Vodafone and Three to get merger blessing with poisonous remedies

Price controls were a common feature of Jacobin France, satisfying the immediate demands of the sans culottes for affordable bread but prolonging the economic death spiral while the heads rolled and the blood flowed. The backdrop in the UK telecom market is a less sanguinary affair, but fixing mobile prices in an industry that complains about return on investment looks equally poisonous as a "remedy" for a merger. It is, nonetheless, one the Competition and Markets Authority (CMA) seems prepared to accept as a condition of the marriage between Vodafone and Three.

A final decision on the controversial tie-up, which would reduce the number of mobile networks from four to three, is not due until December 7. But most analysts had already decided it would happen, and an unusual update this week seems to make it a formality. In its press release, the CMA boldly states that "customer protections" combined with similarly dubious spending commitments by Vodafone and Three "could solve competition concerns" and "allow the merger to go ahead." Kester Mann, an analyst with CCS Insight, reckons the operators can "tentatively order in the champagne."

There are numerous reasons to support a merger. For a start, the devotion of European regulators to maintaining four networks per country has always seemed like religious dogma. Even Ofcom, the UK telecom regulator, appeared to agree with Vodafone and Three that telcos would struggle to recoup the expense of building and upgrading nationwide networks with their lowish customer numbers. Subtract one network, divide phone users by the remaining three, and the math already looks better.

Juicier profits would also put telcos in a stronger position to invest in 5G, a technology once seen as a crucial spur to economic growth. Alongside other countries, and especially some Asian high-flyers, the UK measures up badly on 5G performance, according to Opensignal, an independent monitor. Rather than digging behind sofa cushions for pennies to spend on meeting basic coverage obligations, Vodafone and Three united would be able to pump £11 billion (US$14.3 billion) into a whizzier 5G network, they say, arguing their rivals would be forced to respond.

Bless our merger, or face the consequences

But Vodafone (or VUK) and Three (3UK) have presented an abysmal case for a merger, while the CMA looks on the verge of simply dismissing valid concerns and accepting remedies that could damage the sector. In the absence of clear demand for a 5G over a 4G service, the main strategy of the two operators has been to scaremonger and present their networks as tottering heaps of steel and concrete that can be saved only by amalgamation.

They have cut expenditure to the detriment of customers while effectively holding the country to ransom. Combined capital expenditure fell about a fifth in the last fiscal year, to less than £1.2 billion (or $1.6 billion), using current exchange rates, as submissions to the CMA revealed their networks featured more holes than overworn socks. Let us merge or the UK suffers, was the implied threat.

The CMA, though, has been having none of it. In its detailed findings after phase one of its investigation, it said there was "a realistic prospect that VUK will have both the ability and incentive to continue to compete … absent the merger." It also found no evidence "3UK is facing such financial difficulty that it is at risk of financial failure."

Despite this, it now appears to have shelved any case for the structural remedies that would usually accompany a mobile merger. An imbalance in spectrum assets would not be as great as it is now if Vodafone and Three sell frequencies to Virgin Media O2 (VMO2), a network-sharing partner, as they have offered to do, the CMA previously argued. Unfortunately, details of that proposed deal have not been made public, and it leaves open the possibility of a gross imbalance in specific bands deemed critical to 5G. Between them, Vodafone-Three and VMO2 would control 310MHz in the 3.4–3.8GHz range, for instance, while BT would have just 80MHz.

Vodafone-Three would also occupy many more sites – around 26,000, it estimates, after its planned decommissioning of about 10,000, giving it roughly 7,000 more than BT. When this advantage is combined with Vodafone-Three's spectrum assets, there is concern it would create a huge "capacity" imbalance between telcos. According to one estimate previously shared with Light Reading, Vodafone-Three would have 60% of download capacity, leaving BT and VMO2 on 20% each.

The network-sharing conundrum

BT has another reason to be upset about those sites. In its most recent submission to the CMA, it said Vodafone and Three had already flagged plans "to use the Vodafone network as the basis" of the 26,000-site one. The implication is that decommissioning will mainly happen across Three's footprint of sites, most of which it shares with BT via a joint venture (JV) called MBNL.

Because Vodafone is in a different network-sharing venture with VMO2 called Beacon, there is concern that a merged entity, with a foot in each JV, could turn knowledge of BT's 5G plans to its commercial advantage. The CMA seemed to recognize this in its phase-one investigation. But six months later, it unconvincingly dismissed the concern, arguing that any information is "unlikely" to be of much use to Vodafone-Three.

To complicate matters, Three has already agreed to sell UK sites to Cellnex, a European towers specialist, in a deal reportedly worth €3.7 billion ($4 billion). After 2031, when the MBNL agreement expires, sites will be divided between BT and Three, leaving Cellnex to pick up the rights to Three's share. But contractual obligations with both Cellnex and MBNL prevent Three from just abandoning premises.

In one scenario BT envisages, Three would be replaced by "a different network sharing partner." But it is hard to think of viable candidates, and BT sounds unhappy about the prospect of an unknown quantity that might have a "questionable" long-term commitment to the UK. Another possibility is that Three is forced to continue making payments for up to 10,000 sites it wants to decommission.

Taking Vodafone sites out of service instead seems improbable, largely because Vodafone's network appears in less need of a refresh. It is believed to have completed a switch from Huawei to Ericsson across 3,500 sites and already used Ericsson across the rest of the footprint, excluding 2,500 sites where it is shifting from Huawei to an "open RAN" mix of Samsung, NEC and IT component suppliers. Three, by contrast, juggles Nokia (3G marked for scrap), Samsung (old 4G), Huawei (4G and 5G that must be removed by 2028) and Ericsson (permitted 4G and 5G).

A barking-mad watchdog

Sorting out this mess will inevitably chew into Vodafone-Three's capital expenditure budget, and that is not as impressive as the headline about an £11 billion 5G investment makes it sound. As a footnote in the statement on the deal indicates, the £11 billion is actually the combined capex projection, and telcos must invest in more than just 5G. It also works out at only £42,300 ($55,000) per site each year, assuming a footprint of 26,000 sites. That is just 3% more than Vodafone spent per site, on the same basis, last year.

Besides getting Ofcom to monitor prices, the CMA seems to think holding Vodafone-Three to account on this spending promise is a viable remedy. But an investment amount is the wrong thing to examine, as James Crawshaw, a principal analyst with Omdia (a Light Reading sister company), noted on a recent Telecoms.com podcast.

"Measuring capex is an input," he said. "What the regulator needs to measure is the outcome or the output. What's the benefit for the consumer? The amount of capex is the wrong thing to be looking at. It could be fudged. You could be spending that capex on Nvidia GPUs [graphical processing units], if you wanted."

The outcome or output, moreover, is already supervised under the commitments that telcos are supposed to honor when they receive a license. Failure to hit coverage targets risks a hefty fine or even the revocation of that license. Of course, this never happens because it would inadvertently penalize customers. But the right approach to monitoring already exists. What regulators really need is an answer to the problem of enforcement, and that they have not provided.

Monitoring inputs, as Crawshaw refers to them, might work if Ofcom examines the details of Vodafone-Three's planned investments and intervenes where it disagrees. But this would amount to micromanagement of the operator and implies Ofcom would be a better judge of what's needed than Vodafone-Three's bosses. Buying GPUs might turn out to be wiser than spending money on standalone 5G, a newer version of the technology that has not buoyed telco revenues in markets where it has been made available. Who can say?

Wherever the money goes, Vodafone-Three will not be allowed to raise prices for certain customers, including the mobile virtual network operators that rent capacity on its network, for a specified period. This will clearly affect the tariff plans of non-merging telcos even though UK mobile tariffs are already among the lowest in developed markets. A postpaid phone customer of T-Mobile in the US, a market of three big mobile networks, currently spends about $50 a month. His equivalent at Vodafone spends £18 ($23.40). If economic circumstances change dramatically, prices will not be allowed to rise.

What's curious is the CMA's enthusiasm for behavioral remedies. A former employee of the agency told Light Reading it has never favored them in the past because monitoring is such a challenge. But finding structural remedies acceptable to all parties looks especially difficult in this case, and inviting a fourth player into the market would undermine the rationale for a deal that Vodafone and Three have advanced. Those network-sharing concerns, meanwhile, are especially hard to address. The CMA seemed disinclined to bother.