This article analyses the early experiences of spectrum auction in some advanced countries and distils lessons from them. An issue that has been examined in this context is whether auction necessarily leads to a market-efficient outcome. This is of topical interest in India at the current juncture since the department of telecommunications is on the verge of holding an auction for allotting the spectrum for the third-generation (3G) of mobile phones.
INSIGHTEconomic & Political Weekly EPW september 13, 200833The views expressed are the author’s own.Manas Bhattacharya (manasb2@gmail.com) is with the ministry of tourism, New Delhi.The International Experience of Auctioning SpectrumManas BhattacharyaThis article analyses the early experiences of spectrum auction in some advanced countries and distils lessons from them. An issue that has been examined in this context is whether auctionnecessarily leads to a market-efficient outcome. This is of topical interest in India at the current juncture since the department of telecommunications is on the verge of holding an auction forallotting the spectrum for the third-generation (3G) of mobile phones.This article aims to analyse the early experiences of auctioning spec-trum in advanced countries in order to distil lessons and understand whether an auction necessarily leads to a market efficient outcome. Findings of this study would be relevant in the context of the recent discussions in the Indian tele-com sector about introducing auctions as a method of spectrum assignment for the potential service providers of the 3G mobile technology.The paper is organised as follows. The first section presents the experience of New Zealand, Australia and the United States (US) in auctioning spectrum, how these auctions were designed and what results were achieved. An example is cited in this section, which shows that auction prices can reflect gaming behaviour rather than the intrinsic value of spectrum. The second section cites three instan-ces of auctions conducted by the Federal Communications Commission (FCC) that led to disappointing results. Again, the outcome was due to a number of extra-neous factors rather than the real worth of spectrum itself. These experiences drive home the point that auctions neednot necessarily lead to a win-win situation between the government and service providers.The third section presents the results at the other extreme – the exciting 3G auc-tions of the United Kingdom (UK) that net-ted a huge amount of revenue. This sec-tion also shows that per capita auction revenue earned from 3G licences by seven European countries around the same peri-od varied widely. It appeared that a range of factors, some of which are exogenous can influence auction outcomes. In this context, a view point is presented in this section that high bids, delay in payments and delay in services could be parts of a larger manipulative game.Some arguments and counter arguments have been presented in the fourth section as to whether an auction does necessarily yield competitive results. A reference has been made in this context to a study of the outcomes of the European Universal mobile telecommunications sys-tem/international mobile technology (UMTS/IMT)1 – 2000 auctions.The fifth section concludes the paper by distilling some of the lessons drawn from these international experiences and analysing them from policy perspectives. Several administrations of the world use auctions to award spectrum licences. While the US uses only auctions and no other method to assign licences for com-mercial use, some other countries like Australia, Canada and theUK have used auctions to assign certain bands yet main-taining the legacy of administrative pric-ing of spectrum in one form or another.An auction is a market-based approach, which puts spectrum in the hands of those who values them the most. Unlike other methods, auctions determine allocation and pricing of licence (or spectrum for that matter) simultaneously. In other methods, separate sets of criteria need to be evolved for allocation and pricing.In a simplistic way, the auction process can be described as a fair, transparent, objective and expeditious method of awarding licences. In theory, winning prices of the bidders are supposed to reflect the value the market places on spectrum based on the bidders’ realistic appreciation of the market conditions. One difficulty with this approach is the lack of certainty regarding the continued efficient use of spectrum once an assign-ment is made after the auction is over. Moreover, its applicability is limited to the licensing of spectrum for the services, which are commercially viable. Prices of spectrum used by non-profit agencies including various government depart-ments cannot be directly determined through auction.There are various types of auction mod-els used in different countries like sealed tender auctions, open outcry, simultane-ous ascending multiple round auctions, etc. The same model undergoes variations
INSIGHTseptember 13, 2008 EPW Economic & Political Weekly34in matters of details and designs in differ-ent countries at the implementation stage.Based on the popular auction experi-ences in some of the countries, the present paper seeks to examine whether an auc-tion simulates competitive behaviour and leads to competitive prices that truly reflect the intrinsic value of spectrum con-sistent with bidders’ rational intention. The paper also discusses the extent to which an auction can be described as a objective process and explores the basis of a common apprehension that exaggerated bidding in auction may have catastrophic consequences for the telecom market.International ExperiencesThough certain episodes of auction suc-ceeded in raising large sums of money, analyses of the outcomes of a wider sample of auctions held in different countries do not exhibit consistent and optimistic results to support high revenue expectations on a sustained basis. There are instances of fail-ures of auctions too. Moreover, values gen-erated in auctions and bidders’ rational valuation of intrinsic worth of licences do not necessarily converge.New Zealand’s first auctions of nation-wide ultra high frequencies lots of 8 MHz licence took place in 1990 and were based on sealed tender. For five licences, the total auction revenue was justNZ $ 36 mil-lion against NZ $ 250 million projected by the consultants. Bidders were, however, permitted to buy and sell licences after the auction.Early auction experience in Australia is an interesting case study. Australia’s spec-trum auctions in 1993 followed the sealed tender method. There was no provision of advanced deposits or default penalty. Moreover, there was no restriction against multiple bids by the same bidders. As per the rules, the licence would be offered to the second highest bidder if the highest bidder failed to make payments within a stipulated time. The results were quite unexpected. To quote Milgrom (1995):In April of 1993, two newly formed compa-nies surprised the established players by placing unexpectedly high bids of A$ 212 million and A$ 177 million for two satel-lite television licences. One of the bidders had paid-in capital of just A$100. When the requisite time period had elapsed and neither bidder had paid, the government rejected those bids and announced the sec-ond highest bids, which were submitted by the very same bidders. Under the rules, this was perfectly legal. The bidders were again allowed a period of time to find the neces-sary funding before those bids, too, were rejected. This pattern of default, rejection and announcement of the new winning bid continued for 10 months. While the licence assignment was delayed, the bidders negoti-ated with potential buyers. The first licence was eventually bought by the defaulting bid-der for just A$ 117 and immediately resold for A$ 21 million in profit. The second was acquired for A$ 77 million and resold for an unknown price but presumably at a profit.Between 1927 and 1982, the US followed the “comparative hearing” method to assign spectrum licences based on evaluation of competing proposals from the point of view of public interest. However, this was a time consuming process. In 1982, the Congress authorised the FCC to assign licences through lotteries. Any telephone company, regardless of its qualification or experience was eligible to participate in the lottery. Licence rights were transferable. A second-ary market was therefore allowed.The upshot of this plan was the creation of a new business for Washington law firms: organ-ising “telephone companies” for the sole pur-poses of filing lottery applications to acquire a valuable cellular telephone licences. During the lottery era, more than four hundred thou-sand applications were filed. Most lottery win-ners resold their rights at large profits to older telephone companies. In one case, a “tele-phone company” that had won the right to supply cellular services to Cape Cod promptly resold their licence to the Southwestern Bell Telephone Company for $ 41 million.With estimates of the value of the spectrum given away in tens of billions of dollars and with mounting political pressures to do something about the hugeUS federal budget deficits, the Congress decided to put an end to what the legislation called the “unjust en-richment” of lottery winners. The Omnibus Budget Reconciliation Act passed in July of 1993 authorised the FCC to conduct auctions for rights to use the airwaves and mandated that it begin assigning licences within just 10 months. The licence selling price, this time, was to accrue to the US Treasury, rather than to some randomly determined lottery winner [ibid, McMillan 1995].Among several varieties of auction models, the FCC adopted simultaneous multiple rounds of auction, which allowed bidders to raise their bids in each round after prices were announced. This process continued till no new bid was received. The auction was simultaneous in the sense that a set of licences was auctioned simul-taneously, which allowed bidders to take into account interdependencies among licences and bid for more than one licence together. Another advantage of this meth-od was that a bidder could bid on the basis of his knowledge of his competitors’ bid prices and was not required to bid for more than that was necessary to acquire a licence. An important feature included was “activity rule”, which required each bidder to maintain minimum levels of bidding activity in each round to qualify for later rounds of auction. In order to avoid frivo-lous or mischievous bids a penalty was introduced for the withdrawal of bids. “Nationwide narrowband auction” (for paging and messaging services) of the US which tried out the simultaneous auction in July 1994, raised $ 17 million and set a new US record. In the same year, the second trial of simultaneous multiple rounds auction was launched in October for the sale of “region-al narrowband licences”. A special feature of this auction was the large discount and preferential financing terms offered to small business and woman and minority owned firms in any licences acquired in certain bands. The results showed that the prices in the October auction were significantly higher than the average prices of the com-parable national licences purchased in July.In December 1994, 99 licences, mostly for wireless telephone applications were put to auction. These licences were offered for 51 major trading areas (MTAs).2 In order to increase the pace of auction in an effec-tive manner theFCC introduced stipula-tion on bid increments and retained its discretion to modify the number of rounds per day during the auction. Direct revenue from this auction exceededUS$ 7 billion.Analysts have studied the bidding strategies adopted in the auctions and drawn inferences, which called into question the perceived straightforward-ness of bidding as a tool for buying licences, which were of the greatest value to the respective bidders. In reality, the bidding behaviour seemed more complex than theory. It would be of interest to refer to a report on the bidding behaviour of the general telephone and electronics (GTE), one of the bidders in the December
INSIGHTEconomic & Political Weekly EPW september 13, 200835auction in the US (discussed in the earlier paragraph).Insider David Salant’s account reports that GTE’s goals early in the auction were to avoid bidding on target markets, partly to keep ri-vals from guessing thatGTE’s target markets were Atlanta and Seattle, maintain eligibil-ity and push up the price in non-target mar-kets, partly to avoid pushing up the price in target markets and partly to induce rivals to spend more money on those markets. The point of keeping rivals guessing was that had we not done so, we might have been taken advantage of by others who wished to use up our budget [Milogram 1995].Regardless of whether the prices of licences reflected their opportunity cost or not, the US treasury received substan-tialrevenue. It shows that the auction prices can result from gaming behaviour and need not necessarily reflect the intrin-sic value of spectrum.Disappointing AuctionsAmid success stories there have been cer-tain disappointing auctions too. In 1994, the FCC held auction to sell licences for interactive video and data services. In this auction, licences were sold one at a time. Many of the bidders were speculators and several of the bidders, unhappy that they had paid more for their licences than others paid for similar licences later in the auc-tion, chose to default on their bids(ibid).One of the broadband personal com-munication service (PCS) auctions of the FCC, identified as the C-block auction, held in 1995, ran into serious problems reportedly because of favourable treat-ments earmarked for the “designated entities” (ie, small businesses and women or minority owned firms) to encourage their participation.The FCC has used bidding credits, set-asides and installment payments to encourage the participation and success of designated enti-ties. The idea is that without special treat-ment, these small businesses would find it dif-ficult to compete with the large incumbents. The favoured treatment can serve to “level the playing field” and there by foster innovation and intensify competition....The auction failed largely because of overly attractive instalment payments (10 per cent down and six-year interest only at the risk-free 10-year treasury rate). This encouraged speculative bidding, which led to all the major bidders defaulting and declaring bankruptcy. Even now, years after the auction, much of this C-block spectrum lies unused, tied up in bankruptcy litigation [Cramton 2001b].Another instance, held in disappoint-ment, was the wireless communication services (WCS) auction of the FCC held in April 1997. There was very little competi-tive activity for the 128 licences that were put to auction. Various explanations have been offered for the lack of success of this auction. According to Taylor Simmons of Simmons Associates,3 a combination of factors, including the uncertainty sur-rounding WCS and the initial steep mini-mum bid increment was responsible for lack of competition. The FCC initially set the bid increment at 2 cents per mega hertz per post office protocol,4 which was subse-quently reduced to 0.25 cents for round 7. However, this reduction was of no conse-quence [Carlson 1997]. The total net reve-nue in this auction was US $ 6.2 million.According to Cramton (2001a), the design of this auction did not address dif-ficult trade-offs in effective implementa-tion issues and the auction was conducted in a hurry.The main problem was the stringent out-of-band emission limit. Equipment manufactur-ers warned that this would threaten the com-mercial viability of this spectrum. The low prices at auction and the absence today of ac-tivity in this band appears to confirm that the equipment manufacturers were right. At the time of the decision, the FCC was facing a dif-ficult tradeoff between the rights of prior win-ners of neighbouring licences and WCS use. Such decisions are always difficult but the FCC was under intense time pressure to meet the timetable that the Congress set for the auction. This aggressive timetable may well have led the FCC to make too hasty a decision on interference rules, which damaged the value of this spectrum. The Congress’s desire for receiving revenues according to its fiscal calendar may have resulted in substantially reduced auction revenue.Exciting 3G AuctionsOne of the most exciting auction results of the recent past relates to the auction of five third-generation (3G) mobile wireless licences of the UK held in April 2000, which raised an astonishing £ 22.5 billion. According to Cramton (2001b): “although theUK did not begin auctioning spectrum until 2000, it began with a bang. Its very first auction broke into the record books as the world’s largest auction ever.Cramton offered a set of plausible expla-nations of why prices were so high. Accord-ing to him (ibid): (1) the UK was the first mover in 3G auctions in Europe and there-fore reaped the benefit of first mover’s advantage as the foot in the door to Europe and potentially the world; (2) the auction took place at the peak of an apparent high-tech stock bubble. Bidding companies’ images in the stock market came to be asso-ciated with their abilities to win a licence; and (3) ascending auction format coupled with large excess demand might have led to a competitive spiraling up of prices.The government of theUK auctioned five 3G licences, each of 20 years duration. One large licence was set aside for a new entrant. A “spectrum cap” was introduced through the stipulation that any company (or associated company) can win at most one licence. There were also requirements of a deposit of £ 50 million initially and an additional deposit of £ 50 million when the bid exceeded £400 million. The deposit was refundable to the unsuccess-ful bidders. The payment was to be made in full at the end of the auction or in instal-ments in sufficiently unattractive terms. Other features of the design of the auction included specification of minimum open-ing bids, activity rule, number of rounds per day and the stopping rule. The auction was fully transparent with bidders’ identi-ties known to each other and the bid prices were announced after each round. Withdrawals of bids were not allowed.It may be noted that not all the govern-ments of Europe auctioned 3G licences. Finland, Spain, Norway, Sweden and France chose “comparative hearing” whereas the UK, the Netherlands, Germany, Italy, Austria, Switzerland, Belgium and Denmark chose to use auctions. The timing of auctions differed from country to country with theUK being the first to auction in April 2000 and Denmark held 3G auctions in August 2001. Apart from timing, these auctions also differed from each other in design and details. The following table (p 36) shows the number of 3G licences auctioned by these coun-tries (arranged in sequence of the timing of auction) and per capita revenue gener-ated in these auctions.It has been observed that for the coun-tries, which held auctions later in the period, the timing of auctions coincided with the bubble burst of the stock prices in the tele-com market. This might have had an impact
INSIGHTseptember 13, 2008 EPW Economic & Political Weekly36on auction results [Damme 2002]. Differ-ences in revenue might have also been contributed by other factors such as details of the auction design, policy and regulatory framework, nature of telecom market, level of competition, market share of the existing incumbents, expectations regarding future profit, etc.The lesson that emerges from the varied auction experiences described above is that the auction results depend upon the particular contexts in which they take place. A large number of factors, some are policy induced and others, exogenous and circumstantial, tend to impact the results of auction. Experiences of the 3G auctions were different in Belgium and Singapore. Both the Belgian auctions ended after the submission of the initial bids at the reserve price. Neither auction had excess demand [Crampton 2001b]. Even after such a record breaking experience of 3G auction in the UK, many licences remained unsold in the auction for fixed broadband wire-less access held by the UK government subsequent to 3G auction.5Yet another example, the 3G licence auction in Hong Kong held in September 2001 fetched the minimum price because out of the six incumbent 2G operators, only four bid for a 3G licence and there was no new entrant.Ozanich et al (2002) advanced an alter-native interpretation of the phenomenon of high bid, delay in payments and delay in services. Their explanations go beyond the perceived notion that high auction prices result from exaggerated profit expectations of the bidders.Analysing 3G wireless auction results on the basis of discounted cash flow method to calculate pay back period, the authors concluded that companies overpaid for 3G spectrum. The German and UK auctions stand out in particular. Cash flow require-ments to justify investment in 3G licences in these two countries appeared unrealis-tic. The authors (ibid) have also noted:Perhaps a more interesting element of the auction process is the delay in auction pay-ments. In theUS, of the $ 41 billion bid in wireless auctions, only $ 14 billion has been collected. Evidence shows “a similar pattern is already playing out in Europe” [New York Times 2001]. Companies are also requesting that amount of the winning bids be reduced.…Finally, most of the winners of 3G auctions were incumbents or owners of existing 2G li-cences. These companies continue to gener-ate revenue and positive cash flow from these facilities while 3G is delayed.“Win-at-any-cost” can be a part of a larger strategy to block entry, hoard spectrum, delay payment, delay services and continue to earn profit over the existing network.Customary ex post facto analyses of auction results tend to isolate a number of factors holding them responsible for per-ceived success or failure of auctions. Table: 3G Licence Auctions by West Europe (April 2000-August 2001)Country Number of Licences Per Capita Auction Auctioned Revenue (Euro)UK 5650Netherlands 5 170Germany 6 620Italy 5210Austria 6 105Switzerland 4 20Belgium 4 45Denmark 4 95Sources: Damme Eric van (2002).CENTRE FOR MICRO FINANCEFountain Plaza, Pantheon Road, Egmore, Chennai – 600008 Call for Applications: Microfinance Researchers’ Alliance ProgrammeThe Centre for Micro Finance (CMF), in collaboration with the Ford Foundation, is initiating a Microfinance Researchers’ Alliance, an intensive research capacity development programme targeting Indian researchers currently working on, or interested in working on, microfinance research. This programme, the first of its kind in the country, aims to cultivate and develop a network of top mid-career researchers on microfinance over 3 years. For this programme, CMF will select approximately 20 motivated, committed researchers in their early- to mid-career. The programme extends support to participating researchers over the course of three years in the form of workshops and research seminars; exposure visits to CMF project sites; providing a forum for idea exchange with a network of critical thinkers on microfinance issues; opportunities to work on CMF projects with leading development researchers from MIT, Harvard, Yale, IIM, and other top international institutions; and potential project funding support.Eligibility CriteriaGiven the extent of financial and non-financial support the participating researchers would receive, the selection process will be highly rigorous and competitive. The following is a list of prerequisites for individuals that apply for the programme:i) PhD holder in academic disciplines including economics, political science, sociology, anthropology, women’s studies, statistics and finance. ii) Experience in leading a field research team.iii) Associated with a university, teaching college or an autonomous research institution.iv) Motivated, enthusiastic young mid-career researchers who are able to commit time and effort for 3 years of research.Interested researchers can visit www.ifmr.ac.in/cmf/rap.html and download the necessary documents and instructions for application. All documents, including letters of recommendations, should be sent in one package and arrive at CMF no later than October 31. Applications will be evaluated upon receipt, and short-listed candidates will be invited to the first meeting to be held on November 14-15, 2008 in Delhi. This programme is financially supported by the Ford Foundation.
INSIGHTEconomic & Political Weekly EPW september 13, 200837Credibility of such explanations is difficult to establish in the absence of counter-factuals. By and large, such explanations seek to find reasons with the designing aspects of auctions, which in most part remain judgmental with outcome, always tentative. The Swiss wireless-local-loop auction of March 2000, for example, sold nationwide licences in sequential ascend-ing auctions. While the first two licences for a 28 MHz block were sold at 121 million francs and 134 million francs respectively, the third and largest licence of 56 MHz fold for only 55 million francs. As an explana-tion of these paradoxical results, the blame was put to the fact that the licences were sold in sequence rather than simultaneously [Cramton 2001a].Does Auction Yield Competitive Results?Observations of real life situations have given rise to certain attempted characteri-sation of behavioural outcomes of auctions, which do not support the perceived notion that auction tends to generate the intrinsic value of spectrum that should prevail in a competitive market. Even theoretically, anti-competitive behaviour is possible in certain situations. For instance, in a pre- existing monopolistic market, a monopolist will value a new licence more than a new entrant because the latter, if he wins, will be a threat to the monopolist’s profit. The value that the monopolist places on a new licence will partly reflect his preference for continuing monopoly and partly the per-ceived market value of the new licence. The auction price, in such a situation is not expected to reflect pure competitive price [Gilbert and Newbery 1982].In the context of UMTS auctions, Damme (2002) observed that important asym-metries and value differences do arise between bidders, for example, between the incumbents and the entrants. For an incum-bent, not winning a licence means that he might lose not only his new 3G customers but also his existing 2G business. Further-more, another element of asymmetry arises from the fact that the cost of rolling out a network may be much lower for an incum-bent than for an entrant. Jehiel and Moldovanu (2000), who examined the outcomes of European UMTS/IMT-2000 auctions, observed that incumbent carriers consistently paid more than new entrants. Arguably, this could be the result of the barrier to entry strategies rather than cost advantages and rational bidding. An interesting example of anti-competitive behaviour relates to the Dutch UMTS auc-tions, where a new comer Versatel partici-pated in the auction not to win but to get concession from other players, who were existing incumbent operators. A press release issued by Versatel shortly before the auction revealed their intention: “We would however not like to see that we end up with nothing whilst other players get their licences for free. Versatel invites the incum-bent mobile operators to immediately start negotiations for access to their existing 2G networks as well as entry to the 3G market either as a part owner of a licence or as a mobile virtual network operator”.6 This resulted in Telfort, a British Telecom subsidiary and participant in the auction overbidding Versatel and terminating auc-tion at revenue that was 17 per cent higher. It is, therefore, apparent that an auction may lead to various possible outcomes. Overpricing or undervaluation of spec-trum are possible outcomes in an auction depending upon the circumstances.Noted economist of the FCC, Evan Kwerel has argued that several commonly perceived notions of adverse consequences of spectrum auctions are actually common misconceptions about auctions. One com-mon misconception is that auctions tend to overprice spectrum and thereby under-mine the development of the communica-tion industry, a phenomenon described as “winners curse”. According to Kwerel (2001), the cost of winning a licence in an auction will have no effect on the price of telecommunication services to consumers. The price/or tariff that a consumer will pay will depend upon the market clearing price and forces of supply and demand.Telecommunications experience in theUS has also been consistent with the theory that historic costs do not alter pricing. For example, within a given market, the prices charged by cellular operators who obtained their licences via comparative hearings or lotteries are no lower than the prices of those firms that purchased their cellular licences in the secondary market or firms that obtainedPCS licences in an auction. Similarly, where a US cellular licence has been bought at a significant cost from a party that obtained it at no cost, we have notobserved any increase in consumer prices.A company, who overpays for a licence may be in financial distress or bankruptcy leading to a one time loss in investment and consequent change in the ownership of the licence. However, an appropriate legal frame-work should be in place to ensure that the licence does not get trapped into protracted transition delaying its utilisation to provide services to consumers. It may also be noted that if a winner in any auction has overpaid for a licence, which would not match his expected profit, he would soon resell the licence rather than hold on to it. The new buyer of the licence will invest in infrastruc-ture and the serviceswill not be held up.According to Kwerel, indeed, the US experience to date shows no apparent lack of investment in facilities of cellular orPCS operators that bought their licences rela-tive to firms that acquired their licences in a comparative hearing (ibid).Kwerel has also argued against the pos-sible misconception that an auction might lead to the monopolisation of spectrum. According to him, as long as there is a free-dom to resell, any method of spectrum assignment may potentially lead to monopo-ly. However, there are extraneous methods of preventing monopoly through limits on spec-trum aggregation, setting a cap on spectrum use by a single entity. Auction can also be structured to create or increase competition (ibid). Out of 27 FCC auctions of non-broad-cast spectrum, small businesses have success-fully obtained licences in 22.Kwerel’s arguments are theoretically sound. He had also substantiated his conten-tions by citing the US experiences in auction. One issue that emerges clearly and easier recommended than done, is the framework to facilitate subsequent market corrections in the event of high bids and default. How expeditiously this can be done is an open question. Costs of wasted investment and time to relocate a licence in the event of a market failure should be carefully weighed against possible benefits of auction.ConclusionsThere are certain lessons, which can be dis-tilled from the above portrayal of the inter-national experience of spectrum auctions: One, the revenue outcome in spectrum auctions is tentative. Two, the distinction
INSIGHTseptember 13, 2008 EPW Economic & Political Weekly38between “allocation of spectrum” and “assignment of spectrum” needs to be kept in mind. While allocation defines different characteristics of spectrum such as the fre-quency band, geographic area, time period, restriction on use, etc, assignment refers to the allotment of spectrum. An auction is a method of assigning licences. If spectrum allocation is inefficient, a highly efficient auction will deliver poor results [Cramton 2001a]. Three, a large number of discretion-ary parameters enter into auctions. What variant of auctions is to be used, the details of the design, etc, are to be decided in advance. Experiences in other countries suggest that there is no “one-size-fits-all” approach in auction. There is always an ele-ment of uncertainty regarding what will work and what will not. In the absence of counterfactuals, it is difficult to establish cause-and-effect relationships even in ex post facto analysis of auction results. Four, decisions to be taken at the planning stage of auction may turn out to be complex at times. One tricky area relates to “packaged bidding” where a bundle of licences with large complementarity can potentially earn more auction revenue per licence than if individual licences are put to bid separately and sequentially. Such decisions require careful planning and analysis. Five, auctions may require a framework of efficient clear-ing of encumbered spectrum if the vacant band is not available. Six, winner in an auc-tion may not use licences efficiently. How-ever, the terms and conditions or rights and restrictions in a licence can be framed appro-priately to safeguard public interest. How-ever, these conditions will, no doubt, affect bidders’ willingness to pay for the licence [Kwerel 2001]. Finally, regulators should consider any request by the winning bidders to change the terms of the auction after con-test with scepticism. Such auctions may be a fall-back strategy to the “win-at-any-cost” outcome of wireless telephony auctions [Ozanich 2002]. It may be mentioned that regulators in England and Germany had denied the requests of the companies after the 3G auction to reduce the licence fees.Notes1 UMTS/IMT is one of the 3G cell technologies. 2 MTAs do not overlap with states, but are areas surrounding major business centers.3 Simons Associates is an internationally known consulting firm.4 POP is an Internet standard protocol.5 One plausible reason cited is that the minimum opening bid was set at too high a level [Cramton 2001].6 ‘Versatel Disappointed with Dutch UMTS Auction Tomorrow’, Business Wire, July 6, 2000. http://findarticles.com/p/articles/mi_m0EIN/is_2000_July_6ai_63171362 visited on August 14, 2008.ReferencesCarlson, C (1997): ‘WCS Auction Sees Sluggish Start’, Wireless Week.Cramton, P (2001a): ‘Spectrum Auctions’, University of Maryland, pp 31-32.– (2001b): ‘The Auction of Radio Spectrum for the Third Generation of Mobile Telephones’, the report commissioned by the National Audit Office of the United Kingdom, pp 47.Damme, Eric van (2002), ‘The European UMTS – Auctions’, European Economic Review, Vol 46, pp 846-58.Gilbert, R and D Newbery (1982): ‘Preemptive Patent-ing and the Persistence of Monopoly’,American Economic Review, Vol 72, pp 514-26.Jehiel, P and B Moldovanu (2000): ‘Licence Auctions and Market Structure’, discussion paper, Mannheim University, referred in Ozanich Gray W, et al (2002).Kwerel, E (2001): ‘Auctioning Spectrum Rights’, US Federal Communications Commissions, pp 4.Mcmillan, J (1995): ‘Why Auction the Spectrum?’Tele-communications Policy, Vol 19, No 3, p 191-99.Milgrom, P (1995):Auction Theory for Privatisation, Cambridge University Press, p 11.Ozanich, Gray W, et al (2002): ‘Wireless Auction: A Barrier to New Services for Local Communica-tion’, paper presented to the 23rd Conference and General Assembly, International Association for Media and Communication Research, Law Division, Barcelona, Spain, July 25.International Livestock Research InstituteTata-ILRI Livestock Programme ManagerA senior livestock specialist is required to manage an innovative programme of applied research in support of livestock development for poverty reduction in India. The International Livestock Research Institute (ILRI) and the Sir Ratan Tata Trust (SRTT) will develop a programme, including analysis of investment opportunities in livestock development, action research to provide sustainable solutions to technical, institutional and policy constraints to livestock development and monitoring and evaluation of research and development activities.The post holder will be responsible for the day-to-day management of the programme, including supervision of a team of Research Assistants, liaising with SRTT and partners and ILRI subject matter specialists and management of the programme budget. The selected candidate will have a sound understanding of livestock systems in India and their role in rural livelihoods, particularly in underprivileged communities, experience of project management in the field of livestock research or development, strong leadership skills and excellent communication skills. The post will be based in the ILRI Office in New Delhi, but will require frequent travel to project locations in India and occasional travel to ILRI headquarters in Nairobi and in Addis Ababa.Qualifications: A degree in veterinary or animal science, agricultural economics, or other relevant discipline and a postgraduate qualification (Masters or PhD) in a relevant area.Further information: http://www.ilri.org/research/Jobsshow.asp
INSIGHTEconomic & Political Weekly EPW september 13, 200833The views expressed are the author’s own.Manas Bhattacharya (manasb2@gmail.com) is with the ministry of tourism, New Delhi.The International Experience of Auctioning SpectrumManas BhattacharyaThis article analyses the early experiences of spectrum auction in some advanced countries and distils lessons from them. An issue that has been examined in this context is whether auctionnecessarily leads to a market-efficient outcome. This is of topical interest in India at the current juncture since the department of telecommunications is on the verge of holding an auction forallotting the spectrum for the third-generation (3G) of mobile phones.This article aims to analyse the early experiences of auctioning spec-trum in advanced countries in order to distil lessons and understand whether an auction necessarily leads to a market efficient outcome. Findings of this study would be relevant in the context of the recent discussions in the Indian tele-com sector about introducing auctions as a method of spectrum assignment for the potential service providers of the 3G mobile technology.The paper is organised as follows. The first section presents the experience of New Zealand, Australia and the United States (US) in auctioning spectrum, how these auctions were designed and what results were achieved. An example is cited in this section, which shows that auction prices can reflect gaming behaviour rather than the intrinsic value of spectrum. The second section cites three instan-ces of auctions conducted by the Federal Communications Commission (FCC) that led to disappointing results. Again, the outcome was due to a number of extra-neous factors rather than the real worth of spectrum itself. These experiences drive home the point that auctions neednot necessarily lead to a win-win situation between the government and service providers.The third section presents the results at the other extreme – the exciting 3G auc-tions of the United Kingdom (UK) that net-ted a huge amount of revenue. This sec-tion also shows that per capita auction revenue earned from 3G licences by seven European countries around the same peri-od varied widely. It appeared that a range of factors, some of which are exogenous can influence auction outcomes. In this context, a view point is presented in this section that high bids, delay in payments and delay in services could be parts of a larger manipulative game.Some arguments and counter arguments have been presented in the fourth section as to whether an auction does necessarily yield competitive results. A reference has been made in this context to a study of the outcomes of the European Universal mobile telecommunications sys-tem/international mobile technology (UMTS/IMT)1 – 2000 auctions.The fifth section concludes the paper by distilling some of the lessons drawn from these international experiences and analysing them from policy perspectives. Several administrations of the world use auctions to award spectrum licences. While the US uses only auctions and no other method to assign licences for com-mercial use, some other countries like Australia, Canada and theUK have used auctions to assign certain bands yet main-taining the legacy of administrative pric-ing of spectrum in one form or another.An auction is a market-based approach, which puts spectrum in the hands of those who values them the most. Unlike other methods, auctions determine allocation and pricing of licence (or spectrum for that matter) simultaneously. In other methods, separate sets of criteria need to be evolved for allocation and pricing.In a simplistic way, the auction process can be described as a fair, transparent, objective and expeditious method of awarding licences. In theory, winning prices of the bidders are supposed to reflect the value the market places on spectrum based on the bidders’ realistic appreciation of the market conditions. One difficulty with this approach is the lack of certainty regarding the continued efficient use of spectrum once an assign-ment is made after the auction is over. Moreover, its applicability is limited to the licensing of spectrum for the services, which are commercially viable. Prices of spectrum used by non-profit agencies including various government depart-ments cannot be directly determined through auction.There are various types of auction mod-els used in different countries like sealed tender auctions, open outcry, simultane-ous ascending multiple round auctions, etc. The same model undergoes variations 
INSIGHTseptember 13, 2008 EPW Economic & Political Weekly34in matters of details and designs in differ-ent countries at the implementation stage.Based on the popular auction experi-ences in some of the countries, the present paper seeks to examine whether an auc-tion simulates competitive behaviour and leads to competitive prices that truly reflect the intrinsic value of spectrum con-sistent with bidders’ rational intention. The paper also discusses the extent to which an auction can be described as a objective process and explores the basis of a common apprehension that exaggerated bidding in auction may have catastrophic consequences for the telecom market.International ExperiencesThough certain episodes of auction suc-ceeded in raising large sums of money, analyses of the outcomes of a wider sample of auctions held in different countries do not exhibit consistent and optimistic results to support high revenue expectations on a sustained basis. There are instances of fail-ures of auctions too. Moreover, values gen-erated in auctions and bidders’ rational valuation of intrinsic worth of licences do not necessarily converge.New Zealand’s first auctions of nation-wide ultra high frequencies lots of 8 MHz licence took place in 1990 and were based on sealed tender. For five licences, the total auction revenue was justNZ $ 36 mil-lion against NZ $ 250 million projected by the consultants. Bidders were, however, permitted to buy and sell licences after the auction.Early auction experience in Australia is an interesting case study. Australia’s spec-trum auctions in 1993 followed the sealed tender method. There was no provision of advanced deposits or default penalty. Moreover, there was no restriction against multiple bids by the same bidders. As per the rules, the licence would be offered to the second highest bidder if the highest bidder failed to make payments within a stipulated time. The results were quite unexpected. To quote Milgrom (1995):In April of 1993, two newly formed compa-nies surprised the established players by placing unexpectedly high bids of A$ 212 million and A$ 177 million for two satel-lite television licences. One of the bidders had paid-in capital of just A$100. When the requisite time period had elapsed and neither bidder had paid, the government rejected those bids and announced the sec-ond highest bids, which were submitted by the very same bidders. Under the rules, this was perfectly legal. The bidders were again allowed a period of time to find the neces-sary funding before those bids, too, were rejected. This pattern of default, rejection and announcement of the new winning bid continued for 10 months. While the licence assignment was delayed, the bidders negoti-ated with potential buyers. The first licence was eventually bought by the defaulting bid-der for just A$ 117 and immediately resold for A$ 21 million in profit. The second was acquired for A$ 77 million and resold for an unknown price but presumably at a profit.Between 1927 and 1982, the US followed the “comparative hearing” method to assign spectrum licences based on evaluation of competing proposals from the point of view of public interest. However, this was a time consuming process. In 1982, the Congress authorised the FCC to assign licences through lotteries. Any telephone company, regardless of its qualification or experience was eligible to participate in the lottery. Licence rights were transferable. A second-ary market was therefore allowed.The upshot of this plan was the creation of a new business for Washington law firms: organ-ising “telephone companies” for the sole pur-poses of filing lottery applications to acquire a valuable cellular telephone licences. During the lottery era, more than four hundred thou-sand applications were filed. Most lottery win-ners resold their rights at large profits to older telephone companies. In one case, a “tele-phone company” that had won the right to supply cellular services to Cape Cod promptly resold their licence to the Southwestern Bell Telephone Company for $ 41 million.With estimates of the value of the spectrum given away in tens of billions of dollars and with mounting political pressures to do something about the hugeUS federal budget deficits, the Congress decided to put an end to what the legislation called the “unjust en-richment” of lottery winners. The Omnibus Budget Reconciliation Act passed in July of 1993 authorised the FCC to conduct auctions for rights to use the airwaves and mandated that it begin assigning licences within just 10 months. The licence selling price, this time, was to accrue to the US Treasury, rather than to some randomly determined lottery winner [ibid, McMillan 1995].Among several varieties of auction models, the FCC adopted simultaneous multiple rounds of auction, which allowed bidders to raise their bids in each round after prices were announced. This process continued till no new bid was received. The auction was simultaneous in the sense that a set of licences was auctioned simul-taneously, which allowed bidders to take into account interdependencies among licences and bid for more than one licence together. Another advantage of this meth-od was that a bidder could bid on the basis of his knowledge of his competitors’ bid prices and was not required to bid for more than that was necessary to acquire a licence. An important feature included was “activity rule”, which required each bidder to maintain minimum levels of bidding activity in each round to qualify for later rounds of auction. In order to avoid frivo-lous or mischievous bids a penalty was introduced for the withdrawal of bids. “Nationwide narrowband auction” (for paging and messaging services) of the US which tried out the simultaneous auction in July 1994, raised $ 17 million and set a new US record. In the same year, the second trial of simultaneous multiple rounds auction was launched in October for the sale of “region-al narrowband licences”. A special feature of this auction was the large discount and preferential financing terms offered to small business and woman and minority owned firms in any licences acquired in certain bands. The results showed that the prices in the October auction were significantly higher than the average prices of the com-parable national licences purchased in July.In December 1994, 99 licences, mostly for wireless telephone applications were put to auction. These licences were offered for 51 major trading areas (MTAs).2 In order to increase the pace of auction in an effec-tive manner theFCC introduced stipula-tion on bid increments and retained its discretion to modify the number of rounds per day during the auction. Direct revenue from this auction exceededUS$ 7 billion.Analysts have studied the bidding strategies adopted in the auctions and drawn inferences, which called into question the perceived straightforward-ness of bidding as a tool for buying licences, which were of the greatest value to the respective bidders. In reality, the bidding behaviour seemed more complex than theory. It would be of interest to refer to a report on the bidding behaviour of the general telephone and electronics (GTE), one of the bidders in the December 
INSIGHTEconomic & Political Weekly EPW september 13, 200835auction in the US (discussed in the earlier paragraph).Insider David Salant’s account reports that GTE’s goals early in the auction were to avoid bidding on target markets, partly to keep ri-vals from guessing thatGTE’s target markets were Atlanta and Seattle, maintain eligibil-ity and push up the price in non-target mar-kets, partly to avoid pushing up the price in target markets and partly to induce rivals to spend more money on those markets. The point of keeping rivals guessing was that had we not done so, we might have been taken advantage of by others who wished to use up our budget [Milogram 1995].Regardless of whether the prices of licences reflected their opportunity cost or not, the US treasury received substan-tialrevenue. It shows that the auction prices can result from gaming behaviour and need not necessarily reflect the intrin-sic value of spectrum.Disappointing AuctionsAmid success stories there have been cer-tain disappointing auctions too. In 1994, the FCC held auction to sell licences for interactive video and data services. In this auction, licences were sold one at a time. Many of the bidders were speculators and several of the bidders, unhappy that they had paid more for their licences than others paid for similar licences later in the auc-tion, chose to default on their bids(ibid).One of the broadband personal com-munication service (PCS) auctions of the FCC, identified as the C-block auction, held in 1995, ran into serious problems reportedly because of favourable treat-ments earmarked for the “designated entities” (ie, small businesses and women or minority owned firms) to encourage their participation.The FCC has used bidding credits, set-asides and installment payments to encourage the participation and success of designated enti-ties. The idea is that without special treat-ment, these small businesses would find it dif-ficult to compete with the large incumbents. The favoured treatment can serve to “level the playing field” and there by foster innovation and intensify competition....The auction failed largely because of overly attractive instalment payments (10 per cent down and six-year interest only at the risk-free 10-year treasury rate). This encouraged speculative bidding, which led to all the major bidders defaulting and declaring bankruptcy. Even now, years after the auction, much of this C-block spectrum lies unused, tied up in bankruptcy litigation [Cramton 2001b].Another instance, held in disappoint-ment, was the wireless communication services (WCS) auction of the FCC held in April 1997. There was very little competi-tive activity for the 128 licences that were put to auction. Various explanations have been offered for the lack of success of this auction. According to Taylor Simmons of Simmons Associates,3 a combination of factors, including the uncertainty sur-rounding WCS and the initial steep mini-mum bid increment was responsible for lack of competition. The FCC initially set the bid increment at 2 cents per mega hertz per post office protocol,4 which was subse-quently reduced to 0.25 cents for round 7. However, this reduction was of no conse-quence [Carlson 1997]. The total net reve-nue in this auction was US $ 6.2 million.According to Cramton (2001a), the design of this auction did not address dif-ficult trade-offs in effective implementa-tion issues and the auction was conducted in a hurry.The main problem was the stringent out-of-band emission limit. Equipment manufactur-ers warned that this would threaten the com-mercial viability of this spectrum. The low prices at auction and the absence today of ac-tivity in this band appears to confirm that the equipment manufacturers were right. At the time of the decision, the FCC was facing a dif-ficult tradeoff between the rights of prior win-ners of neighbouring licences and WCS use. Such decisions are always difficult but the FCC was under intense time pressure to meet the timetable that the Congress set for the auction. This aggressive timetable may well have led the FCC to make too hasty a decision on interference rules, which damaged the value of this spectrum. The Congress’s desire for receiving revenues according to its fiscal calendar may have resulted in substantially reduced auction revenue.Exciting 3G AuctionsOne of the most exciting auction results of the recent past relates to the auction of five third-generation (3G) mobile wireless licences of the UK held in April 2000, which raised an astonishing £ 22.5 billion. According to Cramton (2001b): “although theUK did not begin auctioning spectrum until 2000, it began with a bang. Its very first auction broke into the record books as the world’s largest auction ever.Cramton offered a set of plausible expla-nations of why prices were so high. Accord-ing to him (ibid): (1) the UK was the first mover in 3G auctions in Europe and there-fore reaped the benefit of first mover’s advantage as the foot in the door to Europe and potentially the world; (2) the auction took place at the peak of an apparent high-tech stock bubble. Bidding companies’ images in the stock market came to be asso-ciated with their abilities to win a licence; and (3) ascending auction format coupled with large excess demand might have led to a competitive spiraling up of prices.The government of theUK auctioned five 3G licences, each of 20 years duration. One large licence was set aside for a new entrant. A “spectrum cap” was introduced through the stipulation that any company (or associated company) can win at most one licence. There were also requirements of a deposit of £ 50 million initially and an additional deposit of £ 50 million when the bid exceeded £400 million. The deposit was refundable to the unsuccess-ful bidders. The payment was to be made in full at the end of the auction or in instal-ments in sufficiently unattractive terms. Other features of the design of the auction included specification of minimum open-ing bids, activity rule, number of rounds per day and the stopping rule. The auction was fully transparent with bidders’ identi-ties known to each other and the bid prices were announced after each round. Withdrawals of bids were not allowed.It may be noted that not all the govern-ments of Europe auctioned 3G licences. Finland, Spain, Norway, Sweden and France chose “comparative hearing” whereas the UK, the Netherlands, Germany, Italy, Austria, Switzerland, Belgium and Denmark chose to use auctions. The timing of auctions differed from country to country with theUK being the first to auction in April 2000 and Denmark held 3G auctions in August 2001. Apart from timing, these auctions also differed from each other in design and details. The following table (p 36) shows the number of 3G licences auctioned by these coun-tries (arranged in sequence of the timing of auction) and per capita revenue gener-ated in these auctions.It has been observed that for the coun-tries, which held auctions later in the period, the timing of auctions coincided with the bubble burst of the stock prices in the tele-com market. This might have had an impact 
INSIGHTseptember 13, 2008 EPW Economic & Political Weekly36on auction results [Damme 2002]. Differ-ences in revenue might have also been contributed by other factors such as details of the auction design, policy and regulatory framework, nature of telecom market, level of competition, market share of the existing incumbents, expectations regarding future profit, etc.The lesson that emerges from the varied auction experiences described above is that the auction results depend upon the particular contexts in which they take place. A large number of factors, some are policy induced and others, exogenous and circumstantial, tend to impact the results of auction. Experiences of the 3G auctions were different in Belgium and Singapore. Both the Belgian auctions ended after the submission of the initial bids at the reserve price. Neither auction had excess demand [Crampton 2001b]. Even after such a record breaking experience of 3G auction in the UK, many licences remained unsold in the auction for fixed broadband wire-less access held by the UK government subsequent to 3G auction.5Yet another example, the 3G licence auction in Hong Kong held in September 2001 fetched the minimum price because out of the six incumbent 2G operators, only four bid for a 3G licence and there was no new entrant.Ozanich et al (2002) advanced an alter-native interpretation of the phenomenon of high bid, delay in payments and delay in services. Their explanations go beyond the perceived notion that high auction prices result from exaggerated profit expectations of the bidders.Analysing 3G wireless auction results on the basis of discounted cash flow method to calculate pay back period, the authors concluded that companies overpaid for 3G spectrum. The German and UK auctions stand out in particular. Cash flow require-ments to justify investment in 3G licences in these two countries appeared unrealis-tic. The authors (ibid) have also noted:Perhaps a more interesting element of the auction process is the delay in auction pay-ments. In theUS, of the $ 41 billion bid in wireless auctions, only $ 14 billion has been collected. Evidence shows “a similar pattern is already playing out in Europe” [New York Times 2001]. Companies are also requesting that amount of the winning bids be reduced.…Finally, most of the winners of 3G auctions were incumbents or owners of existing 2G li-cences. These companies continue to gener-ate revenue and positive cash flow from these facilities while 3G is delayed.“Win-at-any-cost” can be a part of a larger strategy to block entry, hoard spectrum, delay payment, delay services and continue to earn profit over the existing network.Customary ex post facto analyses of auction results tend to isolate a number of factors holding them responsible for per-ceived success or failure of auctions. Table: 3G Licence Auctions by West Europe (April 2000-August 2001)Country Number of Licences Per Capita Auction Auctioned Revenue (Euro)UK 5650Netherlands 5 170Germany 6 620Italy 5210Austria 6 105Switzerland 4 20Belgium 4 45Denmark 4 95Sources: Damme Eric van (2002).CENTRE FOR MICRO FINANCEFountain Plaza, Pantheon Road, Egmore, Chennai – 600008 Call for Applications: Microfinance Researchers’ Alliance ProgrammeThe Centre for Micro Finance (CMF), in collaboration with the Ford Foundation, is initiating a Microfinance Researchers’ Alliance, an intensive research capacity development programme targeting Indian researchers currently working on, or interested in working on, microfinance research. This programme, the first of its kind in the country, aims to cultivate and develop a network of top mid-career researchers on microfinance over 3 years. For this programme, CMF will select approximately 20 motivated, committed researchers in their early- to mid-career. The programme extends support to participating researchers over the course of three years in the form of workshops and research seminars; exposure visits to CMF project sites; providing a forum for idea exchange with a network of critical thinkers on microfinance issues; opportunities to work on CMF projects with leading development researchers from MIT, Harvard, Yale, IIM, and other top international institutions; and potential project funding support.Eligibility CriteriaGiven the extent of financial and non-financial support the participating researchers would receive, the selection process will be highly rigorous and competitive. The following is a list of prerequisites for individuals that apply for the programme:i) PhD holder in academic disciplines including economics, political science, sociology, anthropology, women’s studies, statistics and finance. ii) Experience in leading a field research team.iii) Associated with a university, teaching college or an autonomous research institution.iv) Motivated, enthusiastic young mid-career researchers who are able to commit time and effort for 3 years of research.Interested researchers can visit www.ifmr.ac.in/cmf/rap.html and download the necessary documents and instructions for application. All documents, including letters of recommendations, should be sent in one package and arrive at CMF no later than October 31. Applications will be evaluated upon receipt, and short-listed candidates will be invited to the first meeting to be held on November 14-15, 2008 in Delhi. This programme is financially supported by the Ford Foundation. 
INSIGHTEconomic & Political Weekly EPW september 13, 200837Credibility of such explanations is difficult to establish in the absence of counter-factuals. By and large, such explanations seek to find reasons with the designing aspects of auctions, which in most part remain judgmental with outcome, always tentative. The Swiss wireless-local-loop auction of March 2000, for example, sold nationwide licences in sequential ascend-ing auctions. While the first two licences for a 28 MHz block were sold at 121 million francs and 134 million francs respectively, the third and largest licence of 56 MHz fold for only 55 million francs. As an explana-tion of these paradoxical results, the blame was put to the fact that the licences were sold in sequence rather than simultaneously [Cramton 2001a].Does Auction Yield Competitive Results?Observations of real life situations have given rise to certain attempted characteri-sation of behavioural outcomes of auctions, which do not support the perceived notion that auction tends to generate the intrinsic value of spectrum that should prevail in a competitive market. Even theoretically, anti-competitive behaviour is possible in certain situations. For instance, in a pre- existing monopolistic market, a monopolist will value a new licence more than a new entrant because the latter, if he wins, will be a threat to the monopolist’s profit. The value that the monopolist places on a new licence will partly reflect his preference for continuing monopoly and partly the per-ceived market value of the new licence. The auction price, in such a situation is not expected to reflect pure competitive price [Gilbert and Newbery 1982].In the context of UMTS auctions, Damme (2002) observed that important asym-metries and value differences do arise between bidders, for example, between the incumbents and the entrants. For an incum-bent, not winning a licence means that he might lose not only his new 3G customers but also his existing 2G business. Further-more, another element of asymmetry arises from the fact that the cost of rolling out a network may be much lower for an incum-bent than for an entrant. Jehiel and Moldovanu (2000), who examined the outcomes of European UMTS/IMT-2000 auctions, observed that incumbent carriers consistently paid more than new entrants. Arguably, this could be the result of the barrier to entry strategies rather than cost advantages and rational bidding. An interesting example of anti-competitive behaviour relates to the Dutch UMTS auc-tions, where a new comer Versatel partici-pated in the auction not to win but to get concession from other players, who were existing incumbent operators. A press release issued by Versatel shortly before the auction revealed their intention: “We would however not like to see that we end up with nothing whilst other players get their licences for free. Versatel invites the incum-bent mobile operators to immediately start negotiations for access to their existing 2G networks as well as entry to the 3G market either as a part owner of a licence or as a mobile virtual network operator”.6 This resulted in Telfort, a British Telecom subsidiary and participant in the auction overbidding Versatel and terminating auc-tion at revenue that was 17 per cent higher. It is, therefore, apparent that an auction may lead to various possible outcomes. Overpricing or undervaluation of spec-trum are possible outcomes in an auction depending upon the circumstances.Noted economist of the FCC, Evan Kwerel has argued that several commonly perceived notions of adverse consequences of spectrum auctions are actually common misconceptions about auctions. One com-mon misconception is that auctions tend to overprice spectrum and thereby under-mine the development of the communica-tion industry, a phenomenon described as “winners curse”. According to Kwerel (2001), the cost of winning a licence in an auction will have no effect on the price of telecommunication services to consumers. The price/or tariff that a consumer will pay will depend upon the market clearing price and forces of supply and demand.Telecommunications experience in theUS has also been consistent with the theory that historic costs do not alter pricing. For example, within a given market, the prices charged by cellular operators who obtained their licences via comparative hearings or lotteries are no lower than the prices of those firms that purchased their cellular licences in the secondary market or firms that obtainedPCS licences in an auction. Similarly, where a US cellular licence has been bought at a significant cost from a party that obtained it at no cost, we have notobserved any increase in consumer prices.A company, who overpays for a licence may be in financial distress or bankruptcy leading to a one time loss in investment and consequent change in the ownership of the licence. However, an appropriate legal frame-work should be in place to ensure that the licence does not get trapped into protracted transition delaying its utilisation to provide services to consumers. It may also be noted that if a winner in any auction has overpaid for a licence, which would not match his expected profit, he would soon resell the licence rather than hold on to it. The new buyer of the licence will invest in infrastruc-ture and the serviceswill not be held up.According to Kwerel, indeed, the US experience to date shows no apparent lack of investment in facilities of cellular orPCS operators that bought their licences rela-tive to firms that acquired their licences in a comparative hearing (ibid).Kwerel has also argued against the pos-sible misconception that an auction might lead to the monopolisation of spectrum. According to him, as long as there is a free-dom to resell, any method of spectrum assignment may potentially lead to monopo-ly. However, there are extraneous methods of preventing monopoly through limits on spec-trum aggregation, setting a cap on spectrum use by a single entity. Auction can also be structured to create or increase competition (ibid). Out of 27 FCC auctions of non-broad-cast spectrum, small businesses have success-fully obtained licences in 22.Kwerel’s arguments are theoretically sound. He had also substantiated his conten-tions by citing the US experiences in auction. One issue that emerges clearly and easier recommended than done, is the framework to facilitate subsequent market corrections in the event of high bids and default. How expeditiously this can be done is an open question. Costs of wasted investment and time to relocate a licence in the event of a market failure should be carefully weighed against possible benefits of auction.ConclusionsThere are certain lessons, which can be dis-tilled from the above portrayal of the inter-national experience of spectrum auctions: One, the revenue outcome in spectrum auctions is tentative. Two, the distinction 
INSIGHTseptember 13, 2008 EPW Economic & Political Weekly38between “allocation of spectrum” and “assignment of spectrum” needs to be kept in mind. While allocation defines different characteristics of spectrum such as the fre-quency band, geographic area, time period, restriction on use, etc, assignment refers to the allotment of spectrum. An auction is a method of assigning licences. If spectrum allocation is inefficient, a highly efficient auction will deliver poor results [Cramton 2001a]. Three, a large number of discretion-ary parameters enter into auctions. What variant of auctions is to be used, the details of the design, etc, are to be decided in advance. Experiences in other countries suggest that there is no “one-size-fits-all” approach in auction. There is always an ele-ment of uncertainty regarding what will work and what will not. In the absence of counterfactuals, it is difficult to establish cause-and-effect relationships even in ex post facto analysis of auction results. Four, decisions to be taken at the planning stage of auction may turn out to be complex at times. One tricky area relates to “packaged bidding” where a bundle of licences with large complementarity can potentially earn more auction revenue per licence than if individual licences are put to bid separately and sequentially. Such decisions require careful planning and analysis. Five, auctions may require a framework of efficient clear-ing of encumbered spectrum if the vacant band is not available. Six, winner in an auc-tion may not use licences efficiently. How-ever, the terms and conditions or rights and restrictions in a licence can be framed appro-priately to safeguard public interest. How-ever, these conditions will, no doubt, affect bidders’ willingness to pay for the licence [Kwerel 2001]. Finally, regulators should consider any request by the winning bidders to change the terms of the auction after con-test with scepticism. Such auctions may be a fall-back strategy to the “win-at-any-cost” outcome of wireless telephony auctions [Ozanich 2002]. It may be mentioned that regulators in England and Germany had denied the requests of the companies after the 3G auction to reduce the licence fees.Notes1 UMTS/IMT is one of the 3G cell technologies. 2 MTAs do not overlap with states, but are areas surrounding major business centers.3 Simons Associates is an internationally known consulting firm.4 POP is an Internet standard protocol.5 One plausible reason cited is that the minimum opening bid was set at too high a level [Cramton 2001].6 ‘Versatel Disappointed with Dutch UMTS Auction Tomorrow’, Business Wire, July 6, 2000. http://findarticles.com/p/articles/mi_m0EIN/is_2000_July_6ai_63171362 visited on August 14, 2008.ReferencesCarlson, C (1997): ‘WCS Auction Sees Sluggish Start’, Wireless Week.Cramton, P (2001a): ‘Spectrum Auctions’, University of Maryland, pp 31-32.– (2001b): ‘The Auction of Radio Spectrum for the Third Generation of Mobile Telephones’, the report commissioned by the National Audit Office of the United Kingdom, pp 47.Damme, Eric van (2002), ‘The European UMTS – Auctions’, European Economic Review, Vol 46, pp 846-58.Gilbert, R and D Newbery (1982): ‘Preemptive Patent-ing and the Persistence of Monopoly’,American Economic Review, Vol 72, pp 514-26.Jehiel, P and B Moldovanu (2000): ‘Licence Auctions and Market Structure’, discussion paper, Mannheim University, referred in Ozanich Gray W, et al (2002).Kwerel, E (2001): ‘Auctioning Spectrum Rights’, US Federal Communications Commissions, pp 4.Mcmillan, J (1995): ‘Why Auction the Spectrum?’Tele-communications Policy, Vol 19, No 3, p 191-99.Milgrom, P (1995):Auction Theory for Privatisation, Cambridge University Press, p 11.Ozanich, Gray W, et al (2002): ‘Wireless Auction: A Barrier to New Services for Local Communica-tion’, paper presented to the 23rd Conference and General Assembly, International Association for Media and Communication Research, Law Division, Barcelona, Spain, July 25.International Livestock Research InstituteTata-ILRI Livestock Programme ManagerA senior livestock specialist is required to manage an innovative programme of applied research in support of livestock development for poverty reduction in India. The International Livestock Research Institute (ILRI) and the Sir Ratan Tata Trust (SRTT) will develop a programme, including analysis of investment opportunities in livestock development, action research to provide sustainable solutions to technical, institutional and policy constraints to livestock development and monitoring and evaluation of research and development activities.The post holder will be responsible for the day-to-day management of the programme, including supervision of a team of Research Assistants, liaising with SRTT and partners and ILRI subject matter specialists and management of the programme budget. The selected candidate will have a sound understanding of livestock systems in India and their role in rural livelihoods, particularly in underprivileged communities, experience of project management in the field of livestock research or development, strong leadership skills and excellent communication skills. The post will be based in the ILRI Office in New Delhi, but will require frequent travel to project locations in India and occasional travel to ILRI headquarters in Nairobi and in Addis Ababa.Qualifications: A degree in veterinary or animal science, agricultural economics, or other relevant discipline and a postgraduate qualification (Masters or PhD) in a relevant area.Further information: http://www.ilri.org/research/Jobsshow.asp 
