Blog Post

High expectations for 5G confront practical realities

The next wave of mobile network innovation is provoking great excitement in the industry. And indeed, there is substantial potential for improvement. However, the exact form of the technology and the appropriate policy support are still far from clear. And we should beware of over-ambitious promises about the impact and uptake of new network technologies.

By: Date: March 14, 2017 Topic: Innovation & Competition Policy

A version of this piece was originally published on Corriere della Comunicazione.


The next generation of mobile technology, 5G, is being developed along markedly different lines from previous generations. In the past, mobile generations were generally characterised by a core technology (or sometimes by two or more core technologies), and were designed to fulfil the requirements of a fairly small number of mobile voice and data applications. By contrast, 5G is being developed to fulfil the needs of multiple use cases.

Most definitions of 5G assume that it will provide some combination of (1) high speed, (2) low latency, (3) the ability to use high frequencies well above 6 GHz, and (4) the ability to support huge numbers of users (some of which will be machines rather than human users) and applications. Some applications require high bandwidth and low latency; many machine-to-machine applications, by contrast, require only modest bandwidth, but imply the need to support huge numbers of devices.

The technical means of implementing this wide range of capabilities is still very much a work in progress.

Functional requirements for 5G are interlinked, but there is a degree of tension between them. To provide higher speeds, bandwidth is needed that is simply not available in the heavily used sub-6 GHz bands; however, the use of higher frequencies implies faster attenuation of the signal, and thus limitations in the distance that can be covered by a base station. (The use of directional antennas might possibly help to overcome this.)

This in turn implies a need for more cells, and thus greater cost. One study found that the cost (CAPEX) of coverage at 3500 MHz using presently available technologies (not 5G) was roughly 6.7 times as great as the cost of coverage at 700 MHz (see Figure 1).

Whether this is a problem depends on the use cases that the network operator is seeking to fulfil. Not every use case or business model requires nationwide coverage.

Indeed, if we are seeking capacity rather than coverage, the rapid attenuation at high frequencies can be a positive rather than a negative factor. It enables greater re-use of the same radio spectrum in the same frequency bands within a given geographic area. For areas where bandwidth demands are high (for instance, dense urban areas), a dense deployment of 5G small cells might be entirely appropriate.

Figure 1. Frequency, cell radius in Km, and the impact on cost (CAPEX).

Source: Simon Forge, Robert Horvitz and Colin Blackman (2014): Is Commercial Cellular Suitable for Mission Critical Broadband?

Figure 1

Not so long ago, many experts assumed that all 5G deployments would entail ultra-fast services at high frequencies – a view that ignores fundamental underlying economic realities. It would however be equally wrong-headed to assume that all deployments will take place in the desirable frequencies below 1 GHz.

The real opportunity that 5G offers is the ability to customise deployments to meet requirements. A deployment that needs extensive coverage might use a mix of widely dispersed sub-1 GHz deployments for rural areas, together with high density deployments of small cells at higher frequencies for dense urban areas. This kind of mixed or hybrid approach might be much more common under 5G than under current technologies.

Source: RSPG (2016), Strategic Roadmap towards 5G for Europe.

The roadmap of the Radio Spectrum Policy Group (RSPG) needs to be understood in this light. Their key recommendations are:

(1) that the 3400-3800 MHz band should be the primary band for the introduction of 5G use in Europe before 2020;

(2) that 5G will also need existing mobile bands, including the 700 MHz band, in order to enable nationwide and indoor 5G coverage;

(3) that additional bands are required above 6 GHz.

Some argue that Network Virtual Functionalisation (NFV), implemented using Software Defined Networking (SDN) and OpenFlow protocols, might be the key to providing the required dynamic configurability.

It is also possible that 5G will co-exist for an extended period of time with earlier technologies (LTE, 2G, and perhaps to a lesser extent 3G), not only to support legacy equipment, but also due to gaps in 5G capabilities. This would be analogous to the situation with LTE, where voice services are most often provided today using Circuit Switched Fall-Back (CSFB) (i.e. to 2G/3G) rather than using Voice over LTE (VoLTE).

Every previous mobile generation has offered the promise of being better, faster, and cheaper. In the early design phase, all things are possible. Sooner or later, however, hard rules of physics, of engineering, and of economics must be confronted. Even so, every generation has indeed been better, faster, and cheaper than the previous generation, but not to the degree that was initially hoped for.

In this sense, 5G is not altogether different from prior generations; however, the gap between hope and reality is even greater than usual, especially in the policy domain. Every possible policy measure is claimed to be necessary to the success of 5G (and we sometimes see this claimed for opposite policy positions). Since nobody can say with certainty today what 5G will be, nobody can refute these claims.

Successive generations of technology have purported to provide a single view of all networks, and to enable all network services to compete with all other network services. Claims along these lines have been advanced for ISDN, for Open Systems Interconnection (OSI) protocols, for ATM, and for the NGN IMS. Alas, none of these have brought us to the promised network nirvana.

There are lessons that we can take from this history:

  • The initial degree of corporate or government support for a technology or standard does not uniquely determine the degree to which it will ultimately be accepted by the marketplace. IP protocols initially had scant corporate or government support, while OSI protocols enjoyed seemingly universal acclaim … but who won? Few readers today will even remember what OSI protocols were.
  • What a technology is capable of does not uniquely determine how it will be used. The electronic communication value chain is comprised of many mutually competing firms … and that is a good thing! The actual deployments will be determined by business needs – technology is an enabler, but it alone is not dispositive.
  • Dystopian forecasts likewise need to be taken with a grain of salt. Claims that it might not be possible to support more than one 5G deployment in a European Member State assume that a huge number of cells would be required. With a hybrid design, however, this might not be the case, especially when one takes into account the possibility to share base station locations, antennas, and perhaps back-haul facilities among competitors.

There are many factors that could delay or impede 5G deployment, even if business and technical circumstances are otherwise favourable, such as:

  • Possible inability of the technology to converge to a single, interoperable standard.
  • Delays in assignment or re-farming of the required spectrum resources (for instance in the 700 MHz or 3400-3800 MHz bands.
  • Challenges in deploying base stations and in obtaining fibre-based backhaul to them, especially where large numbers are needed in dense areas. This may be less of an issue for deployments in the 700 MHz, where existing locations designed with 800 MHz in mind are widespread. This risk is compounded by delays in transposition and implementation of the Cost Reduction Directive, where infringement actions are pending against most of the Member States.

Claims regarding 5G today are surely over-blown; nonetheless, we should expect significant benefits.

Republishing and referencing

Bruegel considers itself a public good and takes no institutional standpoint. Anyone is free to republish and/or quote this post without prior consent. Please provide a full reference, clearly stating Bruegel and the relevant author as the source, and include a prominent hyperlink to the original post.

View comments
Read article Download PDF More on this topic

External Publication

Network Sharing and 5G in Europe: The Potential Benefits of Using SDN or NFV

What technological, economic, and regulatory policy implications might flow from the use of Software Defined Networks (SDN) and Network Function Virtualisation (NFV) technologies, and what are the interactions among these different dimensions? The authors explore the application of SDN/NFV technology to achieve greater flexibility as to how communication networks are used.

By: J. Scott Marcus and Gabor Molnar Topic: Innovation & Competition Policy Date: December 7, 2017
Read article More on this topic

Blog Post

Can roaming be saved after Brexit?

The referendum where UK voters chose to exit the European Union has many unanticipated consequences. One that is gaining visibility in the UK just now is the impact of Brexit on mobile roaming arrangements. How might the UK maintain roaming arrangements with the EU in the event of a hard Brexit?

By: J. Scott Marcus and Robert G. Clarke Topic: Innovation & Competition Policy Date: September 21, 2017
Read article

External Publication

Economic Implications of Further Harmonisation of Electronic Communications Regulation in the EU

One of the ways in which the European Commission has sought over the years to strengthen the European single market is by means of increased harmonisation of the regulation of electronic communications. To the extent that the European Union functions as a confederation of somewhat autonomous member states, however, there are both practical and political limits to the degree of harmonisation that is realistically desirable or achievable.

By: J. Scott Marcus and Christian Wernick Topic: Innovation & Competition Policy Date: August 11, 2017
Read about event More on this topic

Past Event

Past Event

Geo-blocking in the digital single market

Geo-blocking is a discriminatory practice that is wide-spread in EU. It prevents online customers from accessing and purchasing products or services from a website based in another member state

Speakers: Marine Elgrichi, J. Scott Marcus, Fabian Paagman, Bertin Martens, Georgios Petropoulos, Agustin Reyna, Gareth Shier, Werner Stengg and Roza von Thun Topic: Innovation & Competition Policy Location: Bruegel, Rue de la Charité 33, 1210 Brussels Date: May 30, 2017
Read article Download PDF More on this topic

Working Paper

From start-up to scale-up: examining public policies for the financing of high-growth ventures

What are the challenges of financing scale-ups, and how can long-term public policies support the creation of a better scale-up environment?

By: Gilles Duruflé, Thomas Hellmann and Karen E. Wilson Topic: Innovation & Competition Policy Date: April 10, 2017
Read article Download PDF More on this topic

External Publication

Extending the scope of the geo-blocking prohibition: an economic assessment

This paper was prepared for the European Parliament at the request of the Committee on Internal Market and Consumer Protection.

By: J. Scott Marcus and Georgios Petropoulos Topic: Innovation & Competition Policy Date: February 27, 2017
Read article More on this topic More by this author

Blog Post

Big data and first-degree price discrimination

What’s at stake: first-degree price discrimination - or person-specific pricing, had until recently been considered a theoretical case with unlikely real-world application. Yet the increasing availability of big data could make this possible. We review recent contributions on this issue.

By: Silvia Merler Topic: Innovation & Competition Policy Date: February 20, 2017
Read article More on this topic More by this author

Blog Post

How good a shield is Privacy Shield?

Privacy Shield was put in place in 2016 to ensure that transfers of personal data from the EU to the US would be in compliance with European Union privacy law, and thus permissible. The institutional framework of Privacy Shield was weak, and depended on the good will of the US administration. Recent actions by the new administration, including the famous executive order forbidding residents from 7 predominantly Muslim countries to enter the US, may have (presumably unintended) effects on Privacy Shield. To preserve the validity of Privacy Shield in European Courts, strong EU-US cooperation and potentially additional agreements may become necessary.

By: J. Scott Marcus Topic: Innovation & Competition Policy Date: February 7, 2017
Read article More on this topic More by this author

External Publication

Policy and Politics in the Era of the Industrial Internet: How the Digital Transformation Will Change the Political Arena

The digital transformation has already had an impact on policymaking, and this trend will continue in the years to come. How will the political process change and how can influencers guide this change?

By: Giuseppe Porcaro Topic: Innovation & Competition Policy Date: December 7, 2016
Read article Download PDF More on this topic

Policy Contribution

Going local: empowering cities to lead EU decarbonisation

Decarbonisation and digitalisation are reshaping the European energy system, which will become more decentralised and interconnected with other sectors. Cities have the opportunity to be the key drivers of decarbonisation, but this will require the implementation of a new bottom-up governance system. This paper outlines a four-step mechanism in order to achieve decarbonisation at city level.

By: Simone Tagliapietra and Georg Zachmann Topic: Energy & Climate Date: November 30, 2016
Read article More on this topic More by this author

Blog Post

Search engines, big data and network effects

Search engines are intermediaries in a two-way market between users and advertisers. Their huge stocks of data about users and their preferences can help search engines offer better services to all parties. But does this make market entry difficult for new players? And can we see network effects emerging in the search engine market?

By: Georgios Petropoulos Topic: Innovation & Competition Policy Date: November 22, 2016
Read about event More on this topic

Past Event

Past Event

Digital platforms: A policy and research agenda

The number of digital platforms is currently rising in many countries and sectors. What are the opportunities of platforms and which kind of regulation and policy framework is necessary to promote healthy competition?

Speakers: Bruno Basalisco, Diane Coyle, Jacques Crémer, Werner Stengg, Nicolas Petit, Georgios Petropoulos and Simon Wilkie Topic: Innovation & Competition Policy Location: Bruegel, Rue de la Charité 33, 1210 Brussels Date: October 20, 2016
Load more posts