Five key reasons for high spectrum prices in the US

With US C-band spectrum prices remaining high, especially when compared to international benchmarks, the question remains which key reasons drive this seeming inexhaustible appetite for C-band spectrum?

First and foremost, it is clear that C-band spectrum offers excellent capacity at good coverage – with an unmatched bandwidth of spectrum available compared to lower bands (480MHz already in use in the US, as indicated above) and the ability to deploy new antenna solutions such as massive MIMO, there is a jump in capacity compared to other key mid-band spectrum assets like PCS and AWS. As 5G networks are all about capacity, it is clear that C-band is essential for offering real 5G services. Network deployments to date also show that C-band can be integrated into existing macro networks and achieves good coverage.

It is that much better coverage that appears to set it apart from an alternative 5G capacity band that was always touted as a key 5G pioneer band – namely mmWave spectrum. For years, Verizon and others have told the industry that mmWave would be the capacity component of 5G. However, many markets outside of North America have ignored mmWave for mobile, waiting on brave US (and some Asian) pioneers to demonstrate its effectiveness in a mobile network. With Verizon spending big on C-band in Auction 107, it appears that mmWave was, perhaps, just one possible insurance policy until the C-band came along.

Furthermore, there are three key factors that contribute to the value of C-band spectrum in the US being is so much higher than most other countries in the world:

• High-value subscribers: The US has one of the highest ARPU levels in the world. High revenue per subscribers would inevitably increase the benefit of acquiring additional spectrum and offering real 5G to customers (to retain existing customers and possibly attract new ones).
• Costly infrastructure: US mobile operators utilise taller towers (and fewer towers per capita), with higher costs per site than either Europe or Asia. Although small cells are progressing, they have moved more slowly than many hoped – hence the old formula of adding more spectrum to the existing towers remains the route to increased capacity. In Europe, for example, the trend is to use more (but shorter) towers and in some cases to share the radio access network with competitors. In short, it is cheaper to build your way out of a capacity crunch than it is in the US, where mid-band spectrum really is the only answer for now.
• Indefinite licences: The US offers spectrum licenses that can be renewed indefinitely – it is not necessary to spend more to regain your spectrum at an auction 15/20 years down the road like in most other countries.

The final component contributing to high US spectrum prices are the competitive auction processes. In many European countries, the supply of spectrum has historically been more plentiful, with auctions being designed such that operators can acquire relatively large quantities of spectrum at low prices. Spectrum caps often reduce the level of competition between operators further. In addition, the prevalence of spectrum trading in the US (compared to other parts of the world) invites a greater number of speculative bidders to auction, adding to the number of bidders and thereby further intensifying competition and leading to prices being driven higher than they might have otherwise. In the US, more than in the rest of the world, spectrum is seen as a potential investment asset.

It is thus this perfect storm of a prime 5G spectrum asset in a country with a number of factors that lend themselves to high valuations in a competitive auction environment that have contributed to continuously high spectrum prices in the C-Band.

Ultimately, the 5G transition has required substantial spectrum expenditure. For all the innovation that is happening in small cells, spectrum sharing and mobile mmWave, the mobile networks have had little choice but to spend more heavily than ever to secure mid-band spectrum to build the capacity they need for 5G.

And finally, some speculation…

As we have outlined above, the winning bidders will not be known before the conclusion of the assignment phase in January. However, that has not stopped us from considering who we expect to have spent significant money in Auction 110.

Considering the motivation for each bidder, we believe that AT&T has the highest incentive to acquire spectrum and will almost certainly have been willing to spend big in order to reduce its spectrum deficit in comparison to Verizon and T-Mobile (the current spectrum holdings of each operator are shown in Figure 1).

The situation is less clear for T-Mobile – it may have decided that it needs some C-band spectrum to add to its swathe of 2.5GHz assets.

Verizon is likely to have had a lower valuation for the Auction 110 spectrum due to the large amount of C-band spectrum that it acquired at Auction 107 and so may not have contributed significantly to the auction proceeds.

Finally, Dish is notoriously difficult to read, but it raised USD4 billion of financing during the auction, so presumably it might also have felt the need to strengthen its position in the mainstream 5G band.

Figure 1: Existing mid-band spectrum by bidder (weighted average nationwide MHz)

A quick summary of Auction 110

The primary phase of the US Auction 110 for 100MHz of spectrum in the range 3.45-3.55GHz has concluded. This auction has already raised USD21.9 billion (USD0.71/MHz/pop), adding to the USD95.2^[1] billion and the USD4.6 billion spent in the previous C-band auctions – Auction 107 (3.70-3.98GHz) and Auction 105 (CBRS – i.e. 3.55-3.65MHz) – respectively.

The FCC had set a high reserve price of USD14.8 billion for the auction. This represents 110% of the significant cost of freeing up the spectrum from previous incumbents.

The auction used a standard clock format for the primary phase. The clock phase started on 5 October 2021 and ultimately lasted 151 rounds across 42 days, making it one of the longest FCC auctions. As shown in Figure 2 however, most of the price increases happened in the first few 50 rounds. Once the largest PEAs cleared, the auction carried on for another 100 rounds, with only a limited further increase in prices (the dynamics were thus similar to Auction 107 – on which we reported extensively).

Winning bidders will have the opportunity to bid for specific positions within the band during the assignment stage that is expected to start in January. The identity of the winning bidders will only be revealed after the conclusion of the assignment stage.

Figure 2: Round-by-round gross proceeds on primary axis, gross proceed USD/MHz/pop on secondary axis

[1]     Includes USD14 billion of additional fees and incentives for satellite operators to clear the band

Putting US C-band auctions in an international context

Across Auctions 105, 107 and 110, the US has sold 480MHz of mid-band spectrum for USD121.7 billion. Of this, Auction 105 only contributed USD4.6 billion at 0.21USD/MHz/pop due to the cooperative sharing requirements of the Citizens Broadband Radio Service (CBRS). The bulk of the expenditure came from Auction 107, selling 280MHz of spectrum at 1.29USD/MHz/pop for USD95.2 billion. Auction 110 has come in between these two in price, at 0.71USD/MHz/pop for 100MHz of spectrum.

These prices (even for the CBRS spectrum) are high when compared to international C-Band spectrum prices, as shown in Figure 3. With 280MHz of mid-band spectrum already sold in previous auctions, a 40MHz spectrum cap and some cooperative sharing requirements, it might have been expected that the prices at Auction 110 would be well below the Auction 107 price. However, the price achieved is still at a respectable 65% of the Auction 107 price and higher than almost all other global mid-band auctions (the exceptions being Canada’s 2021 C-band auction and a 2020 auction for C-band in Taiwan).

Figure 3: Spectrum prices in USD/MHz/pop for recent global 3.5GHz auctions