The indispensable four per cent

Radio access networks account for just 4% of operators’ cost. So how did it become a public policy discussion?

In one of the more surprising news to come out of Sweden, we read that the CEO of Ericsson was lobbying on behalf of Huawei against Sweden’s decision to block Chinese network equipment. Ericsson’s support for its competitor seems quite counterintuitive, and some have noted how the Wallenberg family – who the Chinese media call “Rockefellers of Sweden” that holds a controlling stake in Ericsson – need to tap into the Chinese investment opportunities.

What the China-critics miss (or deliberately leave out) when they accuse of undue influence, is how this lobbying has fallen to deaf ears: The Swedish Constitution prohibits ministerial interference in administrative decision. The matter is also already referred to the courts, where it’s out of reach for any Cabinet-level influence.

It is pertinent to note that the decision (where the opinions of the Security Service were seminal) was not based on the merits of the Chinese companies, or demerits based on loose speculations. Instead, intelligence agencies based their legal opinion on its reading of China’s controversial National Intelligence Law that may be used to coerce any Chinese citizen or organisation, with little judicial redress. One may speculate that Säpo – our national security agency (not the name of an IKEA table) would deliver a similar assessment of the US Foreign Intelligence Surveillance and Cloud Acts if a US-based company (say, Mavenir) were on the supplier shortlist in the future.

Regulatory capture over 5G

After all, business lobbying – by Chinese or Rockefellers alike – is fair game, if it stays within the remits of the law. It’s how our political decision-makers respond to the lobbying that determines whether the influence is improper by democratic standards. Here, Germany’s handling of the same issue is a striking contrast to Sweden: Politicisation delayed the new German security bill ad infinitum, allegedly allowing a state-owned telco to hoard Chinese 5G base stations. While its final draft gives German security agencies far-reaching technical and legal means to monitor the network’s integrity, the bill also defers the decision on high-risk vendors to a unanimous decision by all cabinet ministers. It’s a very elusive unanimity one might add – in a country united by mercantilist interests and divided by coalition governments.

Despite reaching two very different conclusions, both the Swedish and German decision follows the EU 5G Toolbox. Member States must assess its own risks and apply relevant restrictions to mitigate them. That includes necessary exclusions of vendors for “key assets” considered as critical and sensitive. Such assets include the Radio Access Networks (RAN) that make up most of the costs in a 5G rollout – which happens to be Huawei’s principal business.

But unlike Sweden, the German outcome is politicised to its core: The high-risk vendor designation is a ministerial decision taken by politicians and therefore politicised by default. Regardless of your personal politics agree with either the Swedish or German decision, the episode illustrates a problem of “regulatory capture”: Rather than acting on behalf of the public or national interest, authorities represent the industries they are supposed to be regulating – often without quid pro quo. The largest German multinationals were even given their own 5G telecom licenses and don’t need to worry about the controversy at all. 

As much as there is a “China Inc.”, there is a “Deutschland GmbH” with a revolving door between the ruling party, regulators and state-owned enterprises. The State Secretary in charge of the 5G decision is not just a senior CDU politician – he is also a former Deutsche Telekom manager. The political opposition accuses Merkel’s CDU to let its national champions (led by the state-controlled Volkswagen) dominate its regulatory policy, and its senior ranks were entangled in spy scandals involving telecom interests. 

China’s state-capitalist model and its mystical prowess often foil the western intellect. But the Chinese economy can be remarkably similar to German industrial policy. Even the GDP share of state-owned enterprises in Germany is marginally below China’s.

I spy a national champion

RAN has become a household acronym in recent months – at least if your household happens to include a diplomat, sinologist, or a spy. The German bill also contains a last-minute poison pill that allows the government to interfere in interoperability or compliance of RAN standards. The provision goes against the western practice of open competition between industry protocols by letting operators – small and large – freely determine the technology they want to use. Such provisions allow dominant operators to lobby for private consortia like ORAN to become national mandates, by forcing them on all networks.

Writing “networks” – in the plural – is an important distinction. 5G consists of several network domains, and not just RAN. Sure, a healthy telecom industry is an essential public interest as 5G underpins all critical societal functions. But RAN accounts for just 3-4% of the costs in the mobile business. In comparison, electricity accounts for 5-7% of telco costs according to McKinsey. Electricity is perhaps twice (or at least on par with) the annual budget spent on RAN. In fact, telecom is one of the most energy-intensive industries, and 5G has a severe climate change impact. 

So, the question is whether RAN costs would have been such a singular focus of our debate if it wasn’t for the overhyped trade wars. Any marginal changes in RAN costs are not the most obvious target for public subsidies. Revenues, rather than state aid, can compensate higher costs: It takes just one per cent increase in revenue to pay for a 30% higher cost on RAN. By the same token, the gains of a, say, 30% cheaper RAN are lost if there are cost increases elsewhere. ORAN is no longer a cost-saver if the electricity bill increases by 15%. This, of course, doesn’t take into account negative externalities of rising carbon emissions.

This is not to say that countries should not provide aid to rollout 5G faster to be on the cutting edge of digitalisation. Similarly, telcos rightly expect compensation if laws retroactively declare any of their equipment “untrustworthy”. But non-discriminatory consumer support and compensation for removing hitherto lawful equipment are distinct from discriminatory subsidies and interference where the state picks winners. Neither is it in the public interest to pay subsidies to help them cope with cost changes that amount to just 1%. Let’s also get some perspective here: In the days when China forced western suppliers to stick to its own “open” protocol (TD CDMA) developed by a similar closed-door consortium, it was deemed a serious market access barrier. 

The bottom line is that no company likes costs. All industries want to avoid regulatory action – whether they make test-cheating diesel cars, nuclear power plants, import biofuels or internet services. CEOs and lobbyists are only doing their jobs when they advocate against compliance costs. But only overprivileged national champions in state-capitalist societies make their woes a public policy concern – and get the taxpayers to pay their costs. 

Here is where policymakers – diplomats, sinologists and spies alike – need to reflect whether politics surrounding RAN is a cynical corporate game. 

Some notes on the data:

• RAN costs are based on Dell’Oro data on global RAN spend on all RAN technologies (2G to 5G) for the past ten years (2009 to 2019) to properly take into account any effects from write-offs and cyclical fluctuations. However, the spending does not show any significant year-to-year variances that would alter the conclusions. Total industry cost is based on global mobile revenues in 2020 (1.03 tn USD according to GSMA; 1.07 tn USD by Statista) against a global pre-tax, pre-stock compensation profit margin of 14.4% based on 104 wireless telecom companies from all regions published by researchers at NYU Stern.
• Equivalent estimate for Europe by Dell’Oro and IDC 2019 data on turnover against 13% profit margin (based on 13 EU wireless operators) leads to similar (or lower) share for RAN – approx. 3% – of total costs.
• A recent staff note by the World Bank suggest that the share of SOEs and partnerships in China’s national GDP is 23% or more. Arguably, Chinese SOEs are less internationalised than German SOEs and the annual turnover of the major German SOEs is equivalent to 15% of Germany’s GDP. These SOEs are in transports (Deutsche Bahn, Deutsche Post, DHL, DFS, Hapag Lloyd), telecoms (Deutsche Telecom), banking (KfW Bank, Commerzbank, Hypo), manufacturing (Volkswagen and a minor stake in Airbus), mining and raw materials (RAG, Evonik). In addition, there are many regional and municipal SOEs, pointing to a much larger economic role of SOEs in the German economy.