还剩12页未读,继续阅读
本资源只提供10页预览,全部文档请下载后查看!喜欢就下载吧,查找使用更方便
文本内容:
Wireless CommunicationsbyJoshuaS.Gans,Stephen P.King JulianWrightand
1.IntroductionIn1895,Guglielmo Marconiopened the way formodern wireless communications bytransmittingthe three-dot Morsecode for the letterS overa distanceof threekilometers usingelectromagneticwaves.From thisbeginning,wireless communicationshas developed into akeyelement ofmodern society.From satellitetransmission,radio andtelevision broadcastingto thenowubiquitous mobile telephone,wireless communicationshas revolutionizedthewaysocieties function.This chaptersurveys the economics literature on wireless communications.Wirelesscommunications and theeconomicgoods andservices thatutilise ithave somespecial characteristicsthathave motivatedspecialised studies.First,wireless communicationsrelies ona scarceresource-namely,radio spectrum-the property rights forwhich weretraditionally vestedwith thestate.In ordertofoster the development of wireless communications including telephonyand broadcastingthoseassets wereprivatised.Second,use of spectrum for wireless communicationsrequired thedevelopmentof keycomplementary technologies;especially thosethat allowedhigher frequenciesto beutilised moreefficiently.Finally,because ofits specialnature,the efficientuse of spectrum requiredthe coordinateddevelopmentof standards.Those standards in turnplayed acritical rolein the diffusion oftechnologiesthat reliedon spectrum use.In largepart ourchapter focuseson wirelesstelephony rather than broadcastingand otheruses ofspectrume.g.,telemetry andbiomedical services.Specifically,the economicsliterature onthatindustry hasfocused onfactors drivingthe diffusionof wirelesstelecommunication technologies and onthenature ofnetwork pricingregulation andcompetition in the industry.By focusingon theeconomicliterature,this chaptercomplements othersurveys in this Handbook.Hausman2002focuses ontechnologicaland policydevelopments inmobile telephonyratherthaneconomic researchper se.Cramton2002provides asurvey of the theoryand practiceof spectrum auctions used for privatisation.Armstrong2002a andNoam2002consider generalissues regardingnetwork interconnectionandaccess pricingwhile Woroch2002investigates the potential forwireless technologiesas asubstitutefor localfixed linetelephony.Finally,Liebowitz andMargolis2002provide ageneral surveyof theeconomicsliteratureonnetwork effects.In contrast,we focushere solelyon theeconomic literatureonthe mobile telephony industry.The outlinefor thischapter is as follows.The nextsection providesbackground informationregardingthe adoptionof wirelesscommunication technologies.Section3then considersthe economicissuesassociated withmobiletelephonyincluding spectrumallocation andstandards.Section4surveysrecent economicstudies of thediffusionof mobiletelephony.Finally,section5reviews issuesofregulation andcompetition;in particular,the needfor andprinciples behindaccess pricingfor mobilephone networks.
2.BackgroundMarconis pioneeringwork quicklyled tovariety ofcommercial andgovernment particularlymilitarydevelopments and innovations.In the early1900s,voice and then music无线通信约书亚圣甘斯,斯蒂芬・金和朱利安赖特著Introduction概述
1.1895年,Guglielmo Marconi利用电磁波把3位摩尔斯电码编码的字母S传递到3公里以外的地方,为现代无线通信指明了道路至此,无线通信就慢慢成为了现代社会的重要组成部分从卫星传输,广播电视,到现在无处不在的移动电话,无线通信已经使社会发生了革命性的变化这一章主要是概述经济文献对无线通信的研究现状无线通信为了支持无线通信的发展(包括电话和广播),私有化应运而生其次,基于无线通信的频谱使用要求互补性技术的发展,尤其是及其在经济活动应用中的一些特点刺激了关于无线通信的专门研究首先,无线通信是建立在一种稀有资源之上的,即,无线电频谱,而它往往为国家所掌控要能使那些高频率的频谱得到更有效的利用最后,因为其本身的特性,高效率的频谱使用又要求标准的协调发展这些标准反过来又大大推动了基于频谱使用的技术的扩散这一章,我们的重点是无线电话,而不是广播或其他种类的频谱使用技术(例如,遥测和生物医学服务)而且,这些文献主要关注的是,在这个产业内,推动无线通信技术扩散的因素和网络定价的规则及其竞争力通过关注这些经济文献,本章补充说明了整个手册中其他调查中的一些问题Hausman主要研究的是移动电话技术和政策的发展而不是经济研究本身Cramton侧重于,在频谱私有化过程中,频谱拍卖的一些理论与实践Armstrong和Noam主要研究网络互连和接入定价的有关问题,与此同时Woroch关注的则是无线技术取代本地固定电话的可能性而Liebowitz andMargolis则侧重于网络效应而我们主要是侧重于研究移动电话产业的一些经济文献本章的结构如下第二部分主要是采用无线通信技术的背景信息第三部分讨论的是与移动电话有关的一些经济问题,包括频谱分配和标准第四部分是最近关于移动电话普及的经济学研究最后,第五部分关注的是规则和竞争力的问题,尤其是,移动电话网络接入定价的需求和原则Background背景
2.Marconi开创性的工作很快就引起了一系列商业和政府部门(尤其是军事)的发展和变革20世纪初,声音和音乐先后开始能够通过电磁波传递,而且现代收音机诞生1920年,商用无线电站,底特律WWJ和匹兹堡KDKA建立无限电报也于1900在盎格鲁-布尔战争中由在南非的英国军队首先使用英国海军在迪拉果阿湾的船只上使用由Marconi提供的设备相互联系船只是无线电报早期最重要的客户到1912年,当铁达尼号使用无线电发出求救信号的时候,船只的无线电标准已经形成在此之前,人们已经认识到了,要发挥无线通信的作用必须加强国际合作合作包括两个方面首先是无线电传输中可能的干扰,也就是说至少人们要通过协调以避免信号冲突其次,由于频谱在诸如海上安全和导航等领域的应用牵扯到不同国家,因此必须通过国际协调来保证这些服务的畅通这就要求政府干预来协调无线电频谱的合理分配2频谱分配.1无线电传输涉及电磁频谱的使用电磁能量通过不同的频率传输,而且能量的特点也14142取决于频率例如,可见光的频率在4X10至IJ
7.5X10Hz之间Ultra紫外射线,X一射线和伽玛射线频率较高(或波长教短),而红外辐射,微波和无线电波频率教低(波长较长)无线电频谱中的电磁射线的频率在3000Hz—300GHz之间即使在同一无线电频谱中,不同的频率也具有不同的性能Cave曾经说过,频率越高,信号传播的距离越短,容纳的数据越多国际间协调无线电频谱使用,防止干扰和制定全球标准的任务是由国际电信联盟承担的,它成立于1947年,前身可以追述到1865年左右,是联合国下属的一个特殊机构,由180多个成员国组成其中的无线电通信部按照无线电规则协调全球的频谱使用这些规则是在1906年的柏林国际无线电报会议上第一次设立的无线电频谱的分配需要考虑三方面的因素一一频率,地理位置和用户优先权(为了防止干扰,先到先得)无线电频谱被分为八个频段,从最低(3—30kHz)到最高(30—300GHz)整个世界被分为三个地区ITU就在全球或地区性的基础上分配一定的频段然后,不同国家在ITU分配的基础之上进一步分配本国的频段例如,在美国,联邦通信委员会频率分配表就是基于国际和美国国内的频率分配表共同形成的用户被分为主要和次要两个部分,主要部分的用户免受次要用户的干扰,反之不成立例如,2003年,9kHz以下的波段在国际上和美国都没有被分配无论国际的还是美国的分配表,抑或者是国际上所有地区,9-14kHz都被用于无线电导航,而14-70kHz则用于海上通信和固定无线通信,且为主要用户虽然国际上已经有了一个时间信号,为20kHz但美国又增加了一个额外时间频率,60kHzo从70—90kHz波段,国际上地区间的差异开始出现在无线电导航,无线电定位和水上移动用途中用途和优先权都不相同依此类推,频谱分配一直到300GHz,300GHz以上的频段在美国没有被分配,而国际上没有被分配的则是275GHzITU为预防相互干扰要求各成员国在采用某一频率工作时,例如无线电基站或新的卫星,必须遵守公示和登记程序
2.2无线服务的应用范围无线电频谱有着广泛的用途,可以分为以下几个方面广播服务包括短波,AM和FM电台以及地面电视语音和数据的移动通信包括用于船舶和飞行器上的海上和航空移动通信,陆上固定基站与类似出租车无线电和寻呼机等不定点之间的移动通信,移动用户与固网或移动用户之间的移动通信,例如移动电话固定服务点对点或一点对多点服务卫星用于广播,电信和因特网,特别是长距离通信业余无线电其他用途包括军事,射电天文,气象和科学用途实现以上功能所使用频谱数量因为国家和波段的不同而有所不同例如,在英国从88MHz到1GHz之间这一波段,其中40%用于电视广播,22%用于国防,10%用于GSM移动通信,1%用于海上通信相反,从1GHz到3GHz这一波段,没有安排电视通信,19%用于GSM通信和3G移动电话,17%用于国防,23%用于航空雷达不同种类的无线通信设备的数量正在急剧增加传感器和嵌入式无线控制器越来越多地用于一系列电器和应用工具上PDA和笔记本电脑也越来越多的通过无线通信收发邮件和接入因特网,象机场贵宾休息室这样可供电脑无线上网的地方也不在希奇尽管如此,在过去二十年中,无线通信最重要也是最剧烈的变化还是移动电话的出现
2.3移动电话的出现和普及移动电话的历史可以分为四个时期第一个(前蜂窝)时期移动电话只能在特定地区使用特定的频率电话经常占线,通话质量也不好如果用户在某一地区使用了某一频率,那么其他人就无法使用这一频率而且,美国FCC分配给移动电话使用的频率数量也比较少,限制了同时通话的数量德国也建立了类似的系统,有名的如A-Netz和B-Netz蜂窝技术的使用极大的提高了移动电话频率的使用效率相对于在一个很大的区域一个电话单独使用一个频率,蜂窝技术把一个移动电话的通信分成了儿个小的区域或单元不同的用户在不同的(不相邻)的单元可以使用同一频率进行通话而不会受到干扰第一代蜂窝移动电话在不同的国家采用了相互不兼容的模拟技术例如,20世纪80年代,美国采用了AMPS系统,英国采用了TACS,德国采用的是C-Netz,而斯堪的纳维亚半岛国家使用了NMT系统这就导致了彼此间的不兼容,尤其是在欧洲,尽管整个美国使用的是统一的AMPS系统二代移动电话则使用了数字技术虽然改进了早期模拟技术的一些问题,但二代技术的使用在美国和欧洲还是有很大的不同因为政府干预,欧洲采用了统一的技术标准第一个二代系统GSM在20世纪80年代建立可直到90年代,在ETSI的主持下,GSM才得以标准化标准化GSM能完成国际漫游,自动定位服务,常见的加密和传输相对高品质音频的功能GSM是目前在全球使用最广泛的第二代系统,应用于130多个国家,使用900兆赫的频率范围同时,在美国却出现了一系列不兼容的二代标准包括TDMA,(GSM的近亲)和CDMA,即Time andCodeDivision MultipleAccess respectively.在一个单元内,如何分流电话从而有效使用频谱,这些系统采用了不同的技术由于美国人没有采用统一的2G标准,所以哪个系统最好的争论一直存在,又加上手机漫游和升级的相关利益问题,因此,整个90年代AMPS依然是美国通用移动技术的主流移动电话最新的发展趋势是3G,其最大特点在于通信的高速率和数字服务例如,3G电话收发电子邮件将会更加快捷,而且能从网上下载音乐和视频等内容类似拍照手机等设备拍摄的图片也会传输的更快在IMT-2000计划下,ITU正试图协调建立一个统一的国际3G标准IMT-2000计划决定3G技术本应采用CDMA系统,可CDMA却有至少两个以上的竞争对手,因此最后折中,采用了一套办法,而不是单独的一个系统在ITU2000年全球无线通信大会上,世界范围内IMT-2000系统使用的频率确定下来尽管有很多公司都宣布了在接下来几年进军3G的计划,但只有日本在2002年开始了3G营运移动电话数量一直保持很好的增长势头80年代几乎没有人知道移动电话是什么,可到了2002年,据估计,全世界平均每一百人就拥有
15.57部移动电话当然,国与国之间的普及水平有很大的不同在美国,每一百人拥有
44.2部移动电话,而法国是
60.53部,德国
68.29,芬兰
77.84,英国
78.28因此,美国的电话拥有率是低于欧洲的一些富裕国家的在欧洲和美国之外,澳大利亚的拥有率是
57.75%,新西兰是
62.13%,日本是
58.76%毫无疑问,普及率与经济发展水平密切相关,所以意大利2002每一百人只拥有
0.63部移动电话,肯尼亚为
1.60部,中国
11.17,马来西亚
29.95在一些国家移动电话的数量甚至超过了固定电话,例如,德国,法国,英国,希腊,意大利和比利时但在美国,加拿大和阿根廷固定电话依然超过移动电话2001年,日本两者的普及率相当但整体来看,移动电话的普及速度远远高于固定电话不同国家间移动电话服务的资费难以比较部分是因为汇率的变化,但主要是因为不同国家的价格套餐和定价形式不同最明显的是,不同的国家有不同的收费制度,美国以外的国家主要采取打电话的一方付费但在美国和加拿大,移动电话采用接电话乙方付费的计费方式不同的套餐,绑定服务和通话资费也似的比较难于进行90年代后期,最重要的变革之一就是移动电话收费采用了预付卡这种消费者在打电话之前就付费而不是在之后再付帐单的方式在很多国家取得了成功例如,在瑞典,预付卡采用两年以后就占领了25%的市场分额尽管定价方式不同,但0ECD估计,从92年到98年,平均个成员国移动电话的主要绑定服务成本下降了25%
3.无线通信的经济问题
3.2稀缺的频谱资源无线电频谱是一种具有特殊用途的自然资源就象上文提到的,它的性能不是均匀的,特定的部分最好用于特定用途频谱也是有限的,因为只有部分电磁频谱能用于无线通信,尽管频率是否合适和传输系统的容量取决于技术无线电频谱又是无限的,不是说今天使用的频段以后就不能用了,可它无法储存在ITU的指导下,频谱按不同用途分配,然后又根据合适的用途分配给特定用户以前,用户分配都是靠政府法令而且用户一般也是政府所有80和90年代的私有化以及一些国家移动电话产业内竞争的成功促进了90年代频谱分配更加公平从90年代早期开始,无线电频谱的用户,尤其是2G和3G移动电话频谱的用户一直都是由两种主要的方法之一选出来的,“选美比赛”或是拍卖“选美比赛”需要用户向政府或政府任命的委员会提交商业计划胜利者就在这些提交了计划的用户中产生胜利者可能付政府一笔费用,尽管可能最想购买频谱的潜在用户不一定在胜利者中例如,90年代英国就使用“选美比赛”的方法颁发了2G牌照瑞典和西班牙也用这种方法颁发了3G牌照法国用同样的方法颁发了四个3G牌照国家电信的管理者要求公司必须于2001年一月低之前提交申请这些申请将会被根据事先定好的标准评级,并打分,总分为500这些标准包括就业率(25分)、服务(多达50分)以及部署速度(多达100分)胜出的公司必须支付政府设定的高额牌照使用费最后,只有两家公司提出了申请它们2001年六月拿到了牌照,剩下的两个无人问津利用市场机制来配置频谱使用劝和预防电磁干扰的概念可以最少追述到50年代,先后由Herzel
(1951)和Coase
(1959)提出但当频谱拍卖普及的时候,已经过去三十多年了新西兰1989年修改了法律允许拍卖频谱,到了90年代早期,与移动电话,电视,无线电广播以及其他私人无线服务有关的频谱分配也开始采用拍卖的方式1993年八月分,美国修改法律允许FCC通过拍卖颁发无线频谱牌照,到96年七月分FCC已经举行了七次拍卖,颁发了2100张牌照,包括通过两次拍卖颁发的两张全美范围的2G牌照2000年,英国拍卖了5张3G牌照,共计大约$34b拍卖有一系列不同的形式,有新西兰的第二价格密封出价,美国的改性升序出价和混合升序出价以及英国的荷兰式拍卖竞标者在参加拍卖时必须满足一定的标准,比如服务保障和参与保证金在一定区域一个公司能够拍得的牌照数量是有限的,防止垄断从经济学角度来说,频谱拍卖能够确保频谱在用户手中发挥最大价值虽然频谱能拍卖到不同的用户手中,但依然要优先集中保证一些特殊用途对于一些频谱的获得权从市场角度来说,这可能会导致频谱使用的效率低下一个特定波段的用户可能愿意出更高的价钱购买与其相邻的波段,而这个波段确已经被其他用户购得但是一些波段的优先分配能够保证这些用户无法通过相互交易而获利根据现有的牌照使用章程,即使是双赢,用户也不能随意更改频谱的用途作他用Coase1959,Valletti2001提出了一个频谱所有劝相互贸易的系统,能利用市场在分配频谱用途的同时确定用户电磁干扰问题也能通过相邻频谱所有者之间的产权分配和谈判解决Valletti说过无论是竞争问题还是法定标准问题都只能由频谱使用权的这个市场来解决接下来我们讨论的是标准问题,竞争的问题将会在下面的第5部分讨论Noam1997进一步发展了可交易频谱分配这一概念由于技术进步,信号可以在发送端打包成很多单独的数据信息而在接受端又可以重新聚合,这就意味着永久性的频谱分配在不久将来显的很多余在此基础上,Noam提出特定频段分配可以通过市场完成这样频谱用途的价格就能反映出其用途的实际价值DeVany1998也涉及过基于市场的频谱分配政策,出了未来建立一个开放、商用、非绑定的频谱市场体系的可能美国FCC的频谱拍卖一直是冲突不断参与拍卖的1850-1910MHz和1930-1990MHz波段已经有了私人固定点对点用户FCC最后决定现有的用户有最多三年的时间与新的用户谈判可替代频谱和补偿问题如果谈判失败,现有用户将被强制重新分配Cramton,Kwerel andWilliams1998调查了一系列现有频谱重新分配的替代选择,发现美国重新分配的情况大致相当于讨价还价虽然经济学家一般都提倡频谱分配要通过拍卖,可拍卖依然饱受批评Binmore和Klemperer2002认为很多反对拍卖的观点都被误解了但是Noam1997都Gruber2001b批评说频谱拍卖导致了垄断,竞争力不足而Gruber认为技术进步提高了频谱使用的效率,增加了无线服务的竞争力例如,从频谱使用效率来看,GSM移动电话大约是早期模拟系统的40到50倍频谱拍卖是在频谱分配之后进行的政府通常先分配固定的频段给一些适当服务然后,再决定这一频段内拍卖的牌照的数量所以拍卖的价格和无线服务的竞争力是由频谱数量和政府事先分配给这项服务的牌照数量共同决定的虽然拍卖导致了对于稀缺频谱的竞争,但依然不能靠市场决定竞争的最好形式Noam认为,为了克服人为的市场的不公平,分配体系应具有进入的灵活性
3.3频谱使用补偿作用基于频谱的无线通信服务容易导致服务与服务之间以及地区和地区之间的协合作用Binmore和Klemperer就曾提到英国的3G频谱拍卖可能存在2G和3G移动通信基础设施的协合作用在2G电信产业,相对于后来者,现有的电信企业拥有巨大的优势不仅仅是现存企业2G生意对于3G的补偿,而且运营3G需要的建设基础的成本,现存企业也比后来者少的多,因为他们可以在现有2G的基础设施上进行建设因此,在能否利用现存品牌向消费者提供新的产品和2G以及3G的服务之间存在协合作用地域间的协合作用能从FCC的2G拍卖中很明显的看出来Moreton和Spiller1998研究过美国1995-96年中的两次移动电话拍卖为了获得一定牌照地区的人口数量,有竞争力和规范的环境以及任何协和作用的效果,他们对每个牌照采取了反向递减的拍卖形式同样也有升序拍卖,这样最后的牌照成交价格就大致等于最后一位竞标者的出价显然,相关协和作用是和最后一位的竞标者的网络有关的,在出价的基础上发挥这一网络的最大作用为了获得地域协和作用的效果,Moreton和Spiller假设竞标者事实上拍得的网络与其期待的完全相同他们把地理协同作用分为本地和全球本地协和考虑的是某地区的牌照价值与其相邻地区2G牌照所有者的关系全球协和考虑的是最终意义上的胜出者拍得的全国性网络Moreton和Spiller发现了本地协和的存在“在本地协和中,我们发现对于单独一个竞标者和儿个竟标者来说,若干组两个或两个以上相邻的牌照显然对于前者更有价值当在一个地区内增加想邻的牌照时,本地协调却急剧下降,而全球协调则微弱在Moreton和Spiller的研究当中,现存蜂窝服务的本地覆盖能降低2G牌照的费用这似乎和Binmore和Klemperer关于不同移动电话服务之间的经济领域理论相冲突Moreton和Spiller认为负面关系可能反映竞争的减少公司都不愿意投资更多与现存的模拟移动电话公司竞争,而是愿意把他们有限的资源使用在其他地方可是,这一理论有点牵强在升序拍卖中,竞标者都会根据自己的评估出价,如果说在现存的模拟电话服务和2G服务中存在积极的协和,那么无论其他竞标者出什么价格,最后,牌照的价值只能是中标者的出价正如Moreton和Spiller的发现一样,2G牌照的价格与市场的人口数量和人口增长率成正比,与其服务的地区大小成反比这些发现只是一些感性研究,与Ausubel的发现大体上一致人口和需求好象总是积极相连,所以无论哪个层次的竞争,增加的人口总是会增加利润但无论有多少人口,更大的区域只会增加建设2G蜂窝网络的成本,降低利润Moreton和Spiller的研究还表明,管理者对于现存模拟电话网络税收的管辖权总是会增加2G牌照的价格这表明现有服务的税收政策削弱了竞争,整体上推高了服务价格管理者的通告和有关价格信息的出版物可能削弱竞争的现象,已经在其它产业中表现出来was transmittedand modernradio wasbom.By1920,commercial radiohad beenestablished withDetroitstation WWJand KDKAin Pittsburgh.Wireless telegraphywas firstused by the Britishmilitaryin SouthAfrica in1900during theAnglo-Boer war.The Britishnavy usedequipment suppliedbyMarconi tocommunicate between ships inDelagoa Bay.Shipping was a majorearly clientfor wirelesstelegraphyand wireless was standard for shippingby the timethe Titanicissued itsradioidistress callsin
1912.Early on,it wasquickly recognizedthat internationalcoordination wasrequired forwirelesscommunication to be effective.This coordinationinvolved twofeatures.First,the potential forinterference inradio transmissionsmeant thatat leastlocal coordination was neededto avoidthetransmission ofconflicting signals.Secondly,with spectrum to beused forinternational communicationsand areas such as maritimesafety andnavigation,coordinationwasnecessary between countries toguaranteeconsistency in approach to these services.This drovegovernment interventionto ensurethecoordinated allocation of radio spectrum.
2.1Spectrum AllocationRadiotransmission involvesthe use of partof theelectromagnetic spectrum.Electromagneticenergy istransmitted indifferent frequencies and theproperties of the energydepend on the14142frequency.For example,visible lighthas afrequency between4x10and
7.5x10Hz.Ultra violetradiation,X-rays andgamma rayshave higherfrequencies orequivalently ashorter wavelength whileinfraredradiation,microwaves andradio waveshave lowerfrequencies longerwavelengths.The radiofrequencyspectrum involveselectromagnetic radiationwith frequencies3between3000Hz and300GHz.Even within the radio spectrum,different frequencieshave differentproperties.As Cave2001notes,the higherthe frequency,the shorterthe distancethe signalwill travel,but thegreater thecapacity of thesignal tocarry data.The tasksof internationallycoordinating the useofradiospectrum,managinginterference andsetting globalstandards areundertaken by the InternationalTelecommunication UnionITU.The TTUwas createdby theInternational Telecommunications4Convention in1947but haspredecessors datingback to approximately
1865.It isa specialistagency ofthe United Nationswith over180members.The RadiocommunicationSector of the ITUcoordinates globalspectrum usethrough theRadioRegulations.These regulationswere firstput inplace atthe1906Berlin InternationalRadiotelegraphConference.Allocation of the radiospectrum occursalong threedimensions-the frequency,thegeographic locationand thepriority of the userwith regardsto interference.The radiospectrum isbrokeninto eightfrequency bands,ranging fromVery LowFrequency3to30kHz up to ExtremelyHighFrequency30to300GHz.Geographically,the worldis alsodivided intothree regions.The ITUthenallocates certainfrequencies forspecific useson eithera worldwideor aregional basis.Individualcountries maythen furtherallocate frequencieswithin the ITU internationalallocation.For example,inthe United States,the FederalCommunications CommissionsFCCs table of frequencyallocations isderivedfrom boththe international tableof allocations andU.S.allocations.Users arebroken intoprimary andsecondary services,with primaryusers protectedfrom interferencefrom secondaryusers butnotvice versa.As anexample,in2003,the bandbelow9kHz wasnot allocated in theinternational orthe U.S.table.9to14kHz was allocated toradio navigationin bothtables and all internationalregions while14to70kHz is allocated withboth maritimecommunications andfixed wirelesscommunications asprimaryusers.There isalso an international timesignal at20kHz.But the U.S.table alsoadds anadditional timefrequencyat60kHz.International regionaldistinctions beginto appearin the70to90kHz rangewithdifferences in use andpriority betweenradio navigation,fixed,radiolocation andmaritime mobileuses.These allocationscontinue rightup to300GHz,with frequenciesabove300GHz notallocatedin theUnited States and thoseabove275GHz notallocatedin theinternationaltable.5The ITUdeals withinterference byrequiring membercountries tofollow notificationand registrationprocedures whenever theyplan toassign frequencytoaparticular use,such asa radiostation ora newsatellite.
2.2The range of wireless servicesRadio spectrum is used for awide rangeof services.These can be brokeninto thefollowing broadclasses:•Broadcasting services:including shortwave,AM andFM radioas wellas terrestrialtelevision;•Mobile communicationsof voiceand data:including maritimeand aeronauticalmobile forcommunicationsbetweenships,airplanes andland;land mobilefor communicationsbetween afixedbase stationand movingsites such asataxi fleetand pagingservices,and mobilecommunicationseither betweenmobile users and afixed networkor betweenmobile users,suchas mobile telephone services;•Fixed Services.either pointto pointor pointto multipointservices;•Satellite,,used forbroadcasting,telecommunications andinternet,particularly overlongdistances;•Amateur radio;6•Other Uses:including military,radio astronomy,meteorological andscientific uses.The amountof spectrum allocated to these differentuses differsby countryand frequency band.Forexample,in the U.K.,40%of the88MHz to1GHz bandof frequenciesare used for TVbroadcasting,22%for defense,10%for GSM mobile and1%for maritimecommunications.In contrast,none of the1GHz to3GHz frequencyrange isusedfortelevision,19%isallocatedto7GSM andthird-generation mobile phones,17%to defenseand23%for aeronauticalradar.The number of differentdevices usingwirelesscommunicationsis risingrapidly.Sensors andembeddedwireless controllersare increasinglyused in a variety of appliancesand applications.Personaldigital assistantsPDAs andmobile computers are regularlyconnected toe-mail andinternet servicesthroughwirelesscommunications,and wirelesslocal areanetworks forcomputersarebecomingcommon inpublic areaslike airportlounges.However,by farthe mostimportant anddramatic changeinthe useofwirelesscommunicationsin the pasttwenty yearshas beenthe riseof themobile telephone.
2.3The riseand riseof mobiletelephonyThe historyof mobile telephones canbe brokeninto fourperiods.The firstpre-cellular periodinvolvedmobile telephonesthat exclusivelyused afrequencybandin aparticular area.These telephoneshadsevere problemswith congestionand callcompletion.If onecustomer wasusing aparticularfrequency ina geographic area,no othercustomer couldmake a call onthat samefrequency.Further,thenumber offrequencies allocated bythe FCC in the U.S.to mobiletelephone seiwiceswas small,limitingthe number of simultaneouscalls.Similar systems,known asA-Netz andB-Netz weredeveloped inGermany.The introductionof cellulartechnology greatlyexpanded theefficiency of frequency useof mobilephones.Rather thanexclusively allocatinga bandof frequencyto onetelephone callinalargegeographic area,a celltelephone breaksdown ageographicareainto smallareas orcells.Different usersindifferent non-adjacent cellsare ableto usethe samefrequency for acallwithout interference.First generationcellular mobile telephones developedaround theworld usingdifferent,incompatibleanalogue technologies.For example,in the1980sin the U.S.there was the AdvancedMobile PhoneSystemAMPS,the U.K.had theTotal AccessCommunications SystemTACS,Germany developedC-Netz,while Scandinaviadeveloped theNordic MobileTelephone NMTsystem.The resultwas awiderangeoflargely incompatiblesystems,particularly inEurope,although thesingle AMPS systemwas usedthroughout the U.S.Second generation2G mobile telephones useddigital technology.The adoptionof secondgenerationtechnology differedsubstantially betweenthe United States andEurope andreverses theearlier analoguemobileexperience.In Europe,a commonstandard wasadopted,partly dueto governmentintervention.Groupe SpecialeMobile GSMwas firstdeveloped in the1980sandwas thefirst2G system.But it wasonly in1990that GSMwas standardizedwith thenew nameof GlobalSystem forMobilecommunication under the auspices of theEuropean TechnicalStandards Institute.The standardizedGSMcould allowfull internationalroaming,automatic locationservices,common encryptionandrelatively highquality audio.GSM isnow themost widelyused2G systemworldwide,in morethan130countries,using the900MHz frequencyrange.In contrast,a variety of incompatible2G standardsdevelopedin the UnitedStates.These includeTDMA,a closerelative ofGSM,and CDMA,referring toTime andCode DivisionMultiple Accessrespectively.These technologiesdiffer inhow theybreak downcalls to allow for more efficientuse ofspectrumwithin a single cell.While thereis someargument asto thebetter system,the failureof the U.S.to adopta common2G standard,with theassociated benefitsin terms of roamingand switchingofhandsets,meant thefirst generationAMPSsystemremained themost popularmobile technologyin theU.S.throughout the1990s.The finalstage in thedevelopmentof mobiletelephones isthe moveto thirdgeneration3Gtechnology.These systemswill allowfor significantlyincreased speedsof transmission and areparticularlyuseful fordata services.For example,3G phonescan moreefficiently beusedfore-mailservices,and downloadingcontent suchas musicand videosfrom theinternet.They canalso allowmorerapid transmissionof images,for examplefrom cameraphones.An attemptto establishaninternationalstandardfor3G mobileis beingmoderated through theITU,undertheauspicesofits IMT-2000program.IMT-2000determined that3G technologyshould bebasedon CDMAsystems butthere areat leasttwo alternativecompeting systemsand IMT-2000did notchooseasinglesystem butrather asuite of approaches.At theITUs WorldRadiocommunicationConference in2000,frequencies forIMT-2000systems wereallocated ona worldwidebasis.By2002,the only3G system in operationwas in Japan,although numerouscompanies haveplans toroll out3Gsystems in the nextfew years.The growthinuseof mobiletelephones has been spectacular.From almosta zerobase inthe early1980s,mobile penetrationworldwide in2002is estimatedat
15.57mobile phonesper100peopleworldwide.Of course,the level of penetrationdiffers greatlybetween countries.In the UnitedStates,there were
44.2mobiletelephonesper100inhabitants,with penetrationrates of
60.53in France,
68.29inGermany,
77.84in Finlandand
78.28inthe United Kingdom.Thus,in generalmobile penetrationislower inthe U.S.than inthe wealthierEuropean countries.Outside Europeand theU.S.,the penetrationratein Australiais
57.75,
62.13in NewZealand,and
58.76inJapan.Unsurprisingly,penetration ratesdependon thelevel ofeconomic development,so thatIndia hadonly
0.63mobiletelephonesper100inhabitants in2002,with
1.60for Kenya,
11.17for China,and
29.95for Malaysia.The numberof mobilephonesnow exceedsthe numberof fixed-wire telephonelines ina varietyof countriesincludingGermany,France,the UnitedKingdom,Greece,Italy andBelgium.However,the reverseholds,with fixed-linesoutnumbering mobilesintheUnitedStates,Canada,and Argentina.Penetration rateswere closeto equalinJapan in2001,but inall countries,mobile penetrationis risingmuch fasterthanfixed lines.8The pricefor mobilephone servicesare difficultto comparebetweencountries.Tn partthis reflectsexchangerate variations,but moreimportantly pricingpackages and the formof pricingdifferssignificantly betweencountries.Most obviously,different countrieshave differentcharging mechanisms,with callingparty paysdominating outsidetheUnitedStates.But intheUnitedStatesandCanadareceiving partypays pricingoften appliesfbr callsto mobiletelephones.Different packagesandbundling ofequipment andcall chargesalso makecomparisons difficult.A majorinnovation inmobiletelephone pricinginthelate1990swastheuseof pre-paid cards.This system,where customerspay inadvancefor mobilecalls ratherthan beingbilled ata laterdate,has provedpopular inmany countries.For example,in Sweden,pre-paid cardsgained25%of themobile marketwithin twoyears oftheirintroduction OECD,2000,p.l
1.Despite thechanging patternsof pricing,the OECDestimates thattherewasa25%fall inthe costof arepresentative bundleof mobile services overits membercountriesbetween1992and1998OECD,2000,p.
22.
3.Economic Issuesin WirelessCommunications
3.1Spectrum asa scarceresourceRadio spectrumisanatural resource,but onewith ratherunusual properties.As notedabove,it isnon-homogeneous,with differentparts ofthe spectrumbeing bestusedfordifferent purposes.It isfinitein thesense thatonly partoftheelectromagnetic spectrumis suitableforwirelesscommunications,although boththe availablefrequenciesand the carryingcapacity ofany transmissionsystem dependontechnology.The radiospectrumis non-depletable;using spectrum today does not reduce the amountavailablefor useinthefuture.But itisnon-storable.Under ITUguidance,spectrum hasbeen allocatedto specificuses andthen assignedto particularusersgiven the relevant use.Traditionally,user assignmentwas bygovernment fiat.Not infrequently,the userwas governmentowned.Privatizations inthe1980sand1990s,andthesuccess ofat leastlimited mobiletelephonecompetition insome countries,resulted ina morearms-length processof spectrumallocation developinginthe1990s.Users ofradiospectrum,and particularlyusers of2G and3G mobiletelephone spectrum,have generallybeen chosenby oneof twobroad approachessince theearly1990s-a beauty contest9oran auction.A beautycontest9involves potential users submittingbusiness plansto thegovernment orits appointedcommittee.The winnersare thenchosen fromthose firmssubmitting plans.There maybe somepaymentto thegovernment bythe winners,although thepotentialusermost willingto pay for thespectrumneed notbe amongthe winners.For example,theU.K.used a beautycontestapproach toassign2G mobiletelephone licenses inthe1990s.Sweden andSpain haveused beautycontests toassign3Glicenses.France usedabeautycontest toassign four3G licenses.The nationaltelecommunications regulatorrequiredfirms tosubmit applicationsbytheend ofJanuary
2001.These applicationswere thenevaluatedaccording topreset criteriaand givena markout of
500.Criteria includedemployment worthup to25points,service offeringsup to50points andspeed ofdeployment upto100points.Winningapplicants faceda relativelyhigh licensefee setbythegovernment.As aresult,there wereonly twoapplicants.These firmsreceived theirlicenses inJune2001,with theremaining twolicenses unallocatedPenard,
2002.The concept of usinga marketmechanism toassign propertyrights overspectrum andto dealwithissues suchas interferencegoes backto atleast the1950swhenitwascanvassed byHerzel1951andthen byCoase
1959.But itwas morethan thirtyyears beforespectrum auctionsbecame common.NewZealand alteredits lawstoallowspectrumauctions in1989andintheearly1990s auctionswere used toassign blocksof spectrumrelating tomobiletelephones,television,radio broadcastingand othersmallerservices toprivate managementCrandall,
1998.In August1993,U.S.law wasmodified toallow theFCCto useauctions toassign radiospectrum licensesand byJuly1996theFCChad conductedsevenauctions andassigned over2,100licenses Moreton and Spiller,
1998.This includedthe assignment oftwo new2G mobiletelephone9licensesineach regionoftheU.S.through twoauctions.In2000,theU.K.auctioned offfive3Giolicenses fora totalpayment ofapproximately$34b.Auctions haveinvolved avarietyofformats includingsecond pricesealed bid in NewZealand,modified ascending bidintheU.S.andamixed ascendingbid andDutch auctionformatiiin theU.K.Bidders may have tosatisfy certaincriteria,suchasservice guaranteesand participationdeposits,before theycan participateinthe auctions.Limits mayalso beplaced on the numberof licensesasingle firmcan wininaparticular geographicarea,so that theauctiondoesnotcreate amonopolysupplier.From aneconomic perspective,using anauction toassign spectrumhelps ensurethat thespectrumgoes to the highestvalue user.While auctionshave beenused toassign spectrumto differentusers,they stillinvolve apriorcentralized allocationof bands of spectrumto particularuses.Economically,this can lead to aninefficient useof spectrum.A userofaparticular frequencyband e.g.for3G servicesmight haveamuch higherwillingness-to-payforneighboring spectrumthan thecurrent userof thatneighboringspectrum e.g.a broadcasterorthemilitary.But theprior allocationoffrequencybands means that thesepartiesare unableto benefitfrom mutually advantageous trade.It wouldviolate the existing licenseconditionsto movespectrumallocatedto oneuse intoanother useeven ifthis ismutuallyadvantageous.Building on the workof Coase1959,Valletti2001proposes asystem oftradable spectrumrights,using the market toboth allocatespectrumtouses andsimultaneously assignit tousers.Interference canbedealt withthroughtheassignment ofpropertyrightsand negotiationbetween ownersof neighboringspectrum.Valletti notesthat bothcompetition issuesand issuesof mandatedstandards wouldneed tobeaddressed ina marketfor spectrumrights.We dealwith theissue ofstandards laterinthissection whilecompetitionissues areconsidered insection5below.Noam1997takes the conceptoftradable spectrumassignment onestage further.Technologicaladvancements,suchasthe abilityforasignal tobe brokeninto numerousseparate digitalpackets for thepurposes oftransmissionandthen reassembledon reception,meansthat theconceptof permanentspectrumassignment maybecome redundantinthenear future.As technologyadvances,Noam argues,spot andforward marketscanbeusedtoassign usewithin designatedbandsof spectrum.The priceofspectrum usewould thenalter toreflect congestionof use.DeVany1998also discussesmarket-basedspectrum policies,including thepotentialfora future“open,commoditized,unbundled spectrummarketsystem.p.641Conflicts inthe allocationof spectrumallocation aroseintheFCC auctionsintheU.S.The1850-1910MHz and1930-1990MHz bandstobeallocatedbythese auctionsalready hadprivate fixedpoint-to-point users.The FCCruled thatexisting usershad aperiod ofuptothree yearsto negotiatealternativespectrum locationand compensationwith newusers.If negotiationsfailed,the existingusercould beinvoluntarily relocated.Cramton,Kwerel andWilliams1998examine avarietyofalternativeproperty rightsregimes fbrnegotiated reallocationof existingspectrum andconclude thattheexperience oftheU.S.reallocations isroughly consistentwith simplebargaining theory.While economistshave generallyadvocated theassignmentofspectrum byauction,auctions arenotwithout theircritics.Binmore and Klemperer2002argue thata numberofthearguments againstauctionsare misguided.But bothNoam1997and Gruber2001b makethe criticismthat spectrumauctionsautomatically createa non-competitive oligopolyenvironment.Gruber arguesthattechnological changehas generallyincreased theefficiency ofspectrumuseand increasedthe viabilityofcompetition inwireless services.For example,in termsof spectralefficiency,GSMmobiletelephoneservices areapproximately fourto thirtytimes moreefficient thanearlier analoguesystems Gruber,2001b,Table
1.An auctionofspectrumrights,however,is precededby anallocationofspectrum.Thegovernment usuallyallocates afixed bandofspectrumtothe relevant services.Further,the governmentusuallydecides onthe numberof licensesthat itwill auctionwithin thisband.So theprice paidat theauctionandthelevelofex postcompetition intherelevantwirelessservicesare determinedby theamountofspectrumandthenumberoflicenses thegovernment initiallyallocates tothe service.Whilethe auctioncreates competitionforthescarce spectrum,it doesnot allowthemarketto determinetheoptimal formof competition.Noam arguesthat flexibilityof entryneeds tobe providedby theassignmentsysteminorder toovercome theartificial creationofanon-competitive marketstructure.
3.2Complementarities inspectrum useUsingspectrumtoproduce wirelesscommunications servicescanleadto synergies betweenservices andbetween geographic regions.In theU.K.,3G spectrumauction,thepotentialsynergiesbetween2G and3G mobiletelephone infrastructurewas notedby Binmoreand Klemperer:[T]he incumbentswho arealready operatinginthe2G telecomindustry enjoya majoradvantage overpotentialnew entrants.・・.Not onlyare theincumbents92G businessescomplementary to3G,but thecostsof rollingout theinfrastructure radiomasts andthe likenecessary tooperate a3G industryarevery substantiallyless thanthose ofa newentrant,because theycan piggybackonthe2G infrastructure.2002,p.C80Thus,there aresynergies intermsofbeing ableto supplynew productstoanexistingcustomer baseusing existingbrands,and economiesof scopebetween2G and3G services.Geographic synergiesare evidentfrom theFCC2G auctions.Moreton and Spiller1998examinethe two1995-96mobilephoneauctionsintheU.S.They runa reduced-form regressiononthe winningbid foreach licenseandanumberoffactors designedto capturethe demographicsoftherelevant licensearea,the competitiveand regulatoryenvironment,andthe effects ofany synergies.These wereascendingbid auctionsso thatthewinningprice isapproximately equaltothe second-to-last biddersvaluation forthelicense.As such,therelevantsynergies relatetothenetwork ofthe second-to-last bidder,to captureanyeffect ofthis networkonthe value ofthat bidder.To capturetheeffectof geographicsynergies,Moreton and Spiller assumethattheexpected networkassociated with anybidder isthe sameastheactual post-auction network.They categorizegeographicsynergies aseither localor global.Local synergiesconsider therelationship betweenvalue ofa licenseinone areaand ownershipof2G licensesin neighboringgeographic areas.Global synergieslook atthetotal extentofthesecond-to-last biddersnational network.Moreton andSpiller findstrong evidence of localsynergies.At thelocal level,our resultsindicatethat groupsof twoormoreadjacent licenseswere worthmore toasingle bidderthan to12separate bidders/9p.711These localsynergies appearto fallrapidly asthe geographicarea coveredbyadjacent licensesincreases andevidenceofglobal synergiesis weak.Local coverageby existingcellular servicestended toreducetheprice paidfor2G licensesin theMoreton andSpillerstudy.This appearsto runcounter tothe BinmoreandKlempererargument foreconomiesof scopebetween differentmobiletelephoneservices.MoretonandSpiller arguethat thenegativerelationship mayreflect areduction incompetition.Firms arereluctant tobid stronglyagainstexisting analogue mobiletelephoneincumbents andprefer touse theirlimited resourceselsewhere.Thisargument,however,is weak.In anascendingbidauction,participants willbid upto theirown valuationsandif thereare positivesynergiesbetweenexisting analoguemobileservicesand2G services,thisshould raise thevalue ofthelicense tothesecond-to-last bidderregardless ofany otherparties bids.As expected,MoretonandSpiller findthatthevalueofa2G licenseincreases withmarket populationandpopulation growthrate anddecreases withthe sizeofthearea served.These resultsare broadlyconsistentwith Ausubel,et.al.1997and areintuitive.Population anddemand arelikely tobe positivelycorrelatedso thatfor anygiven levelof competition,increased populationwill tend to increaseexpectedprofits.But increasedgeographicregiontends toraisetheroll-out costofthe2G cellularnetwork foranypopulation size,lowering expectedprofits.The MoretonandSpillerstudy findsome evidencethat thosejurisdictions whereregulators requiretarifffiling fortheexistinganaloguemobilephonenetworkstendtohave highervalues forthe2Glicenses.This suggeststhat tarifffiling onexisting servicesmayhavean anti-competitive effectleadingto higherprices overall.The potentialanti-competitive effectsassociatedwithregulatory notificationandpublication ofprice informationhasbeenshown inother industries.。