Winner’s Curse: A phenomenon that may occur in common value auctions, where the winner will tend to overpay due to incomplete information.
I was planning on writing a short blurb about Rogers and winner’s curse/buyer’s remorse when Industry Canada initially announced the AWS-3 auction. I got into a Twitter conversation about AWS-3, with the thinking that the high cost of Rogers’ 700MHz spectrum would cause them to be uncompetitive against a TELUS-Bell effort to gain the non-set aside block. JF noted that Rogers would still be bidding to drive up their competitors’ costs. It’s a strategic move I agree will be likely, but felt just further reinforced my original observation.
Rogers wouldn’t hesitate to add to their industry-leading spectrum holdings, if it could be acquired at a good price. But with the significant capital outlay for their 700MHz spectrum — some might say, overpaying — Rogers would need to be extra wary of the risks of inflating auction prices beyond what value could be reasonably extracted, not wanting to ‘accidentally’ win over-priced spectrum. Europe’s experience with the winner’s curse surrounding 3G licence costs is a major contributing factor to lagging in LTE investment, something that wireless executives haven’t addressed at all during the recent CRTC wireless wholesale roaming proceedings.
I also noted that this situation would, in fact, be the worst of both worlds. Rogers would be increasing TELUS’ or Bell’s (and, thus, consumer) costs, while likely still not maximizing government revenues.
Anyway, this was an interesting but minor observation regarding Canadian spectrum policy. Which also probably explains the lack of urgency in writing about it. The release of Apple’s iPhone 6 raised the issue of winner’s curse and buyer’s remorse back to top-of-mind.
To give a little background/context, the FCC completed their Digital Dividend auction in 2008, and Apple has been supporting LTE in the 700MHz band since the iPhone 5. When Industry Canada completed Canada’s Digital Dividend auction in 2014, many Canadian consumers already owned devices with the necessary radios to take advantage of the new spectrum as soon as the bands were supported by carriers. With licences being officially awarded April 2nd, 2014 and Bell going live on April 3rd (just in Hamilton, but still, only one day later!) and Rogers deploying the spectrum April 17th (in parts of Vancouver, Toronto, and Calgary), the quick deployment times appear to have validated IC’s decision to use a modified-US band plan over the APT band plan, at least in the near term. On the surface, everything looked great with 700MHz for government and carriers.
But the iPhone 6 announcement in September 2014 adds a serious asterisk to Industry Canada’s band plan choice and Rogers bidding strategy, as the device is still supporting Band 17 and not Band 12 for LTE. To be clear, the iPhone supports a large number of LTE bands, depending on your model and country/carrier. In Canada, it supports AWS spectrum used by regional and national carriers with LTE networks, in addition to most blocks paired in the 700MHz band. Scanning through the list of supported Canadian 700MHz bands, you see Band 13 & 17. (Actually, the iPhone 6 also supports Band 29! This is going to make my analysis even worse for Rogers.)
Initially, the lower A, B & C blocks were meant to be interoperable, classified as Band 12. AT&T — for anti-competitive and/or technological factors, depending on how critical one is — worked with international wireless standards organizations and manufacturers to develop Band 17 as an alternative. This band only supports the B & C blocks that AT&T won in the US 700MHz auction. (Band 13 is the ‘Verizon’ Upper C block, or C1 & C2 in Canada.)
The US A block 700MHz spectrum was initially purchased by smaller, regional carriers, who on their own, and cumulatively, didn’t have enough market size to get handset manufacturers to support Band 12. Note: Verizon also acquired a number of A block licences but were focused on their Band 13 deployment, which was wider (10x10MHz vs 5x5MHz channels) and didn’t ‘risk’ the potential interference issues. The result was the A block spectrum wasn’t being used and appeared potentially undeployable.
Under increasing pressure, AT&T formally announced on September 10, 2013 that they were “committed to working collaboratively with its chipset partners and OEMs to introduce, within a reasonable time frame, new Band 12 capable devices into its device portfolio.” [Emphasis added.] This was good news for US wireless consumers and the small, regional carriers that owned A block licences. I also mused that AT&T’s move may have increased the perceived value of pending 700MHz auction in Canada, and suggested it altered Rogers’ strategy. Certainly, Rogers paid a huge premium to acquire the contiguous Lower A & B blocks. (See here and here for some initial auction analysis.)
The iPhone 5s/5c was released on September 20, 2013. Clearly, it wasn’t going to support Band 12. But the iPhone 6 was released September 19, 2014, and still no support. It’s been over a year since AT&T stated they would support Band 12 and I haven’t been able to find any device supporting it. Since T-Mobile acquired a number of A block licences from Verizon, there seems to be a continuing string of rumours that devices will be released to support it. But even the upcoming Galaxy Note 4 isn’t likely to do so, according to Mobile Syrup‘s Daniel Bader.
Since the 700MHz was meant to be interoperable from the beginning, surely Qualcomm had a roadmap for supporting the A block and Band 12? And I get that radio and network engineering is very, very complicated. But a year (plus) seems more than reasonable to have devices that support a block that was initially awarded (in the US) in 2008.
In the initial aftermath of Industry Canada’s 700MHz auction, CARTT.ca‘s Greg O’Brien argued (Rogers’ point) that — counter to most of the arguments of financial analysts quoted in his article — the acquisition of contiguous A & B blocks meant Rogers didn’t overpay. I haven’t found any analysis out there noting the release of the iPhone 6 makes Rogers appear to have really overpaid, at least in the short term. But until any device, let alone a market leader like Apple’s iPhone or a popular Samsung or Motorola model, supports Band 12, Rogers may just be cursing their ‘win’ of Canada’s 700MHz auction.
The fact the iPhone supports Band 29, unpaired spectrum that US-carriers are still just testing/experimenting with for LTE-broadcast, and not paired spectrum that covers ~60m subs North American wide, seems quite remarkable to me. Alternatively, the unpaired spectrum can be used for carrier aggregation (combining multiple frequencies for increased bandwidth) with LTE-Advanced network upgrades. As Bell & TELUS upgrade their networks, those with iPhone 6s will already have devices capable of accessing the advanced service.
I think it likely that Band 12 will eventually be supported by devices, especially as T-Mobile moves forward with their network upgrades. But as smartphones become ever more commoditized products, it’ll take longer for consumers to actually have devices that can access the A block that Rogers paid such a premium for, increasing the impact of the winner’s curse in the 700MHz auction.