20 Jan 2016

(by Richard Cadman) Last Friday, the Competition and Markets Authority (CMA) gave , the UK’s largest fixed telecommunications provider, of EE, its largest mobile communications provider. The CMA concluded, after a nine-month long investigation, that the merger would not lead to a Substantial Lessening of Competition (SLC) and that neither the wholesale nor retail customers of both companies would suffer any harm from the merger. How much did this depend on product market definition?

Crucially, the CMA determined that mobile and fixed communications are not substitutes for each other and so are in separate relevant markets. This post explores the importance of market definition by asking whether different issues would have been raised if the CMA had adopted two alternative market definitions: first that fixed and mobile broadband are substitutes and so belong in the same relevant market; and secondly that the relevant market consists of a bundle of service including fixed and mobile broadband, telephony and TV content. These alternatives are set forward to illustrate the importance of the definition in identifying the competition issues, rather then to suggest that the CMA is incorrect in its findings.

Under the CMA’s definition, each party has only a small share in the other’s market: BT has only a small share in mobile markets and EE has only a small share in fixed broadband markets. The merger will, therefore, not lead to increased market concentration in either market and consumers will not be faced with any less choice. The incentives on the merged entity to compete as hard in future as the separate companies do today will not be diminished by the merger and they will gain no additional market power.

But what issues would arise if another definition had been adopted?

A single broadband product market

The UK communications regulator, Ofcom, has always found that fixed and mobile broadband are in separate markets. In its decision, the CMA, which is not bound by Ofcom’s market definitions, also found separate markets. Ofcom and the CMA are in the majority in this view as most other regulators agree with them. However, this finding is not unanimous. The Austrian regulator, for example, . This finding was based on the fact that fixed and mobile broadband prices were broadly similar as was the service (download speed) offered. There is less evidence to support a single market definition in the UK: prices and access speeds are more divergent. However, as this post is simply a thought experiment we can suppose this evidence is not decisive and just ask: what issues would have arisen if the CMA had found a single product market?

Separately, the two firms have market shares in a combined fixed and mobile market of 7% and 26% respectively and the merged entity would have a share of 33%. This is not enough for a presumption of dominance, but perhaps enough to ask some questions. It would raise the HHI from a pre-merger level of 1,944 to 2,310: a rise of 366, which is above the level that the suggest would give cause for concern. Would the fact that the two largest rivals have market shares of 26% and 20% be enough to mitigate any single firm lessening of competition such that it is not “substantial”? Could a coordinated effects argument be sustained on the basis that there is little change in the symmetry of the three leaders while a potentially strong fourth party has been removed? What is certain is that these questions would need to be addressed if fixed and mobile broadband were in the same market.

The telecom bundle as the product market

The CMA rejects a suggestion that the market should be defined as a bundle consisting of fixed and mobile services. It suggests that the availability of services outside the bundle are sufficient to place a competitive constraint on a hypothetical monopolist of bundled products such that a SSNIP (Small but Significant Non-transitory Increase in Price) would not be profitable. Again, this conclusion is in line with nearly all regulators and competition authorities, but what if the CMA had found the relevant market to be a bundle of services?

BT/EE will have access to a combined set of resources that no other UK operator has: BT owns the largest fixed network in the UK and has significant content rights, in particular, for English Premier League and UEFA Champions League and, in BT Sports, it has its own TV channels. EE has its own mobile network and the most widely available 4G network in the UK. BT/EE could therefore provide the same content on multiple platforms all within the same bundle. Its competitors are unable to match this set of resources: Sky has content but buys broadband access wholesale from BT; Virgin Media has its own content and fixed cable access network, but relies on EE to provide infrastructure for its Mobile Virtual Network Operation (MVNO). Other competitors similarly lack at least one key asset. These competitors all rely on at least one other party to provide a comparable bundle. If BT/EE were to use its set of assets to offer consumers a unique product, would that result in a SLC? Even if it did, would consumers be worse off or would they benefit from the dynamic efficiency gains from the merger? Again, a different market definition leads to different questions.

Whether a different market definition would lead to a different conclusion is not the purpose of the blog post. Rather it is to point out the importance of getting the market definition right to ensure the correct questions are asked. Looking at the mobile and fixed markets as separate, as the CMA has done, leads to one conclusion. Perhaps if an alternative market definition were deployed a different conclusion would be reached.