Should BT be broken up?

I had two radio stations reach out to me this morning to comment on BT’s announcement of £6bn of fresh investment in broadband and mobile access. Whilst I was unable to oblige their listeners, I do have some thoughts to share on the matter.

It has been an ongoing policy debate whether the UK regulator, Ofcom, should separate BT’s Openreach division from the rest of the company. It is a complicated question to answer, both in general regulatory terms, as well as dealing with specific issues like BT’s pension fund liabilities.

Whilst it is easy (and huge fun) to bash BT, my past experience of working there is that it is a competent and professional organisation. BT does a good job of delivering critical national infrastructure at scale. The quality of network engineering at BT is high. You’d notice quickly if BT wasn’t about.

BT is not a hollowed-out marketing and vendor management shell. Messing with something that broadly works risks unintended consequences. If structural change is imposed onto BT, then there is a high political and legitimacy risk for Ofcom if things don’t work out the way intended. Given the rate of change in network technology and cloud ecosystems, the unexpected is to be expected.

The controversy has several aspects, two of which are figural to my mind. Firstly, is BT’s investment in fibre to the cabinet (as opposed to the premises) really delivering the digital infrastructure the UK needs? Secondly, is the market structured fairly and efficiently to enable effective competition in both infrastructure and services?

My answer to the headline question is a politician’s one: I will ask another question instead, which is the one I want to answer. What would it take for BT to want to break itself up of its own volition? The primary political risk then falls upon BT itself and its shareholders.

For BT itself to take the initiative, we need to see BT for what it is: the result of a century of political interventions into the telecoms market. BT is the genetically-modified organism of UK technology companies. It is highly resistant to regulatory pesticides that would poison other market participants. If you don’t believe me, just watch it masterfully duck the upcoming “hole and pole” access rules through the small print. BT knows how to use its cable ties to restrain any prospective competition!

BT responds to political and market incentives, like any organisation. To get the kind of behaviours we desire, those incentives have to align the interests of different parties. Today those incentives leave many of people disappointed as they find themselves in broadband “notspots” and “grotspots”.

BT also exists in symbiosis with Ofcom. All regulators are necessarily captured to some extent, if only by a shared framing of the nature of the market. By definition “future major new market entrants” don’t yet exist to lobby their position, so everything is couched in terms of the incumbent player’s world.

In this case, both Ofcom and BT see the world in traditional telecoms terms, not as cloud-like ‘on demand’ and ‘software-defined’. That means everything is framed as being circuit-like and statically provisioned. The key metrics are supply-centric: homes where the service is available, and the “speed” of the broadband service.

As a result there is lots of debate among analyst and consumer advocates over the merits of different broadband services and technologies. Is it really “fibre” when the last mile or meter is coax or twisted copper? This focuses attention primarily on the delivery mechanisms, not on the experience on offer.

Given this supply-led framing, BT has rationally responded to the past incentives that Ofcom (guided by its political policy masters) has constructed for it. The unbending laws of supply and demand mean it is not in BT’s interest to create supply superabundance. Why splurge on higher capex just to oversupply the market and crater prices? It is also not in BT’s interest to make it easy to unbundle the service, so demand can easily be satisfied by retail rivals.

An alternative regulatory view of the world is a demand-centric one. What is it that users are broadly attempting to achieve in their lifestyle and work, both now and in the future? How can we tell whether the infrastructure is fit-for-purpose for those use cases? What are the technical parameters that define such fitness-for-purpose? How can we adopt enough flexibility for different plausible demand scenarios?

In an alternative demand-led regulatory regime, Ofcom would create incentives aligned to satisfying user demand, not for creating supply for its own sake. For example, copper-based solutions have “micro outages” and also fail more often with longer time to repair. This has a QoE impact, one that would be accounted for in the value on offer to both the individual and society as a whole.

Rather than “speed”, we would focus on the “quantity of quality” on offer, and what applications it might support, and how many concurrently. This would be prospective too. Readiness for things like 5G (and its rather tight backhaul timing needs) and the Internet of Everything would be rewarded right now with higher price caps.

This shift to a demand-led view would effectively render the debate over access technologies irrelevant. Whether you use hybrid copper plus fibre, gigahertz radios, USB sticks glued on drones, or bionic carrier pigeons would not matter. If you deliver the application outcome, you get the income.

The deal would be open to anyone with the right wholesale interfaces. For instance, local communities could build their own access infrastructure and plug it into an Openreach-operated “national information grid”. Suddenly, Openreach would have serious competition, and could not rely on copper “line rental” forever.

We have covered the potential for changing the “positive” incentives for BT, so it has tastier carrots for doing the right thing. What about the “negative”, so it feels the whack of the regulatory stick for misbehaving?

Many have noted how remarkably innovative and fast to market BT is when someone else overbuilds its network and it feels threatened. Who at BT wants existence proofs of alternative delivery models, let alone real viable competition? It has all too often entered into a guerrilla economic and regulatory battle with its upstart enemies, and won.

The first step would be to enforce existing competition law, so BT can’t “sniper fire” alternative network overbuilders. This is a real risk with the investment announced today. How much will go to duplicate the capital spent by brave (or foolhardy?) entrepreneurs that have risked BT’s infrastructure ire?

However, I suspect that the political will for this is lacking. (After all, the UK doesn’t seem to have the political will to chase down serious corruption, electoral fraud or child rape, so over-brutal broadband competition doesn’t stand much chance.) So what else might Ofcom do?

One possibility is for Ofcom to take a leaf from the mobile world. Spectrum auctions are designed by game theorists, who carefully model the “play space” and its strategies. Ofcom might wish to consider how it wants to force BT’s hand to say in advance where it plans to build, and then make BT stick to those plans.

This creates some greater certainty for rivals who may wish to overbuild BT’s network. It also gives Ofcom a bat with which to threaten to break some of BT’s bones if it tells whoppers. Maybe we might even have a system of “mud auctions” for rights to certain territories? I am not an economist, so I will leave it to professionals to ponder how to design the game and its incentives.

In summary, the question of “should BT be broken up?” is not the right question. For a start, the answer is far from obvious. It depends on myriad technical and economic factors, and the information is dispersed among many actors and roles. A regulatory wrestling match would be a distraction, and the game would be perceived as being rigged no matter what the outcome is.

I am not knowledgable or clever enough to answer the original question with any confidence. I propose that a smarter approach is to ask better questions.

What do we really want from our infrastructure, in terms of the future demand it has to meet? How can BT’s long-term incentives be better aligned to those of society at large? What would it take for BT (or its divested successors) no longer to be “too big to fail in the UK, but too small to succeed globally”?