Evaluating the FCC’s $10 Billion Gamble: Successfully Accelerating Access to Spectrum in Auction 107

ABSTRACT

This research analyzed how much bidders in the record-breaking C-Band spectrum (3.7–4.2 GHz) auction were willing to pay for earlier access to frequency rights and the policy implications of the incentive system employed by the Federal Communications Commission to clear the band on an accelerated timeline. The analysis found that bidders paid 20.7 percent more on average for licenses available two years earlier with no subsequent legal challenges. Even though it did not follow traditional price discovery, Auction 107 laid the groundwork for accelerating the transition of wide swaths of the spectrum.

On September 28, 2018, under the direction of President Trump, Former Federal Communications Commission (FCC) Chairman Ajit Pai released the Facilitate America’s Superiority in 5G Technology plan, also known as the 5G FAST plan.¹ A significant component of this plan was to expeditiously increase the amount of spectrum available for 5G technology in the marketplace in order to support the development of modern-day communications networks.² Since radio spectrum is a limited resource, the FCC often reallocates spectrum for new use by relocating incumbent spectrum users. Determining a relocation timeline involves balancing the service loss of incumbents and the benefits of rapid reallocation to new users and their customers.

The term 5G, standing for fifth generation, refers to a technology standard that supports a wide range of frequencies.³ The set of technical standards that define 5G was created by the Third Generation Partnership Project (3GPP), a global partnership between several standardization organizations.⁴ Unlike its predecessors, the previous four generations of wireless communication standards, 5G promises to unify different connection standards across one high speed, ultra-low latency, and high-density network.⁵ In the frequency allotted for 5G utilization, the standard supports deploying a heterogeneous network (HetNet), a network that uses algorithms to share bandwidth across different connection technologies, even across multiple generations.⁶ In this way, 5G technology allows for higher data rates and reduced latency compared to 4G.⁷ The flexibility of creating an efficient mesh of different connections and technologies is what makes 5G revolutionary.

The flexibility and speed of 5G coverage have the potential to empower the development of modern communication networks across America. In densely packed cities such as New York, fast and flexible short-range coverage is desirable, so 5G can be deployed efficiently and profitably by placing numerous high-band antennas in close proximity.⁸ The flexibility of 5G coverage also has great revolutionary potential in sparsely populated regions, but a different deployment strategy must be used, as densely placing short-range antennas would be inefficient and prohibitively expensive. Instead, using low-band 5G is a better solution for coverage in large sparsely populated regions. Low-band spectrum has a longer range, hundreds of square miles of coverage versus the less than one-mile range of high-band spectrum, which comes at the expense of data rates that are barely faster than 4G. Between these two extremes falls mid-band spectrum, which consists of spectrum in the 2.5–3.7 GHz range.⁹ Mid-band can offer data rates from 100 to 900 Mbps within a few miles of the towers, thus striking a balance between high data rates and long ranges.¹⁰

Following the 5G FAST plan, the FCC planned Auction 107 to offer the C-band, a range of mid-band spectrum with high data rate and longer-range terrestrial properties.¹¹ Prior to 2018, C-band spectrum was primarily occupied by Fixed Satellite Service (FSS) operators. In response to the FCC’s initiative to reallocate C-band spectrum for 5G use, four major providers of C-band satellite services (Intelsat, SES, Eutelsat, and Telesat) formed the C-band Alliance (CBA). The CBA proposed a plan to make the lower 300 MHz of the C-band (3.7–4.0 GHz) available for 5G use through a private secondary market auction with minimal FCC oversight. Citing concerns for transparency and speed in the management of a public resource,¹² the FCC countered with its own plan for a public secondary market auction brokered by the FCC and an accelerated spectrum clearing timeline.

To persuade the incumbent operators to follow the accelerated timeline and avoid bureaucratic hoops such as litigation,¹³ the FCC’s plan offered full reimbursement of incumbent operators’ relocation costs and incentive payments to reward early relocation, all paid for by the winning bidders of Auction 107 (C-Band).¹⁴ By FCC estimates, the decision to resell already-issued licenses shaved two to three years off of the time the transition would have taken without the CBA’s cooperation.¹⁵ Table 1 provides an overview of the dates and corresponding Auction 107 (C-band) milestones.

As seen in Table 1, Auction 107 (C-band) began on December 8, 2020, and lasted through February 17, 2021, with the official results announced on February 24. This auction marked the fifth auction of 5G spectrum in the last two years and grossed more than $81.2 billion in bids, making it the highest-grossing FCC Auction to date.^16,17 This figure increases to $96.1 billion when considering the additional $9.7 billion in accelerated relocation payments and $3.3–$5.2 billion in relocation costs. The spectrum offered in the auction would be cleared for 5G mobile service either one year or three years after the auction’s beginning.

The analysis presented in this work aims to determine how the promise of expedited spectrum availability impacted the monetary value evaluated in the form of closing bids that mobile service providers were willing to pay in Auction 107. Sections “Band Plan,” “Timeline,” and “Auction Mechanics” cover auction terminology, timeline, and mechanics. Section “FCC Offered Incumbents $9.7 Billion in Accelerated Relocation Payments” introduces how Auction 107 was groundbreaking, especially with its use of fixed sum accelerated relocation payments, and section “Auction 107 Results and Comparison to Other 5G Auctions” provides a brief overview of the Auction’s overall results. Section “Demand Side Price Evaluation” contains an analysis of the relationship between Partial Economic Areas (PEA) population and final bid prices, and the resulting percent difference (ratio between price difference and price average) of Category A compared to Category BC licenses. This resulting percent difference was performed on a per PEA basis to minimize any other contributing factors, highlighting the effect due to accelerated access. Lastly, Sections “Supply Side Price Evaluation” and “Conclusion” provide a closing overview of the achievements and struggles of Auction 107.

Band Plan

In Auction 107 (C-band Auction), the FCC sold 5,684 renewable, flexible‐use, fifteen-year licenses for spectrum in the lower 280 MHz of the C-band (3.7–3.98 GHz) across 406 PEAs in the continental United States.¹⁸ As illustrated in Figure 1, the PEAs are 416 areas encompassing all of the United States and Territories as defined by the U.S. Department of Commerce and the FCC in 2014, with 2010 population density overlaid in blue.^19,20

In Auction 107, only PEAs in the contiguous United States were auctioned, thereby excluding ten PEAs: Honolulu, Anchorage, Kodiak, Fairbanks, Juneau, Puerto Rico, Guam-Northern Mariana Islands, U.S. Virgin Islands, American Samoa, and the Gulf of Mexico (PEAs numbers 42, 212, 264, 298, 360, 412–416). PEAs are numbered in order of decreasing population, with PEA1 being New York, NY, PEA2 being Los Angeles, CA, and PEA3 being Chicago, IL, and the top fifty PEAs holding 63 percent of the population of the United States.²¹ For each PEA, the FCC divided its 280 MHz of mid-band spectrum (3.7–3.98 GHz) into three Blocks (A, B, and C) and further subdivided the blocks into auctioned licenses spanning 20 MHz within each Block.²² Figure 2 shows the division of this spectrum. It is important to note that Block (A, B, C) refers to partitions of the C-band, while Category (A, B, C) refers to different clock phase bidding categories during the auction.

Block A and Block B span 100 MHz, while Block C only spans 80 MHz because the upper 20 MHz (3.98–4.0 GHz) is reserved as a guard band to protect satellite operations using 4.0–4.2 GHz. For any given PEA, there are five licenses in each of Block A and Block B and four licenses in Block C. In total, there were fourteen available licenses per PEA as defined by the 20 MHz slices of the A, B, and C Blocks.

Timeline

The FCC employed a public auction with a two-phase accelerated relocation timeline. This two-phase accelerated relocation design prioritized clearing the most populous PEAs on an accelerated timeline (Phase I), and the remaining PEAs on a later, but still accelerated timeline (Phase II). The staged series of deadlines gave incumbent C-band FSS operators sufficient time to migrate their services to the upper portion of the band at 200 MHz (4.0–4.2 GHz) while maintaining comparable and uninterrupted service. Figure 3 shows the important incumbent dates in blue boxes and the auction timeline in red boxes.

In Phase I, incumbent FSS operators cleared the lower 100 MHz (plus 20 MHz guard band) in forty-six of the first fifty PEAs by 2021.²³ The four excluded PEAs are PEA5 (Baltimore, MD–Washington, DC), PEA11 (Atlanta, GA), PEA20 (Denver, CO), and PEA42 (Honolulu, HI). In Phase II, incumbent FSS operators cleared the remaining PEAs for the entire 300 MHz (3.7–4.0 GHz) by 2023.

All eligible FSS operators elected to follow the accelerated relocation timeline designated by the Phase I and Phase II accelerated relocation deadlines on December 8, 2021, and December 5, 2023, respectively. If an FSS Operator could not transition their services by the Phase I or Phase II deadline, they were subject to an incremental reduction in the accelerated relocation payment amount based on the number of months late. The rate of decline increased every two months, so by one month late, there would be a 5 percent reduction, but by six months, there would be no accelerated relocation payment.²⁴ The final deadline to move out was December 5, 2025, after which incumbent licenses for the C-Band would expire without plans for renewal, preventing any further legal use of this band.

Auction Mechanics

This two-phase accelerated relocation design prioritized clearing the most populous PEAs on an accelerated timeline (Phase I), and the remaining PEAs on a later, but still accelerated timeline (Phase II). The staged series of deadlines gave incumbent C-band FSS operators sufficient time to migrate their services to the upper portion of the band at 200 MHz (4.0–4.2 GHz) while maintaining comparable and uninterrupted service. To transition and efficiently use the narrower band, satellite operators installed new ground stations and satellite filters, replaced noncompliant satellites, and updated their modulation software and technology. As shown in red in Figure 3, the auction timeline for Auction 107 (C-band) was divided into two phases: the Clock Phase and the Assignment Phase.²⁵ During the Clock Phase, auction bidders bid for amounts in generic license block Categories. As summarized in Table 2, license blocks were grouped into three categories: A, BC, and ABC.

Category A represents Block A licenses for forty-six of the first fifty PEAs (PEAs numbers 1–4, 6–10, 12–19, 21–41, and 43–50). All of the licenses in Category A are set to be cleared by the Phase I Accelerated Relocation Deadline (December 8, 2021). The Commission used PEA weights determined from previous auctions to estimate that the forty-six PEAs cleared by the Phase I Deadline account for 77 percent of the total value of clearing 3.7–3.8 GHz.²⁶

Category BC represents Block B and Block C licenses for forty-six of the first fifty PEAs. Category ABC holds the licenses in the remaining 360 PEAs (PEAs numbers 5, 11, 20, 51–211, 213–63, 265–97, 299–359, and 361–411) over Blocks A, B, and C. All licenses in Categories BC and ABC are set to be cleared by the Phase II Accelerated Relocation Deadline (December 5, 2023).²⁷

During the Clock Phase, bidders bid on a number of licenses in each Category A, BC, and ABC, but cannot specify a specific frequency. For example, a bidder may place a bid in the Clock Phase for a license in Category A in PEA1 for up to five licenses of 20 MHz each, but may not specify which frequencies.²⁸ In another example, a bidder may bid in the Clock Phase for a license in Category ABC in PEA51 for up to fourteen licenses but may not specify which block or subblock. Winners of the Clock Phase still receive the number of license blocks they won regardless of whether or not they bid in the Assignment Phase.

In the optional Assignment Phase, winners have the opportunity to bid for frequency-specific licenses. Depending on the number of licenses bidders secured in the Clock Phase, bidders have the option to bid for specific 20 MHz licenses as designated in Figure 2. Bidding in the Assignment Phase is useful for winners that want to deploy services using the same frequencies across multiple markets.

After bidding in the Assignment Phase concludes, the FCC assigns frequency to the remaining licenses so that contiguous spectrum is maximized for each winner. Contiguous spectrum allows carriers to make efficient use of their spectrum. For example, if a carrier wins two licenses in the same PEA for the noncontiguous A2 and A4 subblocks, they need to leave guardbands on both edges of their subblocks, whereas if a carrier wins two licenses in the same PEA for the contiguous A2 and A3 subblocks they only need to leave guardbands on the edges of their spectrum to prevent interference with the carriers holding the A1 sub-block license and the A4 sub-block license, thus halving the total amount of spectrum reserved for guardbands.²⁹

FCC Offered Incumbents $9.7 Billion in Accelerated Relocation Payments

Traditionally, reverse auctions are the standard market-price-based method of identifying a select number of incumbents who will voluntarily clear the occupied spectrum for a given amount of incentive payment. However, in clearing spectrum for Auction 107 (C-Band), a reverse auction was infeasible. A reverse auction assumes that incumbents have exclusive licenses in a certain area. Since the incumbents of the C-Band spectrum were FSS operators, they share the same nonexclusive rights to transmit nationwide across the full 500 MHz.³⁰ To make spectrum available for flexible use, all incumbent FSS operators would need to clear their existing services together from a portion of the band.

Thus, instead of using a reverse auction for price discovery, the FCC used the novel strategy of disbursing $3.3–5.2 billion in relocation payments to ensure all incumbents could migrate out of the C-Band without interrupting or impairing satellite service and, to further incentivize quick migration, allotting up to $9.7 billion in accelerated relocation payments for FSS operators that moved out before the Phase I or Phase II accelerated deadlines.^31,32 This relocation fund was paid in full by the mobile operator bidders that won Auction 107 (C-Band), proportional to their expected increase in profit from having earlier frequency access.³³

The FCC’s decision to offer incentive payments for accelerated relocation is regarded as controversial with numerous commentators challenging the amount or validity of these payments.³⁴ In deciding the monetary amount of the relocation fund, the FCC wanted to use a value that would incentivize incumbent FSS operators to move out faster but could not determine what the minimum value should be. So, they decided to use a relocation fund value that was just shy of the FCC’s estimation of mobile operators’ increase in expected profit due to early access to C-Band spectrum. The FCC determined that using this method was the best way to reduce risk, as the determined value was likely higher than the minimum incentive incumbent FSS operators would require to move out faster while still being low enough that the mobile operators would likely make a profit.

The FCC estimated the mobile operators’ early access profits as the accumulated capital costs over time using a weighted average cost of capital of 8.5 percent. The 8.5 percent estimate was borrowed from the 2014 Connect America Phase II Cost Model Order, which applied the methodology from a 2013 analysis on the rate of return analysis for local exchange carriers to data from price-cap carriers, a different purpose than it was developed for.³⁵ This 2014 Connect America Order calculated a “zone of reasonableness” for the cost of money for price-cap carriers between 7.8 percent and 9.2 percent and took the midpoint to arrive at 8.5 percent. This 2014 analysis was intended to be used for five years ending in 2019.^36,37 The FCC’s analysis for the Auction 107 accelerated relocation payment also assumed that the bidder’s overall capital cost would equal their costs of capital for buying spectrum.³⁸

With this rate, the FCC calculated a top bound of $10.52 billion in economic profits and determined that a $9.7 billion accelerated relocation payment to incumbent FSS operators was “reasonable and [would] serve the public interest.”³⁹ The total award value reflected the CBA’s estimation of the cost of moving their operations to higher frequencies, while still maintaining uninterrupted service to their customers.⁴⁰ The FCC divided the $9.7 billion accelerated relocation payment to the top five FSS Operators (Intelsat, SES, Eutelsat, Telesat, and Star One), weighting the size of shares by market-based agreement factor, transponder usage, and coverage, in that order, from highest to lowest priority factor.⁴¹ The market-based agreement factor was based on a CBA report that calculated the market-based value each member would lose if spectrum was converted to flexible use, which is loosely based on 2017 revenue numbers.⁴²

The FCC has two basic models for reimbursing licensees for costs incurred during the process of clearing the spectrum: direct reimbursement and clearinghouse. In the direct reimbursement model, the FCC pays through programs approved by Congress. These programs include the TV Repack program and the Rip and Replace program. For the Rip and Replace program, Congress requires carriers to remove certain foreign-made equipment, primarily Chinese-made Huawei and ZTE items. The estimated time period for this program lasts from December 31, 2021 to December 31, 2022. The TV Repack program is used for the Broadcast Incentive Auction (Auction 1000), which allows TV channels 38–51 to move to new channels.

Since the reimbursement funds are paid by a third party and not the government in Auction 107 (C-Band), the FCC utilized the clearinghouse model with a Relocation Payment Clearinghouse (RPC) to handle all of the financial functions related to this auction. These functions include collecting and distributing payments among parties, reviewing reimbursement claims and managing the reimbursements, overseeing the financial dispute process, and mediating any disputes between parties.⁴³ Using an RPC is advantageous to the FCC because it saves the FCC from liability in processing the many multibillion dollar payments involved in this auction, and by having the stakeholders choose the clearinghouse, it also protects the FCC from claims of bias in handling payments or payment claims. On March 3, 2020, the FCC found that using a clearinghouse to handle financial transactions would serve the public interest, and they created a search committee composed of nine entities that represented the interests of stakeholders.⁴⁴ On July 21, 2020, the search committee unanimously chose CohnReznick to function as the clearinghouse in the C-band auction, and the FCC would later deem that CohnReznick met the set guidelines to do so.⁴⁵

Auction 107 Results and Comparison to Other 5G Auctions

From late 2018 through 2020, the Commission held four higher-band auctions, starting with Auction 101 (28 GHz Band), which awarded 2,965 licenses and raised $703 million in gross bids ($0.011 per MHz-pop).^46,47 At the conclusion of Auction 101, the Commission had already planned two auctions for 2019.⁴⁸ Less than two months later, Auction 102 (24 GHz Band) began, awarding 2,904 licenses and raising $2.02 billion in gross bids ($0.009 per MHz-pop).^49,50 Starting in late 2019, Auction 103 sold 14,142 licenses in the upper 37 GHz, 39 GHz, and 47 GHz bands and grossed $7.57 billion ($0.007 per MHz-pop).^51,52 Finally, in 2020, Auction 105 (3.5 GHz Band) sold 20,625 licenses, grossing $4.59 billion ($0.216 per MHz-pop).^53,54 Four out of five (80 percent) of these 5G auctions grossed over $1 billion, all within two years.⁵⁵ Whereas only ten of the ninety (11.1 percent) spectrum auctions held over twenty-four years between 1994 and 2018 grossed more than $1 billion.⁵⁶ Auction 107 (C-band) grossed a total of $81.2 billion in winning bids ($0.944 per MHz-pop or $1.096 per MHz-pop with relocation costs and accelerated incentives), surpassing the prior record of $44.9 billion to become the highest-grossing spectrum auction to date.^57,58 All 5,684 licenses available for auction were awarded among 21 bidders.⁵⁹ At the end of the Clock Phase, the winning bids accounted for $80.9 billion, more than 99 percent of the final profits.⁶⁰ When considering the results normalized by frequency and population (MHz-pop), Auction 107 had over four times the price for spectrum and users than Auction 105, the second-highest auction. This difference can be mostly explained by Auction 107 offering a larger 280 MHz continuous piece of spectrum, and to a lesser extent, the higher data rates achievable with C-band’s higher frequency ranges.

In Auction 107, the gross license payments for Category A were $23.2 billion, $37.8 billion for Category BC, and $20.0 billion for Category ABC.⁶¹ The gross difference between Category A and BC was about $14.6 billion. Table 3 lists the top five highest-grossing bidders and the bidders who won the largest number of licenses.⁶² The highlighted cells show that NewLevel II, L.P. ranked in the top five for gross winning bids, but not for the number of licenses, and that Canopy Spectrum ranked in the top five for the number of licenses won, but not for gross winning bids.

Cellco Partnership (Verizon Wireless), AT&T, T-Mobile, and U.S. Cellular Corporation ranked first through fourth, respectively, in both total gross winning bids and the number of licenses won. Bids from Cellco Partnership (Verizon Wireless), AT&T, and T-Mobile accounted for over 96 percent of the bidding total. Canopy Spectrum holds fifth place in terms of the number of licenses won, while NewLevel II holds fifth place for total gross winning bids.

Demand Side Price Evaluation

This analysis examines closing bid prices at the end of the Clock Phase, as more than 99 percent of the final profits were accrued during the Clock Phase. All closing bid prices are compared on a per-license basis.

As explained earlier in the Timeline (section “Timeline”), the accelerated relocation timeline states that Category A (Block A licenses in forty-six of the first fifty PEAs) will be cleared by December 8, 2021, and Category BC (Block B and C licenses in the same forty-six of the first fifty PEAs) will be cleared by December 5, 2023. Since Clock Phase bidding occurs on PEA by Category, licenses in Category A effectively have equivalent pairs in Category BC with the most notable difference being that Category A is available two years earlier than Category BC. Since bidders would pay more for commodities that hold higher perceived value, studying how the closing bid prices of Category A compare to those of Category BC gives a measure of how much bidders were willing to pay for early access to cleared spectrum. Confounding differences between the Categories are that A has five licenses per PEA whereas BC has nine, and Category A has slightly lower frequency spectrum blocks compared to BC. Figure 4(a) and 4(b) show the total closing bid prices for the forty-six PEAs in Categories A and BC against (a) PEA and (b) Population. Note that PEAs are ranked in order of increasing population (i.e., highest population PEA is PEA1).

Figure 4(a) shows that closing bid prices exponentially increase as the PEA number (ranked by increasing population) decreases. The relation between closing bid prices and population is even clearer. As shown in Figure 4(b), closing bid prices linearly increase as the population increases. To quantify the relationship of population size on bidders’ estimation of spectrum value, the linear regression fits of closing bid prices in millions of dollars against the 2010 population in millions of people for Category A and Category BC are shown in Figure 5(a) and 5(b), respectively.

For Category A, the estimated base closing price is $11.8 million with a $23.1 million estimated increase for every additional million people in the PEA. For Category BC, the estimated base closing price is $5.2 million with a $22.3 million estimated increase for every additional million people. Both linear fits are strong, with r-squared values of 0.94 and 0.95 for Category A and BC, respectively. The r-squared is a statistical measure of how closely data matches a regression. An r-squared closer to a positive 1 indicates a good fit with a positive rate of increase. Since both linear fits have r-squared values of around 0.94, this indicates that closing bid prices for each PEA region have a strong linear correlation with population. Category A’s base closing price (y-intercept) is more than double that of Category BC, and per every additional million people in the PEA, Category A’s closing bid price rate of increase is $0.78 million higher than that of Category BC.

To visually confirm the consistent effect size between Categories and PEA, Figure 6 plots the clock phase generic license closing bid prices normalized by the population of A and BC Categories across PEAs.

Figure 6 reveals that Category A (solid orange line) tends to cost more than Category BC (dotted blue line) per license. Category A and Category BC closing bid prices follow a similar pattern across PEAs, even after being normalized by population.

Figure 7 shows the price difference in clock phase closing bid price between generic licenses in Categories A and BC for each PEA, not normalized by population. Positive differences mean that Category A closing bid prices are greater than Category BC closing bid prices.

Category A licenses cost more than Category BC licenses in 89.1 percent of PEAs. Only five of forty-six PEAs (with only three of these having a large enough difference to be discernible in Figure 6) have higher Category BC closing bid prices. These five PEAs are PEA1 (New York, NY), PEA₆ (Philadelphia, PA), PEA21 (Tampa, FL), PEA39 (Oklahoma City, OK), and PEA47 (Brownsville, TX).

Even though early access did not have a large effect on variance across PEAs, due to the sheer magnitude of money handled in the auction, the consistently higher closing bid prices for Category A licenses are notable. The total monetary difference in closing bid prices between the two Categories is $444.3 million. The mean price difference between A and BC licenses is $9.7 million, while the median difference is $7.8 million, indicating a positively skewed distribution as reflected in Figure 7.

The percent difference in closing bid prices between Category A and Category BC across PEAs is shown in Figure 8(a), with the corresponding distribution shown in Figure 8(b).

Figure 8(a) shows a general trend where the variation in the percent difference of closing bid prices increases as the PEA number increases and the population decreases. When looking at the price difference on a per PEA basis, bidders paid 15.5 percent more for Category A licenses than Category BC licenses within the same PEA. As shown in Figure 8(b), the Category A versus Category BC percent difference distribution is positively skewed, with the median percent difference being 10.6 percent. When not including the accelerated relocation payment or reimbursement fees (i.e., not fully loaded), industry expert Sasha Javid measured similar results with Category A yielding $1.312893 per MHz-pop, Category BC $1.186914 per MHz-pop, and Category ABC $0.551728 per MHz-pop.⁶³ This results in a 10.6 percent premium for Category A licenses per MHz-pop when compared to Category BC. When taking into account the accelerated relocation payment and reimbursement fees, the fully loaded prices for Category A are $1.619281 per MHz-pop, Category BC are $1.342050 per MHz-pop, and Category ABC are $0.624122 per MHz-pop.⁶⁴ Therefore, the full-loaded price premium paid for Category A licenses when compared to Category BC is 20.7 percent.

Supply Side Price Evaluation

In addition to demand-side factors from bidders that determined final license pricing, the FCC determined the floor price of bids and the accelerated relocation payment bidders had to pay incumbents. The minimum bid price used a tiered approach with a higher floor price according to the population density of the PEAs: PEAs 1–50 had a starting bid of $0.03 per MHz-pop, PEAs 51–100 with $0.006 per MHz-pop, and all other PEAs had a minimum bid of $0.003 per MHz-pop. There was also an overall bottom price set of $1,000.⁶⁵ The FCC reasoned that a minimum bid price would help accelerate the auction toward finding a final price, while the tiered approach reflects the general levels of demand expected for each PEA group.⁶⁶

The accelerated relocation payment was far more controversial due to the contention around the final sum. Intelsat believed that the sum was too low, and smaller satellite operators maintained that accelerated relocation payments shouldn’t exist.⁶⁷ The FCC noted that incumbent satellite operators didn’t have rights to sell their spectrum licenses, yet the FCC still committed to “make them [satellite operators] whole during and after that transition”.⁶⁸ Interestingly, the FCC did not calculate the relocation costs based on the minimum price satellite operators would accept. The FCC reasoned that “eligible space station operators have had every incentive not to disclose precisely how high an accelerated relocation payment must be for them to accept it”.⁶⁹ Instead, the FCC decided to base the value of the accelerated payment on the price “overlay licensees would willingly pay for an earlier transition”; thus, the FCC decided to have cellular providers pay a fraction of their expected profits based on earlier spectrum rights. The predicted profits were based on an estimated $0.50 per MHz-pop in future profits, which worked out to $10.52 billion in total.⁷⁰ The FCC then lowered this amount to $9.7 billion to leave a proper incentive for cellular carriers to buy accelerated licenses.⁷¹ No matter the outcome of the bidding phases, this relocation payment was fixed, which the bidders paid proportionally to their winning share of the total accelerated MHz-pop offered.⁷²

The accelerated payment aligned all of the parties’ interests: the burden of payment is off the FCC, cellular bidders pay a premium lower than their expected revenue for earlier access to spectrum, and the space station operators received a $9.7 billion incentive to clear the spectrum and to follow the accelerated timeline. On the other hand, this $9.7 billion value was not a premium decided by a traditional auction and free market forces, but instead, a mandated fee cellular operators had to pay to access frequency in these accelerated blocks. In the words of the FCC, “Of course, the Commission might have chosen a number lower than $9.7 billion, to gamble that space station operators might accept a lower price. But the smaller the payment the greater the risk that such a payment will be insufficient to incent earlier clearing.”⁷³ This potentially meant sacrificing proceeds that could have either gone to the treasury or to cellular carriers to develop costly physical infrastructure. Altogether, as stated by the FCC, “there is no precise science that allows us to arrive at the ‘right’ accelerated relocation payment total”.⁷⁴ As this auction was the first of its kind, there is no precedent to point to an alternative sum that accomplishes the FCC’s goals without a potential legal and political quagmire.

Interestingly, Verizon took action after the auction had closed to further accelerate its access to its Phase II A-block licenses, originally set to be cleared by December 5, 2023.⁷⁵ Verizon had already focused heavily on A-block licenses in a bid to be the first to offer C-band coverage nationwide, so this step was complementary to those efforts.⁷⁶ Due to Verizon being a private entity, it is not known how much they paid satellite operators, but they received access to 60–100 MHz of A-block spectrum in at least thirty major markets.⁷⁷ This occurred in March 2022, twenty-one months ahead of schedule.

Conclusion

Based on the analysis of the closing bid price of Auction 107 (C-band), this article finds that FCC spectrum license auction participants paid more for licenses available earlier, supporting the FCC’s decision to award accelerated payments to incumbent users to facilitate this transition. As of February 2023, all satellite operators had cleared the accelerated Category A licenses (3.7–3.8 GHz) in forty-six of the top fifty PEAs by the December 5, 2021, deadline.⁷⁸ This demonstrates that the $9.7 billion sufficiently incentivized satellite operators. Furthermore, Verizon paid an additional sum on top of the accelerated relocation payment, which demonstrated demand for earlier access rather than the accelerated relocation payment dissuading bidders from the most valuable PEAs. Overall, bidders paying 20.7 percent more for spectrum available two years earlier suggests that repeating accelerated relocation payments in future auctions could increase value for bidders and sellers and expedite the efficient use of spectrum.

In Auction 107, the FCC prioritized a rapid transition. Without this priority, the FCC could have shifted satellite operators while only paying for relocation costs or could have waited for incumbent’s fifteen-year licenses to expire. Instead, the FCC deemed speed worth the $9.7 billion in accelerated relocation payments. More importantly, choosing a fixed sum provided certainty to all parties involved. Any alternative options, such as those that pay the satellite operators a percentage of auction proceeds, faced the possibility that satellite operators may be less inclined to agree to a variable payout, especially if the auction had underperformed. Overall, the absence of any legal challenges to Auction 107 and the accelerated transition of frequency can be considered a resounding success.

A side question explored was whether the FCC offered accelerated reimbursement payments to all incumbent satellite operators. The FCC made these payments available to five incumbent satellite operators: Intelsat, SES, Eutelsat, Telesat, and Star One. As a result, it may be concluded that other satellite operators were left out, only benefiting the largest incumbent players. Figure 9 represents the number of unique frequency requests for the 3.7–4.2 GHz C-band granted by the FCC as well as the proportion of the requests that belong to each company.⁷⁹ Requests for this figure were filtered for geostationary satellites at longitudes that could provide service to the United States. Eutelsat is represented by Hughes Network System LLC, which designs, builds, and files for satellites on behalf of Eutelsat; hence, why Hughes Network System LLC appears instead of Eutelsat.⁸⁰

Figure 9 suggests that three corporations, Hispamar Satellites S.A., Panamsat Licensee Corp., and Satelites Mexicanos S.A. DE C.V., did not receive accelerated payments despite having licenses to portions of the C-band sold within this auction. Upon further investigation, all three of these operators’ C-band satellites provide services outside of the United States, mostly in South America. Therefore, all FSS satellite operators in the 3.7–4.2 GHz frequency range that operated in the United States qualified to receive accelerated relocation payments.

ACKNOWLEDGMENTS

The authors would like to acknowledge the support of the Amateur Radio Digital Communications (ARDC) Grant, the National Science Foundation (NSF) SII-Center: SpectrumX—An NSF Spectrum Innovation Center Grant (Federal Grant Number 2132700), the Massachusetts Space Grant, and the Henry Luce Foundation through the Olin College of Engineering’s Clare Boothe Luce Fellowship, an initiative to provide women-identifying researchers funding and support, which funded this work. Additionally, the authors would like to acknowledge Olin College of Engineering Professor Zach del Rosario for inspiring the data science conducted in this work and Olin College of Engineering alumni and Olin Satellite + Spectrum Technology & Policy (OSSTP) member Argyris Kriezis.

FOOTNOTES

BIBLIOGRAPHY