The last round of the day for Auction 73, Round 31, was postponed until tomorrow by the FCC late today. The announcement on the Integrated Spectrum Auction System read:

“2/4/2008 04:11:42 PM
Round 31 Postponed
Due to a delay in the availability of complete downloadable reports, Round 31 will be postponed until 9:30 a.m. ET tomorrow, February 5, 2008.

”We will continue with the previously announced five round bidding schedule until further notice. Bidders are reminded to monitor auction announcements for further changes in the bidding schedule."

Sources at the FCC indicate that the system glitched on producing the end-of-round reports from Round 30 and there was insufficient time to locate and correct the bug before Round 31 was scheduled to commence at 4:30 p.m. EST. Round 30 still hasn't been posted on the Integrated Spectrum Auction System as of 4:40 p.m. EST, and it was due at 3:40, which suggests that the bug hunt is a bit more complicated than the FCC initially anticipated, but there's no reason to think the auction won't recommence tomorrow as scheduled.

I'd like to reiterate what fellow Wetmachiner Harold Feld wrote yesterday: the telecoms incumbents' claims of problems arising from anonymous bidding are nonsense, part of a campaign to sow disinformation lest Auction 73 (700 MHz) and its success persuade the FCC to permanently adopt anonymous bidding rules for its auctions.

I call your attention to this table, which compares the number of bids on each license in B Block in rounds 1-26 of Auction 73 to the number of bids on each comparable CMA in Auction 66 (AWS-1) in rounds 1-26 of that auction. Note that in general the smaller the CMA number, the larger the population of the CMA (e.g., CMA001 is New York City and its immediate environs, CMA002 is the Los Angeles area, etc.).

What is striking about the data presented in this table is threefold. First, significantly more bids are being placed in general in rounds 1-26 in Auction 73 than in Auction 66. Second, extraordinarily more bids are being placed on the smaller and intermediate-size CMAs in Auction 73 than in Auction 66. Third, a much smaller percentage of licenses are receiving no bids in the first 26 rounds in Auction 73 than in Auction 66.

I am hard put to find an explanation of this extraordinary increase in competition, particularly for the smaller and intermediate-size licenses, which does not involve the effects of anonymous bidding. I suggest that the data, even though they do not disclose bidder identities, are entirely consistent with a more vigorous presence of new entrant and non-incumbent bidders who are protected from retaliatory and blocking bidding by large incumbents by anonymous bidding and are, therefore, more willing to engage in strong competition.

The telcos and cablecos can wail and moan to Communications Daily about the “risks” of anonymous bidding to the FCC, but the principal risk of anonymous bidding seems thus far to be the risk that fat-cat telecoms incumbents won't be able to get all the spectrum in this auction by their usual bullying and exclusionary tactics. There's no risk at all to the treasury — revenue from the auction is already wildly exceeding pre-auction projections — and there's no risk that competition will be wan, as the data presented here amply demonstrate.

I can't prove it, but the following table is suggestive that there is significant new entrant/non-incumbent activity over a wide range of licenses, most strikingly in B Block. This table tracks numbers of bids per license by block (and the licenses are further sorted by total number of bids per license) over rounds 1-26 (round 27 will open on Monday).

What is interesting is that there are large numbers of bids on a wide variety of licenses of differing population sizes. Since the incumbents traditionally focus on the higher population area licenses in each block, the relatively high levels of competition over licenses of significantly differing population size suggests that anonymous bidding has emboldened new entrants and non-incumbents to compete more intensely than in earlier auctions (which also accounts, in part, for the larger revenue stream indicated by the higher provisionally winning bids).

It's not much evidence, but it is an interesting datum, and it appears to show a pattern consistent with higher levels of competition over a wider range of licenses.

A, B, and C Blocks have exceeded their reserve prices as of round 17 in Auction 73, and E Block has reached 83.73% of its reserve price, while D Block has languished at 26.99% of reserve price since the first round. Unless a great deal of the activity in A and B Blocks is intended to preserve eligibility for later round intervention in C Block, the probable C Block winner has likely made its winning bid in round 17.

The rate at which A and B Blocks have exceeded their reserve prices by the end of round 21 today — 190.51% and 462.50%, respectively — seems unlikely to abate, which may push revenue from Auction 73 to $16-17 billion, perhaps as much as $20 billion, despite the fact that D Block will almost certainly have to be reauctioned if the current pattern holds.

How much better A, B, and C Blocks are doing at this point than even at the end of Auction 66 (AWS-1) is shown in this table, which compares the dollar per MHz per population price each license in those three blocks obtained with the provisionally winning bid as of the end of round 21 to the final dollar per MHz per population price comparable licenses received by the final round of Auction 66. Since the bandwidth is different in each auction, $/MHz/Pop standardises the data for comparison.

Clearly the majority of 700 MHz spectrum on offer in Auction 73 is much more highly valued than the spectrum on offer in Auction 66: the average $/MHz/Pop price of an A Block license at the end of round 21 in Auction 73 is 193.53% of the final $/MHz/Pop price of comparable spectrum in Auction 66; the average $/MHz/Pop price of a C Block license at the end of round 21 in Auction 73 is 623.06% of the final $/MHz/Pop price of comparable spectrum in Auction 66. For C Block, the 50-state package (REAGs 1-8) is reaching 102.65% of the final price of comparable spectrum in REAGs 1-8 in Auction 66, while REAGs 9-11 are averaging 134.10% of what they finally obtained in AWS-1.

From the point of view of the U.S. Treasury Auction 73 is already a hell of a success. What remains to be seen is how well new entrants and smaller competitors did, whether the incumbents ran the table again, and whether we got a national third broadband pipe. But we won't know that until the FCC releases bidder identities and bids at the end of the auction.

The fact that the C Block has dangled on the precipice of reaching its reserve price from round 13 to the close of today's bidding action in round 16 has led to speculation that Google never intended to go seriously for the spectrum, but was merely trying to goad Verizon or ATT into committing on the Block. I grant that we have almost no intelligence on who the C Block bidders are, and it is very, very early to speculate on the auction's ultimate outcome. However, I have a theory, grounded in an understanding of game theory and the auction rules, which calls this latest conventional wisdom into question.

There are at least two, and possibly three, current bidders for the bulk of C Block. Two have been trading off the lead for the 50 state package (REAGs 1-8), let's call them A and B: A in the first round (1 new bid), B in the second (1 new bid), A in the third (1 new bid), B in the fourth (1 new bid), A in the fifth (1 new bid), B in the seventh (1 new bid), A in the eighth (1 new bid), B in the tenth (1 new bid), A in the twelfth (1 new bid), B in the thirteenth (1 new bid). B has been the high bidder since the thirteen round with no need to raise its bid. In the sixth round there were also mid-range bids placed individually on REAGs 1-8. Either the individual bids on REAGs 1-8 in round six were B's response to A's bid on the package in round 5 or another bidder, C, forayed at that point.

B can sit indefinitely on its current bid, waiting for the minimum acceptable bid (MAB) to converge on the reserve price of the Block without requiring activity waivers (the FCC historically reduces MABs in the presence of bidding inactivity). That would allow B to obtain the package for almost $122 million less than the current MAB for round 17. A must bid on REAGs 1-8 either on the package or individually in round 17 or lose eligibility, since it has had to expend three activity waivers to avoid bidding in rounds 14, 15, and 16. That is what we know.

I hypothesize that B is Google, that it is sitting just below the reserve price, and will continue to do so unless another actor bids, until just before the close of the auction, when it will bid the reserve price and save roughly $122 million. I grant that it is also possible that B is Verizon or ATT or some other bidder which I don't know and haven't mentioned. But game theory and the auction rules explain why B is sitting pat. A has to bid in round 17 (the MAB for the 50 state package in round 17 is over the reserve price of the Block, and the sum of MABs for REAGs 1-8 individually in round 17 is equal to the MAB for the 50 state package), or B's strategy is likely to win.

Much of the press surrounding the first two days of the FCC's 700 MHz auction has been like this Information Week story. I confess to being both amazed at the shallowness of the reporting and amused at its gloom and doom tone. To hear the press tell it, it's time to be very bearish on this auction.

A look at historical precedent is salutory. The FCC's Integrated Spectrum Auction System files for Auction 66 and Auction 73 are the places to start.

At the end of round four in Auction 66 (AWS-1), the high bids for the EAs, CMAs, and REAGs were, respectively, 4.15%, 7.09%, and 12.03% of the final net PWB prices with 47.84% of licenses receiving at least one bid. At the end of round 4 in Auction 73 (700 MHz Band) the high bids for EAs (A and E Blocks), CMAs (B Block), REAGs (C Block), and the nationwide D Block license were, respectively, 31.87%, 43.03%, 39.06%, and 26.99% of reserve price with 83.80% of licenses receiving at least one bid.

Auction 66 netted $13.7 billion. Auction 73 has a reserve price threshold of $10,386,011,520. By any objective criteria Auction 73 is off to a much better start generally than Auction 66 was. The fact that the D block has had only one bid in the first four rounds isn't terribly unusual; several licenses which eventually went in Auction 66 for very substantial sums had very little early-round action. It's important to point out that auctions with relatively high reserve prices tend to exhibit slow convergence bidding on reserve price and provide significant incentive to try to obtain the license for as little over reserve price as possible. When this tendency is coupled with the FCC's bidding increment rules, it is rather obvious that the auction is going to take some serious time and that it's rather impressive how close to reserve price the bidding is at so early a stage.

Auction 66 ran 161 rounds. I expect Auction 73 to run at least 100 rounds, and probably significantly longer. It is much too early to announce that the results of Auction 73 are disappointing... unless you appear to know as little about how FCC spectrum auctions actually work as much of the press does.

Finally, again let's begin our analysis of strategic options for major actors in Auction 73, 700 MHz Band, with a look at the footprints established by many of those actors in two previous Lower 700 MHz auctions (Auction 44 and 49) and the AWS-1 auction (Auction 66):
Cellular Market Areas (CMA) Map for Auction 44
Economic Area Groupings (EAG) Map for Auction 44
Cellular Market Areas (CMA) Map for Auction 49
Economic Area Groupings (EAG) Map for Auction 49
Cellular Market Areas (CMA) Map for Auction 66
Economic Areas (EA) Map for Auction 66
Regional Economic Area Groupings (REAG) Map for Auction 66

The Big Guys

There are quite a few major actors who qualify as the genuine big guys in Auction 73. Their participation and fundamental interests in this spectrum ensure that the reserve prices will be met and likely exceeded on all blocks (with some caveats on D Block).

QUALCOMM makes the list of the big guys in the auction if for no other reason than it nearly scored national footprint (minus the Western EAG) in a Lower 700 MHz auction. The 700 MHz Band auction provides a source of spectrum entirely compatible with its acquisition for its MediaFLO datacasting enterprise. It may be a C Block contender, but it is more likely that QUALCOMM will concentrate on E Block to flesh out its national footprint and consolidate. This isn't going to be a QUALCOMM versus the world auction; QUALCOMM will narrowly target specific licenses, go after them tenaciously, and then get out if it looks like the spectrum is going for higher prices than expected.

More below...
[Read More!]

Once again let's begin our analysis of strategic options for major actors in Auction 73, 700 MHz Band, with a look at the footprints established by many of those actors in two previous Lower 700 MHz auctions (Auction 44 and 49) and the AWS-1 auction (Auction 66):
Cellular Market Areas (CMA) Map for Auction 44
Economic Area Groupings (EAG) Map for Auction 44
Cellular Market Areas (CMA) Map for Auction 49
Economic Area Groupings (EAG) Map for Auction 49
Cellular Market Areas (CMA) Map for Auction 66
Economic Areas (EA) Map for Auction 66
Regional Economic Area Groupings (REAG) Map for Auction 66

The Mid-Range Competitors (Continued)

Cablevision is bidding as CSC Spectrum Holdings LLC. In Auction 66 it bid as Dolan Family Holdings and got creamed by incumbent blocking bidding. Cablevision unsuccessfully bid on two EAs, AW-BEA010-B (NYC-Long Is. NY-NJ-CT-PA-MA-VT) and AW-BEA010-C (NYC-Long Is. NY-NJ-CT-PA-MA-VT), and the following CMAs: AW-CMA001-A (New York-Newark, NY-NJ), AW-CMA042-A (Bridgeport-Stamford-Danbury CT), AW-CMA062-A (New Brunswick-Perth Amboy NJ), AW-CMA070-A (Long Branch-Asbury Park NJ), AW-CMA144-A (Orange County NY), AW-CMA151-A (Poughkeepsie NY), AW-CMA551-A (Ocean NJ), and AW-CMA552-A (Sussex NJ). Cablevision unsuccessfully sought all three licenses for the Northest REAG: AW-REA001-D, AW-REA001-E, and AW-REA001-F. The pattern is straightforward: replicate the footprint of their cable service in the NY-CT-NJ region in the A and B Blocks and try for one of the Northeast REAGs. Cablevision didn't get it in AWS-1 and it has to do well in Auction 73 or its triple play options are seriously curtailed. Anonymous bidding helps Cablevision only a bit, because the chief competitors know exactly where they have to bid and it is prime spectrum in the richest market in America.

More below... [Read More!]

To begin our analysis of strategic options for major actors in Auction 73, 700 MHz Band, it is useful to look at the footprints established by many of those actors in two previous Lower 700 MHz auctions (Auction 44 and 49) and the AWS-1 auction (Auction 66):*
Cellular Market Areas (CMA) Map for Auction 44
Economic Area Groupings (EAG) Map for Auction 44
Cellular Market Areas (CMA) Map for Auction 49
Economic Area Groupings (EAG) Map for Auction 49
Cellular Market Areas (CMA) Map for Auction 66
Economic Areas (EA) Map for Auction 66
Regional Economic Area Groupings (REAG) Map for Auction 66

The Classic Pattern

The classic pattern for an RTC and for most CLECs and WISPs in these auctions is that of Agri-Valley: expansion through Auctions 44, 49 and 66 to attempt to match its landline footprint with CMA acquisitions. In Auction 66 Agri-Valley went for and failed to obtained consolidation in Flint, Lansing, Saginaw, Muskegon, Gogebic, Alger, Cheboygan, Roscommon, and Cass. Expect Agri-Valley to continue this pattern in Auction 73. The same holds true for Whidbey Telephone in Maine in Auction 49, for Hemingford Cooperative in Nebraska, Wyoming, and Colorado in Auction 66, Bluegrass Cellular In western Kentucky in Auction 44, Union Telephone in Wyoming and Colorado in Auction 44 and 66, East Kentucky Network in eastern Kentucky in Auction 44, Fidelity Communications in Missouri in Auction 66, KTC AWS in South Dakota in Auction 66, Public Service Wireless in Georgia in Auction 66, Redwood County Telephone Company in Wisconsin and Minnesota in Auction 44, 44, LL License Holdings in Iowa, North Dakota, South Dakota, Minnesota, and Nebraska in Auction 66, Grand River Communications in Iowa in Auction 44, and Iowa Telecommunications in Iowa in Auction 66. This pattern will continue to hold in Auction 73, and will hold for the vast majority of new entrants in Auction 73: their action will be in the CMAs and to a much lesser extent in the EAs.

More below... [Read More!]

Frontline Wireless LLC, which submitted an incomplete application to participate in the FCC's Auction 73 for the 700 MHz band as Licenseco LLC and which was expected to be a major competitor for the D Block nationwide commercial-public safety broadband license, has folded and is “closed for business.”

Industry rumours suggest that Frontline's bidding entity, Licenseco LLC, failed to make a required upfront payment deadline on January 4.

Speculation focuses on several possible explanatory scenarios. Frontline has changed its business plan several times and, frankly, I was never completely convinced that it would bid when push came to shove. Verizon's belated embrace of open attachment rules — the Carterfone condition which the FCC has imposed on Auction 73 — gave many of Frontline's Silicon Valley backers what they wanted without having to hazard the auction or undertake the encumberance of deployment requirements if they prevailed at auction. The possibility that Google might bid the reserve price on C Block to force Verizon and AT&T to concentrate on battling it out for the C Block REAGs while Google seriously bid on the less expensive D Block to acquire a nationwide third broadband pipe and implement its nondiscriminatory, wholesale open access business model may have had something to do with Frontline's decision to pull out. The possibility that AT&T may have been interested in D Block for national backhaul could have presaged a serious challenge has also been mooted as a factor in Frontline's decision.

It's likely that some of Frontline's backers and associates — Fortress Investment Group's Backline bidding entity and Cellular South in particular — will remain in the auction, but Frontline's demise creates extremely interesting possibilities for D Block competition in the auction.

Part III of the 700 MHz series, Bidding Strategies of the Major Actors, coming soon...