Unlocking UK mobile: lessons from a decade of annual licence fees Executive summary

(The executive summary of the report can also be read via Flint's website which can be found here)

The UK urgently needs to find ways to improve economic growth. As the Government’s Wireless Infrastructure Strategy (WIS) recognises, the most advanced form of 5G (standalone 5G, or 5G SA) could unlock a wide range of productivity-enhancing use cases and applications for businesses and the public sector in the UK. The WIS suggests that the rapid and widespread deployment of 5G SA could add £159bn to the UK’s Gross Value Added by 2035.

However, 5G SA requires new investment on a scale unprecedented in the mobile industry. The GSMA estimates that operators will need to spend more than three times their 2018 annual capex levels1 between 2025 and 2027.

The UK mobile market in its current state will not support this scale of investment. Mobile network operators in the UK have seen a decade of declining revenues and are cautious of additional investment requirements when the potential revenue benefits are uncertain – as recognised by both Ofcom and the Government. Ofcom’s analysis shows that over the last 10 years, mobile retail revenues have declined by 35% in nominal terms, ² as household spend on mobile telecoms has fallen, while mobile data usage has increased more than tenfold. During this period, share prices of the parent companies of the UK MNOs have fallen by around 50% or more, ³ reflecting investors’ nervousness about returns in the sector.

Annual Licence Fees (ALFs) are an important part of the problem. As an additional operating cost, ALFs have reduced returns for MNOs. In 2021, MNOs collectively paid £204m in ALFs, and average ROCE for the industry⁴ was only just above Ofcom’s assumed WACC of 7.8%. In 2022, ALFs of over £292m per year (rising to over £300m in later years) will have reduced returns even more, at the same time as the industry WACC is rising significantly, in line with higher risk-free rates. The higher ALFs – growing with inflation – and higher WACC could lead to average industry ROCE below the WACC in the short and medium term. In the context of an industry that is looking to attract significant new investment, this is far from ideal.

ALFs are not necessary to incentivise efficient spectrum allocation

Following a Direction from the UK Government in 2010, Ofcom has set ALFs at full market value. Thirteen years on from the Direction, we are in a position to review and assess whether that approach has achieved the stated objective of securing the optimal use of spectrum. MNOs have responded to the opportunity cost of spectrum without the signal from ALFs, as seen from successful spectrum trades involving spectrum with no ALFs attached, in both the UK and US. ⁵

‍

-----------------------------------------------------

^{1 For reference, in 2018, UK MNOs spent £2.1bn on tangible fixed asset additions.}

^{2 This is a 44% decline in real terms.}

^{3 Over the last 10 years (June 2013-June 2023), Vodafone Group’s share price has declined by 62%, BT’s has declined by 60%, Liberty Global has declined by 46%, and Telefonica’s has declined by 48%. In contrast, the FTSE 100 has grown by 21%.}

^{4 Based on Ofcom’s preferred Economic ROCE basis, which includes the deemed market value of spectrum assets in the MNOs assumed asset base.}

^{5 In the UK, Qualcomm sold its 1400 MHz spectrum to Vodafone and H3G in in 2015, EE sold its 2.6 GHz spectrum to O2 in 2020, and Vodafone and O2 swapped 40 MHz of 5G mid-band spectrum in 2021. There are also several examples of spectrum trading in the US (where there are no ALFs) including Verizon’s acquisition of 2.1 GHz spectrum from Comcast and Time Warner (2011), and T-Mobile’s acquisitions of 700 MHz spectrum in 2013 and 600 MHz spectrum in 2022.}

-----------------------------------------------------

‍

And there is evidence that ALFs distort the optimal use of spectrum by introducing uncertainty into the trading process.⁶

ALFs set using full market value are unnecessary in theory, and in practice may have hindered rather than promoted efficient spectrum trading. The full market value-based approach has shortcomings and introduces significant regulatory judgement. For example, the current approach risks producing inconsistent outcomes for similar bands over time, and does not reflect that full market values can and do change over time. This is particularly problematic in the context of the current macroeconomic environment, where external shocks have likely reduced the real value of spectrum licenses while increasing the nominal levels of ALFs.⁷ The full market value-based approach is also out of line with the approach Ofcom has taken in other sectors, such as broadcasting, ⁸ as well as pricing for 10, 28 and 32 GHz spectrum.

ALFs have a distortionary effect on investment incentives

There is no rationale for retaining ALFs. Crucially, ALFs distort both the incentives and abilities of MNOs to invest in their networks, by acting as a drag on retained earnings and free cash flow. MNOs and investors perceive ALFs as an unavoidable tax, a cost of doing business in the UK. The direct experience of operators, supported by the academic literature, ⁹ suggests that as a result, telecoms groups are likely to assess investment projects in other jurisdictions to be more profitable than in the UK, risking a slower pace of investment in UK 5G SA networks. For example, MNOs in several countries (Spain,¹⁰ Germany,¹¹ Belgium,¹² the US,¹³ Japan,¹⁴ and South Korea,¹⁵ among others) have already launched 5G standalone networks.

A new approach to ALFs could help meet the WIS 5G investment objectives

This is the right time for UK policymakers to review how ALFs are levied and consider whether reforming them provides an opportunity to support economic growth through 5G investment.

This report presents two alternative approaches to mobile spectrum pricing that would help the government deliver its growth objective. While they would depart from the full market value approach of the 2010 Direction, they would promote investment and remain fully consistent with Ofcom’s duty to secure the optimal use of spectrum:

‍

-----------------------------------------------------

^{6 As seen from the unsuccessful attempt, led by Three, to defragment the 3.4-3.8 GHz band in 2019 ahead of the 3.6 GHz auction.}

^{7 Frontier Economics, 2023, Inflation indexation of annual licence fees for mobile spectrum}

^{8 Where Ofcom decided to only apply a regulatory pricing mechanism to recover its on-going costs in managing the spectrum, despite establishing excess demand for the spectrum.}

^{9 See, for example, Devereux and Griffith (1999), The taxation of discrete investment choices; Lewellen and Lewellen (2016), Investment and Cash Flow: New Evidence; Bilicka (2020), Are financing constraints binding for investment? Evidence from a natural experiment.}

^{10 5G Observatory, 2023, Orange Spain launches standalone 5G}

^{11 RCR Wireless, 2023, The current state of 5G in Germany}

^{12 5G Observatory, Orange Belgium deploys Standalone 5G}

^{13 Fierce Wireless, 2020, T-Mobile launches nationwide 5G standalone network}

^{14 Softbank, 2021, SoftBank Corp. First Carrier in Japan to Provide 5G Standalone Commercial Services}

^{15 Fierce Wireless, 2021, Samsung, KT launch 5G SA in South Korea}

-----------------------------------------------------

‍

● Approach 1 – Fee-free licences. Ofcom would not charge additional fees beyond the initial amounts paid for spectrum, with minimal additional commitments. This approach has been adopted in both the US and increasingly in Europe. ¹⁶ It removes distortions to MNOs’ ability and incentive to invest, as well as removing any frictions from the spectrum trading process, thereby promoting optimal use. It also promotes competition between MNOs by ensuring a level playing field for spectrum costs across different bands. These benefits, which could unlock greater investment in 5G, would need to be weighed against concerns around potentially gifting a scarce natural resource to the MNOs.

● Approach 2 – Targeted fee reductions. If Government and Ofcom want to retain some level of spectrum fees, those fees should deliberately be calibrated conservatively to minimise the distortive impact of the ALF tax. This could be done in a number of ways. For example, Ofcom could use an opportunity cost model to estimate MNOs’ valuation of marginal blocks of spectrum, most likely on the basis of avoided costs, or Least Cost Alternative (LCA) – Approach 2a. This is the approach Ofcom used before 2010, and which it currently uses for high bandwidths (including 10 GHz, 28 GHz, and 32 GHz). ¹⁷ It could ensure modelled outputs are meaningfully conservative through the application of an overlay such as an adjustment factor that it keeps under regular review. Alternatively, Ofcom could retain its existing market valuebased approach and apply a blanket downward adjustment to generate a meaningful reduction from current ALF levels – Approach 2b. This would be in line with the Government's (and Ofcom’s) prior approach of applying a 50% reduction to opportunity costs in setting AIPs.¹⁸ While this would retain some of the shortcomings of the existing approach, particularly in relation to distorting investment incentives and the optimal use of spectrum, it would meaningfully reduce the risk of such distortions.

Approach 2 could be taken forward in combination with an investment commitment model. MNOs could have the option of paying the reduced ALF under Approach 2, or face no further charge for existing licences in return for undertaking to deliver specific coverage or investment commitments (a similar approach to the one adopted in France).

As the Government has recognised, the time is right to consider alternative approaches to ALFs, but any reform will have to start with the repeal of the 2010 Government Direction. If it is repealed, the above alternatives could be used by Ofcom to help support the Government’s growth objectives through encouraging 5G investment.

‍

-----------------------------------------------------

^{16 In the recently auctioned 700 MHz band, Spain extended the licence term to 40 years without any financial contribution from licence holders. Spain is now consulting on extending all existing mobile licences at no cost.}

^{17 Ofcom set AIPs for these bands based on current fixed links fees in functionally substitutable bands, which are calculated based on an estimated average incremental cost to an operator of reducing its need for spectrum by adopting more spectrally efficient technology. (Ofcom, 2023, Ofcom’s decision on licence fees for 10 GHz, 28 GHz and 32 GHz spectrum.)}

^{18 The Government set the prices below the opportunity cost values estimated using the Smith-NERA approach (by 50%). A consultation by the RA (Ofcom’s predecessor) in 1998 also supported the general concept of an initial end-point of 50% of own-use opportunity cost. Until the change in policy direction from 2010 onwards, Ofcom continued to develop AIP fees across a number of sectors from a base set by this general 50% reduction. This reflects the assessment that the risks to optimal use posed by setting fees too high are more significant than those associated with setting them too low. See Ofcom, 2009, Policy evaluation report: AIP, para 4.11-4.15.}

-----------------------------------------------------