 |
NTL
|
Dear Sir/Madam,
SPECTRUM PRICING: IMPLEMENTING THE THIRD STAGE AND BEYOND
This document sets out ntl’s initial response to the RA’s consultation document on this issue. We reserve the right to make further detailed submissions on a number of detailed points, as set out below. A copy of this document is being sent in parallel to Cliff Mason.
Ntl recognises that spectrum pricing is a legitimate tool for the Government to use in ensuring that the public gets value for money for a scarce national resource. However, we have been at pains at several stages in the process of implementing spectrum pricing to point out that, unless carefully managed, spectrum pricing can degenerate into nothing more than a business tax – and moreover, a business tax which is likely to impinge upon the delivery of other, important Government objectives.
The latest set of proposals in our view pay insufficient attention to this concern. Indeed, these proposals seem to be driven by the internal logic of the RA’s own spectrum pricing programme, rather than any clear rationale stemming from congestion and/or inefficient use of spectrum in the affected licence categories.
Much reliance is made on the Smith/NERA modelling work. The latter is of course only as good as the assumptions used to build the model. We have raised significant concerns about some of the core Smith/NERA modelling assumptions relating to the definition of congestion throughout the implementation process, without these being (in our view) satisfactorily addressed. But in any event, Smith/NERA have clearly not given due regard to wider, stated objectives of public policy, on the one hand, and the latest market developments on the other, with respect to a number of the licence classes in making their recommendations. The Agency now has a duty to ensure that in following through their recommendations, it does indeed take these wider factors into account.
We deal with those bands where we think the concerns are particularly significant in turn.
The RA’s command paper in 1996, ‘Spectrum Management: Into the 21st Century’ recognised that:
[Satellite earth stations] share spectrum with fixed links so ought, arguably, to be charged on a similar basis. On the other hand, effect on fixed links use is not, in practice, significant at present; and if earth stations were moved elsewhere in Europe in response to increased charges in the UK, this could reduce spectrum efficiency since international co-ordination requirements could have the effect that the frequencies in question would not be usable by any other service in the UK.’
We recognise, of course, that what the previous Government said cannot bind this Government. However, the practical points noted by the RA at that time remain highly relevant. They are not addressed either in this RA consultation, or in the Smith/NERA final report and its recommendations. Smith/NERA have simply applied the existing formula for pricing fixed link bands to satellite uplinking, on the basis that they are occupying the same band, without regard to the issue noted above that there is little if any practical congestion caused by satellite uplinking.
There is, we believe, an erroneous view prevalent within RA that satellite earth stations sterilise large areas around them. The first Smith/NERA report helped perpetuate this view by confusing the co-ordination area round an earth station (produced according to Appendix S7 of the Radio Regulations) with a ‘no-go’ area for other services. The second Smith/NERA report used an erroneous method of comparison between earth stations and fixed links, then compounded the felony by making an elementary error (confusing dBm and dBW) which led to suggested earth station licence fees 1000 times higher than warranted. In fact a satellite earth station in a shared band might affect a region of much the same area (albeit a different shape) as that affected by a comparable fixed link. ntl intend to demonstrate this to the Spectrum Pricing ad hoc group of the RA Satellite Consultative Committee.
If an "area affected" approach to earth station licensing is to be adopted, then RA must take into consideration the fact that a cluster of earth stations using the same band (such as at Crawley Court) will affect an area only slightly larger than that affected by a single earth station. There is no justification for charging the same high fee for the second and subsequent antennas.
More important, however, is the commercial impact of spectrum pricing in this band. The UK is currently a hub for satellite uplinking activity, which by its nature could easily be transferred to the Republic of Ireland or other countries in mainland northwestern Europe. There are very few barriers to such a transfer and the market for uplinking is highly competitive, so there is little scope for operators to simply absorb increased input costs. Of course, the impact depends on the actual level of pricing. However, we can assure the RA that, were the increases in pricing to be anything like the scale of those mentioned by NERA, the UK uplinking market would be rendered uncompetitive vis a vis overseas suppliers.
We urge the RA to review its apparent commitment in principle to apply spectrum pricing to this area of activity. At the very least, the RA should give a commitment that it will not set prices at a level which render UK-based uplinking uncompetitive vis a vis third countries.
Ntl holds a 10GHz licence. We are committed to a national roll-out of the 10GHz network (our licence requires us to reach 68% of the population by 2003).
Again, we would remind the RA of commitments previously given, in particular by the then Telecoms Minister Mrs. Roche in the passage of the Wireless Telegraphy Act, that the Government would have regard to the impact of spectrum pricing on the commercial prospects of RFA. We do recognise that, in principle, RFA is a suitable candidate for spectrum pricing. Our concern is the apparently erroneous assumption about the extent of current network roll-out, and following on from this, an overly sanguine view that the current pricing proposals, as set out in Annex 1 to the paper, would have on the prospects for competitive Radio Fixed Access services.
First, 10GHz networks are only now being deployed. Ntl is, as far as we are aware, the most advanced of the current 10GHz licensees. We have five base stations established, with a further five under construction and scheduled for completion in January. We are in the process of implementing a full scale commercial trial of 10GHz technology. Of the other current licensees, we are told that CWC have decommissioned their trial 10GHz network and Ionica is defunct, we are not aware of any deployment by Scottish Telecom/Thus.
There is clearly an interest on the part of other parties, as recorded in the responses to the RA’s recent consultation on use of 28and 40GHz spectrum, in WLL. But it is very unlikely that this will lead to anyone deploying a WLL network to significant numbers of subscribers ahead of ntl.
As far as ntl’s own use of 10GHz is concerned, the reasons for delay in using the 10GHz spectrum are well known to the RA. In large measure, they arise from problems with the spectrum itself (which took eighteen months to resolve), and the delays which we then incurred in developing a new and unproven technology with our equipment suppliers Netro. The result is that ntl is today where, in its initial business plan for 10GHz, it hoped to be by early in 1998 in its roll-out of this service. If the RA’s proposals are based on a misunderstanding of the extent of 10GHz roll-out to date, then clearly the solution would be a further period of time before moving from the current per-base station pricing formula to a new formula. This would simply recognise that the timescales for 10GHz roll-out have become unavoidably extended.
We note the argument that moving away from per base-station pricing to a fixed-fee per MegaHertz pricing formula will incentivise rapid roll-out of these services. Presumably, the thought process here is that we will move faster than we would otherwise do to build out the network and acquire customers to offset the fixed charges we are incurring. However, this does not recognise that the wider market for RFA services has changed substantially since the initial licensing decisions on 10GHz RFA were taken. The market for broadband services, which the 10GHz explicitly requires us to use the spectrum to deliver, has become actually and potentially much more competitive since the licences were awarded. It is now clear that there will be a host of competing delivery mechanisms for broadband service delivery. These will include ADSL services delivered via the BT copper loop. There is also scope for further entry in the shape of other RFA operators in different bands (as set out in the RA’s recent consultation document).
In addition, ntl has always intended using RFA primarily to service the SME market. For these customers, BT leased line offerings (certainly for 64kb and 2Mb) are an obvious alternative to the kind of service we would want to deliver. The European Commission has just issued a recommendation which, if implemented in the UK, would lead to BT’s prices for these service offerings coming down very significantly. The extent to which this recommendation leads to an actual reduction in leased line pricing here in the UK depends on the outcome of an OFTEL consultation, which is now commencing. But the Commission’s recommended maximum for a 5km E-1 link is 350 Euro/month which is about £2500/annum, compared with quoted figures for BT retail 5km links of 560-608 Euros per month (i.e. £4200-£4600 per annum).
We estimate that the impact of the suggested pricing used in the RA’s Appendix 1 would lead to a per-customer cost of over £300. We arrive at this figure by considering that our 30MHz duplex allows a base station capacity of 64Mbits or 32 E-1s. In our license application we estimated a total of 315 base stations to achieve 68% population coverage. This gives a license fee per base station of £10,268 assuming a fee per MHz of £54,000. If we assume that we can sell the entire base station capacity to customers taking E-1s, the license cost per customer amounts to £321. This compares, using the same assumptions, with a licence fee cost per customer of under £10 with the existing £300 fee. So the increase in our per-customer cost would be a staggering 3324%!
ntl and other RFA operators would potentially face a pincer effect – regulatory intervention to drive down our competitors’ prices, coupled with a spectrum tax which imposes massive increases in our fixed costs per customer. Since the average WLL link is 5km (or less), a spectrum cost of £321 is a very substantial proportion of a proposed ‘benchmark’ BT tariff of around £2500. In these circumstances, WLL would simply not be economic to operate. In fact, this increase would be a very significant increase in unit costs even if BT’s prices were not forced down from their current levels as a result of the ongoing regulatory scrutiny.
If 10GHz were self-evidently a killer technology with no close substitutes, it would be reasonable to impose stringent ‘use or lose’ requirements on operators such as ntl (whilst recognising that the starting point should reflect that ntl’s delayed launch has been largely unavoidable). But in practice, there is a danger that in the current febrile competitive climate, spectrum pricing of this kind will simply overturn the business case what is already a high-risk commercial strategy of 10GHz deployment.
It is not appropriate to set out all the assumptions which underpin this view in a public response to the consultation. However, we would be happy to share with the RA some of our modelling and business assumptions that underpin the view that the current proposed level of fees could kill 10 GHz deployment stone dead. We hope that there will be an opportunity to do so early in the new year before the RA commits irrevocably to pursuing this course of action. The Government has made clear that it wants to see a healthy and competitive broadband market, and believes that RFA has a part to play in this, most recently in the consultation on 28 and 40GHz RFA. The European Commission has also stated, in its recent communication on the 1999 review, that it believes Member States should take further steps to encourage competitive broadband local loops, in particular through appropriate licensing regimes for RFA. Clearly, it would be a serious matter if these objectives were to be prejudiced through an inappropriate application of spectrum pricing.
There are issues around how the RA should extract best use for the country from 10GHz and other spectrum bands which are suitable for broadband WLL. Many of these issues were touched upon in the RA consultation on the 28 and 40GHz bands. It is very important that a coherent policy is developed in respect of broadband WLL, and that spectrum pricing proposals form a part of that. We urge the RA to complete the process of developing a WLL policy and then see how spectrum pricing could be adopted which would support that, rather than the other way round.
Our interest in this is through our involvement in the transmission of analogue and digital TV and radio broadcasts. In many cases, we believe it would be ntl, as the licensee, which would be required to pay the relevant fees. Although these would in practice be passed through to broadcasters, ntl is concerned to ensure that a sensible and rational approach is taken to the difficult policy issues around analogue switch-off and broadcasting policy.
We note that it is intended to issue further, detailed proposals at the end of this year. But we would want to make some general points now. In our view, there is no logical case for applying spectrum pricing to analogue broadcast spectrum for TV. The major constraint on ending analogue transmissions, as the RA partially recognises, are the conditions imposed by the Government on analogue switch-off. It would clearly be contrary to natural justice to apply spectrum pricing to analogue before broadcasters were able to cease using this spectrum as a direct result of the Government’s own policy. In our view, no broadcaster would want, once the green light for switch-off has been given, to continue dual analogue and digital broadcasting. But in any event, there would be scope to consider applying spectrum pricing at that time, or even outright withdrawal of the spectrum.
There is a separate question as to whether spectrum pricing could be applied to digital terrestrial television transmission spectrum. In our view, this should be approached with considerable caution. At present, DTT spectrum is used for free to air and pay TV services. As regards the former, it is an open question at this stage whether any alternative means of delivery will be found, even in the medium term. Any such delivery mechanism would have to be both ubiquitous and sufficiently cheap to be comparable to DTT.
In pay TV, there is scope for substitution by cable and by satellite DTH. We agree that there may be additional scope for substitution either through RFA technologies operating in other bands (such as MVDS), through wire-based technologies such as xDSL, or even through 3rd generation mobile technologies. However, it is too early to say what the scope for that substitution will be, and in particular on what timescale genuinely comparable delivery mechanisms will arrive. For instance, it is clear that pay per view services will be widely available using ADSL over the BT network ( a company called Video Networks is already offering these services in north London), but the current BT ADSL specification will definitely not support a full bouquet of pay TV channels, or even single broadcast streams which have particularly high bandwidth requirements such as live sportscasts.
For the moment, we believe that the clear rationale should be to ensure that DTT develops as viable competitive delivery platform in both pay TV and free to air. This clearly necessitates a very careful and measured approach to the implementation of spectrum pricing.
The RA document does not discuss radio broadcasting and DAB in any detail. For the record, ntl believes that the key constraint here is that analogue will remain the only practical means of delivering service until such time as the price of digital radio equipment comes down significantly in price, thus starting to replace the installed base of analogue apparatus. At present, digital radios retail for up to £1k. It is going to take a considerable period of time for sufficient impetus to develop behind digital radio to even start to contemplate analogue radio switch-off. It would again simply amount to an unavoidable tax on broadcasters to apply spectrum pricing to analogue radio broadcasting at this stage.
Yours sincerely,
ALEX BLOWERS
Regulatory Affairs Manager
|
