Study of Market-based Exclusive Spectrum Rights

Annex 5 – Spectrum Usage Rights

In this Annex, we discuss the application of electrospace model and issues raised in debate on the subject amongst Whittaker (AUS) v. OFCOM (U.K.) v. Matheson (U.S.). The core of the debate is not about movement to electrospace but on implementation and how to address challenges of ex post litigation and enforcement.

As noted earlier in Chapter 5, Section 5.4.10, a necessary first step to enable flexible market trading of exclusive licences among primary licensees is to ensure that all parties clearly understand what it is that they are trading. The rights should be simple and clear to understand and should be stable (unchanging over time) so as to minimize risks of ambiguous interpretation. Uncertainty as to what the rights are and whether they will remain the same over time would deter both buyers and sellers from participating in the market, would raise the costs associated with completing secondary market transactions, and would increase the potential for various types of moral hazard problems. Among the latter, as we shall discuss further below, is the risk of future costs to resolve interference claims.

One approach to addressing this problem is the one followed by Guatemala where they have progressed, perhaps the furthest, toward the goal of full privatization of the RF spectrum. In Guatemala, titles for spectrum use (TUFs) were granted that are analogous to "fee simple" property rights assigned in real estate. These titles specify the band and provide minimal technical limitations including maximum transmit power and energy levels at the boundaries of the license (frequency and area defined).Footnote 127 These titles are freely tradable and may be sub-divided or consolidated. The owner is free to use the spectrum as they wish (with not service or technical restrictions, beyond the minimal technical restrictions needed to control interference to adjacent licensees. To transition to this framework, the Guatemalans granted licences to incumbents and then made the remaining spectrum available on a first come/first served basis, with conflicts settled by auctioning the spectrum. While fans of this approachFootnote 128 point to the growth of wireless services in Guatemala and the lack of significant challenges or disputes arising from interference disputes or competition concerns, it does recommend a radical change from the status quo in most countries, including Canada, and may not be suitable in such larger-economy/more-mature market settings. As an example of the potential problems that may arise, it is worth noting that the TUF for the 2.4GHz spectrum used by WiFi devices is privately held which precludes the sort of end-user-driven deployment of WiFi networks possible in the U.S. and elsewhere where 2.4GHz is unlicensed spectrum.

Relative to the electrospace concept discussed earlier, the Guatemala approach represents a coarse application of dimensioning the spectrum space, focusing solely on the four dimensions of frequency (1) and geography (3). Further flexibility in allocating the spectrum depends on the TUF owners having adequate incentives to sub-divide the spectrum. Also, in light of the challenges for managing spectrum interference, the rights associated with interference protection are ambiguous. So far this has not caused significant problems, but the potential exists for this to change if title holders significantly change the uses and technologies operating in their spectrum (e.g., move to denser, smaller cells or move from a fixed wireless service to a mobile wireless service). As noted, the Guatemalan approach is to force interference disputes first into arbitration and then, as a fallback, have the regulator adjudicate the dispute.

The electrospace approach is intended to resolve the ambiguity by anchoring the definition of the rights more firmly in a technical foundation that defines the rights. The closest current application of this approach has been adopted in Australia for the so-called "spectrum licences" which define the allowed behaviour of Tx and require all Tx to be registered and were first adopted in 1992 for auctioned spectrum. As explained earlier, knowledge of the location and allowed Tx behaviour of every potential Tx in the environment allows potential entrants and existing incumbents to model the existing RF environment. Further refinements to this approach to emphasize service and technical neutrality (subject to the minimal requirements to ensure adequate interference protection) are being advocated by Michael Whittaker of FuturePace.Footnote 129

Analogous (but slightly different approaches, as will be explained) are being advocated in the U.K. by OFCOM as part of its Spectrum Usage Rights (SUR) proposalFootnote 130 and, in the U.S., by Robert Matheson at the NTIA.Footnote 131 In all three cases, the proponents are recommending that spectrum be partitioned into electrospace volumes (cells) which may be occupied by Tx. In either approach, the party interested in deploying a new wireless system needs to acquire sufficient transmission rights to allow his Tx to operate, given the existing set of rights (and resulting interference environment into which the new system will be added). Licensees may trade and consolidate licences but these may not be sub-divided below the minimal granular electrospace volume – each of which may be held by, at most, a single licensee.Footnote 132 This approach provides a clear way to subdivide the exclusive spectrum access rights into sufficiently "bite-size" pieces to allow flexible trading and reconfiguration of licences to address the challenges of allowing flexible usage and technology choices by market participants.

Also, adjacent licensees are free to negotiate changes to address the need to alter the energy emitted into adjacent areas. These private market negotiations and their outcomes are subsequently registered with the license registrar to update the current allocation of licences. This provides further flexibility for addressing changing market and technology conditions.

While all three approaches focus on explicit Tx rights, they imply implicit Rx rights since they both give licensed Tx the right to transmit within their licensed electrospace, but also limit the energy they are allowed to radiate outside of their licensed area. The logic for focusing on Tx is both practical and logical. It is practical because it is easier to locate Tx and control their behaviour directly than it is to locate passive Rx. It is logical because Rx do not radiate energy and so cannot be said to cause interference to others. In any case, a framework of explicit Tx rights (and implicit Rx rights) gives both Tx and Rx the information needed to predict the interference environment. The lack of explicit interference protection rights for Rx poses a challenge for legacy Rx such as traditional television receivers. As an interim/transition mechanism, it may be necessary to provide some explicit protection to select classes of legacy victim Rx in certain bands. If such an approach is adopted, it should be a temporary approach to facilitate the movement to a more simplified framework based solely on Tx-based rights.

Opponents of the electrospace approach argue that adopting too granular an approach may enhance flexibility at the expense of incurring higher transaction costs that would pose a barrier to new entrants, especially ones interested in accessing spectrum for only a short period of time (e.g., dynamic spectrum access). It is likely to be most attractive to wide-area service operators seeking to acquire spectrum to deploy long-lived, wide-area networks.

Proponents respond that transaction costs ought to be relatively low, assuming sufficient spectrum is allocated according to this framework so that the secondary markets are adequately liquid and there are no market power problems (i.e., spectrum is kept artificially scarce by incumbents or speculators to earn monopoly profits). Furthermore, they argue that entrants who are unwilling to pay market prices for spectrum may be better served by accessing unlicensed spectrum or by sub-leasing spectrum on a more dynamic basis from primary licensees.

The key difference in the OFCOM/Matheson vs. Whittaker approach rests in how Tx rights are specified.Footnote 133 Whereas Whittaker recommends specifying the property rights solely in terms of the Tx behaviour at a point in space, where the Tx is located; OFCOM and Matheson recommend specifying the energy flux over the entire service area of a Tx. In both cases, the licences are specified solely in terms of allowed Tx behaviour. Whittaker argues that the Matheson/OFCOM approach, which specifies power limits over an area is more ambiguous because it leaves the owner of the Tx at risk of future litigation associated with interference disputes. To address this challenge, operators may be overly cautious, resulting in excessive allocations of guard bands. In contrast, Whittaker argues that his approach has the virtue of clear and unambiguous rules for allowed Tx behaviour since they focus solely on the power emitted at the candidate Tx. According to him, this renders the risk of subsequent interference disputes mute because it is a simple process to verify that the Tx is in compliance with its license (and hence, is absolved from any interference claims). Under his approach, all that the operator needs to do is engineer his system to address his target receivers, ignoring other possible receivers.

Proponents of the OFCOM/Matheson approach argue that it comes closer to the ideal of mapping the energy flux caused by Tx and provides a richer model for more closely mapping the assignment of property rights to the core goal of limiting interference protection. Under this approach, once an operator has acquired the requisite spectrum licences for the energy flux he expects to generate over his serving area, with whatever configuration of Tx the provider wants, the provider/licensee may be assured that his system will be protected.

In both cases, the application of the approach requires recourse to RF propagation models. These models are substantially better today than they were even a few years ago, and ample data is accumulating to provide detailed terrain maps and incorporate other environment data to populate these models and allow quite realistic predictions of the energy flux emitted by a radio system over a prospective serving area. However, even the best of these models are only approximations and the extent to which they accurately predict actual behaviour varies by model, locale, and radio system. Thus, the granular approach to defining property rights for spectrum access on the basis of technical behaviour retains an element of ambiguity (since propagation modeling and spectrum measurement technology continue to evolve). Whittaker argues that adopting his approach leaves it up to each Tx operator to choose the level of detailed propagation modeling he wishes to adopt, thereby making market-based trade-offs between acquiring excess spectrum or investing in more careful modeling up front. In contrast, OFCOM responds that it can either pre-specify the propagation model to be used and certain parameters to populate the model (e.g., define the reference victim Rx using industry standards from the ITU or elsewhere). Instead of relying on a modeling approach, OFCOM suggests it may be possible to rely on actual spectrum measurements to identify out-of-licences, non-conforming Tx behaviour. Because of the statistical nature of RF propagation behaviour, a measurement protocol identifying the multiple measurements and the assessment of probabilities of fluxes above threshold values would need to be defined as part of a measurement test. The measurement approach has the merit of providing more certain ex post ways to assess actual interference problems, but potentially at the expense of some ex ante risk as to the reliability of earlier propagation modeling. In contrast, the modeling approach substitutes higher up front modeling costs for potentially less accurate identification of interference. Obviously, the two approaches may be combined to balance the relative merits of each. In response to Whittaker, OFCOM suggests that while the problems of using imperfect RF propagation models is real, this problem is not avoided with Whittaker's approach, contrary to his assertions to the contrary. Thus, the OFCOM approach offers greater flexibility with no greater risk of subsequent litigation/costs to resolve interference problems.

Relative to the simplicity of the Guatemalan approach, the Whittaker/OFCOM/Matheson approach suggests a more complex transition but one that, if successful, is likely to be more enduring and to more closely approximate the challenges of designing a flexible framework for market-based spectrum management that will be robust to change.

Choosing between the Whittaker v. OFCOM/Matheson approaches is more difficult. The Whittaker approach is closer to the Australian system which has over two decades of practical experience without major problems, although it is arguable how close that experience is to the challenges we anticipate confronting spectrum managers in coming decades, when we expect substantial growth (and the risk of congestion) from all kinds of wireless networks. Partially due to continuing legacy restrictions and a lack of liquidity (the causes of which are not precisely known), the Australian secondary markets have not been very active – a problem hardly unique to Australia. In contrast, the Matheson/OFCOM approach offers a richer model that appears more flexible in the types of market trading that it might support and engender, which may be important to accommodate the significant growth and increased complexity/heterogeneity expected from future wireless systems and business models.

More importantly, it may be that all of these approaches reflect too aggressive a move towards a tradable property-rights regime for Canada. Both the U.K. and Australia differ in so far as they have a relatively freer hand to manage their national spectrum autonomously. This makes it relatively easier to adopt a major paradigm shift. In contrast, Canada has to closely coordinate much of its spectrum usage with its neighbour to the south since so much of Canada's population resides within 200 miles of the U.S. border. In light of the somewhat slower progress towards full-blown flexibility to trade spectrum (with near-true service and technical neutrality) in the U.S., there may be added risks in Canada that are not faced by U.K. and Australian regulators (or even Guatemalan regulators).

Footnotes

Footnote 127

The TUFs are issued for a period of 15 years and are renewable for an additional 15 years.

Return to footnote 127 referrer

Footnote 128

See, for example, Hazlett, Thomas, Giancarlo Ibarguen, and Wayne Leighton (2006), "Property Rights to Radio Spectrum in Guatemala and El Salvador: An Experiment in Liberalization," mimeo (available from: http://web.si.umich.edu/tprc/papers/2006/652/Hazlett-Ibarguen-Leighton%2009_28_06.pdf).

Return to footnote 128 referrer

Footnote 129

See Whittaker, Michael (2006), Flexible Radio Spectrum Access: Moving from Device-centric to Space-centric Management, FuturePace Solutions, ISBN 0–9975232-0–9, March 2006 (available at: http://www.futurepace.com.au/).

Return to footnote 129 referrer

Footnote 130

See OFCOM (2006), "Technology and usage neutral access to the radio spectrum," Consultation document, December 2006 (available at: http://www.OFCOM.org.uk/consult/condocs/sur/).

Return to footnote 130 referrer

Footnote 131

See Matheson, Robert and Adele Morris (2007), "The Technical Basis for Spectrum Rights," draft paper, May 3, 2007 (an earlier version of this paper was presented at the IEEE DYSPAN conference in April 2007, see www.ieee-dyspan.org).

Return to footnote 131 referrer

Footnote 132

The minimal "unit" spectrum volumes may be different in different locations (e.g., larger in rural areas) and frequency bands (e.g., larger bandwidth at higher frequencies).

Return to footnote 132 referrer

Footnote 133

The discussion that follows seeks to capture the key themes of an active and on-going debate engaging folks representing all three perspectives via email over the past two months. The debate has yet to be resolved and parties have refined their positions over time as they have come to better understand the key differences. Although it appears the OFCOM and Matheson proposals are closest, there is no formal regulatory proceeding or recommendation seeking to implement Matheson's approach in the U.S. akin to the formal Spectrum Usage Rights (SUR) proposal advocated by OFCOM in the U.K. Similarly, Whittaker is advocating his approach which is an extension of the Australian current framework internationally, but it has not yet been formally adopted by a regulatory agency.

Return to footnote 133 referrer

Date modified: