Economists have for a long time taken a keen interest in governments assignment of licenses to use the radio frequency spectrum. Traditionally, governments have managed the radio spectrum in a way that has been compared to the regime of GOSPLAN, the famous central planning office of the former Soviet Union (Hazlett, 2003). Economists since Ronald Coase (1959) have argued that allocating a scarce resource by administrative fiat makes little sense. Establishing a market for spectrum, in which owners could buy, sell, subdivide and aggregate units of spectrum would lead to a much more efficient allocation of it. The use of auctions for assigning usage rights to spectrum have been heralded as one of the major achievements of the economical science, and the challenge of developing auction formats for spectrum auctions has been among the major driving forces behind the surge in research related to auctions in the last two decades.
In parallel with the introduction of auctions, telecommunication markets have been liberalised. This process of liberalisation has exerted an increasing amount of pressure upon spectrum managers to modernize their practices and to implement a more market based approach to radio spectrum management. The challenge is to assign spectrum in a way that maximizes the value of its use, a problem which ranges far wider than merely picking the auction format among those described in economic theory that is most likely to yield a favourable result.
Designing an auction procedure entails making decisions about what to sell, to whom, when, and how; these decisions are, as explained by Milgrom (2004), not independent ones. Among the problems facing the spectrum manager are identifying the appropriate regime of property rights, dividing the spectrum into suitable units for sale, and specifying any additional restrictions on its use and on the set of eligible buyers. The processes of defining the property rights and of dividing it into units are particularly closely interrelated with the choice of auction design. For example, by deciding whether the spectrum should be sold as a single large unit or as a number of smaller units, the spectrum manager determines the range of possible auction formats. The design of the usage right can create complementarities between the lots in the auction, affecting how different formats could be expected to perform. Economical efficiency is often stated as the main objective of spectrum management, although other objectives are frequently formulated by the government as well. The range of policies open to the spectrum manager may be constrained both by national and international law. The spectrum manager is responsible for a resource which is used in several sectors of the economy, and it can be challenging to evaluate how spectrum management policies are best implemented to interact with the other instruments used by the government in pursuance of its objectives in those sectors. The question of whether spectrum management should be used as a tool to affect the market structure in markets where spectrum is being used as an input is a question which illustrates the requirement for coordination between various national government agencies.
The emergence of secondary markets for spectrum affects both how important it is to achieve an optimal initial allocation (Milgrom, 2004) and how different auction formats will work (Krishna, 2002).
In this thesis the main objective is to describe an assignment procedure carried out by the Norwegian Post and Telecommunications Authority (NPT) in 2004 and to evaluate its outcome. My emphasis will be on the process of deciding what to sell, and how. In doing so, I hope to be able to explain more in depth the challenges facing the spectrum manager and to relate these challenges to the development of a novel variant of a successful auction format, the Simultaneous Multiple-Round Auction (SMRA). The new variant, which has been described as SMRA with augmented switching, (Cramton, 2006; OFCOM, 2006b) was proposed by Professor Nils-Henrik M von der Fehr, and was developed and implemented in collaboration between the NPT, Professor von der Fehr and DotEcon Ltd, a UK based consultancy. It has generally been considered as a success and the variant is due to be implemented again in forthcoming Norwegian and Swedish spectrum auctions.
The auction concluded December 1st 2004 after 87 rounds and 8 days of bidding. As will be described more thoroughly below, the auction appeared to work well, as indicated by the relatively high prices that were generated and by the fact that bidders were able to aggregate seemingly efficient combination of lots.
Unlike many other areas where economic theory is applied to real life situations, it is unlikely that all potential readers of this thesis will be able to rely on their previous knowledge about the resource that is being discussed. Most people have quite extensive knowledge about the nature of (and property rights to) other resources such as land, labour and tangible factor inputs. While some have a little knowledge about the electromagnetic spectrum, few know much about the nature of usage rights to spectrum, how it is actually managed or regulated and how its value is determined to a large degree by standardisation bodies, industry forums , and governmental and non- governmental international institutions.
In order to facilitate readers with varying degree of prior knowledge, I have placed some of the more detailed descriptions of e.g. the nature of the spectrum, its uses and how it is regulated into appendixes. I have tried to organize the main sections in accordance with the logical sequence of the spectrum manager s problem solving process , which leads up to a particular assignment procedure and choice of auction format. The main deviation from this approach is that I have inserted a section containing a description of the Simultaneous Multiple Round Auction (SMRA) before I describe the process of dividing spectrum into lots for an auction. The reason for this is that the properties of auctions in general and this format in particular, are often taken into account when decisions are made on the size and bundling of lots for the auction.
Those who know nothing about the radio spectrum and how it is managed could begin by reading some of the appendixes. The analogy between usage rights to spectrum and usage rights to land will be frequently used. Even though land is normally thought of as being owned in distinct and disjoint plots (at least they are disjoint as long as you restrict your attention to the part of the two-dimensional sphere they occupy) and spectrum usage rights normally are designed in such a way that they overlap in the physical dimension in which they are most frequently described, the analogy is still useful for many purposes. The validity of the analogy does (of course) not rely on the two concepts being exactly equal. I believe that lawyers concept of property as a bundle of rights is a useful complement to the general concept of exclusivity and the economists concept of externalities. The three will enable one to think somewhat intuitively about the properties of usage rights to spectrum.
More appendixes could have been included. In his 1959 article The Federal Communications Commission , Ronald Coase presented the argument for an efficient market based on property rights in radio spectrum. This article was an important predecessor for the 1960 article The Problem of Social Cost in which he states that:
This article, although concerned with a technical problem of economic analysis, arose out of the study of the Political Economy of Broadcasting which I am now conducting. The argument of the present article was implicit in a previous article dealing with the problem of allocating radio and television frequencies ("The Federal Communications Commission," J. Law and Econ., II (1959)) but comments which I have received seemed to suggest that it would be desirable to deal with the question in a more explicit way and without reference to the original problem for the solution of which the analysis was developed.
The 1960 article has certainly been very valuable to me, in that it has enabled me to organize my thoughts about spectrum management. If it were permitted, I would have appended Ronald Coase s The Problem of Social Cost in its entirety.