The European Mobile 3G UMTS Process

4. Proper regulation serves to correct market inefficiencies. One of the goals of regulating the mobile market is to encourage competition, which prevents monopolization and cartelization. It also lowers barriers to entry, respects consumer preferences, and reduces costs to consumers. In order to allow for greater competition in the mobile market, Member States of the EU and other European countries have chosen between auctions and beauty contests to allocate licenses. In an auction, assets are transferred from sellers to buyers in a bidding process. Generally, auctions have a number of benefits. They raise substantial amounts of money for the government. Auctions give the government good information about the value of different uses of spectrum, may help to better allocate spectrum allotment in the future, tend to be quick, and are cost effective. Auctions encourage the efficient use of frequency by assigning the spectrum to the eligible party that values it most. They also keep the number of applicants down, which thereby reduces administrative costs. However, auctions may disadvantage consumers if governments fail to devise effective auctions. In contrast, a beauty contest, or comparative hearing, has government regulators selecting a winner based on its own determination of the potential pros and cons of the candidates.

5. Section II of this paper provides an overview of EU law on mobile telecommunications. Section III explores auction theory and its applications for the licensing of spectrum. Section IV examines how auctions and beauty contests for 3G licenses have faired in Europe. Section V offers an analysis of the impact of the licensing process on European telecom operators and consumers while Section VI concludes that a single simultaneous ascending auction for all of the licenses throughout Europe will provide for greater efficiency, and advocates for the creation of a pan-European telecommunications regulatory authority so that future processes of licensing spectrum in Europe prove to be more successful than the 3G process.

II. EU Mobile Telecom Policy

6. The structure of the EU drives the way in which policies become law. The Council of Ministers of the European Union can only act on the basis of a proposal by the European Commission (“EC”). The EC has two functions. First, it proposes legislation to the European Council. Under a delegated powers doctrine, the EC also issues the majority of EU legislation.[5]

7. EU policy in mobile telecommunications originated in 1987 when the Council of Ministers adopted a Recommendation[6] followed by a Directive[7] of the EC to coordinate the introduction of the Global System for Mobile Communications (“GSM”) into the EU.[8]

8. Nevertheless, the EU only adopted a more uniform member-wide approach to mobile telecom in 1994. That year, the EC issued the Green Paper on a Common Approach in the Field of Mobile and Personal Communications in the European Union (“Green Paper”).[9]

9. One purpose of the Green Paper was to identify and implement the steps that needed to be taken to best exploit the mobile sector. The EC hoped to implement in mobile telecom what it sought to do in the other areas in telecom—to both liberalize and harmonize service. The Green Paper noted a number of barriers that needed to be overcome to reach these goals such as: exclusive and special rights that inhibit full-scale market development and prevent equal access; technology-based licensing that causes market fragmentation; nationally-oriented licensing procedures that substantially delay implementation of pan-European systems; and a fragmented approach to development and service provisions that prevents a global approach to the development of personal communications services.[10] As this article will show, these issues remain problems in the present setting.

10. The Green Paper’s main recommendation for overcoming these barriers was to extend EC telecommunications principles to the mobile sector while maximizing commercial freedom. Though the Green Paper made several specific proposals, such as abolishing special rights for mobile network operation, it did not specify a procedure for granting licenses. It also failed to advocate a EU-wide system of overcoming nationally-oriented licensing procedures. Of note, the Green Paper warned of some possible dangers regarding auctions such as “the danger of excessive transfers to the public budget or for other purposes, to the detriment of law tariffs for the users.”[11]

11. The EU Council supported the findings of the Green Paper with a Resolution.[12] Priorities that it listed in the Resolution included: (a) generalized competition for the provision of mobile and personal communications and granting of licenses according to objective, transparent, proportional, and non-discriminatory criteria; and (b) limiting the number of mobile and personal communications licenses granted only on the grounds of essential requirements, such as the efficient use of the frequency spectrum.[13] This Resolution can be contrasted with licensing procedures actually being followed in Europe at the time. Comparative administrative hearings were used to give incumbents (often government-owned) mobile licenses for free or at a fraction of their market price.[14]

12. As the technology to move beyond 2G arrived, the Commission responded with a Communication proposing the full application of the Treaty competition rules to the mobile sector and support for work towards UMTS.[15] To assist in UMTS policy, the EC and its UMTS Task Force set up a UMTS Forum, composed of telecom industry representatives. In concert with a report from the UMTS Forum, the EC issued a Communication that outlined policies to create a hospitable environment for the implementation of UMTS.[16] A key objective in this Communication was to ensure competition in the mobile sector to promote EU-wide harmonization for future systems. The European Radiocommunications Committee determined the particular spectrum space for UMTS in 1997.[17]

13. Two EC documents, the Directive on Licensing[18] and the Decision on the Introduction of UMTS, play a significant role in the current EU regulatory framework.[19] The EC recognized that the allocation of spectrum would play a key role in the success of UMTS. The Decision states the key concerns of the EC for the 3G technology:

[S]pectrum availability and appropriate pricing, coverage and quality will be essential aspects to the success of UMTS development…[A]ny spectrum pricing method should not adversely impact on the competitive structure of the market, and respect the public interest, while ensuring efficient use of the spectrum as a valuable resource.[20]

14. The Decision also reiterated the Mobile Directive of 1997, stating that Member States shall take all actions necessary to introduce UMTS by January 1, 2002 at the latest.[21] The Directive also added that licenses must be granted through “open, non-discriminatory and transparent procedures” and be based on objective criteria, which are articulated ahead of time.[22]

15. The EC provides a framework for pricing spectrum licenses. The prices need to reflect optimal use of the spectrum, be non-discriminatory, and foster the development of innovative services and competition.[23] A proposed directive will grant the EC the opportunity to propose binding harmonization measures using comitology procedures where it considers that diverging national measures constitutes a barrier to the single market.[24]

16. Because each Member State has some discretion as to the specifics of how to implement EU Directives and Decisions under the doctrine of subsidiary, licensing conditions regarding the UMTS licenses vary across states. Part of the problem of subsidiary has been that, traditionally, Member States have preferred minimal harmonization in telecommunications services.[25] Different selection procedures have led certain states to use auctions, while others have used comparative selections (beauty contests) or a combination of the two. Moreover, the UMTS licenses have varying time durations and enter into force at different dates. The deployment of networks and the ability to share some network infrastructure vary as does access to 2G networks.[26] As the EC notes, “Such fragmented conditions will inevitably create distortions in the way 3G networks and services will be implemented throughout Europe.”[27] Unfortunately, by the time that this statement had been written, the distortions it warned about had already been created, specifically in the licensing process and the creation of a sequential system that affected the bidding in both auctions and beauty contests.

III. Auction Theory

A. Overview

17. An auction serves to transfer an asset from a seller to a buyer. They can be defined as market mechanisms with explicit rules to determine efficient allocation and pricing on the basis of bids from market participants.[28] Countries have used auctions for selling licenses for spectrum beginning with New Zealand in 1990. A number of compelling reasons exist for countries to use an auction as a way to award a 3G license. Auctions lead to more efficient outcomes than administrative hearings. They are more transparent and objective than comparative hearings and are a flexible public policy tool. Auctions allow for governments to value a license best since auctions will reveal how valuable bidders believe the licenses to be and which bidder values them most. As a general rule, the social value of the license is equal to its most efficient valuation. Significant cost savings are possible in auctions over administrative hearings. In the United States, the FCC estimated that the cost of administrative hearings (the sum of the cost of filing the applications, the government’s administrative costs, and the public’s losses from delayed services) is six times higher than the cost of an auction.[29]

18. A number of different forms of auctions exist, though auctions can generally be divided into two basic frameworks: open or closed auctions, and ascending or descending auctions. Many variations exist within these auction forms. Within the framework of open auctions, there are two basic forms: English auctions (ascending bid auctions) and Dutch auctions (descending bid auctions).

19. English auctions are the best-known form of auction. In an English auction, the initial auction bid begins at a low bid price and ascends to higher prices until no new bids are made. Since bidding is public, English auctions allow bidders to revise their bids higher after observing the behavior of other bidders. In contrast, a Dutch auction is the opposite of the ascending auction and gets its name from the auctions used in the tulip market in the Netherlands. In a Dutch auction, the price of the auction is set at a high price and the bids descend to lower prices.

20. In closed or “sealed-bid” auctions, bidders submit only one bid. In a first price sealed auction, the highest bidder wins the bid. This type of auction has traditionally been used in government contracts. One problem with the first price sealed bid is that it might not elicit truthful bids or necessarily the bid that is the most efficient. Another form of sealed auction is the second price auction, also known as a Vickrey auction for its creator, under which the highest bidder wins the item but pays the price equal to the bid of the second highest bidder.[30] In sealed-bid auctions, it is more difficult for competitors to collude because it is difficult to respond to the signals of other bidders. However, this concern must be balanced by a concern that sealed-bid auctions will not necessarily lead to the most efficient outcome because it is more difficult to value licenses in the absence of seeing how much value other bidders give to the same licenses. Sealed-bid auctions also may make it easier for potential entrants to enter into the auction, thereby giving weaker firms a chance to win the auction.

21. Sequential versus simultaneous auctions are also an issue of auction theory and design. To the extent that a license in one country is seen as identical to a license in another country, a sequential process will create a sales mechanism in which identical items are sold for different prices based on when in the process they were sold. Licenses that are sold later in the process will go for less than those sold earlier.[31] Second, a sequential auction encourages strategic behavior as some firms might bid to drive prices higher to eliminate rivals that have smaller budgets from bidding in later rounds.[32] Sequential auctions also make it more difficult to predict the prices of future auctions and for this reason may eliminate a number of strategic alternatives for firms. These factors lead to outcomes that are less optimal than those achieved in simultaneous bidding.

22. In contrast, a simultaneous auction will encourage firms to incorporate more information into their bids and allows for backup strategies to be implemented during an auction. It also serves to limit winner’s curse and makes substitute licenses tend to sell for similar prices, since lower priced licenses will be bid up.[33] Other auction forms will not create an efficient aggregation nor will they create a setting in which similar licenses will sell for the same price.[34]

B. Early Spectrum Auctions

23. Though theoretically auction design might not matter, in application, the opposite has been true. Indeed, the design of the auction can be critical. The first spectrum auction—that of New Zealand in 1990—was unsuccessful in that revenue fell far short of expectations. A Vickrey auction was used but no minimum bids or reserve prices were set. As a consequence, strange results emerged. In one extreme case, one firm that bid NZ$7 million paid the second highest bid of NZ$5,000. In another, a New Zealand student bid NZ$1 for a television license for a small city. Since nobody else bid for the license, he won and paid $0 since this was the next highest bid.[35] Even with its failure to raise revenue, the New Zealand auction was still successful as the spectrum was assigned to the users who valued it the most, and the auction process was efficient in that it was quicker and less costly than administrative hearings would have been.[36]

24. A few years later, Australia used a first price sealed-bid auction for two satellite-television spectrum allocations. Because of a poorly designed auction, strategic games by the bidders created opportunities for securing licenses at lower prices. The auction did not have a minimum deposit requirement for bids nor did it have a default mechanism. Thus, the bidders bid high and low, then proceeded to default on the high bids so that the licenses were awarded to the next highest bids—which were also made by them.[37]

25. In contrast with the New Zealand and Australian auctions, the United States was able to create a more successful auction in 1994 by devising a simultaneous ascending auction in which all licenses were opened for bidding at the same time. As a study of the auction illustrates, “The evidence from the FCC auctions is that the simultaneous ascending auction is an effective mechanism for selling independent items.”[38]

C. Auction Design

26. As Part B reveals, auction design matters. The U.S. auctions serve as a good model for auction design, since its simultaneous ascending auctions suggest the best results. Both the simultaneous and the ascending nature of the auctions serve to aid efficient outcomes. Because the bidding is ascending, each bidder can observe how much rivals value the same license and therefore each bidder will know if it will be able to construct its preferred aggregation and how much its aggregation will cost. Since the auction is simultaneous, bidders can switch to back up aggregation strategies if the first choice looks like it will be too expensive.

27. A poorly-designed auction results in a number of different problems: the wrong firms might receive licenses; the government might not receive as much revenue as it could have; or the design might encourage firms to overbid to the extent that it ultimately affects the quality of service that the firm would provide to consumers. If a firm overbids, it is said to have suffered from winner’s curse. The curse arises in common value auctions in which, though unknown at the time of the purchase, the true value of the auctioned good is the same for all potential buyers.[39]

28. Secondary trading of licenses does not necessarily correct the problem of the initial misallocation. Economist Paul Milgrom notes that on a theoretical level, once property rights have been assigned, bargaining after the initial allocation will generally not achieve an efficient rearrangement: “An inefficient initial assignment cannot, in general, be quickly corrected by trading in licenses after the auction is complete.”[40] Particularly in the case of multiple parties, this becomes less likely.[41] A well-designed auction must also reflect the different preferences of bidders. “Even taking account of increases in the cost of capital and the small possibility of overbidding, the welfare gains from an auction are likely to outweigh these losses as long as the government can ensure collusion in the subsequent market is not possible.”[42]

29. It may be that high front-end costs that an auction creates will actually create incentives for a quick roll out of services. As the Financial Times quoted one analyst, “If you have got only 20 years to make the returns on [the investment on a license] stack up, then maybe you have got to roll it out more quickly.”[43] Economic theory supports this contention. A firm has the incentive to build a network because the firm must satisfy the capital markets and must therefore create a return on investment.

30. Another important factor in a well-designed auction is understanding and incorporating the possibility of complimentary licenses. That is, some licenses do more than just serve their area; they can be linked within a larger network of seamless services. For example, a telecom company with a license to serve the Washington, D.C. market would also want complimentary licenses to serve the Philadelphia, New York City, and Boston markets. In the UMTS context, a number of firms hoped to win licenses in all or most of Europe in order to take advantage of roaming services throughout Europe within the same network. Therefore, a consumer could be served via a network of licenses in Berlin, Rome, Paris, and London. Aggregation would also allow the fixed costs of a firm to be spread out more easily. Therefore, licenses in the different European countries are complimentary. It may be that the aggregation of these licenses exceeded the value of the sum of individual licenses in each country and that the lack of a single simultaneous auction prevented these aggregations from occurring. One potential pitfall of allowing combination bids is that it may create a free rider problem with regard to individual licenses. Another problem, as with any auction, is that it might create opportunities for bidders to collude.

D. Default Possibility of Licensees

31. Auctions have the possibility of creating a default situation if bidders for licenses overbid such that it bankrupts them. The NextWave case illustrates this point.[44] The FCC wanted to increase competition in the mobile sector by encouraging new entrepreneurial entrants. The FCC defined an entrepreneurial business as one that had gross revenues of less than $125 million in each of the last two years and total assets of less than $500 million. NextWave Communications (“NextWave”) qualified under this definition. NextWave acquired C Block mobile licenses in the United States at auctions in 1995 through bids that totaled $4.7 billion. Shortly thereafter, the FCC opened auctions for the D, E, and F blocks of PCS broadband. This effectively lowered the value of NextWave’s spectrum and made financing to build its network difficult, particularly given NextWave’s small size. NextWave entered into bankruptcy and after an ensuing legal battle, the FCC reclaimed the licenses under the automatic cancellation provisions of its power and reauctioned them.[45]

32. Bankruptcy concerns are related to auction design. Low minimum prices will encourage firms to view bidding as a call option since it guarantees that if they bet well, they will reap the rewards of significant profit from their license and if they bid poorly, they can easily default. A low minimum price will encourage under-financed firms to join in bidding though they may not be able to pay for the building of the network. This point is exemplified in the NextWave case. Bidders for C Block spectrum needed to put down only five percent of the winning bid price at the end of the auction and another five percent at the time of the license award. The remainder of the money was to come from quarterly installment payments at the ten-year Treasury rate with interest-only payments for the first six years.

33. Auctions are unlikely to lead to substantially higher consumer prices or lower investment levels than those of beauty contests. As a report to the Dutch Finance Ministry argues, “Where firms bid efficiently it is clear that small increases in the cost of capital will be largely outweighed by the gains in allocative and dynamic efficiency.”[46] The report finds that what causes inefficient increases in the price of services is tacit collusion by bidders after the auction has ended, and it concedes that the cost of capital is a possible exception to this corollary.[47]

E. Collusion

34. As noted above, auctions present the problem of possible collusion by bidders. Regarding the 3G licenses in the United States, a series of agreements among major U.S. wireless operators created the option to swap spectrum in key markets. These agreements potentially reduced pressure by the bidders to acquire spectrum at auction if the prices were to go too high since overpayment in one market would be compensated by spectrum in another market.

35. Collusion is also possible in the post-auction setting. If firms believe that through tacit collusion they will be able to raise prices (by asking for special breaks from the government in the cost of the building of services or through continued government support of state owned incumbent operators) that lead to extraordinary profits, this will lead to overbidding over the amount that would lead to an efficient valuation of the licenses based on a non-collusive price.[48]

F. Overbidding

36. Overbidding occurs when someone pays more in an auction than is necessary to achieve an efficient result. Some overbidding will occur based on reputation issues. Some managers will see the attainment of licenses as a must for the perception that a company has emerged from the auction as a “winner.” This is similar to how mergers and acquisitions often reach an amount beyond the amount that would lead to an efficient valuation.[49] Overbidding could have been stopped with an announcement and legislation prior to the bidding in which the European governments would have stated that auction results would be upheld and those that overbid would not be bailed out.

G. Factors for Incumbents and Why They Favor Beauty Contests

37. The differences between incumbents and new entrants are that incumbents face greater pressure than new entrants to obtain a license, often have the advantage of raw spending power, have large revenue streams, have some networks already built, and have a ready, available customer base as well as a branded identity.

38. Martin Bouygues, the CEO of French mobile operator Bouygues, claimed that mobile auctions presented a choice between “sudden death” and “slow death,” since incumbent operators have no other option but to make sure they take home a license. Otherwise they’re dooming themselves to go out of business.[50] There may be some accuracy in such a statement in terms of firms that received previous licenses for less than the real cost of the license and became used to higher than usual profit based on these licenses. Bouygues received its GSM license for relatively small annual payments and in a way that suggests government favoritism.[51] In such cases, the response by adherents to a market efficiency approach would be that there was an initial misallocation and that Bouygues should never have received a license at such a low price in the first place. Firms that had been granted licenses on a first-come, first-served basis had supranormal levels of profits at the consumer’s expense. They retained profits that consumers would have gotten via the payment for the proper value of the license if firms were forced to pay that price. This misallocation can be seen in share price value of telecom companies. A study by the U.S. Department of Commerce calculates that the cellular licenses that the U.S. gave away during the 1980s increased the recipient’s share values by $46 billion.[52]

39. Beauty contests, by their nature, are tacit attempts by the government to provide state aid.[53] There is no evidence to suggest that a firm that wins a license provides a cheaper product. Any rational firm will try to increase its price to what consumers will pay, even if the sunk cost of a license is low. Beauty contests may have pernicious secondary effects. First, the lack of transparency means that government-favored firms will be more easily able to win than firms that would provide the best service. Second, the favored firms (nearly always incumbents) can use the cost savings from the beauty licenses in one market to subsidize their entry into other markets that use the auction system since they will not be as burdened by a higher debt service that payment for licenses in an auction might cause.

40. Some firms fear auctions because they believe that they would hurt the firm’s bottom line. Michael Bon, Chairman of France Telecom, sums up the fear of many incumbent operators:

[T]he auction system is bad because those firms which already have a mobile license are obliged to obtain a license for the new generation. If they don’t manage to do so, it means the end of their mobile telephony activity, and the stock market value of their shares collapses immediately. If they want to avoid such immense devastation in value for their shareholders, they are obliged to participate in the auction. Under such circumstances, the latecomer determines the price of the license, since the others are obliged to follow indefinitely. The system is extremely destructive, since the latecomer can entertain motives unrelated to the industry and can also be totally wrong in its forecasts. It will cost that firm dearly, but the entire market will also bear the brunt.[54]

41. The fear of auctions is that the higher prices reached via auctions will translate into significantly higher end-user prices. The auctions would thus serve as an indirect tax on end-users who will have to pay via higher phone bills. This may have a chilling effect on the embrace of the 3G technology by consumers and could thereby threaten the European edge in the embrace of wireless technology. Yet, the ability to pass the increase onto consumers will be limited by the competition of the different mobile providers and the possibility of virtual operators that may resell mobile service.

42. These public fears by incumbent operators often reflect a skewed understanding of economics. Many analysts of this issue fail to distinguish between fixed and variable costs. The cost of the license is a fixed (sunk) cost because it is not part of the cost of supplying mobile service. The marginal cost of the firm will determine the amount of profit that will be made from the licenses. However, a high fixed cost may affect the marginal cost of a license. “If capital-market frictions mean that the more the firm borrows, the higher the interest rate it must pay, then the extra debt added by the auction price could result in the firm’s investing less and having a higher marginal cost.”[55] Indeed, an increased interest rate reduces a firm’s efficiency enhancing investments, thereby leading to a higher marginal cost. Though one study dismisses the size of this effect, the article was written before the bankruptcies of the 1994 U.S. C Block spectrum auction and did not foresee the massive debt that the UMTS auctions would create. Indeed, the UMTS auction proves that massive debt as a result of overbidding has put a number of telecom firms in a precarious financial situation in which the investment into their networks may suffer.

IV. UMTS Case Study: The European 3G Licensing Process

43. Member states of the EU and other European states began to allocate spectrum for 3G services in 1999. Though the process in each state had certain unique particularities, some common themes emerge.

A. Finland

44. The first country to offer UMTS licenses, Finland took a beauty contest approach in which each license for the 3G system was given away for free, with the only cost that of a negligible fee to cover administrative costs.[56] Finland is an interesting example of the beauty contest approach since, unlike a country such as France where the beauty contest was subject to the bias of the French government to protect its French incumbent operators and to generate government revenue[57], Finland was more concerned with high prices affecting the ability of high speed services to be rolled out. Given that Finland’s mobile penetration is the highest in the world (the number of mobile phones exceeds the number of fixed lines), the government hoped that lower fees would lead to a quicker roll out of the UMTS network than if operators had to pay a higher cost for the licenses.[58]

B. United Kingdom

45. The first auction for UMTS licenses in Europe was conducted by the United Kingdom. After seven weeks and 150 rounds of bidding by 13 companies, five companies emerged victorious in their attempt to win a 3G license, and the auction raised $35 billion.[59] Revenues were almost seven times more than originally expected.[60] The largest license was reserved for a new entrant. It sold for £4.38 billion. The next largest license, which was open to any bidder, sold for £5.96 billion. The remaining three licenses sold for slightly more than £4bn each.[61] Telecom operators have until 2007 to build their networks and must by that time be able to cover eighty percent of the population.[62]

46. A simultaneous ascending auction was used, because, with one more license than the current number of mobile operators, the possibility for competition was increased as some new entrants were enticed to enter the bidding. The timing of the auction was also important. Because this was the first auction, it made it more likely that a large number of entrants would bid for these licenses.[63] Bidders may have also thought that winning a British license would help to position them for subsequent UMTS auctions because the licenses would be complimentary to those that could be won in other European countries. This had the effect of raising more revenue than might otherwise have been raised.

47. A second effect of the British auction was that it created a two-stage game. Obtaining a British license in the first stage meant that it became more important to win a complimentary license in other major European countries in the second stage in order to justify the high price paid for the British license. Similarly, lack of a British license for some telecom companies meant potential preclusion from future license allocations since the synergy of complimentary licenses and pan-European roaming would not be realized. Therefore, successful high bidding in the British auction might mean lower bidding amounts for future licenses since competitors would drop out from future auctions.

C. The Netherlands

48. The sale of licenses through the auction process brought in approximately $2.5 billion for the Dutch government though, initially, the Dutch Finance Ministry had expected the auction to bring in more than $8 billion.[64] One problem of the auction in the Netherlands was that it used the same auction design as that used in Britain under a different set of circumstances. Because there were five licenses and five incumbents, there was little incentive for new bidders to enter bids since the incumbents had the advantage of preexisting 2G networks, some of whose components could be used in the 3G system. Indeed, potential entrants had strong incentives to partner with incumbents, which is what happened.[65] Only one new entrant, Versital, bid for the licenses, though it stopped bidding as a result of the actions of another bidder. Mobile operator Telfort, sent a letter to Versital in which it threatened possible legal action if Versital continued to bid.[66] Though the Dutch competition authority investigated an improper conspiracy to influence the outcome of the auction, no evidence of such a conspiracy was found.[67]

49. Another structural flaw in the auction was the pass mechanism the government set up. The government specified minimum opening bids that would decline in the first three rounds if firms “passed” and did not bid on a license. In each case, the bidders used a pass in each round to bring the reserve price down from 100 million guilders to 0 guilders after the three rounds.[68]

D. Germany

50. After the British auction, Germany increased the minimum prices for obtaining a license more than once, as it began to see the auction process as a potential cash cow for the government to raise revenue.[69] German Finance Minister Hans Eichel stated that the true meaning of UMTS, was “Unerwartete Mehreinnahmen zur Tilgung von Schulden,” or unexpected revenue for paying back the national debt.[70] As one telecom analyst described the result, “The prices reached were madness… The industry is being heavily taxed by what amounts to the [German] government’s shortsightedness and greed. Consumers and shareholders are going to learn quickly that there is no such thing as a free lunch.”[71] Since Germany is the largest country in Europe, the potential payoff for a German 3G license was significant.

51. Though many analysts prior to the auction believed that only five licenses would be awarded, the auction produced six licenses and two blocks of spectra. Each license cost approximately $7.6 billion.[72] Strategic games made the auction process more uncertain and helped to explain how the cost of the licenses skyrocketed to record heights. Each bidder needed to secure only two blocks to obtain a license but securing three blocks would prevent a competitor from securing an additional license. One study suggests that the eventual license costs were so high because of the willingness of the incumbents to try to preempt entry by bidding for three blocks, thereby forestalling entrance by other possible competitors.[73] If so, then the auction design created the possibility of a more concentrated mobile market of only five operators rather than a more competitive one of six. Indeed, two bidders, T-Mobil and Mannesmann, drove a significant portion of the auction. Both of these bidders attempted to win three blocks. However, this plan became untenable as the price of the bids soared beyond the 50 billion euro mark. Once the two changed their strategy and opted for only two blocks, the auction ended, as six licenses were available for the six bidders.[74] A second auction for additional capacity was opened to all bidders that had won at least two blocks in the first auction. Since this included all the winning bidders, this auction predictably proved to be a failure.

52. Post auction, a number of German operators had hoped to jointly construct the networks in order to reduce roll out costs.[75] Once these discussions among operators became public, Reg TP, the German regulator, stressed that each holder of a UMTS license would be required to build its own networks telecom for twenty-five percent of the population by the end of 2003, and for fifty percent by the end of 2005.[76] Reg TP also stated that each network needed to ensure its “competitive independence” during the lifetime of the UMTS license.[77] More specifically, telecom operators could share network elements such as masts, cables, and antennas providing they retained control of aspects of their networks that would allow them to shut down without affecting a competitor. However, the operators would not be allowed to share backbone facilities such as switching centers.[78] Aside from regulatory issues, there are certain physical limitations to network sharing. It is technically tricky to handle roaming from non-shared to shared parts of the network.[79]

E. Italy

53. Italy adopted a hybrid system including elements of both auctions and beauty contests. The first phase was a beauty contest in which the Italian government made preliminary selections of license bidders based on the strength of the operators’ business plans, followed by an auction among them for the licenses. Another unique aspect of the Italian auction design was that if there were not enough serious bidders after the beauty contest greater than the number of licenses, the government would reduce the numbers of licenses auctioned. Yet, such a policy does not necessarily guarantee a successful auction, because it does not properly allocate demand with capacity of spectrum.[80] As Klemperer correctly notes, this potentially served to create an unnecessarily concentrated mobile sector in order for the auction results to look successful.[81] Instead, the regulators should have taken steps to make the auction more attractive to entry to competitors, though perhaps this was more difficult given that the Italian auction occurred subsequent to those in Germany and the United Kingdom, making the financing of licenses more difficult for a number of potential bidders. For example, Deutsche Telekom AG (“DT”), which had paid a staggering amount for licenses in the United Kingdom and Germany, did not enter the Italian auction.[82]

54. Italy had hoped that it would duplicate the huge fortunes generated in Germany and the United Kingdom. Given that Italy’s mobile market is the largest in Europe, high bids might make sense if the belief was that Italians would jump to the 3G technologies. One Italian firm, Omnitel Pronto Italia SpA, revealed that it would pursue an Italian license at virtually any price.[83] After the UK auction, the Italian government concluded that its revenue forecasts for the auction were too low such that revenues were reforecast from $1.9 billion to more than $11 billion. The Italians became caught up in the hype of the increased over-valuations of the preceding auctions. It was also in the wake of the UK auction that Italy switched from a beauty contest format to the hybrid format.[84] Given the timing of the change immediately in the wake of the UK auction, this policy change by the Italian government seems to have been motivated by revenue maximization rather than any efficiency belief in the auction system.

55. Blu[85], one of the bidders, threw a wrench in the Italian hopes for a huge payoff when it withdrew from the auction. Blu’s Italian shareholders had wanted BT to take a higher stake in the company though BT was unprepared to do so, no doubt in part because of its growing debt as a consequence of its license costs in Germany and the United Kingdom.[86]

56. As a result, the Italian auction took place with only five bidders for five licenses. In response to the low revenue results, Italian Prime Minister Giuliano Amato convened a government meeting to determine whether it was possible to void the results and postpone the auction to a later date. To his great disappointment, he discovered that it was legally not possible. In response, the Italian government focused its wrath on Blu and sought to keep the $1.74 billion deposit Blu made before the auction in order to participate in it, which Blu finally won.[87]

F. Sweden

57. The Swedish government used a beauty contest format in which it charged a nominal fee of $11,020 for each 3G license and an additional yearly charge of 0.15 percent of income.[88] Interestingly, only in Sweden did a national incumbent not win a license. Though, Telia, the Swedish national operator, had secured 3G licenses in Norway and Finland, the Swedish Post and Telecom Agency (“PTA”) rejected outright Telia’s application without even considering it as a finalist for one of the four 3G licenses. The PTA made its awards based on two primary factors—the rollout and coverage commitments. Each of the licenses covers two 15 MHz paired spectrum blocks, plus 5 MHz of unpaired spectrum used for one-way data services.[89]

58. On the one hand, the explanation for an award against an incumbent, a national incumbent for that matter, may be that Telia only offered to commit half of the investment of those operators that won licenses. However, if Telia is to be believed, the PTA showed subjective behavior. As its spokesperson stated, “Telia is a leading mobile operator in the Swedish market with full national GSM-coverage. In light of this, it is very peculiar that PTA doesn’t seem to believe we’re worthy of a license.”[90] Had the license been auctioned, Telia would have been more likely to bid to gain a license rather than to believe that the government would support it no matter how low its bid seemed.

59. The decision by the PTA also concerned other parties. The PTA rejected the DT-led consortium based on “financial deficiencies” and the Sonera consortium in light of “technical deficiencies.”[91] DT, at the time, had a high credit rating and Sonera had won other UMTS licenses, so these arguments do not seem so credible. This may point to the general weakness of agency decision-making in license awards over that of an auction process because of the possible use of subjective criteria.

60. Concern has also been raised that the Swedish operators would not be able to roll out the networks in time.[92] If so, this suggests that a beauty contest did not create an opportunity to roll out service faster than an auction would have done. One effect that these significant costs may cause will be increased consolidation in the industry through mergers or joint ventures. In Sweden, two of the four groups that won 3G licenses are attempting to build a joint infrastructure for their networks, rather than to do so individually, to save on costs.[93]

G. Belgium

61. Belgium’s auction proved to be a failure. Its 3G licenses were sold at minimum prices to incumbent operators, each of whom submitted only a single bid of $138 million for the 20-year licenses.[94] This left the government well short of the $1.4 billion it had hoped to raise.[95] No new entrants submitted bids or even applied for the four licenses, only three of which were sold.[96] Since Belgium’s auction came late in the auction process, telecom companies may have been more worried about the significant debt that they had accumulated in earlier auctions. Another factor may have been the excessive fees that the government charged for participation in the auction. French operator Bouygues made a formal complaint to the EC about the Belgian prices, describing the auction process as “discriminatory” and the license prices as “disproportionate.”[97]

H. Portugal

62. In a beauty contest, Portugal awarded licenses to all three existing GSM operators and to one new entrant. The licenses were awarded for $85.58 million each and last 15 years. In contrast with some other countries, Portugal included a provision in its award mandating that an additional $768.4 million must be invested to build up each network before the official date for launching the new service in January 2002. The beauty contest results came as no surprise as the four winners had been ranked highest in the Portuguese regulatory authority’s November evaluation, compiled the month before the beauty contest results were decided.[98]

I. Spain

63. Spain initially allocated licenses in a beauty contest for a total value of $432 million in March 2000, prior to the UK and German auctions. After these auctions, the Spanish government decided to make up for lost revenue it believed it would have gained in an auction by imposing a fee on 3G operators. The Spanish 3G license saga proves an interesting example of how a government can blunder in such a way as to become a double loser.

64. The Spanish government argued that, in the wake of the UK and German auctions, the value of the Spanish licenses had increased significantly and, therefore, the Spanish government should be compensated for the new market value of the licenses. Spain had first believed in the beauty contest approach because it thought that in return for a low license price, it could obtain commitments by the license winners that 3G services would be rolled out ahead of those services in other countries by half a year, August 2001. In response to the new fee, the 3G Spanish telecom license holders argued for postponing the rollout for the new services because of the extra charges.[99] Of course, given the collapse of the UMTS license market elsewhere in Europe, the reasoning of the Spanish government has been turned on its head. If the market value of the licenses decreased to an amount lower than that at which the licenses were sold, this would suggest that the Spanish government should rebate part of the cost of the license so that the license holder would have paid an amount equal to the true market value of the licenses. Equally interesting has been that the other premise behind the Spanish regulators’ policy has fallen asunder. In Spain, service was to have been launched by August 2001 but has been delayed until June 2002.[100]

J. France

65. The French government set the price for its four 15-year licenses at the high price of $4.7 billion each.[101] France’s telecom regulatory agency, the ART decided on a beauty contest, with the official selection criteria based on: (1) the development of the market and new services; (2) coverage of the territory; and (3) the government’s aim to make the Internet widely available.[102] Yet, Jean-Michel Hubert, head of the ART, readily admitted that a more subjective criterion existed as well. He stated that part of the ART’s mission was to promote the interests of French corporations internationally. A beauty contest would better insulate French incumbent mobile operators from costly high bids that an auction would cause. ART feared that by choosing an auction process, current or aspiring French operators might be unable to match the bids of big, well-financed foreign rivals and lose the fastest-growing business in their core market.[103] Therefore, a beauty contest afforded a better opportunity to use subjective criteria—namely the championing of French companies over possible foreign new entrants—than an auction afforded.

66. The French government hoped to capture the financial gains for revenue based on a high beauty contest price but to do so without the uncertainty of the auction results. The French solution was therefore to combine the worst features of beauty contests and auctions. On the one hand, France would still award the licenses on subjective factors determined by the French regulators rather than by the market. On the other hand, a high entrance fee for licenses would make it more difficult for new competitors to enter the market. It would also increase consumer costs as license winners might try to pass on the extra cost of the licenses to consumers.

67. Predictably, the French beauty contest ended in disaster. When the bidding closed on January 31st, only two firms had submitted offers. Because of the failure of the first round to attract more than two bidders, France was forced to promise to hold a second round of bidding. This outcome should have come as no surprise to French regulators. In early June, the Chairman of France Telecom, Michel Bon, had warned that the entrance fee was too high.[104] Indeed, to date, none of the three French operators of 2G mobile systems have broken even on their investment though interest in mobile telecommunications, as measured by the number of subscribers, has soared.[105]

K. Austria

68. Less revenue was raised in Austria’s mobile auction of UMTS licenses than had been expected. In total, the auction yielded 706 billion euros, where double had been expected. Only six bidders bid for the six pairs of spectrum blocks, though bidding lasted for fourteen rounds.[106] The auction, which was an ascending auction and structured similarly to Germany’s, may have generated lower returns than it could have because of suspicious behavior by some mobile operators. The Austrian government twice halted the auction because of possible collusion. Yet, initially the government had said that it would call off the sale if it detected any signs of collusion.[107] The suspicious behavior was potential signaling of some competitors to others of their intent in the auction.[108] Before the auction, MobilCom told the Financial Times that it would welcome the customers of mobile telecom operator Debitel should Debitel fail to secure a license, thereby allowing Debitel to become a virtual network operator on the MobilCom network. This type of signal may have had the effect of prompting Debitel to withdraw from the auction before it started, leaving only six bidders. The Austrian regulators failed to investigate MobilCom’s behavior, perhaps because it would have threatened to remove a bidder from the auction, thereby reducing the government’s revenue.[109] Another 3G bidder, Telekom Austria, announced the week before the auction that it “would be satisfied with just two of the 12 blocks of frequency…and if others behaved similarly, ‘it should be possible to get the frequencies on sensible terms.’”[110] As a further signal, it stressed that it would bid for a third frequency block if other firms did the same.

69. As in the German auction, prices rose above what was needed to reach an equilibrium (the reserve price in this case) of the same number of bidders each winning a license because the auction was structured around pairs of spectrum rather than on licenses. Therefore, competitors tried to buy greater capacity to reduce the number of licenses.[111]

L. Denmark

70. Though the Danish government will hold an auction in the near future, its auction style will differ from that of the other UMTS auctions that have been held in Europe. Denmark plans to hold a sealed-bid auction for the four licenses it will allocate for 3G service. According to a working paper by the Danish IT and Research Ministry and National Telecom Agency, “In a sealed-bid auction, it is possible for an aggressive new entrant to outbid a conservative incumbent...It also has advantages in terms of reducing the potential for collusion.”[112] Yet, these arguments can generally be applied to any type of auction and are not unique to a sealed-bid auction. The auction will be a modified Vickery auction in which the winning bids will be priced at the amount submitted by the lowest winning bidder. To prevent tacit collusion for underbidding, the Danish government will set a floor price for the licenses.[113]

M. Switzerland

71. Switzerland’s auction suffered a defeat given its late start vis-à-vis the other auctions. Particularly, four of the nine qualified bidders dropped out while two of the remaining five bidders announced plans to merge just minutes before the auction, thereby further reducing the number of bidders.[114] The move left only four bidders in the race for four licenses, and after an investigation into whether the merger agreement had contravened auction rules, the Swiss government was forced to proceed with the auction. Though the Swiss government gained SFr205 million from the license sale, this was a mere fraction of the SFr2.5 billion it had originally expected from the process.[115] This low revenue result was also caused by the new financial reality that the UMTS auctions had helped foster, namely running up significant debt made gaining later licenses more difficult.

72. The specter of possible collusion also emerged from the auction process. One story suggests that DT, which did not take part in the UMTS auction, cut a deal with Tele Danmark to leave only four bidders for four licenses.[116] The motivation for such a move would be DT’s desire to acquire part of Tele Danmark in the future after Tele Danmark had acquired a license not inflated by competition.[117] The week before the auction the number of bidders dropped from nine bidders to four bidders. The final average price of the licenses was only 2½ percent higher than the minimum bid.[118]

N. Norway

73. Norway first envisioned 3G services in its White Paper to the government on this topic.[119] Norway allocated its four 3G licenses by means of a beauty contest. Seven telecom-bidding groups applied for the licenses. Each license was awarded for $11.2 million each, plus an additional $2.2 million a year.[120] The beauty contest emphasized geographic coverage and roll out.[121]

O. Poland

74. Mobile licensing in Eastern Europe has been similar to that in the EU. The outcome of the 3G license process in Poland, the largest country of Eastern Europe, yielded poor results. The Polish government was forced to annul its UMTS tender after it determined that it would be received poorly. Poland thereafter awarded its three mobile incumbents 3G licenses through a 650 million euro extension of their GSM licenses. The Polish government had hoped to award five licenses at 650 million euros each but only the three cellular incumbents filed tender papers. Poland had seen the 3G license process as a way to raise revenue. Indeed, the Polish government had included the expected revenue from the license sale in the country’s 2001 budget.[122]

75. The lack of interest may have resulted from a high cost for licenses, given that Poland’s telecom services are less developed than those of Western Europe. Its fixed-line penetration rate is just thirty-five percent while its mobile penetration rate is a mere fifteen percent.[123]

P. Czech Republic

76. The Czech government has opted for a hybrid approach in which it will use separate auctions and beauty contests. The existing 2G operators will be offered UMTS licenses for $135 million each. The fourth license would be sold in an auction. If any of the existing 2G operators choose not to purchase a license, the unsold license will be auctioned in addition to the fourth license.[124]

V. Implications of the UMTS on European Mobile Operators and Customers

77. The cost for building the 3G networks in Europe is estimated to be an outlay of $150 billion. This figure adds an additional fifty to one-hundred percent to the cost of obtaining the license, depending on whether the license winner is an incumbent that can add to its 2G network or a new entrant that will have to build a new network from scratch. [125] This, of course, is in addition to the money that will be spent on securing the 3G licenses as well as the costs for marketing the new services.[126] Some incumbent fixed costs are already sunk since they can use significant parts of their existing 2G network for 3G services.[127] Therefore, there is inherently an uneven playing field between incumbents and new entrants that may further hinder new entrants in markets and put increased pressure on their governments to potentially bail them out or give them implicit concessions.

78. With such high costs, some auction “winners” may turn out to be losers in the long run. The high auction fees might reduce spending on equipment because license holders might not have enough to build their networks without incurring a staggering debt and a possible downgrade of their ratings which would lead to a lower return on investment. Ultimately, 3G may turn out to be a poor investment. Goldman Sachs Internet Analyst Rajeev Gupta argues that 2G services will suffice for most consumers and that 3G services will not catch on once the additional costs of 3G services are factored into a phone bill.[128] It may also be that telecom companies overvalued the benefits of 3G for consumers and the potential gains from its use. The high valuations of the stocks of European telecom operators have crashed dramatically since the auction process, in part due to the consequences of the spectrum process, making it that much more difficult for these companies to raise revenue.

79. Accumulating additional debt to finance the cost of licenses and their building has significant consequences. More debt means poorer credit quality for firms. As telecom companies go further into debt and lose revenue to pay for debt service, their credit ratings drop. The effects can be quite dramatic for these telecom companies. Bonds with a BBB rating pay investors a yield of about a half a percentage point more than those rated A. This means that when a telecom company is downgraded its debt becomes a significant burden, leading to a larger amount of debt that needs to be repaid because of the increased interest on the debt.[129] Equally significant, a drop in the rating of the company has secondary effects. Certain institutional investors cannot hold bonds that carry a rating worse than A, so that telecom company bonds at that level will have to be sold.[130] This selling has occurred at a time in which telecom operators who have won licenses are in need of new bonds to finance the purchase of the UMTS licenses and the roll out of networks. As the Economist notes, “A surge in supply would force up yields and thus financing costs. These have already risen sharply as investors have pre-empted the rating agencies’ move.” [131]

80. For example, British Telecom (“BT”) has been crippled by a $43 billion debt as a result of the cost of the UMTS auctions. Standard & Poor has placed it on credit watch, with other UMTS license winners France Telecom and DT, since February 2001. The current economic situation makes it difficult for the firms to reduce their debt levels. BT has about $9 billion in bonds that have a “step-up” clause that triggers increased interest payments if BT’s credit rating were to drop below A. The consequence of the step up would be additional bond service payments of between $200 million to $250 million per year.[132]

81. Lower bond ratings also affect some previous bond issuances by telecom companies. Some companies such as DT and Vodafone AirTouch have included clauses in recent bond issuance that automatically increase the interest rate on the bonds if the companies are downgraded to a certain level. In the case of DT, the coupon payable on its $14.5 billion bond issues would rise by half a percentage point if Moody’s and S&P were both to downgrade the company to a BBB+ rating.[133]

82. Debt issuances for telecom companies are crucial since equity financing through increased stock no longer is a viable option given current market conditions and the dramatic drop in the valuation of the stock of telecom companies. First, the telecom sector has been battered as a result of the decline of technology-related issues in the United States and Europe. Second, telecom company equity is already a significant part of the makeup of the U.S. and European stock exchanges. Telecom companies have taken on significant debt to pay for their licenses. Only two years ago, telecom companies accounted for 7.4 percent of company debt in Europe, though the figure is now closer to 35 percent.[134] The possibility of a return on the investment for licenses is not promising for a number of years, suggesting overbidding. As Moody’s stated in April 2001 regarding 3G in Europe, “The break-even point for (these investments) is not likely to be reached before the fourth or fifth year of operation in the most optimistic scenario.”[135]

83. Some telecom companies now readily admit that they have overpaid for their licenses. BT’s CEO Peter Bonfield said BT overbid on licenses. A BT spokesman elaborated on these comments by saying that, “BT paid £10 billion more than it should have...especially in auctions in the UK and Germany,” adding that, “spending had a huge impact on the industry [landing BT with] debts of £30 billion.” A consequence of BT’s overpayment is that its interest payments on debt tripled in the third quarter of fiscal 2000.[136]

84. A direct government bailout of a troubled telecom operator is not possible since it would be considered state aid, which would violate EU policy. Indirectly, a government could bail out a mobile operator through the guise of promoting accessibility of broadband.[137] Increasingly, the possibility of an indirect bailout of operators is becoming more likely.[138] One member of the European Parliament has authored a plan that calls for governments that conducted 3G auctions to develop an annual payback scheme for license fees and guarantee loans for licenses, which would result in lower interest rates for the operators. The plan would also provide for the European Investment Bank to grant special loans with lower prices for the mobile sector.[139]

85. The cost of the licenses may also have the effect of greater consolidation within the European telecom sector. Mark Page of Booz Allen & Hamilton, a consulting firm, argues that “continuing consolidation [in Europe is] a strategic and financial necessity.” Given the nature of the telecom industry and the culturally diverse markets of Europe, it is difficult for telecom operators to push down sales and marketing costs. Page advocates returnable and refundable licenses.[140] Such licenses would be more attuned to the market and would allow consolidation to be sped up.

VI. Conclusion

86. There are significant lessons for companies, government leaders, regulators and lawyers to learn from the 3G process. In the new 3G world, one telecom analyst noted, “Wireless operators are realizing they have to work together.”[141] Therefore, the EU must be vigilant that competition remains strong and that the incentives that firms have to collude in ways that hurt consumers are minimized. If encouraging new entrants is the concern, there are non-auction rule mechanisms that will increase this outcome such as license fee payment by installment, and mandated network sharing and roaming.[142] Auction “losers” might still be able to enter the 3G markets by renting some spectrum from auction winners as virtual mobile operators.[143] Contrary to 3G licenses in the United States and Japan, the European UMTS licenses cannot be used to provide basic mobile telephony. Allowing this change might encourage competition by encouraging new entrants.

87. Another key policy issue in 3G licensing conditions which could alleviate the financial burden of mobile operators is the secondary trading of spectrum. Secondary trading, where operators resell part of the spectrum they have been granted by licenses, is not allowed in Europe, unlike in Japan or the United States. Secondary trading serves to recalibrate the licenses to those who value it most.[144] The 1999 Communications Review advocates that Member States allow, but not mandate, secondary trading.[145] So far no country has done so though the United Kingdom is more advanced in discussions regarding allowing secondary trading than other countries.

88. In auctions themselves, steps must be taken to reduce the ability of firms to collude. This should include round number bidding, which will reduce the signaling mechanism that firms would use to protect certain markets from competition.[146] High enough reserve prices must be set to discourage under-financed bidders and to ensure that enough funds are raised through the auction process. Regulators must make sure to set reserve prices that are not too high (as was the case in France) to chill possible competition. One lesson that telecom companies seemed to learn given the high cost of early auctions was that it was possible to get access by buying a stake in a license winner without directly bidding for a license and taking on the cost of building a network, or by buying wholesale access from the license winners. This sort of behavior should be chilled by strong activity rules in auctions to prevent joint agreements and create a setting with the most possible bidders.

89. Because the licenses went up for tender bids at different times in each individual country, in essence, what occurred in Europe was the creation of a sequential auction. Such auctions have significant problems, as noted in Section III. In a sequential auction it becomes easier for parties to collude the more times an action is repeated. Therefore, a single European-wide auction would have reduced the potential for firms to collude across the auctions in a number of different countries. The sequential nature of the licensing process made it difficult for competitors to undertake second best strategies to ensure coverage throughout Europe and led to a discontinuity in pricing from licensing contest to licensing contest, making prediction of actual purchase prices more difficult.

90. The sequential process also hurt licensing contests that were held later and benefited those held earlier both in terms of the number of entrants and the amounts that they thought to bid. Incumbents faced pressure to win a 3G license at almost any cost since a failure to win would signal significantly lower growth potential to investors. At the time of the initial auctions, financing was easy to come by. This fed into a bidding frenzy since both new entrants and incumbents had significant coffers with which to increase their bids. Later on, as funding dried up, a lack of potential bidders meant that it would be difficult to repeat some of the allocative successes of the earlier auctions. The relationship between the financing of telecoms and the ability to provide service based on a roll out of a new technology has been under-explored in the current UMTS discussion. Indeed, it is key since the failure to impose a European-wide auction had the effect of overbidding in some countries and misallocation of resources for telecom operators. As a result, their marginal costs reflect a higher interest rate on debt service, rather than keeping the license costs as sunk costs.

91. For some academics, the auctions offer different lessons. Klemperer, the designer of the British auction, believes that the lesson from the auctions is that a one-size-fits-all approach for auctions does not work.[147] Yet, he does not connect the fact that the market sentiment for 3G auctions was inaccurate in part because possible entrants did not have a clear sense of the best strategy to pursue since the auctions were sequential rather than simultaneous. Understanding that the different licensing processes in the entire EU played a role in the outcome of each individual license process, whether beauty contest or auction, is central in designing a better regulatory framework for the next set of spectrum assignments.

92. Some changes are possible. This inefficient allocation of spectrum suggests that the time may have come for a pan-European telecommunications authority.[148] In its 1999 Communications Review, the EC examined the issue that “[i]nconsistent application of certain provisions of telecommunications legislation is hindering the development of effective competition and the deployment of pan-European services.”[149] At the time, the Commission found that the potential costs of a new agency outweighed benefits conferred. Given the hundreds of billions of dollars at stake regarding 3G licensing and implementation, a European Regulatory Authority might have been able to achieve more efficient outcomes regarding the UMTS licensing process. It could have pushed for a simultaneous auction of all the licenses. At the time, there was not sufficient support for such a step.[150] Yet even in the report, the possibility of pan-European licensing was advocated in some form. “It is essential to recognize the merits of pan-European licensing and possibility of radio spectrum assignment….”[151] The possibility of such an agency remains bleak at the present. An attempt to expand the powers of the EC over those of national telecom regulators failed in the spring of 2001.[152] Even this action is still conceptually a sine qua non of a possible European telecom regulator and could not find the political support for passage.

93. The lack of a pan-European regulator is a combination of a coordination and distribution problem within the dilemma of a common interest framework as described in international relations scholarship on the incentives to establish and create international regimes.[153] In an important sense, almost any pan-European regulator would be an improvement over the current system where no common scheme results in increased transaction costs. The coordination problem can be understood in a “battle of the sexes” in which both partners in a couple would like to do something together but disagree on the preferred outcome. The man may want to go to the country, while the woman may want to go to the beach.

Payoffs for Man/Woman

	Country	Beach
Country	5/3	0/0
Beach	0/0	3/5

94. Both Country-Country and Beach-Beach (shaded in gray) outcomes would lead to a Nash equilibrium for this game. In order to achieve either of these outcomes, the two sides would need to determine how to spend time with each other. Otherwise, if each player would hold out for the equilibrium that they preferred most, the result would be an outcome that neither player desired, where both were worse off than they needed to be.[154] Similarly, in Europe a pan-European regulatory regime would be better than no regime. European governments need to coordinate to decide on the structure and the role of such a regime to implement the type of regulatory supervision on the auction process that is necessary to achieve the best results.[155] As 4G technology, a successor to 3G, becomes a reality, a new system must be put into place for a more efficient, European-wide allocation of spectrum through a simultaneous ascending auction. Only a common regulatory approach through a European telecom regulatory agency can make such a change a reality.

95. As 4G technology, a successor to 3G, becomes a reality, a new system must be put into place for a more efficient, European-wide allocation of spectrum through a simultaneous ascending auction. Only a common regulatory approach through a European telecom regulatory agency can make such a change a reality.

* B.A., Amherst College, MSt. University of Oxford, J.D. University of Chicago. Associate, Swidler Berlin Shereff Friedman, LLP. This paper reflects the views of the author only and not those of the firm or its clients.

[1] The Introduction of Third Generation Mobile Communications in the European Union: State of Play and the Way Forward, COM(01)141 at 4.

[2] A spectrum is a continuous range of frequencies within which waves have some specific characteristic. Newton’s Telecom Dictionary (Flatiron 1998).

[3] The case for a property-based spectrum system is made in Pablo T. Spiller & Carlo Cardilli, Towards a Property Rights Approach to Communications Spectrum, 16 YALE J. ON REG. 53 (1999). The potential of spectrum as a property right, however, has its roots significantly earlier, as Ronald Coase advocated in the late 1950s before the FCC. See R.H. Coase, The Federal Communications Commission, 2 J.L. & ECON. 1 (1959). Although important in its own right, the nature of the spectrum right as a property right does not matter for the purposes of this article.

[4] Arthur De Vany, Implementing a Market-Based Spectrum Policy, 41 J.L. & ECON. 627, 629 (1998).

[5] JOHN A. USHER, EC INSTITUTIONS AND LEGISLATION 34 (Longman 1997).

[6] Council Recommendation 87/371, 1987 O.J. (L 196/81).

[7] Council Directive 87/372, 1987 O.J. (L 196/81).

[8] GSM is the technical mobile standard in Europe, Japan, and Australia for 2G mobile technology. In total, 85 countries use GSM.

[9] Towards the Personal Communications Environment: Green Paper on a Common Approach in the Field of Mobile and Personal Communications in the European Union, COM(94)145 [hereinafter referred to as “Green Paper”]. This Green Paper resulted from the call for a Green Paper as called for by the Council and the European Parliament in their Resolutions on the 1992 Review.

[10] Id.

[11] Id. at 26.

[12] Council Resolution 395Y0722(02), 1995 O.J. (C 188), 3.

[13] Id.

[14] Sixth Report on the Implementation of the Telecommunications Regulatory Package, COM(00)814, Annex 1 at 43-47, lists the way in which mobile licenses were allocated within the EU for first and second-generation licenses. In many cases, initial licenses were allocated on a first-come-first-served basis to incumbent operators. In other cases, 2G licenses were awarded via beauty contests. In either case this did not lead to an efficient allocation. Such processes suffered from the problem of potential government favoritism of its national favorites and a cozy revolving door relationship between regulators and operators. Under a private interest theory framework, this process can be understood as one in which competing interest groups attempt to use state power to capture rents for the successful groups at the expense of less organized groups and the interest groups that do not prevail. See George Stigler, The Theory of Economic Regulation, 2 BELL J. ECON. & MGMT. SCI. 3 (1971); Sam Peltzman, Toward a More General Theory of Regulation, 19 J.L. & ECON. 211 (1976); Gary S. Becker, A Theory of Competition Among Pressure Groups for Political Influence, 98 Q. J. ECON. 371 (1983). This article uses a private interest framework. The alternative framework would be that of public interest theory. Under public interest theory, government regulation serves to maximize social welfare by correcting market failures and protecting consumers from harm. This theory is also known as the positive theory of regulation. Paul Joskow & Roger Knoll, Regulation in Theory and Practice, An Overview, in STUDIES IN PUBLIC REGULATION (Gary Fromm ed., 1981).

[15] Communication from the Commission to the European Parliament and Council on the Consultation on the Green Paper on Mobile and Personal Communications, COM(94). Subsequent to this Communication, the European Parliament adopted a Resolution in support of further liberalization of the mobile sector. See EUR. PARL. RES. of May 19, 1995 on the Green Paper.

[16] Strategy and Policy Orientations with regard to the further development of mobile and wireless communications (UMTS) - outcome of the public consultation and proposals for creating a favorable environment, COM(97)513.

[17] European Radiocommunications Committee Decision ERC/DEC/(97)07 (1997).

[18] Parliament and Council Directive 97/13, 1997 O.J. (L 117).

[19] Council Decision 99/128, 1999 O.J. (L 17) [hereinafter referred to as “UMTS Decision”].

[20] Id.

[21] Id. art. 3. Member states have up to an additional year for the introduction of UMTS services if they have “exceptional” technical difficulties.

[22] Directive 97/13, art. 9, supra note 18.

[23] Sixth Report on the Implementation of the Telecommunications Regulatory Package, supra note 14.

[24] Commission Proposal for a Directive of the European Parliament and of the Council on a Common Regulatory Framework for Electronic Communications Networks and Services, art. 16(2), COM(00)393 final at 25.

[25] As the Commission itself has noted. See Communication from the Commission on the Introduction of Third Generation Mobile Communications in the European Union: State of Play and The Way Forward, COM(01)141 final at 9.

[26] Commission, Mobile Communication: 5-6.

[27] Id. at 6.

[28] R. Preston McAfee & John McMillan, Auctions and Bidding, 25 J. ECON. LITERATURE 699, 701 (1987).

[29] John McMillan, Why Auction the Spectrum?, 19 TELECOMMUNICATIONS POLICY 191, 193-94 (1994).

[30] William Vickrey, Counterspeculation, Auctions and Competitive Sealed Tenders, 16 J. FIN. 8 (1961).

[31] This is demonstrated in Ashenfelter’s conclusions on bidding in the wine and art contexts. See Orley Ashenfelter, How Auctions Work for Wine and Art, 3 J. ECON. PERSP. 23 (1989).

[32] Carolyn Pitchik & Andrew Schotter, Perfect Equilibria in Budget-Constrained Sequential Auctions: An Experimental Study, 19 RAND J. ECON. 363 (1988).

[33] R. Preston McAfee & John McMillan, Analyzing the Airwaves Auction, 10 J. ECON. PERSP. 159, 161 (1996).

[34] Id. at 162.

[35] John McMillan, Selling Spectrum Rights, 8 J. ECON. PERSP. 145, 148 (1994).

[36] Id.

[37] Id. at 149.

[38] McAfee & McMillan, Analyzing the Airwaves Auction, supra note 33 at 173.

[39] See generally RICHARD H. THALER, THE WINNER’S CURSE: PARADOXES AND ANOMALIES OF ECONOMIC LIFE (1992).

[40] Paul Milgrom, Putting Auction Theory to Work: The Simultaneous Ascending Auction, 108 J. POL. ECON. 245, 269 (2000). This does not mean that trading in an aftermarket will not allow for some improvement in efficiency. Rather, aftermarket trading is a less efficient method to achieve an optimal outcome in license allocation and therefore inferior in approach to a well-designed auction.

[41] Hong-Bin Cai, Delay in Multilateral Bargaining Under Complete Information (University of California, Los Angeles, Working Paper, 1997), cited in Milgrom, supra note 41, at 271.

[42] Auctions and Precautions: Overbidding in Spectrum Auctions and its Possible Impact, at 12 (CPB Netherlands Bureau for Economic Policy Analysis, Working Paper No. 127, 2000).

[43] Ed Crooks, Modern “Scourge” Rings in an Economic Boom, FIN. TIMES, June 5, 2000.

[44] In re Federal Communications Commission, 217 F.3d 125 (2d Cir. 2000).

[45] See In re Service Rules for the 746-764 and 776-794 MHz Bands, and Revisions to Part 27 of the Commission’s Rules, WT Dkt. No. 99-168, FCC 00-224 (2000), Federal Communications Commission.

[46] Auctions and Precautions, supra note 42, at 42.

[47] Id. But see Theo Offerman & Jan Potters, Does Auctioning of Entry Licenses Affect Consumer Prices? An Experimental Study (Center for Economic Research, Working Paper No. 2000-53, 2000) (creating models that would predict that auctions might have the potential for an upward effect on prices primarily due to the use of mark up pricing rules in auctions).

[48] Auctions and Precautions, supra note 42, at 11.

[49] See, e.g., BRIAN BURROUGH, BARBARIANS AT THE GATE: THE FALL OF RJR NABISCO (Harper & Row 1990).

[50] Kevin J. Delaney, Wireless Auction Leaves France Up in Air; Government Faces Dilemma: Cut Deficits or Keep Constituents Happy, WALL ST. J. EUR., May 12, 2000.

[51] See McMillan, Selling Spectrum Rights, supra note 35.

[52] U.S. Department of Commerce, U.S. Spectrum Management Policy: Agenda for the Future, Washington, DC, NTIA 1991: D6, cited in McMillan, Why Auction the Spectrum?, supra note 29, at 197.

[53] They could not do so actively because it violates current EU rules.

[54] UMTS Licenses Too Expensive, Says France Telecom Boss, TECH EUR., June 29, 2000.

[55] McMillan, Why Auction the Spectrum?, supra note 29, at 197.

[56] Sheridan Nye, Finland beats U.K. in race to license UMTS, COMM. WEEK INT., Feb. 1, 1999.

[57] As this article will show in its analysis of France’s spectrum allocation.

[58] Finland seems to be one country in which the private interest framework does not fully explain the particular choice of spectrum allocation.

[59] Bruno Giussani, Wireless and Priceless, INDUSTRY STANDARD, May 22, 2000.

[60] 3G Country Information, FIN. TIMES, Aug. 15, 2001.

[61] Id.

[62] Pass the Painkillers, ECONOMIST, May 3, 2001.

[63] The imperative to make the British auction the first auction was demonstrated with the decision to move forward in spite of complications that arose from the Vodafone acquisition of Mannesman, which under other circumstances might have led to postponement. Another factor that also might have caused postponement was the resignation of the head of the U.K. Department of Trade and Industry (which oversaw the auction), Peter Mandelson, shortly before the auction because of scandal.

[64] Italy Eyes Two-Phase Auction; Likely Costs Give Dutch Carriers Pause, WIRELESS TODAY, May 11, 2000.

[65] Peter Crampton, Spectrum Auctions, in HANDBOOK OF TELECOM ECON. 1 (Martin Cave et al., eds., forthcoming 2001).

[66] Paul Klemperer, What Really Matters in Auction Design, at 5 (Oxford University, Working Paper, 2001), available at http://www.nuff.ox.ac.uk/economics/people/klemperer.htm.

[67] Reuters, Dutch Govt Clears Versatel and Telfort in UMTS Probe, TOTAL TELECOM (Feb. 23, 2001), at http://www.totaltele.com/vprint.asp?txtID=37269.

[68] Crampton, supra note 65, at 26.

[69] Giussani, supra note 59.

[70] Id. Klemperer argues that this staggering figure was reached by luck rather than by the auction design of different spectrum blocks and that the auction could have ended in a less competitive mobile environment. See Klemperer, supra note 66.

[71] Diane See Morrison, Germany Gets Greedy with 3Gs, REDHERRING.COM, Aug. 21, 2000.

[72] Id.

[73] Philippe Jehiel & Benny Moldovanu, The European UMTS/IMT-2000 License Auctions (University of Mannheim, Germany, Working Paper, 2001).

[74] 3G Country Information, supra note 60.

[75] Ouida Taaffe, Europe’s UMTS Players in a Flirtatious Mood, TOTAL TELECOM (Feb. 26, 2001), at http://www.totaltele.com/view.asp?articleID=37339&Pub=TT&categoryid=625&kw=taaffe.

[76] Pass the Painkillers, supra note 62.

[77] Ouida Taaffe, German Regulator Insists on UMTS Competition, TOTAL TELECOM (Feb. 28, 2001), at http://www.totaltele.com/vprint.asp?txtID=37432.

[78] Suzanne Kapner, In Shift, Germany Allows Wireless Network Sharing, N.Y. TIMES, June 5, 2001.

[79] Pass the Painkillers, supra note 62.

[80] Klemperer, supra note 66.

[81] Id.

[82] Id.

[83] Air Doesn’t Come Cheap in Europe, available at http://www.teledotcom.com/article/TEL20000100250009.

[84] Id.

[85] An Italian joint venture with BT and the fourth largest mobile operator in Italy.

[86] Diane See Morrison, Italy’s 3G Flop, REDHERRING.COM, Oct. 25, 2000. As the time to the auction neared, the Italians hoped to raise as much as $25 billion from the auction.

[87] Reuters, Blu Wins Reprieve on UMTS Bill, TOTAL TELECOM (Nov.9, 2000), at http://www.totaltele.com/view.asp?articleID=33629&Pub=TT&categoryid=625&kw=italy+and+umts.

[88] 3G Country Information, supra note 60.

[89] Id.

[90] Fred Donovan, Former Nordic Marriage Partners Come Up Short in Sweden, 18 WIRELESS INSIDER 50 (2000).

[91] Id.

[92] Gerard O’Dwyer, 3G Rollout Problems Challenge Swedish Government, TOTAL TELECOM (Mar. 14, 2001), at http://www.totaltele.com/view.asp?ArticleID=37923&pub=tt&categoryid=0.

[93] 3G on Hold, ECONOMIST, Mar. 6, 2001.

[94] Emily Bourne, Belgian Auction Attracts Just Three Bidders, TOTAL TELECOM (Feb. 9, 2001), at http://www.totaltele.com/view.asp?articleID=36798&Pub=TT&categoryid=625&kw=Belgian+3G+auction+attracts+just+three.

[95] 3G Country Information, supra note 60.

[96] Id.

[97] Id.

[98] Anne Young, Portugal Awards Four 3G Licenses, TOTAL TELECOM (Dec. 19, 2000), at http://www.totaltele.com/view.asp?articleID=35014&Pub=TT&categoryid=625&kw=Portugal+awards+four.

[99] Rebecca Fannin, Global Currents: European High Tech Gets Melodramatic, REDHERRING.COM, Oct. 30, 2000.

[100] Pass the Painkillers, supra note 62.

[101] Kenneth Neil Cukier, France’s 3G Beauty Contest Looks Ugly All Around, REDHERRING.COM, June 8, 2000.

[102] Kevin J. Delaney, Wireless Auction Leaves France Up in Air; Government Faces Dilemma: Cut Deficits or Keep Constituents Happy, WALL ST. J. EUR., May 12, 2000.

[103] Id.

[104] UMTS Licenses Too Expensive, Says France Telecom Boss, TECH EUR., June 29, 2000.

[105] Delaney, supra note 102.

[106] Anne Young, Austria Raises 706 Million Euros in UMTS Auction, TOTAL TELECOM (Nov. 3, 2000), at http://www.totaltele.com/view.asp?articleID=33455&Pub=TT&categoryid=625&kw=austria+raises+706.

[107] 3G Country Information, supra note 60.

[108] Klemperer, supra note 66.

[109] Id.

[110] David Crossland, Austrian UMTS Auction Unlikely to Scale Peaks, TOTAL TELECOM (Oct. 31, 2000), at http://www.totaltele.com/view.asp?articleID=33285&Pub=TT&categoryid=625&kw=austrian+umts+auction+unlikely, cited in Klemperer, supra note 66, at 6.

[111] Jehiel & Moldovanu, supra note 73, at 17.

[112] Reuters, Denmark Plans Sealed UMTS Auction, TOTAL TELECOM (Mar. 1, 2001), at http://www.totaltele.com/view.asp?articleID=37454&Pub=TT&categoryid=625&kw=denmark+plans+sealed.

[113] Id.

[114] Ouida Taaffe, Swiss Regulator’s diAx Probe May Threaten UMTS License, TOTAL TELECOM (Jan. 8, 2001), at http://www.totaltele.com/view.asp?articleID=35475&Pub=TT&categoryid=625&kw=swiss+regulator%27s+diAx.

[115] 3G Country Information, supra note 60.

[116] Id.

[117] Id.

[118] Klemperer, supra note 66, at 10. The auction yielded slightly more than the reserve price due to the slight differences among the licenses.

[119] White Paper no. 24 (1999–2000).

[120] Anne Young, Norway Awards Four UMTS Licenses, TOTAL TELECOM (Nov. 29, 2000), at http://www.totaltele.com/view.asp?articleID=34375&Pub=TT&categoryid=625&kw=norway+awards+four.

[121] See Norwegian government press release, at http://www.dep.no/sd/engelsk/aktuelt/pressem/028021-070020/index-dok000-b-n-a.html.

[122] Emma McClune, Polish UMTS Tender Cancelled—But It’s Good News, Say Analysts, TOTAL TELECOM (Dec. 6, 2000), at http://www.totaltele.com/view.asp?articleID=34592&Pub=TT&categoryid=625&kw=polish+umts+tender+cancelled.

[123] Id.

[124] 3G Country Information, supra note 60.

[125] Killer Applications, ECONOMIST, Aug. 24, 2000.

[126] 3G on Hold, supra note 93.

[127] Jehiel & Moldovanu, supra note 73, at 3.

[128] Dan Briody, Wireless Watch: Is 3G worth the extra Gs?, REDHERRING.COM, Dec. 27, 2000.

[129] Id.

[130] Id.

[131] Killer Applications, supra note 125.

[132] Alistair Robertson, BT Will Not Cut Debts at Any Cost, TOTAL TELECOM (Mar. 19, 2001), at http://www.totaltele.com/view.asp?articleID=38108&Pub=TT&categoryid=625&kw=BT.

[133] Killer Applications, supra note 125.

[134] When Big is No Longer Beautiful, ECONOMIST, Dec. 14, 2000.

[135] Aline van Duyn, More Downgrades Expected, FIN. TIMES, Apr. 11, 2001.

[136] BT: We Blew $14 Billion, CNNfn, Feb. 19, 2001, available at http://www.cnnfn.cnn.com/2001/02/19/europe/british_telecom.

[137] Michelle Donegan, EC urges state policies for 3G license debt relief, COMM. WEEK INT., Mar. 19, 2001.

[138] This is reminiscent of the “Too Big to Fail Doctrine” used by governments to prop up industries that it sees as essential for political reasons, thereby creating a moral hazard. See Maureen O’Hara & Wayne Shaw, Deposit Insurance and Wealth Effects: The Value of Being “Too Big To Fail”, 5 J. FIN. 1587 (1990).

[139] Donegan, supra note 137.

[140] Taaffe, Europe’s UTMS Players in a Flirtatious Mood, supra note 75.

[141] Jonathan Collins, Spectrum Spills and Chills, available at http://www.teledotcom.com/article/TEL20001114S0009.

[142] Jehiel & Moldovanu, supra note 73.

[143] Mobile virtual operators may be able to offer: (a) Expanded choices for a complete mobile service; (b) Potentially wider range of services; and (c) Possibly lower retail prices. See, e.g., Director of General Telecommunications, OFTEL, Mobile Virtual Network Operators: Oftel inquiry into what MNVOs could offer consumers, June 1999, available at http://www.oftel.gov.uk/publications/1999/competition/mvno0699.htm#Chapter%202. MVNOs have made progress in some jurisdictions. In Denmark and Sweden, the government regulators have placed MVNOs on the same regulatory terms as fixed operators, which give MVNO entrants the right to national roaming across all networks and the right to interconnect on commercial terms with operators that have been recognized to hold significant market power. In Ireland, the telecom regulator (ODTR) has announced that 3G network operators which are willing to open up their infrastructure to MVNOs will be given preferential licensing terms. See Annie Turner, Mobile Virtual Network Operators: Taking Root, TOTAL TELECOM (Apr. 1, 2001), at http://www.totaltele.com/newcarrier/view.asp?articleID=39455&Pub=NC&categoryid=705&kw=moblie+virtual+network+operators.

[144] Though an optimal initial allocation is still preferred, secondary trading might be a second best alternative to fix some of the allocation problems that the licensing process has caused.

[145] Towards a New Framework for Electronic Communications Infrastructure and Associated Services, The 1999 Communications Review, COM(99)539 at 37 [hereinafter “Towards a New Framework”].

[146] As the Financial Times notes, “One operator [in the German UMTS auction] has privately admitted to altering the last digit of its bid in a semi-serious attempt to signal to other participants that it was willing to accept [fewer spectrum blocks to end the auction.]” FIN. TIMES, Nov. 3, 2000, 21 cited in Klemperer, supra note 66 at 17. In contrast, the FCC has eliminated this type of signaling by its use of click box bidding. Bidders in FCC auctions do not bid in dollar amounts. Instead, a bidder will indicate in a click box the number of bid increments from 1 to 9 that it wishes to bid above the current high bid. Thus, if a current high bid is 100 and the minimum bid increase is 10%, allowable bids would be 110 through 190. This serves to eliminate signal bidding through the last digit.

[147] Klemperer, supra note 66 at 24. In this country-by-country basis he may be right, but misses the larger implications of auctions in a pan-European setting.

[148] EU legislation requires that the issue be considered under Article 22 of the interconnection directive. See Commission Directive 97/33 1997 O.J. (L 199/32).

[149] Towards a New Framework, supra note 145, at ix.

[150] Id. at 35.

[151] Id. at 36.

[152] Alan Osborn, EU Telecoms Ministers Limit EC Powers of Intervention, TOTAL TELECOM (Apr. 5, 2001), at http://www.totaltele.com/view.asp?articleID=38720&Pub=TT&categoryid=625&kw=EU+telecoms+ministers+limit+EC.

[153] Arthur Stein, Coordination and Collaboration: Regimes in an Anarchic World, in INTERNATIONAL REGIMES (Stephan D. Krasner, ed., Cornell 1983); See also Duncan Snidal, Coordination versus Prisoners’ Dilemma: Implications for International Cooperation and Regimes, 79 AM. POL. SCI. REV. 923 (1985).

[154] This game is described in Stephan D. Krasner, Global Communications and National Power: Life at the Pareto Frontier, 43 WORLD POLITICS 336, 338-340 (1991), and in LLYOD GRUBER, RULING THE WORLD: POWER POLITICS AND THE RISE OF SUPRANATIONAL INSTITUTIONS (2000) 76-79.

[155] Given the numerous other problems of European telecom regulation, a pan-European regulator also would help potentially to resolve some of these other issues.

The European Mobile 3G UMTS Process: Lessons From the Spectrum Auctions and Beauty Contests

D. Daniel Sokol*

I. Introduction

D. Default Possibility of Licensees

F. Overbidding

C. The Netherlands

G. Belgium

M. Switzerland

The European Mobile 3G UMTS Process:
Lessons From the Spectrum Auctions and Beauty Contests