Structural Issues in Telecommunications

I cannot make it to all the conferences I would like to attend, and one of these is going to be this year's ACCC Regulatory Conference. The program notes that industry structure issues will feature twice, once on the issue of "Structural design for effective competition" and once specifically on communications covering "Structural models for NBN deployment".

It is perhaps a suitable time for this consideration, or it might alternatively be a conversation being held after the policy world has moved on. The relevant olicy issue is the agitation that followed the exposure draft of the NBN Co legislation and the "wriggle room" over the scope of the NBN Co's "wholesale only" operation.

But elsewhere Telstra seems to have reached agreement on "prospective separation" and Telecom New Zealand has announced that it is reviewing structural separation.

The main speaker at the ACCC conference on the overall question is Professor Martin Cave, who has had a bit both ways on the topic. In a report he wrote for CEDA Cave supposedly "argues that structural separation is an unnecessary risk." Meanwhile the T4 campaign pointed out that Cave had twice written approvingly of separation options.

The other speaker is Dr. Karl-Heinz Neumann, who is the head of a consulting firm WIK. WIK recently hosted a conference on National strategies for ultrabroadband infrastructure
deployment: Experiences and challenges. This was presumably how he put together his presentation, Investment in Broadband and Next
Generation Networks to Foster
Development and Innovation, a couple of weeks later for the ITU. At the conference both Rob lbon from the ACCC and Kris Funston from the NZ Commerce Commission spoke.

I hold little hope that either ACCC conference sessio will shed any new particular light on the issues. One of the critical factors is whether separation is seen as a regulatory solution to a problem or whether separation is seen as a good market issue. I've previously been published arguing that separation is in the interests of incumbent telcos. This contrasts with the usual view that separation is a regulatory instrument and hence designed to confiscate rents from a monopolist.

Is it possible to reconcile these views? The benefits of competition are usually described in two ways. The first is in terms of the price effect, and the idea that competing firms price at marginal cost unlike firms with market power. The second is that competition provides a stimulus for innovation as firms compete on matters other than price.

I'm increasingly convinced that the price effect is largely illusory. I haven't done the full study yet, but the decline in telco prices over the last thirteen years looks to be due to technology effects rather than competition effects. This is consistent with the dynamic market model constructed by Steve Keen and Russell Standish. That model assumes incremental change in firm production decisions rather than the simultaneous market clearing from an all knowing market in the classic model. They conclude;

The simulation results demonstrate that markets can be locked in a spiral of restricted production converging on monopoly pricing levels as though the firms were colluding, even though no interaction between firms takes place. Perfect profit-seeking rationality and the dynamics of the economy tends to lock firms into synchronous behaviour, leading to a global monopoly behaviour.

There is benefit from competition though, but it is the benefit that Hatek noted of the ability of the marketplace to transmit information on preferences. If all we had to do was set price to marginal cost then an all knowing central-planner could do that. But anyone who has been involved in regulatory price setting knows that it is revealling the demand that is the challenge.

In the competitive market model firms competing can all obtain their share of the monopoly rent, but to "outperform" their competitors they need to obtain the additional benefits from growing market share or reducing costs that comes from innovation.

Where structural separation becomes important for incumbents is where vertical integration becomes an impediment to innovation.

Perhaps the most interesting case study is automobile manufacture. One of the economists favourite examples of vertical integration is the case of Fisher Body and General Motors. The man responsible for much modern theory of the firm and a sub branch of institutional economics summarises the case well. We will ignore for now the details of the case, but instead observe that automobile manufacture was held as a case for vertical integration, but has subsequently evolved into a highly unintegrated business. Actual vehicle manufacture is a process of assembling parts made throughout the world.

The particular significance of this in the telco case is that we are in grave danger of imagining that the structural separation of NBN Co alone solves a regulatory issue, and that merely having a regulator set (or accept) a cost based price resolves all. What is most concerning is that no discussion of the NBN in Australia has considered how to create a dynamic element of the discussion between supplier and acquirer - at least that has been occurring meaningfully within the vertical structure of Telstra. (Yes sorry to disappoint theorists, but CEOs don't work like classic central planners either).

It is to be hoped that the ACCC conference discussion will touch on the dynamic aspects of separation, not just the pre-conceived regulatory constructs.