some logical trade off point based on factors such as how much white space it would create/eliminate, how much spectrum could be comfortably cleared up for cellular use without crewing up broadcasting, where the heck do you put every one else who is suing the bands, and other factors you’d need to solve in a rulemaking. Given how DTV works, I’d recommend reallocating channels at the bottom of the dial (the VHF band). That would nix the effort to reclaim Channels 5 & 6 for radio, but so it goes.
Finally, as noted above, even assuming the best possible circumstances, there are enough complications with having an auction for broadcast bands at all that I’m not sure this is a good idea under any circumstances. But I address those separately.
Stay tuned . . . .
