Predicting Upcoming Spectrum Auction Results…

(this post inspired by a series of tweets by Tim Farrar of TMF Associates)

The biggest proclamation we heard at the height of the latest spectrum frenzy is that we need to free up 500MHz of spectrum for mobile communication between now and 2020, with 300MHz of that coming by 2015. President Obama repeated it, the FCC Chairman at the time Julius Genachowski repeated it. But there is a flip side to all this spectrum becoming available.

Name Frequencies Total MHz 5MHz FDD Pair Equiv.
600MHz Incentive 614-698MHz (TV channels 38-51) Max 120MHz, Likely 70MHz 8*
AWS-3 (paired) 1755-1780/2155-2180 (AWS adjacent 1700/2100) 50MHz 5
AWS-2** (unpaired) 1695-1710 (AWS adjacent 1700/2100) 15MHz 1.5
AWS-2 1915-1920/1995-2000 (PCS adjacent) 10MHz 1
Total 145MHz 14.5

* includes the 6 MHz Lower 700MHz “A” block because the block will become usable when TV stations are relocated, this effectively adds more spectrum to the usable inventory than will be auctioned.

** its possible for this spectrum to become paired if the FCC forces broadcasters to give up some spectrum they use currently (predominately backhaul), this would add another 15MHz to the totals. But if I were the broadcasters, I’d be sick of the FCC’s bullshit by now.

Congress has dollar signs in their eyes. They see spectrum auctions as a way to raise dollars for the federal treasury. But their desire to raise funds ultimately conflicts with the FCC’s and the administration’s desire to free up a lot of spectrum because of basic supply and demand – there will be so much spectrum for sale in the next three years, its likely that the spectrum prices will plummet. The price of 69c/MHz-POP will fall, since its not likely the big four will spend over $30 billion on spectrum purchases in the next three years. So what will they spend?

600MHz Incentive

T-Mobile and Sprint desperately need low-frequency spectrum for their networks (for coverage, not capacity), AT&T could use a little more spectrum below 1GHz, especially if it can get it on a nationwide basis. Verizon is good on spectrum (or so they say), and would really only be playing a game of keep-away from the other carriers if it were to aggressively bid during the auctions. Dish network is also a contender here, acquiring one or two 5MHz blocks for rural and in-building coverage. I expect this to be the most contentious auction if there are no hard ownership caps, and moderately contentious if AT&T and Verizon are limited in the amount of spectrum they can purchase.

AWS Band Extensions

The AWS band extensions are valuable because three of the four carriers (plus a host of smaller regional carriers) already have spectrum here. We’re also likely to see another round of musical chairs here after the auction is closed and the big three look to swap spectrum to create larger contiguous spectrum segments. It’s entirely likely that after the upper band extension is auctioned, we could see AT&T, T-Mobile and Verizon all swapping spectrum to squeeze out as many 20MHz FDD LTE channels as they can. Sprint is likely to stay out of the game, as with Dish. Smaller carriers might be able to get smaller bits in mid-tier markets, but not much in the top 25 markets where I expect competition to be fierce. AT&T will likely have the most to spend in this band to catch up given its small amount of AWS holdings (even if the Leap acquisition goes through, it may be 3-4 years before that spectrum is available for 100% LTE).

PCS Band Extension

The 10MHz PCS band extension is likely to be Sprint’s, unless Charlie Ergen wants to play spoiler. Which he might, just for the hell of it. But I think there is less pressure now that it seems like he is going to get LightSquared’s spectrum through bankruptcy – 40MHz of LS spectrum plus the 40MHz he owns would definitely be enough to start his own carrier, along with some 600MHz spectrum he would need to pick up in the incentive auction.

Totals

My carrier spectrum ownership totals post-auction estimates. Existing ownership figures are national averages from my current spectrum ownership database. Expected spectrum purchases in the expansion bands shown in parenthesis.

Carrier 600Mhz 700MHz Cellular/SMR PCS AWS WCS* BRS/EBS* Total
Verizon 10 22 25 20 40 (10) 0 0 117
AT&T 10 18
7†
25 40 40 (25) 20 0 160
T-Mobile 20 0 0 25 50 (10) 0 0 95
Sprint 15 0 15 35 (10) 0 0 50 115
Dish** 15 5† 0 0 0 0 0 100

* “Boutique” spectrum (not a large ecosystem of players worldwide who use that same band)
** Not listed: 40MHz AWS-4*, 40MHz L-Band* (Lightsquared)
† Unpaired spectrum, only usable to supplement download speeds

This thought exercise made me realize a few things

  1. I understand why Verizon is fighting against spectrum caps – if things play out this way, they could fall further behind AT&T for spectrum, though they do have the trade off of not being saddled with any boutique spectrum, unlike AT&T and Sprint.
  2. The race to acquire large regional carriers is on – US Cellular, Cincinnati Bell, nTelos, etc. will likely be bought, and the FCC doesn’t seem to care as long as we keep at least four major carriers. AT&T is throwing its cash around to buy up these carriers, Verizon is a little more diligent.
  3. AT&T is amassing a spectrum ownership lead over the other carriers, yet it doesn’t translate to any meaningful improvements in service – coverage or speeds. They’re a little faster than Verizon in terms of LTE, but I attribute that to 1) hardware differences (RRH) and 2) less LTE unlimited subs than Verizon (c.f. the Reddit user who uses 40GB/mo on Verizon).
  4. Sprint has been heavily spectrum constrained. But they’re poised to break out – CDMA Adv. and LTE in the SMR band should allow them to catch up a little bit to the big players. They have the chance to be just as fast (with their tri-band LTE equipment), but I’m still skeptical of their execution. I hope I’m wrong.
  5. T-Mobile should work to get as much AWS spectrum as they can – 20+20 channels would keep them at the top of the speed race. 600MHz will allow them to get the suburban, exurban and rural coverage they desperately need. But can they afford to buy all this spectrum?

So back to the question of how much they will spend. For the most part, the spectrum to be sold is uniform within each band, unlike the 700MHz auction which had different characteristics for each of the different blocks (Verizon’s “C” block had open access requirements, the lower “A” block had interference issues). I’m venturing to say that the prices are as follows…

Band Price ($/MHz POP)
600MHz 0.60
AWS Exp. 0.60
PCS Exp. 0.50
Name Band Spectrum Cost
AT&T 600MHz 10MHz 1.8B
AT&T AWS Exp. 25MHz 4.5B
AT&T Total 35MHz $6.3B
Sprint 600MHz 15MHz 2.7B
Sprint PCS Exp. 10MHz 1.5B
Sprint Total 25MHz $4.2B
T-Mobile 600MHz 20MHz 3.6B
T-Mobile AWS Exp. 10MHz 1.8B
T-Mobile Total 30MHz $5.4B
Verizon 600MHz 10MHz 1.8B
Verizon AWS Exp. 10MHz 1.8B
Verizon Total 20MHz $3.6B
Dish 600MHz 15MHz 2.7B
Dish Total 15MHz $2.7B
ALL OPERATORS Total 130MHz $23.1B

Smaller operators excluded from this table.

So $23B is roughly the amount I expect the spectrum auction to go for. This includes 70MHz of incentive auction spectrum, 50MHz of AWS expansion band, and 10MHz of PCS expansion band. Not included is the 15MHz unpaired AWS expansion band (this could bring in another $4-5B if it were 15MHz paired).

11 thoughts on “Predicting Upcoming Spectrum Auction Results…”

  1. Even if Dish buys LightSquared to get a hold of its spectrum; why should we assume they’ll have any better luck converting it to cellular usage than LS did.

    1. I believe Dish’s goal is to re-band the spectrum such that they can use the LS spectrum for uplink, instead of uplink and downlink. The issue with LS and GPS were these big cell phone towers emitting high-powered signals (100W or so) overwhelming receivers on GPS devices. The uplink power of most cell phones is much smaller (1W or less).

      With Dish’s AWS-4 spectrum and LS spectrum (40MHz each), they can use the AWS-4 spectrum around 2.1GHz for downlink to the handset, and the LS spectrum around 1.6GHz for uplink back to the towers. The uplink strength should be small enough that it doesn’t interfere with GPS units. I don’t remember enough of my electrical engineering classes anymore to sit and do the math on the signal strength though, so I’d have to leave it to people like Tim Farrar or Steven Crowley to work out the numbers. [Side note: I’d love to go back and do an EM class again in college, I care so much more now than I did 10+ years ago]

      This implementation would be beneficial for Sprint and Dish, because the future H-block to be bid on would be able to be used completely if Dish converted all of its AWS-4 to downlink. This means Sprint could run two 5MHz LTE carriers (G and H blocks) and use LTE-Adv to bind them together to equal the throughput of one 10MHz LTE channel. Likewise, Dish could use all of their AWS-4 allocation (full 40MHz instead of about 35MHz).

  2. RE: Sprint On the conference call today they will be putting 2500 on almost every Sprint site and leaving the 2000+ Clearwire TD-LTE sites up. In addition, they will be adding more sites for better contiguous 2500 coverage. Obviously they aren’t striving for contiguous 2500 in any rural area, but you can get an idea of the scope.

    -Verizon buys more spectrum in the 600 + AWS auctions. 20MHz 600 would make sense if they can.
    -Sprint seems about right, although they could end up cutting back on spectrum purchases.
    -T-Mobile less because they need to capitalize on their high capacity. They should probably focus mostly on the 600 auction.
    AT&T will take what they can get. And I totally agree they will jump back in to AWS in a big way through the AWS auction.

    The 600 reverse auction will be a mess.

    1. It looks like Sprint is counting on TDD LTE 2500 to keep them fast in major cities – if people can use it most of the time and leave 800/1900 for in-building coverage it would keep those other bands fast considering its only 5MHz FDD LTE instead of 10MHz for the other three carriers (T-Mo announced they’re going to bring 10MHz FDD to most major markets by year end).

      And yes I agree the 600MHz auction will be a mess. I expect that T-Mobile’s plan will be what the FCC uses (hard ownership caps that limit Verizon to a single 5MHz block and AT&T to one or two blocks depending on their 700MHz B/C block ownership in that geographic area). And then they wont reach the revenue target, and then all hell breaks loose (I expect AT&T to out-bid Verizon to secure 20MHz and then stop).

  3. I think both Sprint and Tmobile should get 20 MHz of 600 MHz each. Sprint and Tmobile both desperately need more low band spectrum especially since Sprint is still somewhat constrained with its 800 MHz spectrum because of the IBEZ borders for Mexico and Canada. Agreements need to be made with the FCC and Mexican/Canadian government so that part of that 800 MHz spectrum can be used for CDMA and LTE. So far only the US/Mexico agreement has come to fruition but it limits Sprint’s use to 5.5 x 5.5 MHz of that 800 MHz spectrum.

    What is the point of having Sprint get 15 MHz and Dish get 15 MHz? You can’t do anything with the 5 MHz anyways and it would be a complete waste of 10 MHz of spectrum. Dish has plenty of S-band spectrum they have yet to deploy and if anything they just need 10 MHz of 600 MHz spectrum. Besides I see the FCC selling the 600 MHz in 5 MHz paired spectrum total of 10 MHz so I doubt the 15 MHz prediction will stand true.

    1. That might not be far from what happens – Verizon isn’t going to have the kind of money to throw around for spectrum auctions or acquisitions now that it has to pay the $49B worth of bondholder interest.

      From what I publish above, Verizon’s totals may be diminished somewhat.

  4. This is a great article. It’s been almost 6 months now since you last posted an article. I’d like to hear what you think has changed since you wrote this. I know the 600Mhz auction has been posponed. And Verizon made a bunch of money this last quarter. Where do you think the players stand and about the auction being delayed?

    1. Other than the PCS expansion apparently going to Dish instead of Sprint, I don’t think anything major has happened that would impact on the prediction/results. The various acquisitions/spectrum swaps have shifted somethings around the edges, but mostly aren’t big enough to change the fundamentals. TMobile’s getting VZWs 700mhz A band spectrum is the only possible exception; however unless they buy up significantly more of the band from small owners and/or negotiate roaming deals with them they’ll still need to buy a nationwide block of 600mhz for suburban/rural coverage.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.