Few issues have been as controversial in recent years as telecommunications. The widely publicised dispute between Telstra and the previous government received almost daily coverage in the media; with the recent change of government, the future of Australia’s telecommunications policy is under close scrutiny.
At the centre of these issues is high speed broadband – that is, broadband connections that are capable of supporting bit-rates (i.e. speeds) well in excess of 10 megabits per second. Although high speed broadband is available in some parts of Australia’s metropolitan areas, coverage is very patchy and seems likely to remain so without major investment in upgrading the telecommunications network.
In the recent elections, both sides of politics committed to making high speed broadband widely, if not universally, available. Labor campaigned hard on this point, and in that sense, now has a clear mandate to promote the deployment of a high speed broadband network – and to do so quickly. The policy question is how that is best done.
The Optus/G9 proposal
One proposal as to how to proceed has been put by a consortium of access seekers known as the G9. (“Access seekers” are competitors to Telstra that use its network to provide service to end-users.) The essence of this proposal is a complicated split of Telstra’s current, vertically integrated, network into several components, each of which would be run by a different entity.
Whatever one might think of this proposal, no-one can dispute the fact that its management, and more generally governance arrangements, would be extremely complex. Heath Robinson’s famous drawing of the “Professor’s invention for peeling potatoes” springs to mind.
Heath Robinson’s image is especially apt because the G9’s “invention” is an extraordinarily complicated way of seeking to achieve a policy objective that should not be so complicated – the objective of upgrading Australia’s telecommunications infrastructure. Like the Heath Robinson potato peeler, implementing the G9 “invention” would involve very high costs, and would be highly uncertain in its operation.
Those uncertainties translate into enormous risks: for implementing the G9 “invention” would eliminate the network on which we currently rely, with little easy way of going back should the arrangements fail. The current potato peeler would be gone, and the Heath Robinson one would be all that we would have to rely on. We would therefore be adopting an approach that was both highly costly and involved large, irreversible, risks.
My paper
The attached paper, which I prepared as part of the Telstra’s Submission to the ACCC on the G9’s proposal, examines the economics of the G9 proposal, and some of the claims made on its behalf. In essence, the paper makes the following points:
- There are good economic grounds for vertical integration in network industries. Specifically, vertical integration can yield efficiencies in pricing, in the management of service quality and in investment and innovation;
- Experience with structural separation in a wide range of network industries – including rail, ports and energy – highlights how significant these efficiencies are, and how great the costs of vertical separation can be.
- For example, in Australia, the transport of bulk export freight is highly profitable, so that it is surprising that binding capacity constraints would develop and persist. However, they have, most visibly at Dalrymple Bay and Port Waratah. While many factors are at work, the difficulties have been made more intractable by vertical separation, with ongoing coordination failures between mine, track, above-track rail, and port.
- There is a striking contrast here between the East coast, where vertical separation has been imposed, and the highly vertically integrated productions systems in the Pilbara: investment in rail and port capacity in the highly integrated production systems in Western Australia tracked the upswing in world demand almost perfectly, while the response at Dalrymple Bay and Port Waratah occurred only with a very considerable lag. The result has been long queues of ships at Dalrymple Bay and Port Waratah and ever large losses in export revenues.
- Few countries have experimented with vertical separation in telecommunications and none have implemented a model that looks even vaguely like that proposed by the G9. The only country to implement a form of structural separation – the United States – has since completely reversed course and its telecommunications service providers are now every bit as vertically integrated as Telstra. This is unsurprising, given the fact that vertical separation is even more complex, and likely to fail, in telecommunications than in technologically mature, relatively simple, industries such as rail.
- As a result, the economic case for adopting the approach promoted by the G9 is extremely weak.
The mess we’re in
All of this leaves open the question of why we are in the situation we are in, and what can best be done about it.
I have a forthcoming paper on this, and I hope you will have time to take a look at it when it comes out. I won’t steal its thunder but the central point of that paper is simple: at the heart of the problem lie enormous price distortions created by the telecommunications access regime.
In essence, what the ACCC (which administers the telecommunications access regime) has done is to set access charges in a way that discourages investment both by Telstra and by its competitors.
Consider this fact alone (and many others are cited in the paper I refer to): in recent years, regulated access charges have decreased at an annual trend rate (in real terms) of close to 30 per cent for the Line Sharing Service (LSS), of just under 13 per cent for the unbundled local loop services (ULLS), and of 7 per cent for originating and terminating Public Switched Telephone Network (PSTN) access.
These rates of decline are far greater than those observed overseas. Moreover, it is very difficult to believe that costs could have declined at rates anywhere close to those rates of decline.
Thus, even on the ACCC’s own estimate, total factor productivity in Telstra’s fixed network has been increasing at an annual average rate of 5.4 per cent, and (in trend terms, that is, removing outliers) of less than 2 per cent. And even with that total factor productivity growth, input prices (for important items such as copper and trenching) have been rising in nominal terms, as the “China boom” has increased world demand for those inputs. Simulations with a cost model of the Telstra network show that each 1 per cent increase in the price of copper increases the average cost per line of a hypothetical “new build” network by 0.24 per cent. Given that prices for copper have soared in recent years, and that prices for other inputs have also increased substantially, it is difficult to see how cost-based prices could be decreasing at dramatic rates.
Additionally and importantly, the ACCC’s initial cost estimates – from which these very large declines have occurred – were for a fully optimised network: that is, for a network that had already achieved all the efficiencies that could be secured. As a result, productivity growth for this “ideal” network would likely be significantly slower than that achievable in the actual network, as some of the productivity growth in the latter will be “catching up” to best practice. This makes it even more implausible that the declines mandated by the ACCC have any sensible basis in costs.
When regulated prices are set incorrectly, terrible things happen..
These declines in access charges have had two impacts.
First and most obviously, persistent declines in the regulated prices the incumbent is allowed to charge that are not properly based on reductions in costs must deter investment in the regulated network. They will do so all the more given the asymmetric nature of the risk to which an access provider is exposed in the telecommunications regime: if an investment proves to be a failure, it will left to pick-up all the costs; but if it succeeds, it may be regulated on the basis of the pricing approaches set out above.
Second and no less important, the declines in access charges will surely have been far greater than those access seekers could have hoped to obtain for the costs of using networks of their own, compounding the disincentive effect on competing roll-out that setting the initial level of prices below costs would have had. For access seekers, waiting for Telstra to invest (or be forced to invest) has clearly become the best option.
The result has been to completely undermine the expectations that were widely held when the current telecommunications access regime came into operation.
In effect, when Australia entered into the current telecommunications regime in July 1997 Optus had deployed a competing local network, passing over 2 million homes, and other competing network deployments had been announced. Those networks, it was expected, would provide Telstra with “head to head” competition across the board — indeed, in launching the Optus network, the then Prime Minister promised exactly that.
So as to meet that challenge, Telstra was widely expected to upgrade its own facilities, and most notably the Customer Access Network (CAN) (the network of copper pairs and other transmission media that link customer premises to switching centres), putting it on a basis capable of handling very high speed data.
Outcomes have completely belied those expectations.
Optus, rather than further developing its own network so as to compete with Telstra, has relied ever more heavily on Telstra’s. This is in contrast to what has happened in the US and Canada, as well as some European countries, where networks very similar to Optus’ have been very successfully up-graded into full-scale alternatives to the networks of the incumbent operator – often under conditions far less favourable than those faced by Optus. No other large scale networks capable of competing with Telstra have been deployed; rather, other access seekers too have increasingly relied on Telstra’s network.
As this has happened, investment by access seekers has declined, and Telstra’s investment in its heavily regulated public switched network has been largely confined to investment required to meet regulatory obligations. Total factor productivity growth in telecommunications has faltered, relative to earlier levels, meaning that the scope for continued price reductions in future must have diminished.
All of this contrasts with what might have been. Had Optus faced real incentives to upgrade its network, rather than make ever greater use of Telstra’s, there would likely have been an “investment race” between Telstra and Optus to deploy ever faster broadband; but far from there being such a race, the situation has degenerated into a stand-off, where all parties would rather wait than risk large-scale investment.
Back to G9
Optus’ failure to develop its network to anywhere near its full potential highlights the flaw inherent in the proposals (such as the G9’s) for structural separation of Telstra. Those proposals start from the assumption that it is persistent anti-competitive behaviour by Telstra that is the problem, and that structural separation (or its near cousin) would prevent such behaviour from occurring.
But this seems completely at odds with what has actually been happening. In effect, were the fundamental problem that Telstra was impeding the use of its network by competitors, surely the last thing competitors would do is to rely ever more heavily on Telstra’s network in circumstances where they could readily ramp up their own. Yet the experience across Australia is that third party use of Telstra’s network and services has increased dramatically, while the scope competitors had to aggressively use, expand and upgrade their own networks has been not been taken up. It is unclear how making it even easier and more attractive for competitors to use Telstra’s network (which seems to be the intention of these proposals) could be a sensible policy response to this situation.
Rather, what is needed is genuine reform of the current, deeply flawed, access regime. The price distortions that have thwarted investment in, and the upgrading of, our telecommunications networks need to be set straight. Until that happens, our telecommunications policy will merely fail time and again.