Close
About
FAQ
Home
Collections
Login
USC Login
Register
0
Selected
Invert selection
Deselect all
Deselect all
Click here to refresh results
Click here to refresh results
USC
/
Digital Library
/
University of Southern California Dissertations and Theses
/
Essays on the economics of radio spectrum
(USC Thesis Other)
Essays on the economics of radio spectrum
PDF
Download
Share
Open document
Flip pages
Contact Us
Contact Us
Copy asset link
Request this asset
Transcript (if available)
Content
ESSAYS ON THE ECONOMICS OF RADIO SPECTRUM by Ergin Bayrak A Dissertation Presented to the FACULTY OF THE USC GRADUATE SCHOOL UNIVERSITY OF SOUTHERN CALIFORNIA In Partial Fulllment of the Requirements for the Degree DOCTOR OF PHILOSOPHY (ECONOMICS) December 2009 Copyright 2009 Ergin Bayrak Dedication Dedicated to my family with much love and gratitude, the book of my life, my father, Burhan Bayrak, the light of my life, my mother, G¨ ulcan Bayrak, the joy of my life, my sister, Belgin Bayrak, without whose unconditional support I could not have accomplished this ii Acknowledgements During the development of this dissertation, I have been fortunate to benefit from the encouragement, guidance and support of many distinguished professors, mentors and friends. I would like to take this opportunity to recognise those who have helped make this experience one that I will cherish forever. I would like to extend my deepest gratitude to Simon Wilkie for the inspiration and guidance he provided throughout the development of this dissertation. He has inspired and encouraged me to tackle the questions I have always longed to try and answer. He has also been supportive in every step of the way, with his intellectual guidance, with the generous fellowship he extended at the Center for Communication Law and Policy, and with the professional experiences and insights he was kind enough to share. Thank you Simon. I would like to thank Jonathan Aronson, Francois Bar, Isabelle Brocas, Harrison Cheng, Svetlana Pevnitskaya and Goufu Tan for their input at various stages of the development of this dissertation. I am also deeply grateful to Caroline Betts for the encouragement, patience and trust she has shown when my steps faltered many times in this long journey. She has been an exceptionally diligent and resourceful colleague and a great mentor. iii I have been fortunate to spend these fruitful years of my graduate work at a great institution. IwouldliketothankthestaffoftheDepartmentofEconomicsattheUniver- sity of Southern California. Morgan Ponder and Young Miller have provided exceptional support and simply made things work. I am also grateful to the Center for Risk and Economics Analysis of Terrorism Events at USC and the distinguished colleagues there for the excellent research environment, and collaboration they extended. I would like to thank the Google Policy Fellowship Program for the generous support they provided during the summer of 2009. I am grateful to the Wireless Future Program and the Open Technology Initiative at the New America Foundation for their hospitality during my fellowship. I have also benefited from discussions with Pierre De Vries and Michael Calabrese who introduced me to the Capitol Hill dynamics and the realities of the policy arena. I would like to thank many friends who have helped make this a great experience. Thosewhofueledmyintellectualcuriosity,andthosewhohelpedmestaysaneandecstatic outside the academic world; your friendship will be cherished. Finally I would like to thank my family from the bottom of my heart for their love and patience. Although we have been oceans apart, I felt their unconditional support and trust every single moment. None of this would be possible without their invaluable presence. This work is dedicated to them with much love and gratitude. iv Table of Contents Dedication ii Acknowledgements iii List of Tables vii List of Figures viii Abstract ix Chapter 1: Introduction 1 Chapter 2: Economics of Spectrum Allocation 7 2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 2.2 Brief History of Spectrum Technology and Regulation . . . . . . . . . . . 11 2.3 Status Quo and Looking Ahead . . . . . . . . . . . . . . . . . . . . . . . . 18 2.4 Simple Economics Spectrum Allocation: Valuing Alternatives . . . . . . . 20 2.5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32 Chapter 3: Valuing Time Intensive Goods: An Application to Wireless and Wired Internet 34 3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34 3.2 Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39 3.3 Data and Estimation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42 3.4 Welfare . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46 3.5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48 Chapter 4: Welfare Effects of Spectrum Management Regimes 50 4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50 4.2 An Oligopoly Model of a Communications Device Market . . . . . . . . . 57 4.2.1 Model Preliminaries. . . . . . . . . . . . . . . . . . . . . . . . . . . 57 v 4.2.2 The Device Market . . . . . . . . . . . . . . . . . . . . . . . . . . . 59 4.2.2.1 Quantity Competition . . . . . . . . . . . . . . . . . . . . 60 4.2.2.2 Specification of Quality . . . . . . . . . . . . . . . . . . . 62 4.2.3 Equilibrium Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . 64 4.3 Simulations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66 4.4 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72 References 74 Appendices 76 Appendix A: Appendix to Chapter Two . . . . . . . . . . . . . . . . . . . . . . 76 Appendix B: Appendix to Chapter Three . . . . . . . . . . . . . . . . . . . . . 77 Appendix C: Appendix to Chapter Four . . . . . . . . . . . . . . . . . . . . . . 84 vi List of Tables 2.1 Consumer Surplus from Licensed Allocations . . . . . . . . . . . . . . . . 25 2.2 Cordless phone market data and consumer surplus calculations . . . . . . 27 2.3 Wi-Fi Chipset Shipments . . . . . . . . . . . . . . . . . . . . . . . . . . . 28 3.1 Summary Statistics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43 3.2 Regression of ln((1L i )=L i ) on ln(W) . . . . . . . . . . . . . . . . . . . 45 3.3 Consumer Surplus as a percentage of full income . . . . . . . . . . . . . . 47 A.1 Corless phone market data. . . . . . . . . . . . . . . . . . . . . . . . . . . 76 A.2 Cordless phone regression results . . . . . . . . . . . . . . . . . . . . . . . 76 B.1 ln(W) regression results without controls using lower bounds . . . . . . . . . . 78 B.2 ln(W) regression results controlling for Internet time use for work using lower bounds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79 B.3 ln(W) regression results controlling for Internet at work, assest and education using lower bounds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80 B.4 ln(W) regression results without controls using mid points . . . . . . . . . . . . 81 B.5 ln(W) regression results controlling for Internet time use for work using mid points 82 B.6 ln(W) regression results controlling for Internet at work, assest and education using mid points . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83 C.1 Simulation Algorithm in Pseudocode . . . . . . . . . . . . . . . . . . . . . 84 C.2 Model Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84 vii List of Figures 2.1 Basic economics of spectrum allocation . . . . . . . . . . . . . . . . . . . . 21 2.2 Basic economics of spectrum license prices . . . . . . . . . . . . . . . . . . 23 2.3 Cordless phone market data time series . . . . . . . . . . . . . . . . . . . 27 2.4 Home network composition (% of households) . . . . . . . . . . . . . . . . 29 2.5 Marginal social value of alternative allocations revisited . . . . . . . . . . 31 C.1 Consumer Surplus at SNR=0dB (Licensing in gray, Commons in color) . . . . . 85 C.2 Consumer Surplus at SNR=10dB (Licensing in gray, Commons in color) . . . . 86 C.3 Consumer Surplus at SNR=20dB (Licensing in gray, Commons in color) . . . . 87 C.4 Consumer Surplus at SNR=30dB (Licensing in gray, Commons in color) . . . . 88 C.5 Consumer Surplus at SNR=40dB (Licensing in gray, Commons in color) . . . . 89 C.6 Consumer Surplus at SNR=50dB (Licensing in gray, Commons in color) . . . . 90 C.7 Consumer Surplus at SNR=60dB (Licensing in gray, Commons in color) . . . . 91 C.8 Consumer Surplus at SNR=70dB (Licensing in gray, Commons in color) . . . . 92 C.9 Consumer Surplus at SNR=80dB (Licensing in gray, Commons in color) . . . . 93 C.10 Consumer Surplus at SNR=90dB (Licensing in gray, Commons in color) . . . . 94 C.11 Regions where Consumer Suplus is greater under Commons for all SNRs . . . . 95 C.12 The boundary of welfare dominance . . . . . . . . . . . . . . . . . . . . . . . 96 C.13 Cournot Equilibria when Native SNR is 0dB) . . . . . . . . . . . . . . . . . . 97 C.14 Cournot Equilibria when Native SNR is 10dB . . . . . . . . . . . . . . . . . . 98 C.15 Cournot Equilibria when Native SNR is 20dB . . . . . . . . . . . . . . . . . . 99 C.16 Cournot Equilibria when Native SNR is 30dB . . . . . . . . . . . . . . . . . . 100 C.17 Cournot Equilibria when Native SNR is 40dB . . . . . . . . . . . . . . . . . . 101 C.18 Cournot Equilibria when Native SNR is 50dB . . . . . . . . . . . . . . . . . . 102 C.19 Cournot Equilibria when Native SNR is 60dB . . . . . . . . . . . . . . . . . . 103 C.20 Cournot Equilibria when Native SNR is 70dB . . . . . . . . . . . . . . . . . . 104 C.21 Cournot Equilibria when Native SNR is 80dB . . . . . . . . . . . . . . . . . . 105 C.22 Cournot Equilibria when Native SNR is 90dB . . . . . . . . . . . . . . . . . . 106 viii Abstract Radio spectrum is a fundamental input in communications technology and the allocation of spectrum has gained ever increasing importance with technological advances. In this dissertation we start by recoupling the accurate scientific definition and progression of spectrum with the economic and policy analysis of spectrum. We then identify and pro- vide insights to the current valuation, allocation and management problems in spectrum equipped with this improved understanding. We go on to provide consumer surplus esti- mates for unlicensed spectrum that had been overlooked in the literature and show that the consumer surplus from the utilization of unlicensed spectrum is commensurate with thatoflicensedspectrumfoundinpreviousstudies. Toassessthepossiblenegativeeffects of excessive entry under unlicensed allocations, we continue with a model of a commu- nications device market whereby the allocation of spectrum determines the technological environment, the competitive environment and consequently the quality and the variety of the devices in the market. We then go on to simulate the welfare consequences of instituting the alternative management regimes. We conclude that the regime choice has to be thoroughly informed by the preference and technology structure and the optimal allocation must be an interior solution with a finely tuned mix of the two regimes. ix Chapter 1 Introduction Spectrum is an apparition, an image, or a sudden appearance of an unusual sight ac- cording to its Latin etymological roots. However, today the word represents arguably the most fundamental building block of contemporary communications technology. Al- though the physics and the technology of spectrum are understood to a great extent, much has been said in the economic and policy analysis of spectrum without a thorough understanding of what spectrum really is. The metaphors used to describe spectrum such as a scarce resource or real estate, are far from the reality of spectrum, and have resulted in misinformed economic analysis and policy approaches. In this dissertation we strive to recouple the accurate scientific definition and progression of spectrum with the economic and policy analysis of spectrum. We then identify and provide insights to the current valuation, allocation, and management problems in spectrum equipped with this improved understanding. Spectrum in its contemporary understanding refers to a range of electromagnetic wavesthatinclude,amongothers,lightwaves,radiowaves,infraredandultravioletwaves, X-rays and gamma rays. All matter in the universe emits electromagnetic radiation and 1 produceselectromagneticwavesatvariouswavelengthsorequivalentlyatvariousfrequen- cies. Also, it is possible to artificially create these electromagnetic waves by modifying the electrical and magnetic properties of matter. Spectrum refers to the whole range of these naturally occurring and artificially created electromagnetic waves. In essence, the very small waves between 400 and 700 nanometers in size that we perceive as various colors of light are no different than the radio waves between 1 millimeters and 100000 kilometers in size that we artificially create to be utilized by radios, TVs, cell phones or Wi-Fi routers. Therefor a more suitable metaphor to facilitate thinking about spectrum is the colors of light; some that we see and some we do not. In fact, as we discuss in chapter two, everything about spectrum we know today, and the wireless technologies we enjoy have all stemmed from the curiosity of a few brilliant minds, starting with Isaac Newton in 1671, about how the spectrum of colors emerged from white light. In radio communications, spectrum is utilized to carry information in the absence of physical wired connections and is fundamental in the operation of countless devices and servicesinoureverydaylives,fromcellphonestoTVs,fromwirelessnetworkingtogarage door openers, from intergalactic telescopes to baby monitors. Spectrum is a broadcast medium in the sense that the devices utilizing it emit or receive and interpret modulated electromagnetic waves for an intended recipient, but consequently, many devices trans- mitting information in close frequencies might make it difficult for the intended recipient to filter and decode the information intended for itself among other transmissions. As the number of communications systems and devices has grown significantly over time, more and more attention needed to be paid to the efficient use of spectrum in a way that 2 minimizes interference of equivalently maximizes the error free transmission of informa- tion. One way that this can be achieved is by dedicating certain frequencies to certain transmitter receiver pairs so that no unintended transmission would disturb the flow of information. This approach mostly reveals itself in the regulation of spectrum as the licensing regime. Another way of efficiently using spectrum relies on improving transmit- ters and receiver instead of dedicating certain frequencies to certain transmissions. As the technology continues to develop, transmitters become sophisticated enough to detect transmissions already in progress and stop transmitting before causing interference, and also receivers become advanced enough to filter and decode the intended transmission among other unintended flow of information. Technologies like software defined radios, cognitive radios, listen before talk protocols are few examples of such technologies. And therevelationofthislatterapproachinregulationofspectrumismostlyintheunlicensed allocation regime. There are merits to both approaches in different dimensions and the optimal allocation of spectrum needs to be thoroughly informed by both approaches. Since the invention and widespread adoption of radio communications in early 1900s, Government has been managing the radio spectrum in the public interest with its power vested in the Federal Communications Commission (FCC) and the National Telecommu- nicationsandInformationAdministration(NTIA).Thesespectrumregulatoryauthorities divide spectrum into blocks with distinct geographic and frequency boundaries and ded- icate these blocks to specific usages, or dedicate certain frequency blocks to common use of all prospective users. The rights to operate devices in these blocks are then allocated to competing users, either exclusively by employing auctions, or collectively through an unlicensed allocation. Most of the radio spectrum today is licensed to specific users and 3 uses such as mobile communication or broadcast licenses. The primary means of allo- cating these licenses have mostly been through comparative hearings under a command and control approach until as recently as 1994 when auctions have started to become the increasingly preferred allocation mechanism of the licensing regime. On the other hand, some of the radio spectrum is left open to use for everyone subject to technological rules and usage etiquettes under an unlicensed allocation regime. Wi-Fi and Bluetooth technologies and cordless phones are the most obvious product of unlicensed allocations. Spectrum technology is in a rapid progress with innovations in both new and more effi- cient uses of radio spectrum. However, this brings about new challenges and a need for improvedunderstandingofthealternativespectrumallocationandmanagementregimes. The simple economics behind efficient allocations requires the spectrum to be allocated such that the marginal social welfare of choosing licensed or unlicensed allocations are equal. However, calculating the marginal social welfare of prospective allocations is not an easy task, especially for unlicensed allocations. Therehavebeenstudiesthattrytoestimatethewelfarederivedfromtheutilizationof licensedspectrumbands. Hausman(1997)hasestimatedthattheintroductionofcellular communicationsresultedin$30to$50billionconsumersurplusperyearcorrespondingto between $14 and $24 per MHz per subscriber. Hazlett (2005) on the other hand, has esti- mated the consumer surplus from the Commercial Mobile Radio Services (CMRS) bands in 2003 to be $81 billion or around $2.17 per MHz per subscriber. However, the welfare gainfromtheuseofunlicensedspectrumbandshasnotreceivedcomparableinterestfrom economists. In particular this is a difficult exercise because unlike licensed spectrum that is used to provide few and relatively homogeneous services within a given band, there are 4 numerous and highly heterogeneous devices and services that use unlicensed spectrum bands like Wi-Fi and Bluetooth devices, cordless phones, radio frequency identification (RFID) devices, garage door openers, microwave ovens and so on. More importantly price and consumptions histories of such devices or services are rarely available. In fact the price is by definition zero for most services that use unlicensed spectrum. Realizing that it is not feasible to calculate and aggregate the welfare gains from all the goods and services that use unlicensed spectrum, in chapter three we take wireless networking for it is one of the most widespread uses of unlicensed spectrum, as a first attempt to calculate a lower bound on the welfare gains from goods and services that use unlicensed spectrum. We look at the consumer surplus derived from Internet con- sumption and we restrict our attention to those consumers who report to have some type of home network. We estimate the incremental consumer surplus that accrues to the consumers who connect to the Internet through wireless networks, over those who con- nect via wired Ethernet, power-line or HomePNA networks. The incremental consumer surplus is attributable to characteristics of networking devices that are brought about by the utilization of unlicensed spectrum. After establishing in chapter three that there is marginal value to be gained in unli- censed allocations, the second issue to be addressed is whether an unlicensed allocation would lead to a tragedy of commons result with excessive entry that degrades quality of devices or services because of interference. To root our paper in the recent debate on how to use the empty channels between broadcast bands, i.e. the white spaces that have become vacant with the digital transition, we will use the white spaces analogy but our model is applicable to any measure of spectrum on which prospective management 5 approaches are to be evaluated. We address the interference concern in chapter four by modelingacommunicationsdevicemarketwherebytheallocationofspectrumdetermines the technological environment, the competitive environment and consequently the qual- ity and the variety of the devices in the market. We then go on to simulate the welfare consequences of instituting the alternative management regimes. We show that the rel- ative importance of consumer’s preferences for variety versus increased interference due to increased variety is the determining factor for the welfare outcomes. As for informing the policy decisions we provide the insight that licensed allocations are likely to produce greaterwelfareinenvironmentswhereconsumerdonotcareforvarietyandinterferenceis rapidly increasing with entry, whereas, unlicensed allocations are likely to perform better in environments where consumers do have preferences for greater variety and interference isnotrapidlyincreasingwithentryofnewfirms. Thewelfaredominancebetweenthetwo regimesisdeterminedbyafinebalanceofthesetwofactors, thereforeallocationdecisions needtobeinformedbythepreferenceandtechnologicalcharacteristicsinthemarket. We conclude that favoring one alternative over the other without regard for the technology andpreferencestructureislikelytoaddtotheinefficienciesthatcurrentallocationssuffer from. 6 Chapter 2 Economics of Spectrum Allocation 2.1 Introduction Spectrum is arguably the most important foundation of contemporary communications technology. Although the physics and the technology of spectrum are understood to a great extent, much has been said in the economic and policy analysis of spectrum without a thorough understanding of what spectrum really is. In this chapter we provide a brief account of the early scientific and the technological progression of spectrum as well as a brief history of concurrent regulation and policy. We also present the current valuation and allocation problems in spectrum management which we further analyze in the following chapters. Spectrum in its contemporary understanding refers to a range of electromagnetic waves that include light waves, radio waves, X-rays and gamma rays, among others. All matter in the universe emits electromagnetic radiation and produces electromagnetic waves at various wavelengths or equivalently at various frequencies. Also, it is possible to artificially create these electromagnetic waves by modifying the electrical and magnetic 7 propertiesofmatter. Spectrumreferstothewholerangeofthesenaturallyoccurringand artificially created electromagnetic waves. In essence, the very small waves between 400 and700nanometersinsizethatweperceiveasvariouscolorsoflightarenodifferentthan the much larger radio waves we artificially create that radios, TVs, cell phones or Wi-Fi routers tune into. The caveat is that the human eye evolved to perceive only a small part of the electromagnetic spectrum that we call light but human intelligence evolved to develop the technology to produce devices that harness other parts of the spectrum. Therefore, despite the various metaphors used to describe it as space, as land, or as a scarce resource, the most appropriate metaphor to facilitate thinking about spectrum is nothingbutthecolorsoflight; thosewecanseeandthosewecannot. Indeed, everything about spectrum we know today, and the wireless technologies we enjoy are due to the curiosity of a few brilliant minds about how the spectrum of colors emerged from white light. Suppose you take a source of white light and shine it through a mist of water or a glass prism. What results is a spectrum of colors, a rainbow as we came to call it in its naturally occurring state: with red, orange, yellow, green, blue, indigo, and violet. Isaac Newton observed this phenomenon and started thinking about it as early as 1671. 1 He was the first to use the word spectrum to refer to the emerging range of colors and paved thewaytoitscontemporarymeaning. StartingwithNewton,thelongquesttostudylight and its properties had produced some of the most fascinating scientific discoveries, and culminated in the understanding that light and many other wave forms like radio waves, microwaves, infrared waves and x-rays are subsumed and unified by a larger range of 1 Glass, Lily, ”Color Theory: a Short History” available at: http://lilyglass.net/school/Art200/Handouts/colortheory1.pdf 8 electromagnetic waves the totality of which we today call spectrum. This scientific quest to understand light brought with it the understanding and the technology to harness other parts of spectrum that are not visible. Although we do not stop to think about the physics behind it, all the wireless technologies we enjoy today, from TVs to radios, from cellphonestoWi-Firouters, theyallutilizeelectromagneticwavesatvariousfrequencies, a different color of spectrum, but a color that the human eye can not perceive. Now, to understand spectrum policy in this context, suppose that licenses to use the spectrum of colors were up for auction except red and blue which are reserved for specific government uses. Suppose that those who have sufficient resources were allowed to bid for the licenses, with the understanding that in a properly designed auction, the market forces would efficiently allocate the licenses to the bidders who value it the most. Furthermore, winners of the auction would be free to implement appropriate technologies and business plans to sell devices and services to consumers that deliver the shades of yellow, orange, green, indigo, and violet that they own as long as they did not mix up the colors and interfere with each other. Since the licenses to use those colors belong to auction winners, anyone else using them would be illegally infringing on license owners’ rights or interfering with their operation. But perhaps a light shade of green would be free for everyone to use as long as it’s not too bright. With the objective of making the best use of these colors in mind, the problem is to decide which colors to auction, which colorstoreserveordedicatetospecificusersorusesandwhichcolorstoleaveforeveryone to enjoy freely. Spectrum policy today is concerned with exactly this type of problem, with the only caveat that the part of the spectrum in question is not light waves and the various frequencies we perceive as colors but rather the relatively lower frequencies that 9 can be harnessed with the current technology for communication purposes. We refer to the latter range of frequencies as the radio spectrum. Radiospectrumreferstotherangeofelectromagneticwaveswithfrequenciesbetween 3 Hz and 300 GHz or equivalently, between 1 millimeters and 100000 kilometers in size. These are the frequencies that almost all practical communication technologies utilize. Since the invention and widespread adoption of radio communications in early 1900s, Government has been managing the radio spectrum in the public interest with its power vested in the Federal Communications Commission (FCC) and the National Telecom- munications and Information Administration (NTIA). Most of the radio spectrum today is licensed to specific users and uses such as mobile communication licenses, broadcast licenses or military and federal uses. The primary means of allocating these licenses have mostly been through comparative hearings under a command and control approach until as recently as 1994 when auctions have started to become the increasingly preferred allo- cationmechanismofthelicensingregime. Ontheotherhand, someoftheradiospectrum is left open to use for everyone subject to technological rules and usage etiquettes under anunlicensedallocationregimelikeWi-FiandBluetoothtechnologiesorcordlessphones. Thetechnologyisinarapidprogresswithinnovationsinbothnewandmoreefficientuses of radio spectrum. However, this brings about new challenges and a need for improved understandingofspectrumregulationandpolicyaswell. Thefollowingsectionsareabout how we got to the status quo of spectrum technology and policy from a curiosity about light and the colors therein. And it is an analysis of where to go from here, with a focus on consumer welfare consequences of licensed and unlicensed allocation alternatives. 10 2.2 Brief History of Spectrum Technology and Regulation Spectrum is an apparition, an image, sudden appearance of an unusual sight in its orig- inal Latin meaning. 2 First recorded use of the word that paved the way to its contem- porary understanding was by Isaac Newton in 1671. When experimenting with optics, he came across the unusual colored bands of light that emerged from a glass prism and called it spectrum. This marked the beginning of the technological and scientific evolu- tion of electromagnetic spectrum. For centuries, other scientist tried to understand the physics behind light and the colors therein. About two centuries later in 1864, John Clerk Maxwell, postulated the existence of a continuum of electromagnetic waves with different wavelengths that subsume light along with other waves like radio waves, mi- crowaves, X-rays and so on. In 1888 Heinrich Hertz validated Maxwell’s theories and successfully demonstrated the existence of electromagnetic waves with experiments that produce, transmitanddetectelectromagneticwaves. 3 WhenHertz’sstudentsaskedwhat use might be made of this phenomenon; ”It’s of no use whatsoever,” he replied. ”This is just an experiment that proves Maestro Maxwell was right, we just have these mysterious electromagnetic waves that we cannot see with the naked eye. But they are there.” Indeed, whole universe is filled with electromagnetic waves. These waves are the revelations of a ubiquitous phenomenon called electromagnetic radiation. All matter in the universe at temperatures above absolute zero emits electromagnetic radiation. We do not see all of them, yet we understand, sort, and group the electromagnetic waves into 2 Schwartzman, Steven, The Words of Mathematics: An Etymological Dictionary of Mathematical Terms Used in English, Mathematical Association of America, Washington, D.C., 1994. p204 3 Hertz, H.R. ”Ueber die Ausbreitungsgeschwindigkeit der electrodynamischen Wirkungen”, Annalen der Physik, vol. 270, no. 7, p. 551-569, May, 1888. Also see: Hertz, Heinrich Rudolph. (1893). Electric waves: beingresearchesonthepropagationofelectricactionwithfinitevelocitythroughspace(translated by David Evans Jones). Ithaca, New York: Cornell University Library. 11 categories with similar characteristics according to their wavelength or equivalently their frequency in cycles per second or Hertz(Hz). 4 Today we refer to the range of all possible frequencies of electromagnetic waves as spectrum. In principle spectrum is very large if not infinite. Its size is bounded by our ability to measure electromagnetic waves with very large and very small wavelengths. An elec- tromagnetic wave can be as long as the length of the universe, or as short as the Planck length, smallest possible length we can perceive. If you divide the former by the latter youget5:410 61 possiblewavelengths, thatis54followedby60zeros. 5 So, wecansafely say spectrum is not scarce but on the contrary it is ubiquitous. Furthermore, spectrum is instantaneous and inexhaustible. It exist in its totality every nanosecond and every nanosecond that passes without making use of spectrum is an unrecoverable waste of its potential. The conceivable potential of spectrum is almost limitless; however, our ability to realize its potential is limited by two factors. The first one is technology, and we are getting better everyday in harnessing parts of spectrum that we already use as well as tapping into unexplored parts of spectrum. The second and more binding limitation is securing permission to actually use the spectrum from government that manages the spectrum in the public interest. As the technology of utilizing spectrum evolved, government control and the regulation of spectrum evolved as well. This intertwined history of spectrum technology and regulation takes us back to early 20th century. 4 Frequency of an electromagnetic wave is measured in cycles per second, or Hertz, and has an inverse relationshiptoit’swavelength. Thefrequencyisequaltothespeedoflight-atwhichallwavespropagate- divided by the wavelength, that is f = c . 5 The diameter of the observable universe is 93 billion light years, corresponding to 5:510 23 miles. The smallest distance about which anything can be known is Planck length, which is6:310 −35 inches. Although Electromagnetic waves can have a continuum of wavelengths in principle, assuming increments in Planck length, the number of possible wavelengths, or frequencies is 5:410 61 . 12 When Heinrich Hertz proved the existence of the electromagnetic spectrum in 1888, he thought that is was of no use whatsoever. However, Guglielmo Marconi a young man in his twenties happened to read Hertz’s article and was one of the many enthusiast who tried to harness the potential of spectrum. He took the idea of transmitting signals with electromagnetic waves and went on to develop the radiotelegraphy system. Despite prior demonstrations by Nicola Tesla , Marconi received the first patent for Radio in 1896. 6 In 1897, he established the first ever wireless communication over open sea. In 1903, President Roosevelt was able to send greetings wirelessly over the Atlantic to King Edward VII of England. And in 1912 Marconi radios were on board the Titanic during its transatlantic voyage. The unfortunate disaster of Titanic on April 1912 left its mark in history on many levels. But one of the most important impacts was that April the 15th 1912 was the day that radio communications came of age. The foundations of the regulation of radio communications starting with the Radio Act of 1912 were the outgrowths of the Titanic disaster because the underlying problems were incorrectly attributed to interference re- sulting from spectrum scarcity and were used to justify centralized control and allocation ofspectrum. Butacloserlookattheeventsofthedayrevealsthattherealproblemswere the dominance of commercial interests over safety concerns, and the lack of cooperation and protocols between competing radio companies and their operators. 6 Marconi was awarded the British Patent 12039, Improvements in Transmitting Electrical Impulses and Signals and in Apparatus There-for. (For a reproduction see: http://www.radiomarconi.com/marconi/popov/pat763772.html) However, in 1943, shortly after his death, Nicola Tesla was awarded the patent for Radio by the U.S. Supreme Court. With a decision based on the fact that prior art existed before Marconi’s patent, Tesla’s patent (number 645576) was reinstated as holding priority in the ”invention” of modern radio. 13 Themainpurposeoftheinstallationofradiotransmittersintransatlanticshipsatthe time was profiting from commercial transmission and receipt of messages including stock exchange quotations, business, and private communications, and news services. Titanic was equipped with a Marconi wireless system primarily for handling of passengers’ mes- sage traffic for revenue and the responsibility of the wireless operator was transmitting and receiving messages known as MarconiGrams which cost 12 shillings and 6 pence for 10 words, and 9 pence for any additional word. Using the system for signaling distress was secondary to commercial purposes; hence, commercial message traffic of passengers was prioritized by the Marconi operators over information sharing between vessels. Al- though Titanic received a total of 21 iceberg warnings from nearby ships earlier that day none of them ever made it to the Captain’s desk because Marconi operators were busy with a backlog of passenger messages. Furthermore, some of the ships were fitted with Telefunken radios, Marconi’s main competitor. It was forbidden for Marconi operators to “talk” to Telefunken operators, which was part of the reason that some iceberg warnings were dismissed. However, the very last iceberg warning Titanic received was from SS Californian, a Marconi ship within sight of the Titanic a mere 15 miles away. At 11pm, 45 minutes before Titanic struck an iceberg the Marconi operator of SS Californian Cyril Evans relayed the message: “Say, old man, we are stopped and surrounded by ice.” to Titanic. Even though they both were Marconi operators the response he received from Jack Phillips, Titanic’s operator was: “Shut up, shut up, I am busy; I am working Cape Race.” since he was trying to send passenger messages to the land station at Cape Race, 14 Newfoundland. 7 While Phillips was working Cape Race, Titanic struck an iceberg and started to sink. Followingtheimpact,Phillipsstartedsendingdistresssignalandwasheardbyover20 othervesselsandthesignalswerereceivedeveninNewYork. Accordingtosomeaccounts David Sarnoff, a junior Marconi operator at the time working at the top of Wanamaker Hardware Building was one of the people to intercept the distress calls. For 72 hours he did not take off his earphone and conveyed to the rest of the world the disaster that wasunfolding. ThisaccidentalinterceptionofTitanic’smessagesbroughtSarnoffintothe publiceyewholaterwentontoestablishNationalBroadcastingCompany(NBC)andled Radio Corporation of America (RCA). 8 Unfortunately, Cyril Evans, the radio operator of the closest ship Californian had retired for the day and the crew of the Californian did not receive any of the distress calls, going through the night without knowing the disaster taking place within their sight. Carpathia was about 50 miles away and rushed to the rescue, but it was too late when she arrived at the scene at 4am, only 706 of the 2223 people could survive. What this story amply demonstrates is that the disaster of Titanic could easily be avoided, and it was not due to scarcity of spectrum, or lack of licenses protecting operators from interference, but rather due to lack of cooperation and established communication protocols. In light of the Titanic Disaster, Radio Act of 1912 was the first serious attempt to regulate radio communications. All ships were required to have 24 hour radio watch with time and frequency division rules, as well as priority protocols for distress transmissions. 7 United States Senate Inquiry into the Titanic Disaster Day 8, Testimony of Cyril F. Evans. see: http://www.titanicinquiry.org/USInq/AmInq08EvansCF01.php 8 ForanaccountofSarnoff’sinfluenceontheindustrysee: Bilby, Kenneth., 1986. TheGeneral: David Sarnoff and the Rise of the Communications Industry. New York: Harper and Row 15 Commercial radio communications were confined to wavelengths between 200 and 600 meters with license requirements. Wavelengths between 600 meters and 1600 meters were reserved for government use. The most important part was the wavelengths below 200 meter or frequencies higher than 1500 KHz. These waves were thought to be too small and useless thus were left open for amateur radio operators, however, most of the innovationsofthedayinradiocommunicationshappenedinthisopenrangeofspectrum. As the technology progressed and smaller wavelengths or equivalently higher frequencies became viable, regulation also extended to higher frequencies. Today, electromagnetic waves as small as one millimeter (300GHz) are regulated by the FCC. Anotherexampleoftheclashbetweentechnologyandregulationfollowedshortlyafter the 1912 act. The commercial radio stations that were confined to wavelengths between 200 meters and 600 meters were using amplitude modulation (AM Radio). Erwin H. Armstrong chose to experiment with shorter waves in the open spectrum and invented Frequency Modulation (FM Radio) that provided much better sound quality. 9 His work on demonstrating how bandwidth can be exchanged for noise immunity also laid out the foundations of Information Theory which was the next breakthrough in the technology of communications. However, heavy lobbying on the part of RCA led by David Sarnoff to prevent FM technology from becoming dominant was successful and the FCC moved the FMspectrumtowavelengthsbetween2.7and3.5meterscorrespondingtothe88-108MHz 9 AtanFCCdemonstration,ArmstrongplayedajazzrecordoverconventionalAMradio,thenswitched to an FM broadcast. ”If the audience of 50 engineers had shut their eyes they would have believed the jazz band was in the same room. There were no extraneous sounds,” noted one reporter. He added that several engineers described the invention ”as one of the most important radio developments since the first earphone crystal sets were introduced.” United Press report, ”Radio Set-up Eliminates All Noise,” Ogden Standard-Examiner, June 18, 1936, p1. 16 frequencies which we still use today. 10 This change rendered Armstrong’s technology useless overnight since his radios were built to work with waves between 6 and 7 meters (42-50 MHz). Armstrong committed suicide in 1954 after the events brought him to financial ruin and psychological breakdown. Due to the blocking efforts of RCA, it took another 20 years for the superior technology he invented to prevail, however, it was RCA againthatclaimedtheinventionofFMradioandsettledwithArmstrong’sestateafterhis death for the patentsto the FM technologyand made record profits from his innovations, attheexpenseofdeterringconsumersfor20yearsfromenjoyingthissuperiortechnology. The nominal reason for delaying this technology was preventing ionospheric disturbance, a misguided argument similar to the scarcity one. Despite the fact that the plenty of open spectrum in the shorter wavelengths was deemed useless, there were brilliant minds that figured out how to make good use of it. However, once again it was not in the best commercial interest of the dominant radio company at the time to adopt the technology. Hence, technology was stifled and consumers suffered. The history of telecommunications is full of many similar examples. Although the technology for cellular communications was developed as early as 1947, long before any spectrum was allocated for its use, incumbent broadcasters successfully delayed its widespread adoption until early 80s by lobbying the regulator to not allocate spectrum to cellular systems. A very recent example of the clash between legacy and new spectrum technologies, as well as the efforts of incumbents to deter or delay technological progress that does not fit in their business interests revealed itself in the debates surrounding the 10 For more on lobbying efforts of dominant incumbents see: Lennett, Benjamin, 2008, The Lobby that Cried Wolf Broadcaster Campaign Against Using TV ’White Space’ Follows a Familiar Script, New America Foundation Issue Brief, 23 17 white spaces. White spaces are the unassigned vacant frequencies placed between the television broadcast bands to prevent interference. The gradual completion of the transi- tion from analog to digital broadcast which left increasingly more unused spectrum raises thequestionofhowandtowhattypeofusesshouldthewhitespacesbeallocated. Again, incumbentsexertedagreatdealofefforttoclaimtheunusedfrequenciesoratleastretain the status quo despite the demonstrations that showed that the white spaces could be utilizedinanunlicensedfashionwithoutdisruptingtheoperationoftheincumbents. The technologies that enable such unlicensed utilization have been around for a decade but the regulation has again been slow to catch up. It was only in early November of 2008 that a decision was made and white spaces were allocated to unlicensed use. The debate surrounding white spaces has been one of the motivations for our work, among others. The arguments we put forth in the following chapters are centered around white spaces as it is the most recent example of the spectrum allocation problem and are aimed at correctly identifying the consumer welfare consequences of licensed and unlicensed allo- cation alternatives. We first provide an account of the status quo in the next section and then proceed by describing comprehensive ways to analyze the welfare consequences of the allocation alternatives. 2.3 Status Quo and Looking Ahead FCC through various mechanisms ascertains which users can obtain access to spectrum and what type of services can be provided using the spectrum subject to certain rules. There have been three main approaches to the management and allocation of spectrum. 18 The first; command and control approach, whereby the FCC assigned spectrum to spe- cific users and uses subject to pre-engineered technological rules and on an as needed basis by comparative hearings, had been the dominant approach until recently. However, command and control approach is unanimously regarded as highly inefficient and is al- mostcompletelyabandoned. Mostoftherecentdebateonspectrumpolicyfocusesonthe remaining two approaches which are the licensing approach and the unlicensed allocation approach. The licensing alternative entails the creation and auctioning off of exclusive licenses to use the spectrum with protection from interference. Unlicensed allocation al- ternative, on the other hand, entails specification of usage etiquettes and technological specifications but leaves access to spectrum open to any prospective user willing to ac- cept and manage interference. Commercial Mobile Radio Services (CMRS) bands that are used by the mobile communications industry are the major examples of the licensed allocation. Industrial, ScientificandMedical(ISM)frequencybandsarethemostobvious products of the unlicensed allocation alternative, particularly the band around 2.45GHz which houses Wi-Fi and Bluetooth devices being the most densely used one for functions like wireless networking among others. Proponents of the licensing approach argue that the efficient way to allocate the spec- trum is to assign exclusive licenses to use the spectrum with minimal restrictions on possible uses. It is argued that protection from interference granted by the exclusivity of the licenses, and the flexibility on the end use is the key to efficiency and incentives to invest. Previous allocations that most resemble the property rights approach like the Commercial Mobile Radio Services and Advanced Wireless Services frequencies (used mostly by cellular systems) are referenced as successes in terms of consumer surplus 19 and government revenue created by the licensed allocation. The competing view cred- its unlicensed allocation regime for the success of unlicensed applications like Wi-Fi and Bluetooth and argues that access to spectrum should be granted to any user provided that the user meets industry standards and usage etiquettes. Opponents of unlicensed allocationsalsoarguethatthelackofprotectionfrominterferencewouldleadtoatragedy of commons scenario where a large number of interfering devices would degrade device or service quality. This argument is countered by the technological developments that embedintelligenceintothedevicesandletthemmanageinterferencebyabidingprotocols like listen before talk, or taking preventive action like frequency hopping. Furthermore, the diversity in the device market and the distributed bottom up innovation resulting from open access in an unlicensed allocation are argued to create considerable welfare. Evidently, the choice of allocation regime shapes the nature of communications by de- termining the variety and technological characteristics of the communication devices and the structure of the market. In light of these arguments, the need for economic analysis that can enlighten the trade-offs faced between the two alternatives and highlight the consumer welfare consequences is eminent. 2.4 Simple Economics Spectrum Allocation: Valuing Alternatives The right choice of allocation requires the evaluation of the marginal welfare resulting from allocating an additional piece of spectrum under the two regimes. The simple economics behind efficient allocation dictates that the spectrum be allocated such that 20 the marginal welfare of choosing licensing or unlicensed to assign an additional piece of spectrum is equal as illustrated in Figure 2.1. 0 20 40 60 80 100 % allocated to Licensed Marginal Social Value MSV of Unlicensed Allocation Case1 MSV UL Allocation Case2 MSV of Licensed Allocation X Figure 2.1: Basic economics of spectrum allocation An interior solution to such allocation problems require a certain percentage (X per- cent in Figure 2.1) to be allocated to licensed and remainder (100-X percent) to unli- censed. However there is the possibility that the marginal social value is lower under one of alternatives for all allocation possibilities, as illustrated by the dashed MSV curve in the figure. In this degenerate case, optimum is achieved by a corner solution where all marginal allocation favors the allocation with the higher marginal social value. The empirical question then is whether the marginal social value of a licensed allocation is always above that of an unlicensed allocation. It is hard to predict the market structures andthetypesofdevicesandservicesthatwillemergeusingthenewlyallocatedspectrum. 21 But one can get an idea of the marginal social values by looking at past auction results or market outcomes. If the welfare gain that could result from one alternative is inferior to the other for all possible allocations, then efficiency requires that all of the available spectrum should be allocated under the alternative with higher welfare gain. However, if there is commensurate welfare to be gained in both alternatives, then close attention should be paid to marginal allocations and a balanced allocation is most likely to be optimal. One simple approach to valuing the alternative allocations, specifically the marginal socialvalueoflicensedallocations,istolookattheeffectsofspectrumlicensinginmarkets where spectrum is a key input, and looking at the price that spectrum attained in recent auctions. Consider the market for a communications service like CMRS cellular services. In a competitive market without capacity constraints, the market price of the service is determinedbytheconditionthatsupplyequalsdemanddenotedbythepointEinFigure 2.2. At the equilibrium point E, service price is exactly equal to the marginal cost and therearenosupernormalprofitstobemade. However,inthecaseofalicensedallocation, the amount of spectrum to be licensed and auctioned off effectively becomes the capacity constraint in this otherwise competitive market. The quantity of the service to be offered fallsbelowtheequilibriumquantityandthiscreatesawedgebetweenthemarginalcostof the service and the market price. This difference is the scarcity rent created by licensing. Furthermore, if the spectrum auction is competitive, this amount is exactly equal to the amount that a firm is willing to bid and pay for the license. This amount is also equal to the difference between marginal value of the service to the consumer and the marginal 22 Quantity Service Price Supply (Marginal Cost) Demand Capacity Contraint X MHz Marginal Cost Scarcity Rent = License Price E Figure 2.2: Basic economics of spectrum license prices cost of providing the service which is equal to the marginal social value. Therefore, the marginal social value of an additional MHz of licensed spectrum is exactly equal to the lowest price paid for a block of spectrum on a per MHz basis. The results of Auction 73, the last auction of spectrum in the 700 MHz band reveals that the price paid for an A block license which is a pair of 6MHz bands in the lower 700 MHzwas$1.16perMHzperpopulation. Therefore,themarginalsocialvalueofallocating an additional white space, i.e. a 6 MHz block of spectrum unused buffer spectrum band is approximately $7 per person covered by the license under licensing. If the marginal socialvalueoftheunlicensedallocationofasimilar6MHzblockofspectrumismorethan $7, then it can not be efficient to allocate the entire spectrum under licensing. To put the comparison in more context, imagine if it would be worth paying a one time $7 fee for 23 the possibilityof allocating a 6MHz of spectrum to unlicensed use, perhaps extending the range of your Wi-Fi connection or Bluetooth headset to your backyard. It is important to stress that this is a one time capitalized fee, not a monthly or annual fee and most people would say yes to the above proposition. However, aggregating the valuations of those who would accept such a proposition and taking collective action to bid for the spectrum to be dedicated to unlicensed usage is not feasible. Another approach to valuing alternative allocation regimes is to look at past allo- cations that most resemble typical licensed and unlicensed allocations, estimating the consumer welfare resulting from these allocations and extrapolating the observations to prospective allocations. The valuation of consumer welfare derived from the services that use licensed spectrum has been a fairly standard exercise with the availability of price and consumption histories and the quantity of licensed spectrum allocated to a particu- lar service. For example, Hausman (1997) has estimated that the introduction of cellular communicationscreated$30to$50billionconsumersurplus. Atthesubscriptionlevelsof 1997 of 41 million, and allocated spectrum of 50 MHz, Hausman’s estimates correspond to between $14 and $24 per subscriber per MHz. On the other hand, Hazlett (2005) has estimated that the consumer surplus that accrues to about 195 million cellular service subscribers to be around $80billion in 2003. Commercial Mobile Radio Services bands, which are exclusive assigned flexible licenses used to offer these cellular services include a total of 189 MHz of spectrum, so the consumer surplus appears to be down to $2.17 per subscriber per MHz in 2003. A similar study commissioned in 2006 by Ofcom, the regulatory authority in United Kingdom found that the consumer surplus from mobile 24 telephony to be$25.1 billion. At the subscription levels of 70 million and allocated spec- trum of 432 MHz, the consumer surplus per MHz per person turns out to be$0.7. These estimatesaresummarizedinTable2.1andprovideanideaaboutthesizeoftheconsumer surplus resulting from licensed allocations. Allocated Consumer Consumer Surplus Study Service Spectrum Subscribers Surplus per MHz per person Hausman 1997 Cellular 50 MHz 41 million $30-$50 billion $14-$24 Hazlett 2005 CMRS 189 MHz 195 million $80 billion $2.17 Ofcom 2006 (UK) Mobile Telephony 432 MHz 70 million $21.7 billion $0.7 Table 2.1: Consumer Surplus from Licensed Allocations A similar exercise to obtain a measure of consumer surplus derived from the use of unlicensed spectrum has been overlooked. In particular estimating the welfare from un- licensed allocations is a difficult exercise because unlike licensed spectrum that is used to provide few and relatively homogeneous services in a given band like cellular com- munications, there are numerous and highly heterogeneous devices and services that use unlicensed spectrum like many heterogeneous Wi-Fi and Bluetooth devices that us the ISM band around 2.45 GHz as well as cordless phones, garage door openers, baby moni- torsandmanyothers. Moreimportantlypriceandconsumptionshistoriesofsuchdevices arerarelyavailable. Furthermore,beyondtheinitialfixedcost,thepriceonmarginalcon- sumption is by definition zero for most services that utilize unlicensed spectrum. Cellular services for example are priced by the minute above and beyond subscription fees and price-consumption histories are recorded for billing purposes making them readily avail- able. For wireless networking devices on the other hand, the only data point is the price of the device itself and the price-consumption levels are rarely available. However, the 25 available data does provide some insight about the size of the consumer surplus derived from unlicensed uses of spectrum. Carter et. al. (2003) in FCC’s Office of Strategic Planning and Policy Analysis working paper 39 provide an overview of devices that utilize unlicensed spectrum. Two of the main categories of unlicensed devices identified and discussed by OSP are Cordless Phones, Wireless Local Area Network (LAN) Devices. We will follow the same order in providing insight into the consumer welfare that resulted from these uses of unlicensed spectrum. According to Carter et. al. (2003) Cordless phones had a penetration rate of 81% with 130 million installed bases corresponding to 1.5 units per household. The market data on cordless phones from 1997 to 2004 show the steady increase in sales from about 28.2 million units in 1997 to 51.6 million units in 2004 with a corresponding decline in prices from $74.4 to $47.4 during the same period. The data is presented in Figure 2.3. It is important to note that this data does not correspond to the actual demand but rather a historical presentation of market conditions. However, as presented in the appendixintablesA.1andA.2,aloglineardemand,controllingfortheaverageincomein the corresponding years can easily be estimated that reveals the assumed constant price elasticity of demand to be -1.2. With a simple triangulation at each price-quantity pair, theconsumersurplusisapproximatelyequaltothetotalexpendituredividedbytwicethe absolute value of the price elasticity (CS = pq 2" ). The average consumer surplus between 1997 and 2004 turns out to be $960 million per year with a corresponding $.21 per MHz per person as presented in table 2.2. It is important to note that the ISM frequencies usedbycordlessphonesaresharedwithmanyotherdevices, therefore, toobtainthetotal 26 40 45 50 55 60 65 70 75 80 25 30 35 40 45 50 55 Units sold (millions) Price 1997 1998 1999 2000 2001 2002 2003 2004 Figure 2.3: Cordless phone market data time series consumersurplusperMHzperpersonintheunlicensedbands,theconsumersurplusfrom various uses need to be aggregated. 1997 1998 1999 2000 2001 2002 2003 2004 Average Price $74.4 $69.6 $54.9 $55 $52.9 $51 $49.1 47.4 $56.8 Units (millions) 28.2 31.3 40.1 43.3 45.3 47.4 49.5 51.6 42 Consumer Surplus (billions)* $0.87 $0.91 $0.92 $0.99 $1 $1.01 $1.01 $1.02 $0.96 CS per MHz per person** $0.28 $0.26 $0.21 $0.21 $0.20 $0.19 $0.18 $0.18 $0.21 * Assuming a price elasticity of -1.2 ** Assuming 108 MHz of spectrum at 900MHz and 2400MHz bands Table 2.2: Cordless phone market data and consumer surplus calculations EstimatingtheconsumersurplusfromWirelessLANdevicesismoreinvolvedasprice and consumption histories for Wi-Fi devices are harder to come by. Most of the time, unlicensed wireless capabilities are embedded in complex devices like laptop computers 27 and it is difficult to isolate the value that accrues to the consumers because of the pres- ence of wireless LAN capabilities. However, the data on Wi-Fi chipset shipments provide insight into the widespread use of this technology embedded in other more complex de- vices. Table 2.3 shows the widespread adoption of wireless LAN chipset embedded in other devices. With an average price of about $7, size of the chipset market was around $2.7 billion in 2008. 2007 2008 growth Cellular Wi-Fi Phones 37 56* 51.4% Stationary Consumer Electronic Devices, (gaming consoles, digital televisions, set-top boxes, printers) 32 48 50% Portable Consumer Electronic Devices, (handheld games, cameras, portable music players) 53 71 34% Notebook PC’s, Mini Notebooks, Ultra Mobile Devices, Mobile Internet Devices 117 144 23.1% Total** 307 387 26.1% *Projected 141million in 2009, 520million in 2014 **includes other minor categories of devices Source: Reconstructed data from Wi-Fi Alliance and In-Stat Table 2.3: Wi-Fi Chipset Shipments What is more important than the size of the Wi-Fi chipset market is the services that are created by the devices with embedded Wi-Fi chips. One significant example of such services is Wi-Fi home networking. According to Forrester Research’s North American TechnographicsSurveyofanationallyrepresentativesampleof53000householdsin2008, one in every six households has adopted Wi-Fi home networking. Figure 2.4 depicts the composition of home networks as reported by the survey respondents. The growth in home networking is dominated by the growth in wireless home networks. It is apparent that the conjoint growth of wireless and wired home networking between 2005 and 2006 is tilting towards wireless only households in 2007 and 2008 at the expense of wired home 28 networking. In 2008, 22 percent of all household report to having some type of home network and out of the 22 percent, 17.8 percent report to having a wireless only network. The remaining 4.2 percent report to having a wireless/wired mix home network with essentially no wired only networks apparently demonstrating the switch from wired home networks to wireless home networks. Figure 4. Home Network Composition (% of Households) 0% 5% 10% 15% 20% 25% 2004 2005 2006 2007 2008 2009 Year All Home Networks Wireless Networks Wireless/Wired Mix Networks Wired Only Networks Figure 2.4: Home network composition (% of households) Sinceitswidespreadintroductioninearly2000,wirelessLANapplicationshavegained tractionandinlessthanadecadealmostallhomenetworkingisdeliveredthroughwireless networks. Together with the chipset shipment data, this is a testament to the widespread adoption of Wi-Fi as the dominant means of in-home or short range connectivity in all types of consumer electronics. 29 Thewelfarederivedfromtheuseofthisunlicensedspectrumapplicationisnotreadily computablewithtraditionaldemandestimationmethods,becauseofthelackofgoodprice and consumption time series. However, Goolsbee and Klenow (2006) has introduced an innovative way of computing the consumer welfare from the use of time intensive goods, like the Internet, by using the opportunity cost of time valued at the wage as a price proxy. SinceWi-FinetworkingismainlyusedtointroducealocalmobilitytotheInternet consumptionathome,equippedwiththismethod,onecancomputetheconsumersurplus fromtheconsumptionofInternetthataccruestowirelessandwiredhomenetworkowners. The incremental consumer surplus that wireless network owner enjoy is attributable to the availability of Wi-Fi home networking application of the unlicensed spectrum. Using the North American Technographics data of 2005, in the next chapter, we estimate that the lowest incremental consumer surplus that wireless network owners enjoy above and beyondthatofwirednetworkownersis$824perhouseholdperyear. Atapenetrationrate of 18 percent of 117 million households, the total consumer surplus that is attributable to unlicensed wireless LAN use of Wi-Fi in home networking is about $17.3 billion per year. At an allocated spectrum of 82 MHz, the consumer surplus per household per MHz turns out to be $10. These relatively simple consumer surplus measures from the use of unlicensed spec- trum enable us to revisit the basic economics of spectrum allocations and argue that the marginal social value of unlicensed allocations is likely to be commensurate with that of the licensed allocations. Even with only the two relatively widespread applications of unlicensed spectrum, we can argue that the optimal allocation is likely to be an interior solution with spectrum allocated to both alternatives. From the estimate of Hausman 30 (1997) of $24 per MHz per person in 1997 corresponding to 50 MHz allocated to licensed cellular communications, the additional 139 MHz allocated to CMRS by 2003 brings the consumer surplus per MHz per person down to $2.17 according to Hazlett (2005). On the other hand, our preliminary calculations of consumer surplus accruing to users of unlicensed spectrum from Wi-Fi home networking and cordless phones turns out to be $10.3. Out of a total X MHz spectrum available, plugging these values to the marginal social value figure suggest that MSV curves to intersect resulting in an interior solution as illustrated in figure 2.5. MHz allocated to Licensed Marginal Social Value MSV of Unlicensed Allocation MSV of Licensed Allocation 50 MHz $24 189 MHz (X-108) MHz $2.17 $10.3 Hausman (1997) Hazlett (2005) Bayrak (2009) Figure 2.5: Marginal social value of alternative allocations revisited As these conservative back of the envelope calculations illustrate, it is highly un- likely that the marginal social value of prospective licensed allocations are always higher 31 than unlicensed allocations. Therefore, unilaterally favoring licensed allocations over un- licensed allocations with poorly justified interference or investment incentive concerns is likely to result in a sub-optimal allocation. 2.5 Conclusion A thorough understanding of the physics and the technology of spectrum is the first step toaccurateeconomicandpolicyanalysis. Inthischapterwetryandrecouplethescientific andtechnologicaldefinitionofspectrumwitheconomicandpolicyanalysis. Thecommon metaphors for spectrum such as a resource or real estate imply inherent scarcities which tendtofavorlicensedallocations. However, wearguethatspectrumisnothingmorethan a placeholder word for the range of all possible frequencies of electromagnetic radiation. Within this range, most familiar phenomenon to be taken as an example which also turns out to be a relatively more accurate metaphor is light and the colors therein. This understanding helps to highlight the fact that spectrum is not scarce but on the contrary that it is ubiquitous. It also highlights that interference is not an inherent property of electromagnetic waves, as they pass through each other without loosing form, but rather that interference is a result of the inability of transmitters and receivers to accurately filter interpret the information encoded in the electromagnetic waves. Advances in spectrum technology improve the ability of communication devices to filter and interpret the transmissions intended for them, and consequently, strict geo- graphical and frequency boundaries between communication channels become less and less required for efficient spectrum utilization. This brings about the need to reconsider 32 legacy spectrum allocation approaches. A careful look at the consumer welfare from al- ternative allocations also suggest that an interior solution to the allocation problem is likely to be optimal since unlicensed allocations seem to create considerable welfare com- mensurate with that of licensed allocations. Favoring one allocation regime unilaterally over the other is likely to be sub optimal. With this understanding, in the next chapter we go on to take a deeper look at the incremental welfare that results from one of the most popular implementations of unlicensed spectrum, wireless home networking. 33 Chapter 3 Valuing Time Intensive Goods: An Application to Wireless and Wired Internet 3.1 Introduction Economic theory has a range of techniques for estimating the value derived from an economicactivity. Inthecaseofvaluingapublicresource,calculatingthecostofrecycling or the net present value of exploiting the resource are popular methods. In the case of theconsumptionofexistingortheintroductionofnewgoodsandservices, estimatingthe price elasticities from expenditure data and inferring the consumer surplus is typical. However,notallpublicresourcesorgoodsandservicespermittheapplicationofthese techniques. Some resources go into the production of numerous goods and services the (present)valueofwhicharehardtoquantifyandaggregate. Forsomegoodsandservices, on the other hand, expenditure data is either not available or does not represent the true cost of consumption, especially when consumption is highly time intensive. The conven- tional method of estimating the elasticity from expenditure data could be misleading in the case of highly time intensive goods, since the market price constitutes a minuscule 34 part of the total cost of consumption compared to the opportunity cost of time spent using the good. An alternative to the traditional method in the case of time intensive goods is to estimate the elasticity and welfare from the variation in the opportunity cost of time. Internetisanimportantexampleofatimeintensivegoodforwhichthemarketexpen- diture is minuscule compared to the opportunity cost of the time spend in consumption. According to National Income and Product Accounts, out of the 8.7 trillion dollars that went to personal consumption expenditures in 2005, only 18 billion dollars went to In- ternet service providers, which amounts to (scaling up for the 37% non subscribers) 0.33 percent of the total expenditure. Whereas, consumers report to be spending around 10 percent of their non-sleep time on the Internet. The time share of the Internet is about 30 times higher than the expenditure share which illustrates the highly time intensive nature of Internet consumption. Observing this peculiarity, Goolsbee and Klenow (2006) provide a simple utility model that includes consumption in the form of market expendi- tures and time. The model allows a more concrete estimation of the welfare gains from the Internet that takes into account the time intensive nature of consumption. In this paper as we look at the welfare derived from the use of wireless and wired Internet, we highlight the importance of this new approach to estimating welfare from the consumption of time intensive goods as introduced by Goolsbee and Klenow (2006). Taking data on wages as the opportunity cost of time, and Internet time use data, we estimate the elasticity in a more accurate way that takes into account the time intensive characteristic of the Internet. Furthermore, we restrict the attention to consumers with homenetworksandhighlightthedifferentdemandcharacteristicsandwelfareattainments 35 of consumers who connect to the Internet through wireless networks from those who connect through other(wired) types of networks. The incremental consumer surplus from usingwirelessnetworksrelatestoanotherimportantproblem: thevaluationofunlicensed radio spectrum. The problem of valuing radio spectrum has difficulties that lie in the intersection of thetwotypesofdifficultiesmentionedabove. Asapublicresource,itisusedbynumerous devices for providing numerous services. Radio and TV broadcasting, mobile communi- cations and data networking are few notable examples besides garage door openers, baby monitors, microwave ovens, police radars, the list goes on. There are an estimated 15 million TVs using over the air broadcast, 800 million cell phones, over 250 million Wi-Fi devices, and about 1 billion Bluetooth devices in the market and the value derived from theconsumptionofthesedeviceswhichinturnispartlyduetotheabilityofthesedevices to use radio spectrum, is quite challenging. On the other hand, as an intermediate good, estimating the demand for and the price elasticity of radio spectrum itself is also difficult because of the high heterogeneity with respect to the characteristics of different frequen- ciesandthelackofanestablishedmarketwherespectrumistradedinlargevolumesfrom which variation in expenditure data can be observed to estimate elasticities. Attachingavaluetoradiospectrumorthewelfarederivedfromtheuseofithasgained importance recently because of the increased attention to white spaces i.e. the unused frequencies in the digital television broadcast bands. The gradual completion of the transition from analog to digital broadcast which will leave even more unused spectrum raises the question of how and to what type of uses should the white spaces be allocated. First of the three approaches to managing spectrum, command and control, whereby the 36 allocation of spectrum is done by a regulatory authority based on comparative hearings and pre-engineered technical rules, is unanimously regarded as inefficient and almost completely abandoned. Most of the debate on white spaces is centered around the other two regimes that came to be called Licensing and Commons. Licensing is the approach to spectrum management whereby the spectrum is divided into spectral frequency bands and the exclusive rights to operate in a band at a certain geographicalareaislicensedtoanentitywiththelicensesbeingsoldatauctions. Inmost of the recent allocations, licenses are flexible meaning that the licensee has flexibility on the choice of end use offered and the technology employed. Commercial Mobile Radio Services (CMRS) bands around 0.8, 0.9 and 1.9 GHz, Advanced Wireless Service (AWS) bands around 1.7 and 2.1 GHz are examples of exclusively licensed bands for flexible use by the licensee. Most notable uses of these bands consist of cellular voice and data communicationsservices. Commons,ontheotherhanddoesnotlimittherighttooperate to a number of licensees. Unlimited number of users share spectrum subject to some maximum power restrictions to counter interference. Industrial, Scientific and Medical (ISM) bands around 2.4 and 5.8 GHz are examples of Commons spectrum with the most notable uses being the Wi-Fi and Bluetooth wireless networking devices. There have been studies that try to estimate the value of licensed spectrum bands and promote Licensing regime for further allocations. Hazlett (2005) has estimated the consumer surplus from the Commercial Mobile Radio Services (CMRS) bands in 2003 to be $81 billion or around $500 per subscriber. Hausman (1997) on the other hand, has estimated that the introduction of cellular communications resulted in $30 to $50 billion consumer surplus per year. However, to our knowledge, there has not been a 37 study that estimates the welfare gains from unlicensed spectrum bands. In this paper, we take on the task of estimating part of the welfare gains associated with the use of unlicensed spectrum, keeping in mind the difficulties that prevent the application of the typical methods. Realizing that it is difficult to calculate and aggregate the welfare gains from all the goods and services that use unlicensed spectrum, we take wireless networking for it is one of the most popular uses of unlicensed spectrum, as a first attempt to calculate a lower bound on the welfare gains from goods and services that use unlicensed spectrum. We look at the consumer surplus derived from Internet consumption and we restrict our attentiontothoseconsumerswhoreporttohavesometypeofhomenetwork. Weestimate the incremental consumer surplus of the consumers who connect to the Internet through wireless networks, over those who connect via wired Ethernet, powerline or HomePNA networks. We find as a very conservative estimate, that the consumers who connect to the Internet through wireless networks obtain $824 more consumer surplus per year than those consumers who use wired networks. The paper is organized as follows. In the next section we present the model as in- troduced by Goolsbee and Klenow (2006). In section three we talk about the data and the estimation of elasticities. Section four presents welfare calculations. Section five concludes. 38 3.2 Model Consumers maximize the following utility function which incorporates the time intensive nature of the Internet as a consumption good. U =(C i L 1− i ) 1 +(1)(C o L 1− o ) 1 where C i is the consumption of Internet services and L i is the time spent using the Internet. All other goods and services consumed form the composite good C o with the time spenton the composite good beingL o : representsthe utility weightof the Internet bundle compared to that of the composite bundle. Finally (1) and (1) represent the time intensities of the Internet and the composite bundles respectively. The following is the budget that constrains consumers in their utility maximization. P i C i +F +P o C o =W(1L i L o ) whereW isthewage,P i andP o arethepricesoftheInternetserviceandthecomposite bundle respectively. F is the fixed cost of access to the Internet including the network setup. P i can be interpreted as the price on marginal consumption which is zero in practice since Internet access is usually priced as a flat monthly fee. The combined Cobb-Douglas bundles are denoted asY i =C i L 1− i andY o =C o L 1− o : Letting the price on the bundles i and o to be the weighted average of the market price and the price of time (i.e. the wage), we have 39 i = ( P i ) ( W 1 ) 1− and o = ( P o ) ( W 1 ) 1− Then, the optimal choices for the bundles become Y i = W F i (1+∆) and Y o = W F o (1+1=∆) where ∆= ( i o ) −1 ( 1 ) Breaking down the bundles into their consumption and time counterparts gives the optimal choices as C i = i Y i P i C o = o Y o P o L i = (1) i Y i W L o = (1) o Y o W Using the optimal choices on the Internet bundle and the time spent on Internet we can get the following expression for ∆: ∆= (1)(1 F W )L i L i 40 ObservingthatthecostoftheaccesstotheInternetisasmallflatfee(F=W 0)and there is no marginal use pricing ( = 0), the above expression becomes approximately equal to the time spent on activities other than Internet relative to the time spent on Internet. ∆ 1L i L i Using the prices of the bundles and rearranging, we get another expression for ∆ in terms of the wage ∆=AW (−)(−1) ( 1 ) Where A = [ (P i =) (1−) (Po=) (1−) ] −1 : Equating the two expressions and taking the natural logarithm gives. ln ( 1L i L i ) ln(A)+()(1)ln(W)+ln ( 1 ) The left hand side of the equation is the log of time spent on non-Internet activities relative to time spent on the Internet which can be found in the data. ln(A) is a constant across consumers. ln(W) is the log of the wage and can be found in the data as well. The difference between the time intensities of the two bundles () can also be measured from the data. So, from the estimation of this equation, assuming that the error term arises from the individual variation in the utility weight parameter ; the coefficient on 41 the wage can be translated into an estimate of the elasticity of substitution between the two bundles which in turn can be used to calculate the consumer surplus. 3.3 Data and Estimation We use the North American Consumer Technographics data from Forrester Research. The data comes from a survey conducted with a nationally representative sample of 60000 households. The survey includes various questions on ownership and use of various goods and services with a focus on telecommunications. Demographic, attitudinal and behavioral variables are present as well. We take the sample of 4865 respondents who report to have some type of home network and are online at least monthly. 2991 of the 4865 respondent have a wireless network and the remaining 1874 have other types of networks. WeusedataonthehoursperweekspentbytherespondentsontheInternetfor personal reasons, income of the respondent and ownership and type of home networking devices. We also use data on the time spent on the Internet for work related reasons to contrast the implications of the model. We also include some demographic controls to refine the results in some of the regressions. In the survey, answers to the questions regarding Internet time use are grouped as 1-4 hours, 5-9hoursandsoon. Foraconservativeestimatewetakethelowerboundsofthese intervals for our main results in the text but present the results taking the midpoints as well. The respondents with wireless networks spend an average of 10.66 hours per week on the Internet for personal reasons. This corresponds to 9.5 percent of the respondents’ non-sleep time off 112 hours, assuming 8 hours of sleep per day. For respondents with 42 Mean Internet Use (1) (1) Mean full income ∗ Wireless Network Owners 10.66 hrs(9.5 %) 0.9877 0.6060 $239295 Wired Network Owners 11.04 hrs(9.8 %) 0.9881 0.6045 $190280 Wireless Network Owners (mp) 12.54 hrs(11.1 %) 0.9895 0.5986 $234904 Wired Network Owners (mp) 12.92 hrs(11.5 %) 0.9898 0.5970 $186762 (mp) : taking midpoints for time use calculations: work and leisure time valued at wage Table 3.1: Summary Statistics wired networks, time spent on the Internet for personal reasons is 11.04 hours on average or 9.8 percent of non-sleep time. The numbers in the case where we use midpoints for time use calculations become 12.54 hours (11.1%) for wireless network owners and 12.92 hours (11.5%) for wired network owners. The time intensities of the two bundles can be calculated as one minus the ratio of market expenditures on the bundle to market expenditures plus time expenditures (1)=1 P i C i P i C i +WL i and (1)=1 P o C o P o C o +WL o Dividing the numerator and the denominator by W(1 L i L o ) and letting the expenditure shares of the Internet and the composite bundle be E i and E o respectively gives (1)=1 E i E i + L i 1−L i −Lo and (1)=1 E o E o + Lo 1−L i −Lo We substitute 0:0033 for the expenditure share of the Internet (E i ), 0:9967 for the expenditure share of the composite (E o ) and 0:3570 for the share of work time in the non-sleep time of the consumer (1L i L o ). The consumers with wireless networks spend 0:0951 of their time on the Internet and the remaining 0:5476 on leisure activities 43 other than the Internet. These yield time intensities of (1)=0:9877 for the Internet and (1) = 0:6060 for the composite in the case of wireless network owners. Owners of wired networks spend 0:0986 of their time on the Internet and 0:5442 of their time on other leisure activities. These yield time intensities of 0:9881 for the Internet and 0:6045 for the composite in the case of wired network owners. These statistics along with their counterparts in the case where we take midpoints for time use calculations are presented in table one. We use the time intensities in the calculation of the elasticities from the coefficient on the log of wage. The results of the regressions are reported in table two. The positive coefficients show that the respondents with higher incomes report spending less time on the Internet. Wireless network owners are more responsive to changes in the opportunity cost of time with an elasticity of 1:6381. Wired network owners on the other hand have an elasticity of 1:5222: As a contrast, in the third and eighth row of table one, we replicate the regression taking the time spent on the Internet for work related reasons as the independent variable. It can be assumed that respondents have little or no control on the time spent on the Internet for work related reasons, thus the coefficient need not be consistent with the models prediction for personal Internet use. As indicated by the negative coefficient, people report to spend more time on the Internet for work related reasons as the wage increases, but in this scenario it is not natural to think of the wage as the opportunity cost of the time spent on the Internet for work related reasons. These results are consistent with the findings of Goolsbee and Klenow (2006) that as the opportunity cost of time increases, people spend less time on the Internet for personal reasons but this is not true for the time spent on the Internet for work related 44 Coefficient Standard Error R 2 Implied Elasticity Wireless Network Owners 0.2436 0.0327 0.0182 1.6381 Wired Network Owners 0.2003 0.0404 0.0129 1.5222 Internet for Work -0.1507 0.0334 0.0055 N/A Wireless Network Owners (c) 0.3131 0.0452 0.1219 1.8190 Wired Network Owners (c) 0.2558 0.0568 0.1439 1.6685 Wireless Network Owners (mp) 0.1893 0.0246 0.0194 1.4841 Wired Network Owners (mp) 0.1626 0.0305 0.0149 1.4139 Internet for Work (mp) -0.1152 0.0253 0.0056 N/A Wireless Network Owners (c) (mp) 0.2408 0.0340 0.1275 1.6150 Wired Network Owners (c) (mp) 0.1985 0.0428 0.1511 1.5066 (mp) mid points (c) : controlling for value of assets, education and time spent on the Internet for work related reasons Table 3.2: Regression of ln((1L i )=L i ) on ln(W) reasons. However, conditional on having some type of network, we find the elasticities to be higher than those found in Goolsbee and Klenow (2006). Our benchmark regressions give estimates of the elasticity of 1:68 and 1:52 whereas Goolsbee and Klenow’s larger sampleofallrespondentswhoareonlineatleastmonthlygivesanelasticityof1:32which is not surprising since the larger sample includes respondents who rarely go online. As a second attempt to refine the estimates we include some control variables. We include education level of the respondent, number of hours spent on the Internet for work related reasons and the combined value of owned assets of the respondent. The implied elasticities go up slightly. The regression including the wireless network owners posit an elasticity of 1:8190 compared to 1:6685 for wired network owners. Furthermore, in the last five rows we report the results of the regressions where we use midpoints for time use calculations. Elasticities go down slightly but the effect on welfare estimates is quite pronounced as we will illustrate in the next section. 45 3.4 Welfare The consumer surplus can be approximated by equivalent variation. We use the expen- diture function E (P o ;P i ;F;W; u j Y i >0)= F + o (1+1=∆) 1=(1−) ( u 1 ) 1 and its counterpart in the case when Internet consumption is not available E (P o ;W; u j Y i =0)= o ( u 1 ) 1 to calculate the equivalent variation as a percentage of full income EV W = E (P o ;W; u(P o ;P i ;F;W j Y i >0) j Y i =0) W = [(1+ 1 ∆ ) 1 1 (1 F W )1] The consumer surplus naturally depends on the elasticity of substitution between theInternetbundleandthecompositebundle. RevokingtheassumptionthattheInternet hasasmallflatsubscriptionfee(F=W 0)andnopriceformarginalconsumption( =0) the equivalent variation becomes EV W =(1L i ) 1 1 1 Using the elasticity estimates we can calculate the equivalent variation relative to income. In table three we present the results of these calculations. For respondents with wireless networks, the consumer surplus turns out to be 16 percent of full income 46 EV/W EV/W (l) EV/W EV/W Difference at median income at average income Wireless Owners 1.6381 16% 3.2% $6755 $7684 Wired Owners 1.5222 22% 3.5% $6009 $6840 $844 Wireless Owners(c) 1.8190 13% 2.9% $6342 $7285 Wired Owners(c) 1.6685 16% 3.2% $5723 $6461 $824 Wireless Owners (mp) 1.4841 27% 4.2% $8762 $9980 Wired Owners (mp) 1.4139 34% 4.6% $7570 $8618 $1362 Wireless Owners(c) (mp) 1.6150 21% 3.9% $8404 $9642 Wired Owners(c) (mp) 1.5066 26% 4.2% $7415 $8399 $1242 (l): linearized (c):controlling for value of assets, education and Internet time use for work related reasons Table 3.3: Consumer Surplus as a percentage of full income (wage income plus the value of leisure). For those respondents with other wired types of networks the consumer surplus is 22 percent of full income. However, since the utility of consumption of the first unit is very high with a log demand the above calculations tend to overestimate the consumer surplus. To counter this effect and get a conservative estimate of the consumer surplus, we linearize the demand as in Hausman (1999) and use the fact that consumer surplus relative to full income is equal to the expenditure share divided by twice the elasticity which is equal to 0:5L i =(1L i (1F=W)) in our model. The calculations yield a consumer surplus of 3.2 percent of full income for wireless network owners. This is corresponds to $6755 per year for the wireless network owner with the median full income in the sample. On averagewirelessnetworkownersrealizea consumer surplusof $7648. The medianincome consumer with a wired network realizes $6009 of consumer surplus per year which is 3.5 percent of the full income and on average wired network owners realize $6840 consumer surplus per year. The difference in the average consumer surplus is $844 per year in the case without controls. Controlling for the value of assets, time spent on the Internet 47 for work and education, slightly decrease the welfare estimates. The average consumer surplus of wireless network owners goes down to $7285 per year. The average consumer surplusofconsumerswithawirednetworkgoesdownto$6461peryear. Theincremental consumer surplus realized by wireless network owners on average goes down to $824 per year with the controls. The welfare estimates go up across the board in the case where midpoints are used in the calculation of time use. The Incremental consumer surplus that the wireless networks owners realize goes up to $1362 in the benchmark case and is slightly lower at $1242 when controlling for assets, education and Internet time use for work related reasons. It is important to point out that these estimates, although taking into account the time-intensities in a more accurate way, still have to be viewed with caution. First reason to be cautious is that all of the non-sleep time for the consumer is valued at the wage. If consumers value their leisure time less than the wage, then we would be overestimating the welfare gains. We also do not take into account other time-intensive substitutes to the Internet except the composite. Taking into account other time intensive substitutes like watching TV or gym membership would increase elasticities and would reduce the welfare estimates. 3.5 Conclusion Hightimeintensityofuseforagoodorserviceisoneofthereasonsthatmaketraditional approaches to estimating elasticities difficult to apply. However, if there is enough varia- tion in the time use data, and the opportunity cost of time, elasticities can be estimated 48 from data on time use and wage instead of market prices. In the case highly time inten- sivegoods, thetruecostofconsumptionincludestheopportunitycostoftimebesidesthe minuscule market expenditures. Using the variation in time use and wage data is likely to give more accurate estimates of the elasticities and welfare than using market price and consumption data. We use a simple utility model proposed by Goolsbee and Klenow (2006)andestimatethewelfaregainsfromusingtheInternetforconsumerswithdifferent types of home networks. We find that the consumer surplus from the Internet is around $7000. With the most conservative estimate, consumers with wireless networks are found to be realizing on average $824 more consumer surplus from the use of the Internet com- pared to wired network owners. However we note that these estimates, although taking into account the time intensity explicitly, have to be viewed with caution since all leisure is valued at the wage in the model and no other time intensive substitutes to the Internet are considered. 49 Chapter 4 Welfare Effects of Spectrum Management Regimes 4.1 Introduction Radio spectrum is one of the building blocks of the contemporary communications tech- nology. Itisutilizedtocarryinformationintheabsenceofphysicalwiredconnectionsand is fundamental in the operation of countless devices and services in our everyday lives, from cell phones to TVs, from wireless networking to garage door openers, microwave ovens, baby monitors. Spectrum is a broadcast medium in the sense that the devices us- ing it emit modulated electromagnetic waves for an intended recipient, but consequently, manydevicestransmittinginformationinthesamefrequencymaycauseproblemsforthe intended recepients. It is important to note that, this phenomenon called interference is due to the recepient’s inability to decode the information intended for itself among other transmissions, and not an inherent deficiency of spectrum. As the number of communi- cations systems and devices has grown significantly over time, more and more attention needed to be paid to the efficient use of spectrum in a way that minimizes interference of equivalently maximizes the error free transmission of information. One way that this can 50 be achieved is by dedicating certain frequencies to certain transmitter receiver pairs so thatnounintendedtransmissionwoulddisturbtheflowofinformation. Thisapproachre- vealsitselfintheregulationofspectrumasthelicensingregime. Anotherwayofefficiently usingspectrumreliesonimprovingtransmittersandreceiverinsteadofdedicatingcertain frequencies to certain transmissions. As the technology continues to develop, transmit- ters become sophisticated enough to detect transmissions already in progress and stop transmitting before causing interference, and also receivers become advanced enough to filter and decode the intended transmission among other unintended flow of information. Technologies like software defined radios, cognitive radios, listen before talk protocols are few examples of such technologies. And the revelation of this latter approach in regula- tionofspectrumistheunlicensedallocationregime. Therearemeritstobothapproaches depending on the nature of the transmission as will be discussed below, and the optimal allocation of spectrum needs to be thororoughly informed by both approaches. Spectrum regulatory authorities such as the Federal Communications Commission (FCC) in the US or the European Telecommunications Standards Institute (ETSI) in Europe handle the allocation and management of radio spectrum. These authorities divide spectrum into blocks with distinct geographic and frequency boundaries and ded- icate these blocks to specific usages, or dedicate certain frequency blocks to common use of all prospective users. The rights to operate devices in these blocks are then al- located to competing users, either exclusively by employing auctions, or inexclusively through an unlicensed allocation. There are solid theoretical and empirical studies that have focused on perfecting the details of the competitive bidding process for the rights 51 to use a spectrum block. 1 However, the details of the unlicensed allocation alternative and the economic consequences have been mostly overlooked. Moreover, the spectrum management process prior to the allocation, by which the rules governing the access to and the use of spectrum are created, therefore effectively the choice between licensed or unlicensed allocation alternatives have been made, has not achieved a significant interest from economists. Theprocessofspectrummanagementascertainswhichuserscanobtainaccesstospec- trum and what type of services can be provided using the spectrum subject to certain rules. There are three approaches to the management of spectrum. The first; command andcontrolapproach,wherebytheregulatoryauthorityassignsspectrumtospecificusers and uses subject to pre-engineered technological rules and on an as needed basis by com- parative hearings, has been the dominant approach until recently. However, command and control approach is unanimously regarded as highly inefficient and is almost com- pletely abandoned. Most of the recent debate on spectrum management regimes focus on the remaining two approaches which are the licensing approach and the commons (unlicensed allocation) approach. The licensing alternative entails the creation and allo- cation of exclusive licenses to use the spectrum with protection from interference. The cell phone and the broadcast industry are the major examples of the products of a licens- ing regime and are proponents for further allocations in the same vein. Commons regime on the other hand entails specification of usage etiquettes but leaves access to spectrum open to any prospective user willing to accept and manage interference. Wireless net- working through Wi-Fi and Bluetooth are the most obvious products of these unlicensed 1 SeeCramton(2002),Klemperer(2002),KwerelandWilliams(2002)forexcellentaccountsofspectrum allocation mechanisms and spectrum auctions in particular. 52 allocations, as well as cordless phones, wireless microphones, baby monitors and many more. Evidently, the choice of management regime greatly affects the final variety and technological characteristics of the communication devices in the market and shapes the nature of communications. Another motivation for our work is the increased attention to white spaces, i.e. the unused frequencies in the digital television broadcast bands. The gradual completion in early 2009 of the transition from analog to digital broadcast left even more unused spectrum beyond the currently unused or unassigned frequencies. This raised the ques- tion of how and to what type of uses should these white spaces be allocated. Motivated by this question, we provide a framework highlights the tradeoffs between licensed and unlicensed allocation regime for any prospective allocation problem. One view is that the efficient way to allocate the spectrum is to assign exclusive licenses to use the spectrum with minimal restrictions on possible usages. It is argued that protection from inter- ference granted by the exclusivity of the licenses, and the flexibility on the end use is the key to efficiency and incentives to invest. Previous allocations that most resemble the property rights approach like the Commercial Mobile Radio Services and Advanced WirelessServicesfrequencies(usedmostlybycellularsystems)arereferencedassuccesses in terms of consumer surplus and government revenue created by the licensed allocation. The competing view credits commons approach for the success of unlicensed applications in certain spectrum bands like Wi-Fi and Bluetooth and argues that access to spectrum should be granted to any user provided that the user meets industry standards and us- age etiquettes. One of the main concerns when thinking about management regimes is interference. It is argued that in a commons regime, interference would lead to a tragedy 53 of commons scenario where a large number of interfering devices would degrade device or service quality. On the other hand, the variety and the distributed bottom up inno- vation resulting from open access in a commons scenario is argued to create considerable welfare. Therefore, the need for economic analysis that can enlighten the trade-offs faced in these dimensions is eminent. There are two main issues to be considered. The first is an empirical question about whether the marginal social value of an unlicensed alloca- tion is commensurate with that of licensed allocations so that an equally weighted future allocation is justified. the second question is more theoretical and concerns the welfare effects of quality degrading free entry under unlicensed allocation and compares that to a static oligopoly market with only licensed firms. After establishing in the previous chapters that there is marginal value to be gained in unlicensed allocations, the second issue to be addressed is whether an unlicensed al- location would lead to a tragedy of commons result with excessive entry that degrades quality of devices or services because of interference. To root our paper in recent de- bate, we will use the white spaces analogy but our model is applicable to any measure of spectrum on which prospective management approaches are to be evaluated. We ad- dress this concern by modeling a communications device market whereby the allocation of spectrum determines the technological environment, the competitive environment and consequently the quality and the variety of the devices in the market. We then go on to simulate the welfare consequences of instituting the alternative management regimes. The market is modeled as a differentiated goods oligopoly in three stages. In the first stage firms decide whether or not to incur the entry cost enter the market. In the second stage firms choose the technological attributes of the devices they will produce which in 54 equilibrium along with the choices of other firms results in a quality level. And in the third stage firms compete in quantities in the device market. We incorporate vertical differentiation through the indirect quality choice and horizontal differentiation by intro- ducing imperfect substitutability among devices. In such a model the welfare effects of the alternative management regimes result from their effect on the final product offering of the industry. Under the licensing regime the access to and the use of any given white spaceisassignedtoasinglefirm,thusthereareasmanyfirmsastherearelicenses. These firms also have complete control on the choice of device quality since they are not subject to anyinterferencethat can degrade quality. Consequently, the industry becomes a static differentiated goods oligopoly with as many firms as there are licenses. On the other hand, if the commons regime is chosen, any firm can access any given white space and deploy devices. This non exclusivity may result in interference for firms using the same white space and consequently the device quality might be degraded. Also, the number of firms becomes endogenous and is determined by zero profit condition subject to free entry. Assuming that consumers have preferences for both the quality and the variety of devices,thetrade-offisthenbetweenalicensingregimewithhighqualitydevicesbutwith fewer varieties, and a commons regime with a greater variety of lower quality devices. Our work complements and extends the oligopolistic markets literature. Many of the previous works carry out welfare evaluations between price and quantity competition for a given number of firms. The conventional wisdom in this literature upholds that price competition dominates quantity competition from a consumer welfare perspective. Hackner (2000) shows that the prices may be lower under quantity competition contrary to the prevailing wisdom in Singh and Vives (1984), and Vives (1985) when the market 55 is bigger than a duopoly with vertical differentiation. Hsu and Wang (2005) on the other hand, show that consumer surplus under price competition in Hackner’s model is still higher that quantity competition. Symeonidis (2003) shows that an increase in quality heterogeneity for a given number of firms increases consumer surplus regardless of the mode of competition. We complement these works by showing that with free entry, and endogenous quality choice subject to a negative externality, consumer surplus ceases to be monotonically increasing in the number of firms. In cases where free entry lowers quality rapidly because of a high interference elasticity, free entry may be detrimental to welfare. However, for high enough horizontal differentiation, the welfare enhancing effects of the increase in variety more than offsets the welfare degrading effects of the drop in quality as a result of interference. Ultimately, the choice on the management regime has to be informed bythe preference and the technologystructure. In cases where interference does not increase rapidly with entry, and consumers have a strong preference for variety, commons regime with no restriction on entry is welfare enhancing. Whereas, if interference is a concern and consumers have weak preferences for variety, restricting entry with licensing regime creates greater welfare. In the rest of this chapter we provide a differentiated oligopoly model of a communi- cations device market and analyze the welfare consequences of alternative management regimes. 56 4.2 An Oligopoly Model of a Communications Device Market Having established in previous chapters that unlicensed spectrum can create consider- able welfare even with a very conservative valuation approach, we now go on to analyze whether interference resulting from free entry and the possibility of a tragedy of com- mons scenario in an unlicensed regime would be detrimental to welfare. We develop a modelofacommunicationsdevicemarketwherebytheallocationofspectrumdetermines thetechnologicalenvironment,thecompetitiveenvironmentandconsequentlydetermines the quality and the variety of the devices in the market. We then go on to simulate the welfare consequences of instituting the licensed and unlicensed allocation alternatives. 4.2.1 Model Preliminaries. We start by considering a market for communications devices. There are M consumers with preferences summarized by the following quasilinear utility function defined over a homogenous good and n varieties of differentiated communications devices: U(q 0 ;q 1 ;:::;q n )= n ∑ i=1 (q i q 2 i T 2 i ) ∑ i ∑ j<i q i T i q j T j +q 0 (4.1) where q i and T i are respectively the quantity and quality of variety i=1;:::;n and q 0 is the quantity of a homogenous good that we also chose to be the numeraire. This spec- ification incorporates both vertical and horizontal dimensions of product differentiation. The vertical attribute T i measures the inherent objective quality of a device of variety i. Ontheotherhand,thenvarietiesofdevicesdifferaccordingtoahorizontalcharacteristic 57 which is captured by the parameter 2 [0;2]: This parameter is exogenously given and measures the extent to which any two of the n varieties are substitutable. As goes to zero, devices become independent and as goes to 2, devices become perfect substitutes. The utility function is strictly concave and decreasing in which reflects that the con- sumer has a taste for variety. The consumer also supplies one unit of labor inelastically for the production of the homogenous good in a perfectly competitive industry with con- stant returns to scale. One unit of labor is required to produce a homogenous good and marginal cost pricing under perfect competition equalizes labor income to unity. The quasilinearity isolates the utility in (1) from income effects and assuming that the con- sumer has a large endowment e q 0 of the homogenous good we focus on interior solutions. Following a standard utility maximization, the demand and the inverse demand are given respectively by q i T i = [2+ (n2)]T i (1p i ) ∑ j̸=i T j (1p j ) (2 )[2+ (n1)] (4.2) p i =1 2q i T 2 i T i ∑ j̸=i q j T j (4.3) The differentiated communications devices are produced by oligopolistic firms under increasing returns to scale. Marginal costs of devices are zero, however a fixed cost F is required to produce any level of output of any quality. This fixed cost represents the cost of obtaining access to a key resource or technology, in our case the access to spectrum. The presence of the fixed cost also implies a one-to-one mapping between varieties and firms. Thus, the subscript i interchangeably refers to a firm and the single variety it 58 produces. This model subsumes the standard horizontal differentiation model that can obtained by setting T i = 1 which reduces the model to the standard quadratic utility ∑ i (q i q 2 i ) ∑ i ∑ j<i q i q j with the inverse demand p i =12q i ∑ j̸=i q j . 4.2.2 The Device Market We model the device market as a three stage game. In the first stage, firms decide on whether incur the fixed cost and enter the market. In the second stage, those firms that participate, noncooperatively choose the quality of their devices T i () and incur the associated costs K(T i ). We will allow quality to be a direct choice variable as well as a function of an underlying choice variable, namely device design. This will later allow us to incorporate quality externalities that are caused be free entry. In the third stage firms compete in quantities. The fixed cost and the rules of entry in the first stage represents theregulatoryauthority’schoiceonthemanagementregime. Licensingregimeentailsthe creation of a license to access and use the spectrum white space. Abstracting from any optimizing behavior on the part of the regulatory authority over the number of licenses andthefixedfee, weassumethatasinglelicenseforeachofthew whitespacesiscreated. These licenses are assumed to be allocated with a second price auction for the sake of simplicity. Assuminganoutsideoptionofzero, allfirmswouldbiduptotheirprospective symmetricprofit. Thistieisthenbrokeninfavorofarandomallocationofthelicensesto w of the firms and those firms do participate in this case of indifference. The regulatory authority extracts all the profit with the access priceF = i (w) where i (w) is the profit of each firm when there are w firm that are active. On the other hand, under commons regime, the access fee f is set close to zero and firms enter the market until net profit 59 falls to zero. In both cases, firms end up with zero profit and we only focus on consumer surplus. We solve for the subgame perfect equilibria of the model for a given number of firms. After solving the model for a given number of firms, we provide equilibrium analysis and comparative statics as well as a simulation of the free entry equilibria under the two regimes. 4.2.2.1 Quantity Competition We solve the model by backward induction. In the last stage firms choose quantities simultaneouslytomaximizeprofits, takingthequantitychoicesoftherivalfirmsasgiven. max q i i =M 1 2q i T 2 i T i ∑ j̸=i q j T j q i K(T i ) (4.4) Taking the first order conditions and rearranging gives the reaction functions as q i T i = 1 4 T i ∑ j̸=i q j T j We impose symmetry and solve for the Cournot quantities and the resulting prices q c i = T i ( [4+ (n2)]T i ∑ j̸=i T j ) [4+ (n1)](4 ) p c i = 2 ( [4+ (n2)]T i ∑ j̸=i T j ) T i [4+ (n1)](4 ) Plugging the Cournot quantities and prices back into the profit function gives the second stage profit as a function of qualities. In the second stage firms maximize profit by choosing the quality. The maximization problem at the second stage is: 60 max T i i = 2M a 2 b 2 (a )T i ∑ j̸=i T j 2 K(T i ) where a=[4+ (n1)] and b=(4 ):Differentiating with respect to quality gives the first order condition: 4M a 2 b 2 (a )T i ∑ j̸=i T j (a )T ′ i K ′ (T i )=0 Summing up over i, imposing symmetry, rearranging and substituting for a and b yields T ′ c ()T c () ( 4 M [4+ (n2)] [4+ (n1)] 2 (4 ) ) =K ′ (T c ) (4.5) The solutions to this set of equations provide the best response correspondence, the fixed point of which obtains the equilibrium quality T c . We can then characterize the subgame perfect equilibrium by plugging quality back into the Cournot quantities and prices. For the symmetric case at hand, these become: q c = (T c ) 2 [4+ (n1)] p c = 2 [4+ (n1)] The equilibrium profit of each firm is given by c = 2M(T c ) 2 [4+ (n1)] 2 K(T c ) 61 The equilibrium number of firms will be determined in the first stage by the rules and the fixed cost of entry. When licensing regime is implemented number of firms will be exogenous and equal to the number of licenses n ∗ = w. Under commons regime, equilibrium number of firms will be determined by zero profit condition c (n ∗ )f = 0. The consumer surplus is then obtained by calculating how much consumers would have to be paid to forgo the entire consumption bundle, which is expressed as CS c (n ∗ )=n ∗ M ( q c ( q c T c ) 2 (n ∗ 1) 2 ( q c T c ) 2 p c q c ) 4.2.2.2 Specication of Quality Quality specification is motivated by the engineering aspects of communications devices. We treat the quality choice as a convex problem in its own stage. Firms indirectly choose quality by choosing the design of their devices with a decreasing marginal quality. When carried over to the quantity competition stage, this configuration results in a concave net profit as a function of design. We assume design is increasingly costly with a convex marginal cost. This amounts to assuming that cost increases faster than gross profit and admits an interior optimum design. For the specification of quality we turn to Shannon- Hartley theorem of Information Theory. In this configuration the quality of a device is measured by the information throughput of the device in kilobits per second. The amount of information that can be reliably transmitted over a communications channel by any device is bounded above by a channel capacity. This upper bound is specified by the Shannon-Hartley theorem as C =W log 2 (1+ S N ) where C is the channel capacity in kilobits per second (kbps),W is the bandwidth of the channel in kilohertz (kHz),S is the 62 total signal power in watts over the bandwidth and N is the total noise power over the bandwidth. Shannon (1949) calls the system that transmits without errors at rate C an ideal system and goes on to add that such a system can not be achieved with any finite encodingprocessbutcanbeapproximatedascloselyasdesiredbyefficientlydesigningthe device and the encoding properties. Firms in our model aim to approximate the Shannon upper bound by investing in the design d i 0 of the devices they produce. Thus the quality T i of a device of variety i; measured by the attained information throughput is given by T i (d i ) = (1 e −d i ) C with T ′ i > 0;T ′′ i < 0 and lim r i →∞ T i (d i ) = C. The cost of design is K i (d i ) = e d i d i 1; with K i (0) = K ′ i (0) = 0 and K ′ i > 0;K ′′ i > 0: The convexity of the design cost reflects the idea that as one aims to get closer to the Shannon upper bound, the necessary encoding becomes increasingly complex resulting in higher marginal cost. 2 This assumption guarantees the existence of a symmetric interior equilibrium. The quality degrading effects of interference are captured through the noise power over the bandwidth. Interference is an increase in noise that a system becomes subject to with the operation of another system in the same bandwidth. Assuming that the n firms will be distributed uniformly over the white spaces, the number of firms that operate on the same white space at any given time is defined as m = n−(nmodw) w where w is the number of white spaces. Because of the discreteness of the number of firms we use a modularapproachandweassumethattotalnoisepoweroverthebandwidthisincreasing in m : @N(m)=@m > 0; and that the interference elasticity is dlogN(m)=dlogm = ": With these specifications, the quality of a device is expressed as 2 See James Martin’s book Telecommunications and the Computer, ch 16 for an account of encoding and channel capacity. 63 T i (d i jW;w;S;N;n)=(1e −d i ) W log 2 (1+ S Nm " ) (4.6) 4.2.3 Equilibrium Analysis Setting quality equal to one reduces the model to the standard horizontally differentiated Cournot model with the inverse demand p i = 1 2q i ∑ j̸=i q j . The equilibrium quantities and prices are given by: q c = 1 [4+ (n1)] p c = 2 [4+ (n1)] For any positive substitutability, the profit c =2M=[4+ (n1)] 2 is monotonically decreasing in the number of firms. Furthermore the consumer surplus nM(2 + (n 1)=2[4 + g(n 1)] 2 is monotonically increasing in the number of firms. Therefore, a reduction in the fixed cost of entry that allows more firms to enter would increase the consumer surplus in a scenario without vertical differentiation. This conventional wis- dom that greater competition leads to greater consumer surplus fails to hold in scenarios with vertical differentiation and the threshold where free entry starts to decrease con- sumer surplus depends on the substitutability and the speed of decrease in quality due to interference that comes about with free entry. Substituting the quality T i (d i ) = C(1e −d i ) where C = W log 2 (1+ S Nm " ) and the cost of design K(d i )=e d i d i 1 into the first order condition yields 4MC 2 (a ) 2 a 2 b 2 (e −d i ) (1e −d i ) 4MC 2 (a ) a 2 b 2 (e −d i ) ∑ j̸=i (1e −d j )=e d i 1 64 Summing over i and solving for the symmetric design gives the equilibrium design as d c = 1 2 ln ( 4 M C 2 [4+ (n2)] [4+ (n1)] 2 (4 ) ) (4.7) With this closed form expression for device design we can characterize the subgame perfectequilibriumbysettingthedesignchoicetoitsequilibriumlevelwhichthenimplies the equilibrium quality T c (d c jW;w;S;N;n)=(1e −dc ) W log 2 (1+ S Nm " ) (4.8) Cournot output and price then becomes q c = (T c ) 2 [4+ (n1)] p c = 2 [4+ (n1)] The symmetric equilibrium profit is given by c = 2M(T c ) 2 [4+ (n1)] 2 K(d c ) Finally the consumer surplus is obtained by finding how much the consumers would havetobepaidtoforgotheentireconsumptionbundle. ThisisgivenbyCS c =M(U(Q) np c q c ) where Q=fq i ji21;:::;n;q i =q c g:Then the consumer surplus becomes CS c =nM ( q c ( q c T c ) 2 (n1) 2 ( q c T c ) 2 p c q c ) 65 4.3 Simulations Therichpreferenceandtechnologyspecificationcomesatthecostofcomplexclosedform comparative statics. However, we can simulate the equilibria and analyze the effects of key parameters and regime choices. Having solved for the subgame perfect equilibria, we go on to simulate these equilibria for a range of parameter choices. There are seven parameters of the model. These are the number of white spaces w, the bandwidth of white spaces W, signal power over the bandwidth S, noise power over the bandwidth N, substitutability , interference elasticity ", and the fixed cost F: For a given number of firms n and a set of parameter values fw;W;S;N; ;";Fg we can calculate the equilibria of the second and the last stages. The free entry condition of the first stage can be simulated by iterating the number of firms until profit is exhausted. Under the licensing regime, the equilibrium number of firms is always equal to the number of white spaces since there is a single license to obtain access to each white space. The profit of the activefirms will be extracted in the entrystage by the regulatory authority withF = i obtained by the auction mechanism. So given the set of parameter values except F, the equilibrium under licensing regime is obtained simply by pinning down n = w. The fixed cost parameter is obtained in equilibrium by setting i = F. Under the commons regime, access cost is close to zero and number of firms will increase until profit equals this minimal fixed entry cost. So, the equilibrium under commons regime for a given set of parameter values is obtained by setting the fixed cost sufficiently low at F =10 −6 and iterating over the number of firms until equilibrium profit is zero. 66 We pin down the bandwidth W of a given white space to 6000kHz since this is the bandwidth of a TV white space that is under question. And we pin down the number of white spaces w to 10, since this is roughly the average number of white spaces in a given geographic area. Another convenient way to reduce the parameter space without loss of generality is to make use of the fact that the noisiness of a communication channel is expressed in decibels(dB) by Signal to Noise Ratio, SNR = 10log 10 ( S N ). A typical signal to noise ratio for a wireless networking device is about 30dB which corresponds to a signal power that is 1000 times the noise power over the channel. To be able to analyze equilibria in different native noise environments in the simulations we let this parameter vary within the set S = f0;10;20;30;40;50;60;70;80;90g. In the same vein, we want to analyze different values for how fast interference in a given white space in- creases with the entry of a firm. So we let the interference elasticity" vary within the set E = f0:2;0:4;0:6;0:8;1;1:2;1:4;1:6;1:8;2g. For an interference elasticity of " = 0:2, the increaseinnoiseduetotheentryofnewfirmswouldberelativelyslow,whereasfor"=1it wouldbelinear,andfor"=2itwouldbequadratic. Wealsodonotwanttopindownsub- stitutabilitytobeabletoallowdifferentpreferencesonthedemandside. Weletthesubsti- tutability parameter vary within the set Γ=f0:1;0:3;0:5;0:7;0:9;1:1;1:3;1:5;1:7;1:9g. To avoid degenerate equilibria, we choose not to include the extreme cases of perfect substitutability at = 2 and no substitutability at = 0: Table C.2 in the appendix C summarizes these parameter choices. Consistent with the parameter choices there are three levels of iterations in the simu- lation algorithm. Given the number and the bandwidth of the white spaces and the fixed 67 cost of access, first the algorithm picks the native signal to noise ratio. Then, the algo- rithm picks the interference elasticity at the next level. Finally, the algorithm picks the substitutability in the innermost level. It then continues by calculating the equilibrium design, qualities, quantities, prices, profits and consumer surplus as it iterates over the number of firms. To find the equilibria under licensing regime one iteration suffices since number of firms equals that of the white spaces with a single firm operating in a given white space. Effectively, licensing regime creates an industry with Cournot competitors that operate without interference and that are equal in number to that of white spaces. To simulate the commons regime, the algorithm iterates over the number of firms until the profit per firm is equal to the access fee. The algorithm is summarized in pseudocode in table C.1 in appendix C. With these specifications we simulate the model in 10 cases of native signal to noise ratios, and in each case, one hundred times for the combinations of substitutability and interference elasticity parameters. We obtain one thousand equilibria for each parameter combinationinSEΓ. AlloftheequilibriaareincludedintheappendixinfiguresC.13 through C.22 in appendix C. Since the fixed cost is lower under commons than it is under licensing, and since profit is decreasing in the number of firms, there are always at least as many firms under commons regime as there are under licensing. So, moving from an oligopolywithafixednumberoffirmstoafreeentryscenariowouldbewelfareenhancing according to the conventional wisdom. This would ideally lead to a welfare improvement throughtoughercompetition, lowerpricesandagreatervariety. However,thepresenceof interference complicates the conventional comparative statics. In environments like ours 68 that are prone to congestion or interference, care must be taken when evaluating welfare consequences of policies that shape the market. Figures C.1 through C.10 show the consumer surplus under commons and licensing regimes for each of the ten native signal to noise rations that we simulated the model for. In each figure the x and y axis are the combinations of substitutability and interference elasticity, and the two dimensional manifolds plot the consumer surplus under licensing regime and the consumer surplus under commons regime measured at the vertical axis. The gray surface is the consumer surplus under licensing regime , and the colored surface is the consumer surplus under commons. Starting at the lowest native SNR, we observe inonlyoneoftheonehundredcombinationsofsubstitutabilityandinterferenceelasticity that the commons regime with free entry dominates the licensing regimes with number of firms restricted to the number of licenses. Except the combination of the lowest sub- stitutability at 0.1 and lowest interference elasticity at 0.2, all other combinations result in higher consumer surplus under licensing. So in environments that are inherently noisy, and with low return to design investment, instituting a commons regime with free entry would result in higher welfare only in cases where interference does not increase quickly with entry and consumers have strong preferences for variety. As we go to higher na- tive signal to noise ratios, the combinations of substitutability and interference elasticity for which commons regime dominates expands. At higher native SNRs firms have more room to respond to increases in noise due to entry through their choice on device de- sign. As native SNR increases, the marginal quality improvement for a unit of design investment becomes higher and more firms prevail in a free entry equilibrium. At the highest native signal to noise ratio, about half of the combinations result in commons 69 dominating licensing in terms of consumer surplus. The regions where commons domi- nates can more clearly be seen in the contour plots of the consumer surplus ratio in figure C.11 in the appendix. The tension between the preference for quality and the preference for variety determines the thresholds in the parameter space where the dominance goes from one regime to the other. The regions where commons regime creates more surplus fall towards the regions with lower substitutability and low interference elasticity and the region grows as one move towards higher native SNR. The boundary in the space of substitutability, interferenceelasticity and nativeSNR, where welfare dominance changes from one regime to the other is depicted in figure C.12. At any combination that lies above this manifold, commons creates higher consumer surplus whereas any combination below favors licensing. Comparison of the free entry and static equilibria for a given native SNR, and com- binations of substitutability and interference elasticity can be carried out by analyzing figures C.13 through C.22. These figures present the key variables for each of the 100 equilibria that is simulated for each of combination of substitutability and interference elasticity parameters. Licensing equilibria on the right all have 10 firms since that is the numberoflicenses. Commonsequilibriaontheleftblockhavehighernumberoffirmspre- vailing in equilibrium, thus higher varieties. Not surprisingly, prices are uniformly lower under the free entry equilibria, so are the qualities. Quantities on the other hand show a non monotonic pattern as one moves from licensing to commons. For high interference elasticities, as one moves to a free entry equilibria, the drop in quality because of inter- ference is pronounced. The number of firms that enter is not very high and each active firm produces less compared to the licensed firms. Consequently, the welfare degrading 70 effects of the drop in quality and quantity more than offset the welfare enhancing in- crease in varietiesand the decrease in prices. However, as one moves to lower interference elasticities, a switch from licensing to commons allows more and more firms to prevail in equilibrium and causes the usual drop in price and qualities. The drop in prices becomes sharper and the drop in qualities become less pronounced. At the same time, along the substitutability dimension, lower substitutability leaves more market power to each pre- vailing firm and allows higher number of firms to prevail in equilibrium. Depending on the intensity of entry, prices may decrease or increase. For high interference elasticities, the intensity of entry is low, thus prices increase as substitutability declines, however, for lower interference elasticities, intensity of entry is high, and prices go down. The final effect on welfare of going from a static equilibrium to a free entry equilibrium is a priori ambiguous. We observe that for low enough interference elasticity, there are low levels of substitutabilityforwhichfreeentryequilibriaandthewelfareenhancingeffectsofgreater varieties therein, more that offsets the welfare degrading effects of lower qualities. As we movetohighernativeSNRs,thedecreaseinthenoisinessofthechannelallowsmorefirms to operate for any given parameter combination without degrading quality. This pushes pricesdownandincreasesquantitiesleavingtheconsumerwithahighergaininconsumer surplus for any combination of substitutability and interference elasticity. Therefore, the regions of substitutability and interference elasticity plane where commons dominates in terms of consumer surplus expands as one moves to less noisy environment indicated by higher native SNRs. This can be seen in the contour plots in figure C.11. We can summarize all these observations in figure C.12. This figure shows the manifold in the three dimensional parameter space that is obtained by combining the contour plots for 71 each native SNR. The separating manifold divides the space into two regions and any pa- rameter combination above this manifold the free entry equilibria of the commons regime creates more consumer surplus than the static equilibria of licensing regime. 4.4 Conclusion After establishing that there is welfare to be gained in unlicensed spectrum allocations in previous chapters, in this chapter, to address the concerns about a tragedy of commons scenario we provide a model in which interference is incorporated as a quality degrading externality and in which the choice of spectrum management regime affects consumer welfare through basic market mechanisms. We model consumers as having preferences, over a range of varieties, for communications devices and device qualities. Devices are produced and sold by competitive firms. Choice of management regime affects the wel- fare outcome in two ways. First it directly affects the level of competition in the market and secondly by indirectly affecting the quality and variety of devices produced. We find that,ifinterferenceincreasesrapidlywithnumberoffirmsinthemarketandconsequently quality is degraded quickly with free entry, than restricting entry with a licensing regime might be welfare enhancing. Otherwise, under an unlicensed allocation, despite lower qualities, the increase in variety and the consequent increase in utility more than offsets the drop in quality. In addition to the empirical finding that establishes commensurate values for similar licensed and unlicensed spectrum bands, the oligopoly model also pre- scribes a mix of regime choices informed by the technological (interference) and economic (substitutability) characteristics of the market and application at hand. Favoring one 72 regime unanimously over the other without regard to the technological and economic environment would likely add to the inefficiencies that current allocations suffer from. 73 References [1] Y. Benkler. Overcoming agoraphobia: Building the commons of the digitally net- worked environment. Harvard Journal of Law and Technology, 11(2):287–400, 1998. [2] K. R. Carter, A. Lahjouji, and N. McNeil. Unlicensed and unshackled: A joint osp- oet white paper on unlicensed devices and their regulatory issues. SSRN eLibrary, 2003. [3] R. Cellini, L. Lambertini, and G. Ottaviano. Welfare in a differentiated oligopoly with free entry: a cautionary note. Research in Economics, 58(2):125–33, 2004. [4] P. Cramton. Spectrum auctions. In M. Cave, S. Majumdar, and I. Vogelsang, editors, Handbook of Telecommunications Economics, chapter 14, pages 605–639. Elsevier Science B.V., Chapter 14, 605-639, 2002, 2001. [5] A. K. Dixit and J. E. Stiglitz. Monopolistic competition and optimum product diversity. American Economic Review, 67(3):297–308, 1977. [6] L. Glass. Color theory: a short history. available at: http://lilyglass.net/school/Art200/Handouts/colortheory1.pdf. [7] A. Goolsbee and P. J. Klenow. Valuing consumer products by the time spent using them: An application to the internet. American Economic Review, 96(2):108–113, 2006. [8] J. Hackner. A note on price and quantity competition in differentiated oligopolies. Journal of Economic Theory, 93(2):233–239, 2000. [9] J. Hausman. Valuing the effect of regulation on new services in telecommunica- tions. Brookings Papers on Economic Activity, Economic Studies Program, 28(1):1– 54, 1997. [10] J. Hausman. Cellular telephone, new products, and the cpi. Journal of Business and Economic Statistics, 17(2):188–194, 1999. [11] T. Hazlett. Spectrum tragedies - avoiding a tragedy of the telecommons: Finding therightpropertyrightsregimefortelecommunications. YaleJournalonRegulation, 22, 2005. [12] J.JaskoldGabszewiczandJ.F.Thisse. Entry(andexit)inadifferentiatedindustry. Journal of Economic Theory, 22(2):327–338, 1980. 74 [13] P. Klemperer. What really matters in auction design. Journal of Economic Perspec- tives, 16(1):169–189, 2002. [14] E. Kwerel and J. Williams. A proposal for a rapid transition to market allocation of spectrum. Technical report, FCC OPP Working Paper Series, 2002. [15] E.R.KwerelandG.L.Rosston. Aninsiders’viewoffccspectrumauctions. Journal of Regulatory Economics, 17(3):253–89, 2000. [16] J. Martin. Telecommunications and the Computer. Prentice-Hall, Englewood Cliffs, N. J, 3 edition, 1990. [17] W. Novshek. Cournot equilibrium with free entry. Review of Economic Studies, 47(3):473–86, 1980. [18] W. Novshek. Finding all n-firm cournot equilibria. International Economic Review, 25(1):61–70, 1984. [19] K. Okuguchi. Quasi-competitiveness and cournot oligopoly. Review of Economic Studies, 40(1):145–48, 1973. [20] J. M. Peha. Spectrum management policy options. IEEE Communication Surveys,, 1(1):2–8, 1998. [21] G.Rosston. Thelongandwindingroad: thefccpavesthepathwithgoodintentions. Telecommunications Policy, 27(7):501–515, 2003. [22] S. Schwartzman. The Words of Mathematics: An Etymological Dictionary of Math- ematical Terms Used in English. The Mathematical Association of America, Wash- ington, D.C., 1994. [23] A.ShakedandJ.Sutton. Relaxingpricecompetitionthroughproductdifferentiation. Review of Economic Studies, 49(1):3–13, 1982. [24] C.E.Shannon. Communicationinthepresenceofnoise. Proceedings of the Institute of Radio Engineers, 37(1):10–21, 1949. [25] N. Singh and X. Vives. Price and quantity competition in a differentiated duopoly. RAND Journal of Economics, 15(4):546–554, 1984. [26] J. Sutton. One smart agent. Rand Journal of Economics, 28:605–628, 1997. [27] J. Sutton. Technology and Market Structure. MIT Press, Cambridge, MA, 1998. [28] G. Symeonidis. Quality heterogeneity and welfare. Economics Letters, 78(1):1–7, 2003. [29] X. Vives. On the efficiency of bertrand and cournot equilibria with product dif- ferentation. Journal of Economic Theory, 36(1):166–175, 1985. [30] X. H. Wang and J. Hsu. On welfare under cournot and bertrand competition in dif- ferentiated oligopolies. Working Papers 0514, Department of Economics, University of Missouri, 2005. 75 Appendix A Appendix to Chapter Two Units sold (millions) Price Income per capita 1997 28.2 $74.4 $29910 1998 31.3 $69.6 $30620 1999 40.1 $54.9 $32260 2000 43.3 $55.0 $34410 2001 45.3 $52.9 $34830 2002 47.4 $51.0 $35240 2003 49.5 $49.1 $37520 2004 51.6 47.4 41160 Source: FCC OSP working paper 39 Table A.1: Corless phone market data R 2 =0:98 Coefficients Standard error t Stat P-value 19.91** 2.77 7.18 0.00008 -1.2*** 0.13 -8.58 0.0003 0.23 0.21 1.08 0.32 ln(Q)=+ln(P)+ ln(I)+" Q=units sold P = Price I = Income "= Standard nomral error Table A.2: Cordless phone regression results 76 Appendix B Appendix to Chapter Three Data Description The data used for the analysis comes from Forrester Research. The data includes responses to surveys on telecommunication device ownership and use as well as general technology attitudes and about ownership, now or in the future, of various technology products. Survey responses are supplemented with rich demographics. Surveys are con- ducted with a nationally representative sample of 68000 North American households, of which 53000 are US households. Data files are available by request from the author. 77 Benchmark Case ALL CASES Regression Statistics Multiple R 0.1267 R Square 0.0160 Coefficients Standard Error t Stat Intercept 1.9650 0.0898 21.8729 LN(W) 0.2233 0.0251 8.9042 WIFI Regression Statistics Multiple R 0.1349 R Square 0.0182 Coefficients Standard Error t Stat Intercept 1.8826 0.1201 15.6785 LN(W) 0.2436 0.0327 7.4457 NOWIFI Regression Statistics Multiple R 0.1137 R Square 0.0129 Coefficients Standard Error t Stat Intercept 2.0570 0.1390 14.7936 LN(W) 0.2003 0.0404 4.9527 Table B.1: ln(W) regression results without controls using lower bounds 78 Controlling for Internet timeuse for work ALL CASES Regression Statistics Multiple R 0.3404 R Square 0.1159 Coefficients Standard Error t Stat Intercept 2.2458 0.1056 21.2579 LN(W) 0.2839 0.0288 9.8707 LIWhr -0.0410 0.0020 -20.1418 WIFI Regression Statistics Multiple R 0.3310 R Square 0.1096 Coefficients Standard Error t Stat Intercept 2.1402 0.1386 15.4446 LN(W) 0.3025 0.0371 8.1613 LIWhr -0.0391 0.0026 -15.3026 NOWIFI Regression Statistics Multiple R 0.3582 R Square 0.1283 Coefficients Standard Error t Stat Intercept 2.3740 0.1672 14.1976 LN(W) 0.2655 0.0471 5.6416 LIWhr -0.0444 0.0034 -13.1610 Table B.2: ln(W) regression results controlling for Internet time use for work using lower bounds 79 Controlling for internet at work, assets and education ALL CASES Regression Statistics Multiple R 0.3582 R Square 0.1283 Coefficients Standard Error t Stat Intercept 2.0475 0.1250 16.3848 LN(W) 0.2846 0.0352 8.0942 LIWhr -0.0421 0.0023 -18.5140 asset value 0.0000 0.0000 -2.9352 education 0.0506 0.0236 2.1422 WIFI Regression Statistics Multiple R 0.3491 R Square 0.1219 Coefficients Standard Error t Stat Intercept 1.8563 0.1678 11.0609 LN(W) 0.3131 0.0452 6.9213 LIWhr -0.0393 0.0029 -13.7125 asset value 0.0000 0.0000 -2.8944 education 0.0591 0.0306 1.9291 NOWIFI Regression Statistics Multiple R 0.3794 R Square 0.1439 Coefficients Standard Error t Stat Intercept 2.2505 0.1922 11.7096 LN(W) 0.2558 0.0568 4.5012 LIWhr -0.0467 0.0037 -12.4998 asset value 0.0000 0.0000 -1.0572 education 0.0473 0.0375 1.2608 Table B.3: ln(W) regression results controlling for Internet at work, assest and education using lower bounds 80 Benchmark Case (MP) ALL CASES Regression Statistics Multiple R 0.1328 R Square 0.0176 Coefficients Standard Error t Stat Intercept 1.7501 0.0677 25.8415 LN(W) 0.1766 0.0189 9.3428 WIFI Regression Statistics Multiple R 0.1391 R Square 0.0194 Coefficients Standard Error t Stat Intercept 1.6983 0.0904 18.7775 LN(W) 0.1893 0.0246 7.6814 NOWIFI Regression Statistics Multiple R 0.1222 R Square 0.0149 Coefficients Standard Error t Stat Intercept 1.8071 0.1050 17.2159 LN(W) 0.1626 0.0305 5.3254 Table B.4: ln(W) regression results without controls using mid points 81 Controlling for Internet timeuse for work (MP) ALL CASES Regression Statistics Multiple R 0.3486 R Square 0.1215 Coefficients Standard Error t Stat Intercept 2.0266 0.0797 25.4121 LN(W) 0.2210 0.0216 10.2341 LIWhr -0.0311 0.0015 -20.6501 WIFI Regression Statistics Multiple R 0.3382 R Square 0.1144 Coefficients Standard Error t Stat Intercept 1.9503 0.1045 18.6607 LN(W) 0.2334 0.0278 8.3867 LIWhr -0.0296 0.0019 -15.6650 NOWIFI Regression Statistics Multiple R 0.3677 R Square 0.1352 Coefficients Standard Error t Stat Intercept 2.1208 0.1264 16.7819 LN(W) 0.2092 0.0353 5.9197 LIWhr -0.0337 0.0025 -13.5220 Table B.5: ln(W) regression results controlling for Internet time use for work using mid points 82 Controlling for internet at work, assets and education (MP) ALL CASES Regression Statistics Multiple R 0.3667 R Square 0.1345 Coefficients Standard Error t Stat Intercept 1.8657 0.0944 19.7575 LN(W) 0.2196 0.0265 8.3003 LIWhr -0.0319 0.0017 -18.9558 asset value 0.0000 0.0000 -3.1219 education 0.0438 0.0178 2.4646 WIFI Regression Statistics Multiple R 0.3571 R Square 0.1275 Coefficients Standard Error t Stat Intercept 1.7174 0.1267 13.5495 LN(W) 0.2408 0.0340 7.0755 LIWhr -0.0298 0.0021 -14.0228 asset value 0.0000 0.0000 -3.0396 education 0.0502 0.0230 2.1797 NOWIFI Regression Statistics Multiple R 0.3887 R Square 0.1511 Coefficients Standard Error t Stat Intercept 2.0250 0.1453 13.9360 LN(W) 0.1985 0.0428 4.6438 LIWhr -0.0355 0.0028 -12.8275 asset value 0.0000 0.0000 -1.1859 education 0.0414 0.0282 1.4664 Table B.6: ln(W) regression results controlling for Internet at work, assest and education using mid points 83 Appendix C Appendix to Chapter Four Algorithm:Equilibria Input: fw;W;SNR; ;";Fg Output: design d, quality T(d), quantity q, price p, profit , consumer surplus CS for each SNR in the set f0;10;20;30;40;50;60;70;80;90g for each " in the set f0:2;0:4;0:6;0:8;1;1:2;1:4;1:6;1:8;2g for each in the set f0:1;0:3;0:5;0:7;0:9;1:1;1:3;1:5;1:7;1:9g while i >F for i:1,2,...nmax calculate design d i ,quality T i (d i ), quantity q i , price p i , profit i ,Consumer Surplus CS if i =F stop record output terminate Table C.1: Simulation Algorithm in Pseudocode Parameter Value W Bandwidth of a whitespace 6000kHz w Number of whitespaces 10 F Fixed cost of access 10 −6 under commons; i under licensing SNR Signal to Noise Ratio(dB) f0;10;20;30;40;50;60;70;80;90g " Interference elasticity f0:2;0:4;0:6;0:8;1;1:2;1:4;1:6;1:8;2g Substitutability f0:1;0:3;0:5;0:7;0:9;1:1;1:3;1:5;1:7;1:9g Table C.2: Model Parameters 84 Figure C.1: Consumer Surplus at SNR=0dB (Licensing in gray, Commons in color) 85 Figure C.2: Consumer Surplus at SNR=10dB (Licensing in gray, Commons in color) 86 Figure C.3: Consumer Surplus at SNR=20dB (Licensing in gray, Commons in color) 87 Figure C.4: Consumer Surplus at SNR=30dB (Licensing in gray, Commons in color) 88 Figure C.5: Consumer Surplus at SNR=40dB (Licensing in gray, Commons in color) 89 Figure C.6: Consumer Surplus at SNR=50dB (Licensing in gray, Commons in color) 90 Figure C.7: Consumer Surplus at SNR=60dB (Licensing in gray, Commons in color) 91 Figure C.8: Consumer Surplus at SNR=70dB (Licensing in gray, Commons in color) 92 Figure C.9: Consumer Surplus at SNR=80dB (Licensing in gray, Commons in color) 93 Figure C.10: Consumer Surplus at SNR=90dB (Licensing in gray, Commons in color) 94 Figure C.11: Regions where Consumer Suplus is greater under Commons for all SNRs 95 Figure C.12: The boundary of welfare dominance 96 Consumer Consumer Surplus 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 Surplus 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 1.9 67325 52848 99520 113818 227799 336524 480466 960001 1.8E+6 4.0E+6 1.9 7.7E+6 7.7E+6 7.7E+6 7.7E+6 7.7E+6 7.7E+6 7.7E+6 7.7E+6 7.7E+6 7.7E+6 1.7 30906 58696 111007 126530 177739 291106 535676 966121 1.9E+6 4.2E+6 1.7 8.4E+6 8.4E+6 8.4E+6 8.4E+6 8.4E+6 8.4E+6 8.4E+6 8.4E+6 8.4E+6 8.4E+6 1.5 34763 66169 70453 93758 200908 262803 517507 910476 2.0E+6 4.5E+6 1.5 9.1E+6 9.1E+6 9.1E+6 9.1E+6 9.1E+6 9.1E+6 9.1E+6 9.1E+6 9.1E+6 9.1E+6 1.3 39772 39473 80924 107520 169118 246810 452395 966699 2.0E+6 5.0E+6 1.3 10.0E+6 10.0E+6 10.0E+6 10.0E+6 10.0E+6 10.0E+6 10.0E+6 10.0E+6 10.0E+6 10.0E+6 1.1 22374 46263 58080 87320 150845 241534 472006 919286 2.1E+6 5.6E+6 1.1 11.1E+6 11.1E+6 11.1E+6 11.1E+6 11.1E+6 11.1E+6 11.1E+6 11.1E+6 11.1E+6 11.1E+6 0.9 27006 32003 45918 76353 143055 248286 462105 934677 2.3E+6 6.4E+6 0.9 12.4E+6 12.4E+6 12.4E+6 12.4E+6 12.4E+6 12.4E+6 12.4E+6 12.4E+6 12.4E+6 12.4E+6 0.7 18223 24931 39891 72442 118290 205001 446888 975617 2.5E+6 7.5E+6 0.7 14.1E+6 14.1E+6 14.1E+6 14.1E+6 14.1E+6 14.1E+6 14.1E+6 14.1E+6 14.1E+6 14.1E+6 0.5 14373 22147 28612 59858 95991 198620 425037 1.0E+6 2.8E+6 9.5E+6 0.5 16.2E+6 16.2E+6 16.2E+6 16.2E+6 16.2E+6 16.2E+6 16.2E+6 16.2E+6 16.2E+6 16.2E+6 0.3 9049.3 12192 20907 42884 78238 169610 398068 1.1E+6 3.5E+6 13.4E+6 0.3 18.8E+6 18.8E+6 18.8E+6 18.8E+6 18.8E+6 18.8E+6 18.8E+6 18.8E+6 18.8E+6 18.8E+6 0.1 3111 5572.9 10714 23561 48855 125310 359070 1.2E+6 5.3E+6 28.3E+6 0.1 21.7E+6 21.7E+6 21.7E+6 21.7E+6 21.7E+6 21.7E+6 21.7E+6 21.7E+6 21.7E+6 21.7E+6 Prices Prices 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 1.9 0.0228 0.0206 0.0202 0.0172 0.0155 0.0138 0.0116 98.3E-4 79.7E-4 61.9E-4 1.9 0.0948 0.0948 0.0948 0.0948 0.0948 0.0948 0.0948 0.0948 0.0948 0.0948 1.7 0.0229 0.0225 0.0195 0.0189 0.0165 0.0145 0.012 98.6E-4 78.8E-4 61.2E-4 1.7 0.1036 0.1036 0.1036 0.1036 0.1036 0.1036 0.1036 0.1036 0.1036 0.1036 1.5 0.0253 0.0235 0.0216 0.0186 0.0165 0.0145 0.0122 0.0101 79.1E-4 60.2E-4 1.5 0.1143 0.1143 0.1143 0.1143 0.1143 0.1143 0.1143 0.1143 0.1143 0.1143 1.3 0.0279 0.0248 0.0223 0.0205 0.0177 0.0151 0.0126 0.0102 80.1E-4 59.6E-4 1.3 0.1274 0.1274 0.1274 0.1274 0.1274 0.1274 0.1274 0.1274 0.1274 0.1274 1.1 0.029 0.0273 0.025 0.022 0.019 0.0161 0.0132 0.0105 80.9E-4 59.3E-4 1.1 0.1439 0.1439 0.1439 0.1439 0.1439 0.1439 0.1439 0.1439 0.1439 0.1439 0.9 0.0334 0.0303 0.0266 0.0238 0.0201 0.0169 0.0139 0.0109 82.5E-4 59.3E-4 0.9 0.1653 0.1653 0.1653 0.1653 0.1653 0.1653 0.1653 0.1653 0.1653 0.1653 0.7 0.0382 0.0337 0.0302 0.0263 0.0225 0.0185 0.0148 0.0116 85.4E-4 59.7E-4 0.7 0.1942 0.1942 0.1942 0.1942 0.1942 0.1942 0.1942 0.1942 0.1942 0.1942 0.5 0.046 0.0408 0.0342 0.0303 0.0255 0.0208 0.0165 0.0125 89.5E-4 60.7E-4 0.5 0.2353 0.2353 0.2353 0.2353 0.2353 0.2353 0.2353 0.2353 0.2353 0.2353 0.3 0.0593 0.0504 0.0438 0.0382 0.0314 0.0252 0.0194 0.0141 97.3E-4 62.8E-4 0.3 0.2985 0.2985 0.2985 0.2985 0.2985 0.2985 0.2985 0.2985 0.2985 0.2985 0.1 0.1005 0.0873 0.0743 0.0612 0.0489 0.0378 0.0274 0.0188 0.0119 69.1E-4 0.1 0.4082 0.4082 0.4082 0.4082 0.4082 0.4082 0.4082 0.4082 0.4082 0.4082 Quantities Quantities 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 1.9 137795 107920 203145 231627 462750 682488 972164 1.9E+6 3.6E+6 8.0E+6 1.9 1.7E+6 1.7E+6 1.7E+6 1.7E+6 1.7E+6 1.7E+6 1.7E+6 1.7E+6 1.7E+6 1.7E+6 1.7 63261 120090 226428 257926 361438 590754 1.1E+6 2.0E+6 3.9E+6 8.5E+6 1.7 1.9E+6 1.9E+6 1.9E+6 1.9E+6 1.9E+6 1.9E+6 1.9E+6 1.9E+6 1.9E+6 1.9E+6 1.5 71332 135526 144020 191070 408569 533365 1.0E+6 1.8E+6 4.0E+6 9.1E+6 1.5 2.1E+6 2.1E+6 2.1E+6 2.1E+6 2.1E+6 2.1E+6 2.1E+6 2.1E+6 2.1E+6 2.1E+6 1.3 81829 80953 165534 219538 344319 501173 916337 2.0E+6 4.1E+6 10.1E+6 1.3 2.3E+6 2.3E+6 2.3E+6 2.3E+6 2.3E+6 2.3E+6 2.3E+6 2.3E+6 2.3E+6 2.3E+6 1.1 46085 95122 119143 178569 307537 490994 956650 1.9E+6 4.3E+6 11.2E+6 1.1 2.6E+6 2.6E+6 2.6E+6 2.6E+6 2.6E+6 2.6E+6 2.6E+6 2.6E+6 2.6E+6 2.6E+6 0.9 55881 66002 94349 156426 291984 505111 937190 1.9E+6 4.7E+6 12.8E+6 0.9 3.0E+6 3.0E+6 3.0E+6 3.0E+6 3.0E+6 3.0E+6 3.0E+6 3.0E+6 3.0E+6 3.0E+6 0.7 37895 51603 82263 148794 242038 417717 907226 2.0E+6 5.1E+6 15.2E+6 0.7 3.5E+6 3.5E+6 3.5E+6 3.5E+6 3.5E+6 3.5E+6 3.5E+6 3.5E+6 3.5E+6 3.5E+6 0.5 30132 46180 59249 123456 197002 405692 864301 2.1E+6 5.7E+6 19.1E+6 0.5 4.2E+6 4.2E+6 4.2E+6 4.2E+6 4.2E+6 4.2E+6 4.2E+6 4.2E+6 4.2E+6 4.2E+6 0.3 19240 25677 43727 89178 161548 347996 811854 2.2E+6 7.0E+6 26.9E+6 0.3 5.4E+6 5.4E+6 5.4E+6 5.4E+6 5.4E+6 5.4E+6 5.4E+6 5.4E+6 5.4E+6 5.4E+6 0.1 6917.2 12212 23150 50192 102734 260467 738398 2.5E+6 10.7E+6 57.1E+6 0.1 7.3E+6 7.3E+6 7.3E+6 7.3E+6 7.3E+6 7.3E+6 7.3E+6 7.3E+6 7.3E+6 7.3E+6 Qualities Qualities 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 1.9 517.92 457.8 627.96 669.48 945.37 1147.1 1367.4 1929.4 2640.2 3915.5 1.9 5996.5 5996.5 5996.5 5996.5 5996.5 5996.5 5996.5 5996.5 5996.5 5996.5 1.7 332.35 457.79 627.5 669.48 791.4 1010.6 1367.1 1831.8 2572.6 3807.4 1.7 5996.5 5996.5 5996.5 5996.5 5996.5 5996.5 5996.5 5996.5 5996.5 5996.5 1.5 332.41 457.66 471.2 541.69 791.06 902.7 1263.3 1671.8 2455.1 3717 1.5 5996.5 5996.5 5996.5 5996.5 5996.5 5996.5 5996.5 5996.5 5996.5 5996.5 1.3 332.49 329.97 471.03 541.77 677.16 815.51 1100.8 1604.6 2310.9 3639.5 1.3 5996.6 5996.6 5996.6 5996.6 5996.6 5996.6 5996.6 5996.6 5996.6 5996.6 1.1 230.05 330.07 368.77 450.44 589.81 743.67 1035.8 1440.7 2193.6 3531.5 1.1 5996.7 5996.7 5996.7 5996.7 5996.7 5996.7 5996.7 5996.7 5996.7 5996.7 0.9 230.31 249.65 297.61 382.32 520.81 683.25 928.43 1315.4 2066.5 3410.6 0.9 5996.8 5996.8 5996.8 5996.8 5996.8 5996.8 5996.8 5996.8 5996.8 5996.8 0.7 168.26 195.58 246.17 329.97 419.49 549.17 806.83 1186.9 1896.4 3276.9 0.7 5997 5997 5997 5997 5997 5997 5997 5997 5997 5997 0.5 128 157.69 177.51 255.31 321.09 458.83 667.06 1026.2 1708 3102.8 0.5 5997.2 5997.2 5997.2 5997.2 5997.2 5997.2 5997.2 5997.2 5997.2 5997.2 0.3 80.524 92.167 119.47 169.68 226.83 330.92 502.6 817.82 1463 2859.7 0.3 5997.5 5997.5 5997.5 5997.5 5997.5 5997.5 5997.5 5997.5 5997.5 5997.5 0.1 29.331 38.365 52.034 75.491 106.57 167.69 279.31 513.36 1045.4 2405.2 0.1 5997.8 5997.8 5997.8 5997.8 5997.8 5997.8 5997.8 5997.8 5997.8 5997.8 Varieties Varieties 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 1.9 45 50 51 60 67 75 90 106 131 169 1.9 10 10 10 10 10 10 10 10 10 10 1.7 50 51 59 61 70 80 97 118 148 191 1.7 10 10 10 10 10 10 10 10 10 10 1.5 51 55 60 70 79 90 108 130 167 220 1.5 10 10 10 10 10 10 10 10 10 10 1.3 53 60 67 73 85 100 120 149 190 256 1.3 10 10 10 10 10 10 10 10 10 10 1.1 60 64 70 80 93 110 135 170 222 304 1.1 10 10 10 10 10 10 10 10 10 10 0.9 63 70 80 90 107 128 157 200 266 371 0.9 10 10 10 10 10 10 10 10 10 10 0.7 70 80 90 104 122 150 188 242 330 474 0.7 10 10 10 10 10 10 10 10 10 10 0.5 80 91 110 125 150 185 236 314 440 652 0.5 10 10 10 10 10 10 10 10 10 10 0.3 100 120 140 162 200 252 332 460 673 1050 0.3 10 10 10 10 10 10 10 10 10 10 0.1 160 190 230 288 370 490 690 1027 1644 2855 0.1 10 10 10 10 10 10 10 10 10 10 Colors indicate whether the value is or than its counterpart under the alternative management regime LOWER HIGHER Cournot Equilibria when native SNR is 0dB Interference Elasticity Interference Elasticity LICENSING COMMONS Interference Elasticity Interference Elasticity Substitutability Interference Elasticity Interference Elasticity Interference Elasticity Interference Elasticity Interference Elasticity Substitutability Substitutability Substitutability Substitutability Substitutability Substitutability Substitutability Substitutability Substitutability Interference Elasticity Figure C.13: Cournot Equilibria when Native SNR is 0dB) 97 Consumer Consumer Surplus 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 Surplus 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 1.9 398826 584124 960001 1.4E+6 2.4E+6 4.1E+6 7.7E+6 15.2E+6 30.2E+6 60.0E+6 1.9 92.4E+6 92.4E+6 92.4E+6 92.4E+6 92.4E+6 92.4E+6 92.4E+6 92.4E+6 92.4E+6 92.4E+6 1.7 284982 457869 811574 1.3E+6 2.3E+6 3.9E+6 7.7E+6 15.5E+6 32.0E+6 65.3E+6 1.7 1.0E+8 1.0E+8 1.0E+8 1.0E+8 1.0E+8 1.0E+8 1.0E+8 1.0E+8 1.0E+8 1.0E+8 1.5 321611 517507 710136 1.2E+6 2.0E+6 3.7E+6 7.5E+6 15.8E+6 34.2E+6 72.0E+6 1.5 1.1E+8 1.1E+8 1.1E+8 1.1E+8 1.1E+8 1.1E+8 1.1E+8 1.1E+8 1.1E+8 1.1E+8 1.3 246810 430493 643910 1.2E+6 1.8E+6 3.7E+6 7.5E+6 16.5E+6 36.6E+6 80.4E+6 1.3 1.2E+8 1.2E+8 1.2E+8 1.2E+8 1.2E+8 1.2E+8 1.2E+8 1.2E+8 1.2E+8 1.2E+8 1.1 200667 375962 607319 982949 1.7E+6 3.5E+6 7.6E+6 16.9E+6 39.9E+6 91.8E+6 1.1 1.3E+8 1.3E+8 1.3E+8 1.3E+8 1.3E+8 1.3E+8 1.3E+8 1.3E+8 1.3E+8 1.3E+8 0.9 173427 267596 492365 896199 1.6E+6 3.3E+6 7.4E+6 17.8E+6 44.4E+6 1.1E+8 0.9 1.5E+8 1.5E+8 1.5E+8 1.5E+8 1.5E+8 1.5E+8 1.5E+8 1.5E+8 1.5E+8 1.5E+8 0.7 161400 268346 438510 778934 1.5E+6 3.1E+6 7.5E+6 19.0E+6 50.8E+6 1.3E+8 0.7 1.7E+8 1.7E+8 1.7E+8 1.7E+8 1.7E+8 1.7E+8 1.7E+8 1.7E+8 1.7E+8 1.7E+8 0.5 97635 194745 325725 625529 1.3E+6 3.0E+6 7.5E+6 20.7E+6 60.5E+6 1.7E+8 0.5 1.9E+8 1.9E+8 1.9E+8 1.9E+8 1.9E+8 1.9E+8 1.9E+8 1.9E+8 1.9E+8 1.9E+8 0.3 79574 132001 249614 470219 1.1E+6 2.7E+6 7.5E+6 23.3E+6 79.1E+6 2.6E+8 0.3 2.3E+8 2.3E+8 2.3E+8 2.3E+8 2.3E+8 2.3E+8 2.3E+8 2.3E+8 2.3E+8 2.3E+8 0.1 29378 58981 117168 285688 726291 2.1E+6 7.4E+6 29.9E+6 1.4E+8 6.2E+8 0.1 2.6E+8 2.6E+8 2.6E+8 2.6E+8 2.6E+8 2.6E+8 2.6E+8 2.6E+8 2.6E+8 2.6E+8 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 1.9 0.013 0.0116 98.3E-4 86.9E-4 73.0E-4 61.2E-4 49.9E-4 39.9E-4 31.8E-4 25.3E-4 1.9 0.0948 0.0948 0.0948 0.0948 0.0948 0.0948 0.0948 0.0948 0.0948 0.0948 1.7 0.0129 0.0116 0.0104 89.6E-4 74.3E-4 61.5E-4 50.2E-4 39.8E-4 31.3E-4 24.8E-4 1.7 0.1036 0.1036 0.1036 0.1036 0.1036 0.1036 0.1036 0.1036 0.1036 0.1036 1.5 0.0142 0.0122 0.011 93.5E-4 77.7E-4 63.0E-4 51.0E-4 40.1E-4 31.1E-4 24.3E-4 1.5 0.1143 0.1143 0.1143 0.1143 0.1143 0.1143 0.1143 0.1143 0.1143 0.1143 1.3 0.0151 0.0131 0.0116 96.1E-4 80.1E-4 66.0E-4 52.1E-4 40.4E-4 31.1E-4 24.0E-4 1.3 0.1274 0.1274 0.1274 0.1274 0.1274 0.1274 0.1274 0.1274 0.1274 0.1274 1.1 0.0161 0.0141 0.0122 0.0104 85.5E-4 69.0E-4 53.7E-4 41.1E-4 31.2E-4 23.7E-4 1.1 0.1439 0.1439 0.1439 0.1439 0.1439 0.1439 0.1439 0.1439 0.1439 0.1439 0.9 0.018 0.0155 0.0136 0.0112 91.3E-4 73.2E-4 56.5E-4 42.4E-4 31.4E-4 23.5E-4 0.9 0.1653 0.1653 0.1653 0.1653 0.1653 0.1653 0.1653 0.1653 0.1653 0.1653 0.7 0.0204 0.0175 0.0149 0.0124 0.01 78.3E-4 59.9E-4 44.2E-4 32.0E-4 23.4E-4 0.7 0.1942 0.1942 0.1942 0.1942 0.1942 0.1942 0.1942 0.1942 0.1942 0.1942 0.5 0.024 0.0211 0.0176 0.0144 0.0115 87.9E-4 65.5E-4 47.0E-4 33.0E-4 23.4E-4 0.5 0.2353 0.2353 0.2353 0.2353 0.2353 0.2353 0.2353 0.2353 0.2353 0.2353 0.3 0.0317 0.0272 0.0223 0.0179 0.0141 0.0105 75.8E-4 52.4E-4 35.1E-4 23.8E-4 0.3 0.2985 0.2985 0.2985 0.2985 0.2985 0.2985 0.2985 0.2985 0.2985 0.2985 0.1 0.0573 0.0472 0.0378 0.0295 0.0222 0.0158 0.0107 67.9E-4 41.2E-4 25.2E-4 0.1 0.4082 0.4082 0.4082 0.4082 0.4082 0.4082 0.4082 0.4082 0.4082 0.4082 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 1.9 808140 1.2E+6 1.9E+6 2.8E+6 4.8E+6 8.3E+6 15.4E+6 30.4E+6 60.6E+6 1.2E+8 1.9 20.4E+6 20.4E+6 20.4E+6 20.4E+6 20.4E+6 20.4E+6 20.4E+6 20.4E+6 20.4E+6 20.4E+6 1.7 577400 926492 1.6E+6 2.6E+6 4.7E+6 7.8E+6 15.5E+6 31.1E+6 64.2E+6 1.3E+8 1.7 22.3E+6 22.3E+6 22.3E+6 22.3E+6 22.3E+6 22.3E+6 22.3E+6 22.3E+6 22.3E+6 22.3E+6 1.5 652510 1.0E+6 1.4E+6 2.4E+6 4.1E+6 7.5E+6 15.1E+6 31.7E+6 68.6E+6 1.4E+8 1.5 24.6E+6 24.6E+6 24.6E+6 24.6E+6 24.6E+6 24.6E+6 24.6E+6 24.6E+6 24.6E+6 24.6E+6 1.3 501173 872450 1.3E+6 2.3E+6 3.7E+6 7.4E+6 15.2E+6 33.2E+6 73.5E+6 1.6E+8 1.3 27.4E+6 27.4E+6 27.4E+6 27.4E+6 27.4E+6 27.4E+6 27.4E+6 27.4E+6 27.4E+6 27.4E+6 1.1 407919 762704 1.2E+6 2.0E+6 3.5E+6 7.1E+6 15.2E+6 33.9E+6 80.0E+6 1.8E+8 1.1 31.0E+6 31.0E+6 31.0E+6 31.0E+6 31.0E+6 31.0E+6 31.0E+6 31.0E+6 31.0E+6 31.0E+6 0.9 353212 543613 998303 1.8E+6 3.2E+6 6.7E+6 15.0E+6 35.7E+6 89.1E+6 2.2E+8 0.9 35.6E+6 35.6E+6 35.6E+6 35.6E+6 35.6E+6 35.6E+6 35.6E+6 35.6E+6 35.6E+6 35.6E+6 0.7 329540 546226 890288 1.6E+6 2.9E+6 6.3E+6 15.0E+6 38.1E+6 1.0E+8 2.6E+8 0.7 41.8E+6 41.8E+6 41.8E+6 41.8E+6 41.8E+6 41.8E+6 41.8E+6 41.8E+6 41.8E+6 41.8E+6 0.5 200062 397866 663134 1.3E+6 2.6E+6 6.1E+6 15.0E+6 41.5E+6 1.2E+8 3.4E+8 0.5 50.7E+6 50.7E+6 50.7E+6 50.7E+6 50.7E+6 50.7E+6 50.7E+6 50.7E+6 50.7E+6 50.7E+6 0.3 164358 271377 510639 957584 2.2E+6 5.4E+6 15.1E+6 46.9E+6 1.6E+8 5.2E+8 0.3 64.3E+6 64.3E+6 64.3E+6 64.3E+6 64.3E+6 64.3E+6 64.3E+6 64.3E+6 64.3E+6 64.3E+6 0.1 62329 123801 243544 588770 1.5E+6 4.4E+6 14.9E+6 60.3E+6 2.8E+8 12.4E+8 0.1 87.9E+6 87.9E+6 87.9E+6 87.9E+6 87.9E+6 87.9E+6 87.9E+6 87.9E+6 87.9E+6 87.9E+6 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 1.9 1247.7 1507.7 1929.4 2315.9 3031.6 3991.4 5428.8 7618.9 10750 15133 1.9 20753 20753 20753 20753 20753 20753 20753 20753 20753 20753 1.7 998.17 1263.5 1679.9 2096.5 2824.8 3646.6 5139.8 7290.3 10469 14937 1.7 20753 20753 20753 20753 20753 20753 20753 20753 20753 20753 1.5 998.25 1263.3 1477.8 1909.9 2480.4 3357.2 4768.6 6913.4 10162 14739 1.5 20753 20753 20753 20753 20753 20753 20753 20753 20753 20753 1.3 815.51 1074.6 1311.7 1749.5 2205.1 3110.7 4455.1 6590.5 9795.7 14498 1.3 20753 20753 20753 20753 20753 20753 20753 20753 20753 20753 1.1 677.84 925.49 1173.4 1489.5 1980.8 2802.4 4104.8 6121.1 9403.3 14256 1.1 20753 20753 20753 20753 20753 20753 20753 20753 20753 20753 0.9 571.85 708.02 958.03 1288.5 1713 2475.7 3684.5 5690.9 8976.3 13959 0.9 20753 20753 20753 20753 20753 20753 20753 20753 20753 20753 0.7 488.6 627.45 799.32 1061.8 1444.8 2107.3 3256.5 5186.6 8472 13609 0.7 20754 20754 20754 20754 20754 20754 20754 20754 20754 20754 0.5 323.12 454.47 584.91 806.94 1161.9 1758.6 2759 4576.6 7816 13145 0.5 20754 20754 20754 20754 20754 20754 20754 20754 20754 20754 0.3 228.86 292.78 399.75 545.08 822.1 1281.1 2144.2 3770.7 6924.1 12486 0.3 20754 20754 20754 20754 20754 20754 20754 20754 20754 20754 0.1 83.768 116.77 162.15 249.95 394.05 670.83 1233.1 2471.7 5282.6 11141 0.1 20754 20754 20754 20754 20754 20754 20754 20754 20754 20754 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 1.9 80 90 106 120 143 171 210 263 330 415 1.9 10 10 10 10 10 10 10 10 10 10 1.7 90 100 112 130 157 190 233 294 374 473 1.7 10 10 10 10 10 10 10 10 10 10 1.5 92 108 120 141 170 210 260 331 427 547 1.5 10 10 10 10 10 10 10 10 10 10 1.3 100 115 131 158 190 231 293 379 493 640 1.3 10 10 10 10 10 10 10 10 10 10 1.1 110 126 147 172 210 261 336 440 581 766 1.1 10 10 10 10 10 10 10 10 10 10 0.9 120 140 160 195 240 300 390 521 704 944 0.9 10 10 10 10 10 10 10 10 10 10 0.7 135 159 187 225 280 360 472 642 889 1218 0.7 10 10 10 10 10 10 10 10 10 10 0.5 160 183 220 270 341 448 604 844 1204 1700 0.5 10 10 10 10 10 10 10 10 10 10 0.3 198 233 286 360 460 620 867 1260 1885 2789 0.3 10 10 10 10 10 10 10 10 10 10 0.1 310 385 490 638 862 1226 1834 2907 4810 7893 0.1 10 10 10 10 10 10 10 10 10 10 Colors indicate whether the value is or than its counterpart under the alternative management regime LOWER HIGHER Substitutability Prices Substitutability Quantities Substitutability Qualities Substitutability Varieties Substitutability Interference Elasticity Prices Interference Elasticity Interference Elasticity Quantities Interference Elasticity Substitutability Interference Elasticity Qualities Interference Elasticity Varieties Interference Elasticity Substitutability Substitutability Substitutability Substitutability Interference Elasticity Interference Elasticity Interference Elasticity LICENSING COMMONS Cournot Equilibria when native SNR is 10dB Figure C.14: Cournot Equilibria when Native SNR is 10dB 98 Consumer Consumer Surplus 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 Surplus 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 1.9 2.6E+6 4.4E+6 7.1E+6 11.7E+6 20.1E+6 35.6E+6 63.1E+6 1.1E+8 1.8E+8 2.8E+8 1.9 3.4E+8 3.4E+8 3.4E+8 3.4E+8 3.4E+8 3.4E+8 3.4E+8 3.4E+8 3.4E+8 3.4E+8 1.7 2.3E+6 3.5E+6 6.3E+6 11.1E+6 19.3E+6 35.2E+6 64.8E+6 1.1E+8 1.9E+8 3.1E+8 1.7 3.7E+8 3.7E+8 3.7E+8 3.7E+8 3.7E+8 3.7E+8 3.7E+8 3.7E+8 3.7E+8 3.7E+8 1.5 2.1E+6 3.4E+6 5.7E+6 10.1E+6 18.9E+6 35.7E+6 66.9E+6 1.2E+8 2.1E+8 3.4E+8 1.5 4.0E+8 4.0E+8 4.0E+8 4.0E+8 4.0E+8 4.0E+8 4.0E+8 4.0E+8 4.0E+8 4.0E+8 1.3 1.7E+6 3.0E+6 5.3E+6 9.4E+6 18.3E+6 35.2E+6 68.9E+6 1.3E+8 2.3E+8 3.9E+8 1.3 4.4E+8 4.4E+8 4.4E+8 4.4E+8 4.4E+8 4.4E+8 4.4E+8 4.4E+8 4.4E+8 4.4E+8 1.1 1.6E+6 2.7E+6 4.7E+6 9.2E+6 17.7E+6 35.3E+6 72.7E+6 1.4E+8 2.6E+8 4.5E+8 1.1 4.9E+8 4.9E+8 4.9E+8 4.9E+8 4.9E+8 4.9E+8 4.9E+8 4.9E+8 4.9E+8 4.9E+8 0.9 1.4E+6 2.2E+6 4.4E+6 8.3E+6 16.7E+6 35.6E+6 75.9E+6 1.6E+8 3.1E+8 5.4E+8 0.9 5.5E+8 5.5E+8 5.5E+8 5.5E+8 5.5E+8 5.5E+8 5.5E+8 5.5E+8 5.5E+8 5.5E+8 0.7 1.1E+6 2.0E+6 3.8E+6 7.4E+6 15.9E+6 35.3E+6 81.2E+6 1.8E+8 3.7E+8 6.8E+8 0.7 6.2E+8 6.2E+8 6.2E+8 6.2E+8 6.2E+8 6.2E+8 6.2E+8 6.2E+8 6.2E+8 6.2E+8 0.5 781930 1.5E+6 3.0E+6 6.5E+6 14.6E+6 35.4E+6 87.3E+6 2.1E+8 4.7E+8 9.1E+8 0.5 7.2E+8 7.2E+8 7.2E+8 7.2E+8 7.2E+8 7.2E+8 7.2E+8 7.2E+8 7.2E+8 7.2E+8 0.3 570591 1.1E+6 2.2E+6 5.1E+6 12.8E+6 34.3E+6 97.0E+6 2.7E+8 6.7E+8 14.3E+8 0.3 8.4E+8 8.4E+8 8.4E+8 8.4E+8 8.4E+8 8.4E+8 8.4E+8 8.4E+8 8.4E+8 8.4E+8 0.1 239131 512690 1.2E+6 3.2E+6 9.3E+6 31.1E+6 1.1E+8 4.3E+8 14.5E+8 37.9E+8 0.1 9.6E+8 9.6E+8 9.6E+8 9.6E+8 9.6E+8 9.6E+8 9.6E+8 9.6E+8 9.6E+8 9.6E+8 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 1.9 69.7E-4 59.8E-4 51.1E-4 43.5E-4 36.2E-4 30.0E-4 24.9E-4 20.8E-4 17.6E-4 15.2E-4 1.9 0.0948 0.0948 0.0948 0.0948 0.0948 0.0948 0.0948 0.0948 0.0948 0.0948 1.7 72.9E-4 61.5E-4 53.1E-4 44.3E-4 36.6E-4 30.1E-4 24.9E-4 20.6E-4 17.3E-4 14.8E-4 1.7 0.1036 0.1036 0.1036 0.1036 0.1036 0.1036 0.1036 0.1036 0.1036 0.1036 1.5 76.8E-4 66.1E-4 55.2E-4 45.7E-4 37.8E-4 30.7E-4 24.9E-4 20.4E-4 17.0E-4 14.5E-4 1.5 0.1143 0.1143 0.1143 0.1143 0.1143 0.1143 0.1143 0.1143 0.1143 0.1143 1.3 80.1E-4 69.3E-4 58.5E-4 47.8E-4 39.0E-4 31.3E-4 25.1E-4 20.4E-4 16.8E-4 14.2E-4 1.3 0.1274 0.1274 0.1274 0.1274 0.1274 0.1274 0.1274 0.1274 0.1274 0.1274 1.1 88.4E-4 74.9E-4 62.1E-4 50.4E-4 40.7E-4 32.3E-4 25.5E-4 20.4E-4 16.7E-4 13.9E-4 1.1 0.1439 0.1439 0.1439 0.1439 0.1439 0.1439 0.1439 0.1439 0.1439 0.1439 0.9 96.9E-4 81.3E-4 67.0E-4 54.4E-4 43.3E-4 33.7E-4 26.3E-4 20.7E-4 16.6E-4 13.7E-4 0.9 0.1653 0.1653 0.1653 0.1653 0.1653 0.1653 0.1653 0.1653 0.1653 0.1653 0.7 0.011 91.7E-4 74.9E-4 60.2E-4 46.8E-4 36.0E-4 27.4E-4 21.1E-4 16.7E-4 13.6E-4 0.7 0.1942 0.1942 0.1942 0.1942 0.1942 0.1942 0.1942 0.1942 0.1942 0.1942 0.5 0.013 0.0109 87.5E-4 68.6E-4 52.7E-4 39.4E-4 29.3E-4 21.9E-4 16.8E-4 13.5E-4 0.5 0.2353 0.2353 0.2353 0.2353 0.2353 0.2353 0.2353 0.2353 0.2353 0.2353 0.3 0.0173 0.0141 0.0111 85.2E-4 63.7E-4 46.2E-4 32.8E-4 23.5E-4 17.4E-4 13.5E-4 0.3 0.2985 0.2985 0.2985 0.2985 0.2985 0.2985 0.2985 0.2985 0.2985 0.2985 0.1 0.0312 0.0247 0.0189 0.014 99.0E-4 67.1E-4 43.8E-4 28.3E-4 19.0E-4 13.8E-4 0.1 0.4082 0.4082 0.4082 0.4082 0.4082 0.4082 0.4082 0.4082 0.4082 0.4082 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 1.9 5.3E+6 8.8E+6 14.2E+6 23.4E+6 40.3E+6 71.4E+6 1.3E+8 2.2E+8 3.6E+8 5.6E+8 1.9 75.6E+6 75.6E+6 75.6E+6 75.6E+6 75.6E+6 75.6E+6 75.6E+6 75.6E+6 75.6E+6 75.6E+6 1.7 4.7E+6 7.1E+6 12.6E+6 22.3E+6 38.7E+6 70.7E+6 1.3E+8 2.3E+8 3.9E+8 6.2E+8 1.7 82.7E+6 82.7E+6 82.7E+6 82.7E+6 82.7E+6 82.7E+6 82.7E+6 82.7E+6 82.7E+6 82.7E+6 1.5 4.3E+6 6.9E+6 11.4E+6 20.2E+6 38.0E+6 71.6E+6 1.3E+8 2.4E+8 4.2E+8 6.9E+8 1.5 91.2E+6 91.2E+6 91.2E+6 91.2E+6 91.2E+6 91.2E+6 91.2E+6 91.2E+6 91.2E+6 91.2E+6 1.3 3.4E+6 6.0E+6 10.7E+6 19.0E+6 36.8E+6 70.7E+6 1.4E+8 2.6E+8 4.7E+8 7.8E+8 1.3 1.0E+8 1.0E+8 1.0E+8 1.0E+8 1.0E+8 1.0E+8 1.0E+8 1.0E+8 1.0E+8 1.0E+8 1.1 3.3E+6 5.4E+6 9.5E+6 18.4E+6 35.5E+6 70.8E+6 1.5E+8 2.9E+8 5.3E+8 9.1E+8 1.1 1.1E+8 1.1E+8 1.1E+8 1.1E+8 1.1E+8 1.1E+8 1.1E+8 1.1E+8 1.1E+8 1.1E+8 0.9 2.8E+6 4.5E+6 8.9E+6 16.7E+6 33.5E+6 71.5E+6 1.5E+8 3.2E+8 6.1E+8 10.8E+8 0.9 1.3E+8 1.3E+8 1.3E+8 1.3E+8 1.3E+8 1.3E+8 1.3E+8 1.3E+8 1.3E+8 1.3E+8 0.7 2.3E+6 4.1E+6 7.6E+6 14.8E+6 32.0E+6 70.8E+6 1.6E+8 3.6E+8 7.3E+8 13.5E+8 0.7 1.5E+8 1.5E+8 1.5E+8 1.5E+8 1.5E+8 1.5E+8 1.5E+8 1.5E+8 1.5E+8 1.5E+8 0.5 1.6E+6 3.1E+6 6.0E+6 13.1E+6 29.3E+6 71.0E+6 1.8E+8 4.2E+8 9.3E+8 18.3E+8 0.5 1.9E+8 1.9E+8 1.9E+8 1.9E+8 1.9E+8 1.9E+8 1.9E+8 1.9E+8 1.9E+8 1.9E+8 0.3 1.2E+6 2.2E+6 4.5E+6 10.2E+6 25.8E+6 69.0E+6 1.9E+8 5.4E+8 13.5E+8 28.7E+8 0.3 2.4E+8 2.4E+8 2.4E+8 2.4E+8 2.4E+8 2.4E+8 2.4E+8 2.4E+8 2.4E+8 2.4E+8 0.1 493639 1.1E+6 2.4E+6 6.4E+6 18.8E+6 62.5E+6 2.3E+8 8.7E+8 29.0E+8 75.9E+8 0.1 3.3E+8 3.3E+8 3.3E+8 3.3E+8 3.3E+8 3.3E+8 3.3E+8 3.3E+8 3.3E+8 3.3E+8 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 1.9 3170.8 4108 5210.1 6685.4 8766 11666 15530 20384 26152 32706 1.9 39946 39946 39946 39946 39946 39946 39946 39946 39946 39946 1.7 2841.8 3491.2 4636.4 6175.7 8127.6 10982 14884 19769 25732 32476 1.7 39946 39946 39946 39946 39946 39946 39946 39946 39946 39946 1.5 2559.4 3233.2 4155.3 5524.3 7571.5 10381 14206 19157 25232 32217 1.5 39946 39946 39946 39946 39946 39946 39946 39946 39946 39946 1.3 2103.6 2795.2 3747.9 4980.5 6932.6 9606 13428 18497 24727 31944 1.3 39946 39946 39946 39946 39946 39946 39946 39946 39946 39946 1.1 1918.7 2440.3 3244.2 4520.7 6263.2 8847.1 12685 17768 24168 31630 1.1 39946 39946 39946 39946 39946 39946 39946 39946 39946 39946 0.9 1613.1 2022.2 2839.2 3896.7 5505.7 8042.3 11730 16889 23491 31268 0.9 39946 39946 39946 39946 39946 39946 39946 39946 39946 39946 0.7 1272.5 1703.4 2320.3 3236.1 4750.2 7058.8 10703 15863 22681 30828 0.7 39946 39946 39946 39946 39946 39946 39946 39946 39946 39946 0.5 900.41 1257.1 1752.4 2569.9 3843.5 5978.9 9381.2 14578 21650 30247 0.5 39946 39946 39946 39946 39946 39946 39946 39946 39946 39946 0.3 599.92 826.39 1172.5 1766.7 2797.5 4569.6 7666.5 12758 20123 29389 0.3 39947 39947 39947 39947 39947 39947 39947 39947 39947 39947 0.1 229.25 332.34 503.43 812.81 1384.8 2517.2 4823.5 9361.7 17050 27596 0.1 39947 39947 39947 39947 39947 39947 39947 39947 39947 39947 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 1.9 150 175 205 241 290 350 422 505 596 692 1.9 10 10 10 10 10 10 10 10 10 10 1.7 160 190 220 264 320 390 472 570 679 793 1.7 10 10 10 10 10 10 10 10 10 10 1.5 172 200 240 290 351 433 534 651 781 920 1.5 10 10 10 10 10 10 10 10 10 10 1.3 190 220 261 320 392 490 610 753 913 1082 1.3 10 10 10 10 10 10 10 10 10 10 1.1 203 240 290 358 444 560 709 887 1088 1301 1.1 10 10 10 10 10 10 10 10 10 10 0.9 226 270 328 405 510 655 841 1071 1334 1613 0.9 10 10 10 10 10 10 10 10 10 10 0.7 254 307 377 470 606 790 1039 1350 1711 2098 0.7 10 10 10 10 10 10 10 10 10 10 0.5 300 360 450 576 752 1008 1360 1821 2368 2958 0.5 10 10 10 10 10 10 10 10 10 10 0.3 373 461 590 770 1034 1431 2020 2830 3830 4926 0.3 10 10 10 10 10 10 10 10 10 10 0.1 603 771 1020 1392 1981 2942 4530 7032 10472 14423 0.1 10 10 10 10 10 10 10 10 10 10 Colors indicate whether the value is or than its counterpart under the alternative management regime LOWER HIGHER Substitutability Prices Substitutability Quantities Substitutability Qualities Substitutability Varieties Substitutability Interference Elasticity Prices Interference Elasticity Interference Elasticity Quantities Interference Elasticity Substitutability Interference Elasticity Qualities Interference Elasticity Varieties Interference Elasticity Substitutability Substitutability Substitutability Substitutability Interference Elasticity Interference Elasticity Interference Elasticity LICENSING COMMONS Cournot Equilibria when native SNR is 20dB Figure C.15: Cournot Equilibria when Native SNR is 20dB 99 Consumer Consumer Surplus 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 Surplus 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 1.9 16.1E+6 24.5E+6 40.7E+6 64.7E+6 1.1E+8 1.7E+8 2.5E+8 3.7E+8 5.2E+8 7.1E+8 1.9 7.7E+8 7.7E+8 7.7E+8 7.7E+8 7.7E+8 7.7E+8 7.7E+8 7.7E+8 7.7E+8 7.7E+8 1.7 14.7E+6 23.9E+6 37.9E+6 64.3E+6 1.1E+8 1.7E+8 2.7E+8 4.0E+8 5.7E+8 7.8E+8 1.7 8.3E+8 8.3E+8 8.3E+8 8.3E+8 8.3E+8 8.3E+8 8.3E+8 8.3E+8 8.3E+8 8.3E+8 1.5 12.5E+6 20.8E+6 36.1E+6 63.5E+6 1.1E+8 1.8E+8 2.8E+8 4.3E+8 6.3E+8 8.8E+8 1.5 9.0E+8 9.0E+8 9.0E+8 9.0E+8 9.0E+8 9.0E+8 9.0E+8 9.0E+8 9.0E+8 9.0E+8 1.3 11.0E+6 19.9E+6 35.2E+6 62.6E+6 1.1E+8 1.9E+8 3.1E+8 4.8E+8 7.1E+8 10.0E+8 1.3 9.9E+8 9.9E+8 9.9E+8 9.9E+8 9.9E+8 9.9E+8 9.9E+8 9.9E+8 9.9E+8 9.9E+8 1.1 10.0E+6 17.4E+6 32.9E+6 60.9E+6 1.1E+8 2.0E+8 3.3E+8 5.4E+8 8.1E+8 11.7E+8 1.1 11.0E+8 11.0E+8 11.0E+8 11.0E+8 11.0E+8 11.0E+8 11.0E+8 11.0E+8 11.0E+8 11.0E+8 0.9 8.9E+6 16.1E+6 30.1E+6 58.2E+6 1.1E+8 2.1E+8 3.7E+8 6.2E+8 9.6E+8 14.1E+8 0.9 12.3E+8 12.3E+8 12.3E+8 12.3E+8 12.3E+8 12.3E+8 12.3E+8 12.3E+8 12.3E+8 12.3E+8 0.7 7.3E+6 13.8E+6 27.8E+6 55.2E+6 1.1E+8 2.3E+8 4.2E+8 7.3E+8 11.8E+8 17.8E+8 0.7 14.0E+8 14.0E+8 14.0E+8 14.0E+8 14.0E+8 14.0E+8 14.0E+8 14.0E+8 14.0E+8 14.0E+8 0.5 5.9E+6 11.6E+6 24.1E+6 52.2E+6 1.1E+8 2.5E+8 5.0E+8 9.2E+8 15.6E+8 24.3E+8 0.5 16.1E+8 16.1E+8 16.1E+8 16.1E+8 16.1E+8 16.1E+8 16.1E+8 16.1E+8 16.1E+8 16.1E+8 0.3 3.9E+6 8.5E+6 19.0E+6 45.7E+6 1.1E+8 2.8E+8 6.4E+8 13.1E+8 23.8E+8 39.1E+8 0.3 18.7E+8 18.7E+8 18.7E+8 18.7E+8 18.7E+8 18.7E+8 18.7E+8 18.7E+8 18.7E+8 18.7E+8 0.1 1.8E+6 4.3E+6 11.1E+6 31.9E+6 99.9E+6 3.3E+8 10.1E+8 26.8E+8 58.4E+8 1.1E+10 0.1 21.6E+8 21.6E+8 21.6E+8 21.6E+8 21.6E+8 21.6E+8 21.6E+8 21.6E+8 21.6E+8 21.6E+8 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 1.9 39.1E-4 33.8E-4 28.8E-4 24.4E-4 21.0E-4 18.1E-4 15.8E-4 13.9E-4 12.4E-4 11.2E-4 1.9 0.0948 0.0948 0.0948 0.0948 0.0948 0.0948 0.0948 0.0948 0.0948 0.0948 1.7 40.5E-4 34.6E-4 29.3E-4 24.9E-4 21.0E-4 18.0E-4 15.6E-4 13.6E-4 12.1E-4 10.9E-4 1.7 0.1036 0.1036 0.1036 0.1036 0.1036 0.1036 0.1036 0.1036 0.1036 0.1036 1.5 42.8E-4 35.9E-4 30.2E-4 25.4E-4 21.3E-4 18.0E-4 15.5E-4 13.4E-4 11.9E-4 10.6E-4 1.5 0.1143 0.1143 0.1143 0.1143 0.1143 0.1143 0.1143 0.1143 0.1143 0.1143 1.3 45.0E-4 38.0E-4 31.5E-4 26.0E-4 21.6E-4 18.1E-4 15.4E-4 13.3E-4 11.7E-4 10.4E-4 1.3 0.1274 0.1274 0.1274 0.1274 0.1274 0.1274 0.1274 0.1274 0.1274 0.1274 1.1 48.8E-4 40.2E-4 33.1E-4 27.1E-4 22.2E-4 18.4E-4 15.4E-4 13.2E-4 11.5E-4 10.2E-4 1.1 0.1439 0.1439 0.1439 0.1439 0.1439 0.1439 0.1439 0.1439 0.1439 0.1439 0.9 53.2E-4 43.8E-4 35.6E-4 28.7E-4 23.1E-4 18.8E-4 15.6E-4 13.2E-4 11.4E-4 10.0E-4 0.9 0.1653 0.1653 0.1653 0.1653 0.1653 0.1653 0.1653 0.1653 0.1653 0.1653 0.7 60.2E-4 48.9E-4 39.0E-4 30.9E-4 24.5E-4 19.5E-4 15.9E-4 13.2E-4 11.3E-4 9.9E-4 0.7 0.1942 0.1942 0.1942 0.1942 0.1942 0.1942 0.1942 0.1942 0.1942 0.1942 0.5 70.7E-4 56.7E-4 44.7E-4 34.7E-4 26.7E-4 20.8E-4 16.5E-4 13.4E-4 11.3E-4 9.7E-4 0.5 0.2353 0.2353 0.2353 0.2353 0.2353 0.2353 0.2353 0.2353 0.2353 0.2353 0.3 92.3E-4 72.8E-4 55.9E-4 42.1E-4 31.2E-4 23.2E-4 17.6E-4 13.9E-4 11.4E-4 9.7E-4 0.3 0.2985 0.2985 0.2985 0.2985 0.2985 0.2985 0.2985 0.2985 0.2985 0.2985 0.1 0.0167 0.0127 93.5E-4 66.5E-4 45.8E-4 31.1E-4 21.4E-4 15.5E-4 12.0E-4 9.8E-4 0.1 0.4082 0.4082 0.4082 0.4082 0.4082 0.4082 0.4082 0.4082 0.4082 0.4082 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 1.9 32.2E+6 49.2E+6 81.7E+6 1.3E+8 2.1E+8 3.3E+8 5.0E+8 7.4E+8 10.4E+8 14.2E+8 1.9 1.7E+8 1.7E+8 1.7E+8 1.7E+8 1.7E+8 1.7E+8 1.7E+8 1.7E+8 1.7E+8 1.7E+8 1.7 29.5E+6 47.9E+6 76.0E+6 1.3E+8 2.2E+8 3.4E+8 5.3E+8 7.9E+8 11.4E+8 15.7E+8 1.7 1.9E+8 1.9E+8 1.9E+8 1.9E+8 1.9E+8 1.9E+8 1.9E+8 1.9E+8 1.9E+8 1.9E+8 1.5 25.1E+6 41.7E+6 72.3E+6 1.3E+8 2.2E+8 3.6E+8 5.7E+8 8.6E+8 12.6E+8 17.6E+8 1.5 2.0E+8 2.0E+8 2.0E+8 2.0E+8 2.0E+8 2.0E+8 2.0E+8 2.0E+8 2.0E+8 2.0E+8 1.3 22.1E+6 39.9E+6 70.7E+6 1.3E+8 2.2E+8 3.8E+8 6.1E+8 9.6E+8 14.2E+8 20.1E+8 1.3 2.3E+8 2.3E+8 2.3E+8 2.3E+8 2.3E+8 2.3E+8 2.3E+8 2.3E+8 2.3E+8 2.3E+8 1.1 20.2E+6 35.0E+6 66.1E+6 1.2E+8 2.2E+8 4.0E+8 6.7E+8 10.7E+8 16.3E+8 23.4E+8 1.1 2.6E+8 2.6E+8 2.6E+8 2.6E+8 2.6E+8 2.6E+8 2.6E+8 2.6E+8 2.6E+8 2.6E+8 0.9 17.8E+6 32.4E+6 60.4E+6 1.2E+8 2.3E+8 4.2E+8 7.4E+8 12.3E+8 19.2E+8 28.2E+8 0.9 3.0E+8 3.0E+8 3.0E+8 3.0E+8 3.0E+8 3.0E+8 3.0E+8 3.0E+8 3.0E+8 3.0E+8 0.7 14.6E+6 27.7E+6 55.7E+6 1.1E+8 2.3E+8 4.5E+8 8.5E+8 14.7E+8 23.6E+8 35.6E+8 0.7 3.5E+8 3.5E+8 3.5E+8 3.5E+8 3.5E+8 3.5E+8 3.5E+8 3.5E+8 3.5E+8 3.5E+8 0.5 11.9E+6 23.3E+6 48.5E+6 1.0E+8 2.3E+8 5.0E+8 10.0E+8 18.5E+8 31.2E+8 48.7E+8 0.5 4.2E+8 4.2E+8 4.2E+8 4.2E+8 4.2E+8 4.2E+8 4.2E+8 4.2E+8 4.2E+8 4.2E+8 0.3 7.9E+6 17.1E+6 38.1E+6 91.9E+6 2.3E+8 5.6E+8 12.8E+8 26.2E+8 47.6E+8 78.2E+8 0.3 5.3E+8 5.3E+8 5.3E+8 5.3E+8 5.3E+8 5.3E+8 5.3E+8 5.3E+8 5.3E+8 5.3E+8 0.1 3.6E+6 8.7E+6 22.4E+6 64.2E+6 2.0E+8 6.6E+8 20.3E+8 53.6E+8 1.2E+10 2.2E+10 0.1 7.3E+8 7.3E+8 7.3E+8 7.3E+8 7.3E+8 7.3E+8 7.3E+8 7.3E+8 7.3E+8 7.3E+8 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 1.9 7843.1 9686.1 12476 15725 20035 25111 30999 37458 44475 51920 1.9 59800 59800 59800 59800 59800 59800 59800 59800 59800 59800 1.7 7100.3 9041.1 11383 14831 19146 24180 30139 36756 43978 51678 1.7 59800 59800 59800 59800 59800 59800 59800 59800 59800 59800 1.5 6156.8 7926.6 10436 13838 18046 23130 29220 36034 43482 51424 1.5 59800 59800 59800 59800 59800 59800 59800 59800 59800 59800 1.3 5377.9 7217.4 9608.8 12798 16825 22112 28279 35267 42941 51131 1.3 59800 59800 59800 59800 59800 59800 59800 59800 59800 59800 1.1 4729.7 6225.7 8548.2 11613 15665 20888 27168 34373 42317 50810 1.1 59800 59800 59800 59800 59800 59800 59800 59800 59800 59800 0.9 4023.6 5421.1 7396.2 10272 14301 19502 25889 33333 41571 50435 0.9 59800 59800 59800 59800 59800 59800 59800 59800 59800 59800 0.7 3215.8 4418.7 6265.7 8830.9 12663 17858 24372 32080 40690 49971 0.7 59800 59800 59800 59800 59800 59800 59800 59800 59800 59800 0.5 2456.6 3433.5 4943 7259.5 10737 15776 22424 30447 39533 49370 0.5 59801 59801 59801 59801 59801 59801 59801 59801 59801 59801 0.3 1552.3 2279.3 3399.7 5270.6 8299.3 12985 19641 28069 37821 48474 0.3 59801 59801 59801 59801 59801 59801 59801 59801 59801 59801 0.1 609.7 944.06 1509.6 2550.2 4500.1 8120.3 14273 23193 34240 46594 0.1 59801 59801 59801 59801 59801 59801 59801 59801 59801 59801 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 1.9 268 310 365 430 500 580 667 757 848 940 1.9 10 10 10 10 10 10 10 10 10 10 1.7 289 339 400 471 558 652 755 861 970 1080 1.7 10 10 10 10 10 10 10 10 10 10 1.5 310 370 440 524 625 740 861 990 1122 1254 1.5 10 10 10 10 10 10 10 10 10 10 1.3 340 403 487 589 710 847 998 1156 1317 1480 1.3 10 10 10 10 10 10 10 10 10 10 1.1 370 450 546 669 816 987 1176 1375 1579 1783 1.1 10 10 10 10 10 10 10 10 10 10 0.9 414 504 620 770 959 1179 1423 1683 1950 2217 0.9 10 10 10 10 10 10 10 10 10 10 0.7 470 580 728 920 1162 1459 1795 2154 2523 2893 0.7 10 10 10 10 10 10 10 10 10 10 0.5 559 698 888 1146 1490 1920 2424 2971 3534 4097 0.5 10 10 10 10 10 10 10 10 10 10 0.3 710 903 1180 1573 2126 2862 3768 4778 5826 6872 0.3 10 10 10 10 10 10 10 10 10 10 0.1 1160 1537 2100 2970 4330 6400 9305 12850 16648 20435 0.1 10 10 10 10 10 10 10 10 10 10 Colors indicate whether the value is or than its counterpart under the alternative management regime LOWER HIGHER Substitutability Prices Substitutability Quantities Substitutability Qualities Substitutability Varieties Substitutability Interference Elasticity Prices Interference Elasticity Interference Elasticity Quantities Interference Elasticity Substitutability Interference Elasticity Qualities Interference Elasticity Varieties Interference Elasticity Substitutability Substitutability Substitutability Substitutability Interference Elasticity Interference Elasticity Interference Elasticity LICENSING COMMONS Cournot Equilibria when native SNR is 30dB Figure C.16: Cournot Equilibria when Native SNR is 30dB 100 Consumer Consumer Surplus 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 Surplus 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 1.9 67.5E+6 1.0E+8 1.5E+8 2.3E+8 3.3E+8 4.6E+8 6.2E+8 8.2E+8 10.6E+8 13.4E+8 1.9 13.6E+8 13.6E+8 13.6E+8 13.6E+8 13.6E+8 13.6E+8 13.6E+8 13.6E+8 13.6E+8 13.6E+8 1.7 63.9E+6 1.0E+8 1.5E+8 2.3E+8 3.4E+8 4.8E+8 6.7E+8 8.9E+8 11.7E+8 14.9E+8 1.7 14.8E+8 14.8E+8 14.8E+8 14.8E+8 14.8E+8 14.8E+8 14.8E+8 14.8E+8 14.8E+8 14.8E+8 1.5 61.6E+6 97.9E+6 1.5E+8 2.4E+8 3.6E+8 5.2E+8 7.3E+8 9.9E+8 13.0E+8 16.8E+8 1.5 16.1E+8 16.1E+8 16.1E+8 16.1E+8 16.1E+8 16.1E+8 16.1E+8 16.1E+8 16.1E+8 16.1E+8 1.3 56.4E+6 94.5E+6 1.5E+8 2.5E+8 3.8E+8 5.6E+8 8.0E+8 11.0E+8 14.7E+8 19.2E+8 1.3 17.7E+8 17.7E+8 17.7E+8 17.7E+8 17.7E+8 17.7E+8 17.7E+8 17.7E+8 17.7E+8 17.7E+8 1.1 51.4E+6 89.9E+6 1.5E+8 2.5E+8 4.0E+8 6.1E+8 9.0E+8 12.6E+8 17.1E+8 22.4E+8 1.1 19.6E+8 19.6E+8 19.6E+8 19.6E+8 19.6E+8 19.6E+8 19.6E+8 19.6E+8 19.6E+8 19.6E+8 0.9 47.5E+6 85.6E+6 1.5E+8 2.6E+8 4.3E+8 6.9E+8 10.3E+8 14.7E+8 20.3E+8 27.1E+8 0.9 21.9E+8 21.9E+8 21.9E+8 21.9E+8 21.9E+8 21.9E+8 21.9E+8 21.9E+8 21.9E+8 21.9E+8 0.7 42.4E+6 78.5E+6 1.5E+8 2.7E+8 4.8E+8 7.8E+8 12.2E+8 18.0E+8 25.4E+8 34.4E+8 0.7 24.9E+8 24.9E+8 24.9E+8 24.9E+8 24.9E+8 24.9E+8 24.9E+8 24.9E+8 24.9E+8 24.9E+8 0.5 35.4E+6 70.5E+6 1.4E+8 2.8E+8 5.3E+8 9.3E+8 15.3E+8 23.5E+8 34.1E+8 47.4E+8 0.5 28.6E+8 28.6E+8 28.6E+8 28.6E+8 28.6E+8 28.6E+8 28.6E+8 28.6E+8 28.6E+8 28.6E+8 0.3 25.7E+6 56.9E+6 1.3E+8 2.9E+8 6.1E+8 12.0E+8 21.5E+8 35.1E+8 53.5E+8 76.9E+8 0.3 33.3E+8 33.3E+8 33.3E+8 33.3E+8 33.3E+8 33.3E+8 33.3E+8 33.3E+8 33.3E+8 33.3E+8 0.1 12.6E+6 32.6E+6 90.1E+6 2.6E+8 7.4E+8 19.2E+8 42.8E+8 82.3E+8 1.4E+10 2.2E+10 0.1 38.4E+8 38.4E+8 38.4E+8 38.4E+8 38.4E+8 38.4E+8 38.4E+8 38.4E+8 38.4E+8 38.4E+8 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 1.9 24.4E-4 21.1E-4 18.4E-4 16.2E-4 14.4E-4 12.9E-4 11.7E-4 10.6E-4 9.8E-4 9.0E-4 1.9 0.0948 0.0948 0.0948 0.0948 0.0948 0.0948 0.0948 0.0948 0.0948 0.0948 1.7 25.0E-4 21.4E-4 18.6E-4 16.3E-4 14.3E-4 12.8E-4 11.5E-4 10.4E-4 9.5E-4 8.8E-4 1.7 0.1036 0.1036 0.1036 0.1036 0.1036 0.1036 0.1036 0.1036 0.1036 0.1036 1.5 25.6E-4 22.0E-4 18.9E-4 16.3E-4 14.3E-4 12.6E-4 11.3E-4 10.2E-4 9.3E-4 8.6E-4 1.5 0.1143 0.1143 0.1143 0.1143 0.1143 0.1143 0.1143 0.1143 0.1143 0.1143 1.3 26.9E-4 22.7E-4 19.3E-4 16.5E-4 14.4E-4 12.6E-4 11.2E-4 10.1E-4 9.1E-4 8.4E-4 1.3 0.1274 0.1274 0.1274 0.1274 0.1274 0.1274 0.1274 0.1274 0.1274 0.1274 1.1 28.3E-4 23.8E-4 19.9E-4 16.9E-4 14.5E-4 12.6E-4 11.1E-4 9.9E-4 9.0E-4 8.2E-4 1.1 0.1439 0.1439 0.1439 0.1439 0.1439 0.1439 0.1439 0.1439 0.1439 0.1439 0.9 30.7E-4 25.3E-4 20.9E-4 17.5E-4 14.8E-4 12.7E-4 11.1E-4 9.9E-4 8.9E-4 8.1E-4 0.9 0.1653 0.1653 0.1653 0.1653 0.1653 0.1653 0.1653 0.1653 0.1653 0.1653 0.7 33.9E-4 27.6E-4 22.4E-4 18.4E-4 15.3E-4 12.9E-4 11.2E-4 9.8E-4 8.8E-4 7.9E-4 0.7 0.1942 0.1942 0.1942 0.1942 0.1942 0.1942 0.1942 0.1942 0.1942 0.1942 0.5 39.4E-4 31.4E-4 24.9E-4 19.9E-4 16.1E-4 13.4E-4 11.4E-4 9.9E-4 8.7E-4 7.8E-4 0.5 0.2353 0.2353 0.2353 0.2353 0.2353 0.2353 0.2353 0.2353 0.2353 0.2353 0.3 50.4E-4 39.1E-4 30.0E-4 23.0E-4 17.9E-4 14.3E-4 11.8E-4 10.0E-4 8.7E-4 7.7E-4 0.3 0.2985 0.2985 0.2985 0.2985 0.2985 0.2985 0.2985 0.2985 0.2985 0.2985 0.1 89.3E-4 66.0E-4 47.4E-4 33.5E-4 23.8E-4 17.3E-4 13.3E-4 10.7E-4 9.0E-4 7.7E-4 0.1 0.4082 0.4082 0.4082 0.4082 0.4082 0.4082 0.4082 0.4082 0.4082 0.4082 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 1.9 1.4E+8 2.1E+8 3.1E+8 4.6E+8 6.6E+8 9.2E+8 12.4E+8 16.4E+8 21.2E+8 26.9E+8 1.9 3.0E+8 3.0E+8 3.0E+8 3.0E+8 3.0E+8 3.0E+8 3.0E+8 3.0E+8 3.0E+8 3.0E+8 1.7 1.3E+8 2.0E+8 3.1E+8 4.7E+8 6.8E+8 9.7E+8 13.4E+8 17.9E+8 23.3E+8 29.8E+8 1.7 3.3E+8 3.3E+8 3.3E+8 3.3E+8 3.3E+8 3.3E+8 3.3E+8 3.3E+8 3.3E+8 3.3E+8 1.5 1.2E+8 2.0E+8 3.1E+8 4.8E+8 7.2E+8 10.4E+8 14.6E+8 19.7E+8 26.0E+8 33.5E+8 1.5 3.6E+8 3.6E+8 3.6E+8 3.6E+8 3.6E+8 3.6E+8 3.6E+8 3.6E+8 3.6E+8 3.6E+8 1.3 1.1E+8 1.9E+8 3.1E+8 4.9E+8 7.6E+8 11.2E+8 16.0E+8 22.1E+8 29.5E+8 38.4E+8 1.3 4.0E+8 4.0E+8 4.0E+8 4.0E+8 4.0E+8 4.0E+8 4.0E+8 4.0E+8 4.0E+8 4.0E+8 1.1 1.0E+8 1.8E+8 3.0E+8 5.1E+8 8.1E+8 12.3E+8 17.9E+8 25.2E+8 34.1E+8 44.9E+8 1.1 4.6E+8 4.6E+8 4.6E+8 4.6E+8 4.6E+8 4.6E+8 4.6E+8 4.6E+8 4.6E+8 4.6E+8 0.9 95.2E+6 1.7E+8 3.0E+8 5.2E+8 8.7E+8 13.7E+8 20.6E+8 29.5E+8 40.7E+8 54.3E+8 0.9 5.3E+8 5.3E+8 5.3E+8 5.3E+8 5.3E+8 5.3E+8 5.3E+8 5.3E+8 5.3E+8 5.3E+8 0.7 85.1E+6 1.6E+8 3.0E+8 5.4E+8 9.5E+8 15.7E+8 24.4E+8 36.0E+8 50.8E+8 68.9E+8 0.7 6.2E+8 6.2E+8 6.2E+8 6.2E+8 6.2E+8 6.2E+8 6.2E+8 6.2E+8 6.2E+8 6.2E+8 0.5 71.0E+6 1.4E+8 2.9E+8 5.6E+8 10.6E+8 18.7E+8 30.7E+8 47.0E+8 68.3E+8 94.8E+8 0.5 7.5E+8 7.5E+8 7.5E+8 7.5E+8 7.5E+8 7.5E+8 7.5E+8 7.5E+8 7.5E+8 7.5E+8 0.3 51.7E+6 1.1E+8 2.6E+8 5.7E+8 12.3E+8 24.1E+8 43.0E+8 70.4E+8 1.1E+10 1.5E+10 0.3 9.5E+8 9.5E+8 9.5E+8 9.5E+8 9.5E+8 9.5E+8 9.5E+8 9.5E+8 9.5E+8 9.5E+8 0.1 25.5E+6 65.7E+6 1.8E+8 5.2E+8 14.8E+8 38.4E+8 85.8E+8 1.6E+10 2.8E+10 4.4E+10 0.1 13.0E+8 13.0E+8 13.0E+8 13.0E+8 13.0E+8 13.0E+8 13.0E+8 13.0E+8 13.0E+8 13.0E+8 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 1.9 16059 19970 24248 29588 35421 41728 48567 55895 63520 71465 1.9 79724 79724 79724 79724 79724 79724 79724 79724 79724 79724 1.7 14777 18618 22951 28210 34093 40634 47702 55201 63035 71226 1.7 79724 79724 79724 79724 79724 79724 79724 79724 79724 79724 1.5 13636 17183 21602 26841 32812 39500 46754 54427 62537 70959 1.5 79724 79724 79724 79724 79724 79724 79724 79724 79724 79724 1.3 12144 15720 20077 25373 31390 38214 45665 53615 61953 70670 1.3 79724 79724 79724 79724 79724 79724 79724 79724 79724 79724 1.1 10675 14106 18330 23649 29821 36797 44467 52680 61314 70341 1.1 79724 79724 79724 79724 79724 79724 79724 79724 79724 79724 0.9 9280 12454 16499 21723 27993 35183 43093 51579 60574 69961 0.9 79724 79724 79724 79724 79724 79724 79724 79724 79724 79724 0.7 7740.2 10522 14416 19531 25858 33191 41388 50264 59665 69493 0.7 79724 79724 79724 79724 79724 79724 79724 79724 79724 79724 0.5 5978.9 8435.3 11983 16821 23086 30629 39195 48533 58477 68885 0.5 79724 79724 79724 79724 79724 79724 79724 79724 79724 79724 0.3 3955.3 5877.2 8807.2 13157 19230 26935 35965 45984 56716 67983 0.3 79725 79725 79725 79725 79725 79725 79725 79725 79725 79725 0.1 1611 2579.1 4274.4 7240 12196 19630 29331 40646 53010 66082 0.1 79725 79725 79725 79725 79725 79725 79725 79725 79725 79725 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 1.9 431 498 570 647 729 813 900 988 1076 1163 1.9 10 10 10 10 10 10 10 10 10 10 1.7 470 548 630 722 820 920 1024 1129 1233 1337 1.7 10 10 10 10 10 10 10 10 10 10 1.5 519 605 705 814 930 1053 1177 1303 1429 1554 1.5 10 10 10 10 10 10 10 10 10 10 1.3 570 675 795 928 1070 1220 1372 1527 1681 1835 1.3 10 10 10 10 10 10 10 10 10 10 1.1 640 761 910 1072 1251 1440 1632 1826 2020 2214 1.1 10 10 10 10 10 10 10 10 10 10 0.9 720 875 1060 1270 1500 1745 1997 2251 2505 2757 0.9 10 10 10 10 10 10 10 10 10 10 0.7 838 1030 1270 1550 1867 2204 2552 2904 3255 3603 0.7 10 10 10 10 10 10 10 10 10 10 0.5 1008 1266 1598 2002 2471 2982 3514 4051 4586 5114 0.5 10 10 10 10 10 10 10 10 10 10 0.3 1310 1691 2212 2890 3716 4649 5634 6635 7628 8606 0.3 10 10 10 10 10 10 10 10 10 10 0.1 2200 2990 4179 5929 8382 11500 15020 18660 22266 25784 0.1 10 10 10 10 10 10 10 10 10 10 Colors indicate whether the value is or than its counterpart under the alternative management regime LOWER HIGHER Substitutability Prices Substitutability Quantities Substitutability Qualities Substitutability Varieties Substitutability Interference Elasticity Prices Interference Elasticity Interference Elasticity Quantities Interference Elasticity Substitutability Interference Elasticity Qualities Interference Elasticity Varieties Interference Elasticity Substitutability Substitutability Substitutability Substitutability Interference Elasticity Interference Elasticity Interference Elasticity LICENSING COMMONS Cournot Equilibria when native SNR is 40dB Figure C.17: Cournot Equilibria when Native SNR is 40dB 101 Consumer Consumer Surplus 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 Surplus 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 1.9 2.1E+8 3.0E+8 4.1E+8 5.5E+8 7.2E+8 9.3E+8 11.8E+8 14.7E+8 18.0E+8 21.8E+8 1.9 21.3E+8 21.3E+8 21.3E+8 21.3E+8 21.3E+8 21.3E+8 21.3E+8 21.3E+8 21.3E+8 21.3E+8 1.7 2.1E+8 3.0E+8 4.2E+8 5.7E+8 7.7E+8 10.0E+8 12.8E+8 16.1E+8 19.9E+8 24.2E+8 1.7 23.1E+8 23.1E+8 23.1E+8 23.1E+8 23.1E+8 23.1E+8 23.1E+8 23.1E+8 23.1E+8 23.1E+8 1.5 2.1E+8 3.0E+8 4.3E+8 6.1E+8 8.2E+8 10.9E+8 14.1E+8 17.9E+8 22.3E+8 27.3E+8 1.5 25.1E+8 25.1E+8 25.1E+8 25.1E+8 25.1E+8 25.1E+8 25.1E+8 25.1E+8 25.1E+8 25.1E+8 1.3 2.0E+8 3.1E+8 4.5E+8 6.4E+8 8.9E+8 12.0E+8 15.7E+8 20.2E+8 25.3E+8 31.3E+8 1.3 27.6E+8 27.6E+8 27.6E+8 27.6E+8 27.6E+8 27.6E+8 27.6E+8 27.6E+8 27.6E+8 27.6E+8 1.1 2.0E+8 3.1E+8 4.7E+8 6.9E+8 9.8E+8 13.4E+8 17.9E+8 23.2E+8 29.5E+8 36.7E+8 1.1 30.6E+8 30.6E+8 30.6E+8 30.6E+8 30.6E+8 30.6E+8 30.6E+8 30.6E+8 30.6E+8 30.6E+8 0.9 1.9E+8 3.2E+8 5.0E+8 7.5E+8 11.0E+8 15.4E+8 20.9E+8 27.5E+8 35.4E+8 44.5E+8 0.9 34.3E+8 34.3E+8 34.3E+8 34.3E+8 34.3E+8 34.3E+8 34.3E+8 34.3E+8 34.3E+8 34.3E+8 0.7 1.8E+8 3.2E+8 5.2E+8 8.3E+8 12.6E+8 18.2E+8 25.3E+8 34.0E+8 44.4E+8 56.6E+8 0.7 38.8E+8 38.8E+8 38.8E+8 38.8E+8 38.8E+8 38.8E+8 38.8E+8 38.8E+8 38.8E+8 38.8E+8 0.5 1.7E+8 3.1E+8 5.6E+8 9.4E+8 15.1E+8 22.8E+8 32.7E+8 45.2E+8 60.3E+8 78.2E+8 0.5 44.7E+8 44.7E+8 44.7E+8 44.7E+8 44.7E+8 44.7E+8 44.7E+8 44.7E+8 44.7E+8 44.7E+8 0.3 1.4E+8 2.9E+8 5.9E+8 11.2E+8 19.6E+8 31.8E+8 48.3E+8 69.6E+8 96.0E+8 1.3E+10 0.3 52.0E+8 52.0E+8 52.0E+8 52.0E+8 52.0E+8 52.0E+8 52.0E+8 52.0E+8 52.0E+8 52.0E+8 0.1 82.2E+6 2.1E+8 5.6E+8 14.0E+8 31.6E+8 62.3E+8 1.1E+10 1.7E+10 2.6E+10 3.7E+10 0.1 60.0E+8 60.0E+8 60.0E+8 60.0E+8 60.0E+8 60.0E+8 60.0E+8 60.0E+8 60.0E+8 60.0E+8 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 1.9 16.7E-4 14.9E-4 13.4E-4 12.2E-4 11.1E-4 10.2E-4 9.4E-4 8.8E-4 8.2E-4 7.7E-4 1.9 0.0948 0.0948 0.0948 0.0948 0.0948 0.0948 0.0948 0.0948 0.0948 0.0948 1.7 16.8E-4 15.0E-4 13.4E-4 12.1E-4 11.0E-4 10.0E-4 9.2E-4 8.6E-4 8.0E-4 7.5E-4 1.7 0.1036 0.1036 0.1036 0.1036 0.1036 0.1036 0.1036 0.1036 0.1036 0.1036 1.5 17.2E-4 15.1E-4 13.4E-4 12.0E-4 10.9E-4 9.9E-4 9.1E-4 8.4E-4 7.8E-4 7.3E-4 1.5 0.1143 0.1143 0.1143 0.1143 0.1143 0.1143 0.1143 0.1143 0.1143 0.1143 1.3 17.6E-4 15.4E-4 13.5E-4 12.0E-4 10.8E-4 9.8E-4 8.9E-4 8.2E-4 7.6E-4 7.1E-4 1.3 0.1274 0.1274 0.1274 0.1274 0.1274 0.1274 0.1274 0.1274 0.1274 0.1274 1.1 18.3E-4 15.8E-4 13.7E-4 12.1E-4 10.8E-4 9.7E-4 8.8E-4 8.1E-4 7.5E-4 7.0E-4 1.1 0.1439 0.1439 0.1439 0.1439 0.1439 0.1439 0.1439 0.1439 0.1439 0.1439 0.9 19.3E-4 16.4E-4 14.1E-4 12.3E-4 10.9E-4 9.7E-4 8.8E-4 8.0E-4 7.4E-4 6.8E-4 0.9 0.1653 0.1653 0.1653 0.1653 0.1653 0.1653 0.1653 0.1653 0.1653 0.1653 0.7 20.9E-4 17.5E-4 14.8E-4 12.7E-4 11.1E-4 9.8E-4 8.8E-4 8.0E-4 7.3E-4 6.7E-4 0.7 0.1942 0.1942 0.1942 0.1942 0.1942 0.1942 0.1942 0.1942 0.1942 0.1942 0.5 23.6E-4 19.2E-4 15.9E-4 13.3E-4 11.4E-4 10.0E-4 8.8E-4 7.9E-4 7.2E-4 6.6E-4 0.5 0.2353 0.2353 0.2353 0.2353 0.2353 0.2353 0.2353 0.2353 0.2353 0.2353 0.3 29.0E-4 22.8E-4 18.1E-4 14.7E-4 12.2E-4 10.4E-4 9.0E-4 8.0E-4 7.2E-4 6.5E-4 0.3 0.2985 0.2985 0.2985 0.2985 0.2985 0.2985 0.2985 0.2985 0.2985 0.2985 0.1 48.8E-4 35.7E-4 26.0E-4 19.2E-4 14.7E-4 11.7E-4 9.7E-4 8.3E-4 7.3E-4 6.5E-4 0.1 0.4082 0.4082 0.4082 0.4082 0.4082 0.4082 0.4082 0.4082 0.4082 0.4082 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 1.9 4.2E+8 5.9E+8 8.2E+8 11.0E+8 14.5E+8 18.6E+8 23.6E+8 29.4E+8 36.1E+8 43.7E+8 1.9 4.7E+8 4.7E+8 4.7E+8 4.7E+8 4.7E+8 4.7E+8 4.7E+8 4.7E+8 4.7E+8 4.7E+8 1.7 4.2E+8 6.0E+8 8.4E+8 11.5E+8 15.3E+8 20.0E+8 25.7E+8 32.2E+8 39.8E+8 48.5E+8 1.7 5.1E+8 5.1E+8 5.1E+8 5.1E+8 5.1E+8 5.1E+8 5.1E+8 5.1E+8 5.1E+8 5.1E+8 1.5 4.1E+8 6.0E+8 8.7E+8 12.2E+8 16.5E+8 21.8E+8 28.2E+8 35.8E+8 44.6E+8 54.7E+8 1.5 5.7E+8 5.7E+8 5.7E+8 5.7E+8 5.7E+8 5.7E+8 5.7E+8 5.7E+8 5.7E+8 5.7E+8 1.3 4.0E+8 6.1E+8 9.0E+8 12.9E+8 17.8E+8 24.0E+8 31.5E+8 40.4E+8 50.7E+8 62.7E+8 1.3 6.3E+8 6.3E+8 6.3E+8 6.3E+8 6.3E+8 6.3E+8 6.3E+8 6.3E+8 6.3E+8 6.3E+8 1.1 4.0E+8 6.3E+8 9.4E+8 13.9E+8 19.6E+8 26.9E+8 35.8E+8 46.5E+8 59.0E+8 73.5E+8 1.1 7.1E+8 7.1E+8 7.1E+8 7.1E+8 7.1E+8 7.1E+8 7.1E+8 7.1E+8 7.1E+8 7.1E+8 0.9 3.9E+8 6.3E+8 10.0E+8 15.0E+8 21.9E+8 30.7E+8 41.7E+8 55.0E+8 70.8E+8 89.1E+8 0.9 8.2E+8 8.2E+8 8.2E+8 8.2E+8 8.2E+8 8.2E+8 8.2E+8 8.2E+8 8.2E+8 8.2E+8 0.7 3.7E+8 6.3E+8 10.5E+8 16.7E+8 25.2E+8 36.4E+8 50.6E+8 68.1E+8 88.9E+8 1.1E+10 0.7 9.6E+8 9.6E+8 9.6E+8 9.6E+8 9.6E+8 9.6E+8 9.6E+8 9.6E+8 9.6E+8 9.6E+8 0.5 3.4E+8 6.3E+8 11.2E+8 18.9E+8 30.2E+8 45.6E+8 65.5E+8 90.5E+8 1.2E+10 1.6E+10 0.5 11.7E+8 11.7E+8 11.7E+8 11.7E+8 11.7E+8 11.7E+8 11.7E+8 11.7E+8 11.7E+8 11.7E+8 0.3 2.8E+8 5.9E+8 11.8E+8 22.4E+8 39.2E+8 63.6E+8 96.7E+8 1.4E+10 1.9E+10 2.6E+10 0.3 14.8E+8 14.8E+8 14.8E+8 14.8E+8 14.8E+8 14.8E+8 14.8E+8 14.8E+8 14.8E+8 14.8E+8 0.1 1.7E+8 4.3E+8 11.3E+8 28.1E+8 63.2E+8 1.2E+10 2.2E+10 3.5E+10 5.2E+10 7.3E+10 0.1 20.3E+8 20.3E+8 20.3E+8 20.3E+8 20.3E+8 20.3E+8 20.3E+8 20.3E+8 20.3E+8 20.3E+8 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 1.9 28242 33615 39382 45692 52454 59516 67015 74803 82823 91116 1.9 99654 99654 99654 99654 99654 99654 99654 99654 99654 99654 1.7 26732 31965 37885 44240 51110 58412 66081 74070 82341 90865 1.7 99654 99654 99654 99654 99654 99654 99654 99654 99654 99654 1.5 24928 30137 36119 42724 49750 57237 65114 73327 81836 90607 1.5 99654 99654 99654 99654 99654 99654 99654 99654 99654 99654 1.3 22946 28212 34318 40995 48177 55867 64019 72478 81254 90309 1.3 99655 99655 99655 99655 99655 99655 99655 99655 99655 99655 1.1 20887 26258 32210 39077 46437 54400 62762 71535 80609 89986 1.1 99655 99655 99655 99655 99655 99655 99655 99655 99655 99655 0.9 18692 23878 29960 36814 44480 52640 61332 70426 79852 89597 0.9 99655 99655 99655 99655 99655 99655 99655 99655 99655 99655 0.7 16049 21054 27121 34207 42046 50533 59571 69073 78936 89131 0.7 99655 99655 99655 99655 99655 99655 99655 99655 99655 99655 0.5 12995 17729 23676 30789 38897 47787 57288 67307 77737 88521 0.5 99655 99655 99655 99655 99655 99655 99655 99655 99655 99655 0.3 9261.4 13317 18872 25947 34316 43726 53907 64693 75957 87615 0.3 99655 99655 99655 99655 99655 99655 99655 99655 99655 99655 0.1 4084.4 6583.1 10637 16808 25180 35370 46845 59204 72207 85706 0.1 99656 99656 99656 99656 99656 99656 99656 99656 99656 99656 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 1.9 630 704 783 864 947 1030 1115 1199 1283 1367 1.9 10 10 10 10 10 10 10 10 10 10 1.7 697 785 879 974 1072 1172 1272 1372 1473 1572 1.7 10 10 10 10 10 10 10 10 10 10 1.5 775 880 992 1109 1227 1347 1468 1589 1709 1829 1.5 10 10 10 10 10 10 10 10 10 10 1.3 870 1000 1135 1277 1422 1570 1718 1866 2013 2160 1.3 10 10 10 10 10 10 10 10 10 10 1.1 990 1149 1320 1497 1680 1866 2052 2239 2424 2608 1.1 10 10 10 10 10 10 10 10 10 10 0.9 1146 1349 1568 1800 2039 2281 2525 2769 3010 3250 0.9 10 10 10 10 10 10 10 10 10 10 0.7 1360 1630 1930 2248 2579 2914 3252 3588 3922 4252 0.7 10 10 10 10 10 10 10 10 10 10 0.5 1690 2074 2513 2993 3496 4009 4523 5035 5542 6043 0.5 10 10 10 10 10 10 10 10 10 10 0.3 2287 2911 3670 4535 5461 6416 7375 8327 9266 10189 0.3 10 10 10 10 10 10 10 10 10 10 0.1 4060 5566 7660 10373 13570 17011 20508 23965 27351 30656 0.1 10 10 10 10 10 10 10 10 10 10 Colors indicate whether the value is or than its counterpart under the alternative management regime LOWER HIGHER Substitutability Prices Substitutability Quantities Substitutability Qualities Substitutability Varieties Substitutability Interference Elasticity Prices Interference Elasticity Interference Elasticity Quantities Interference Elasticity Substitutability Interference Elasticity Qualities Interference Elasticity Varieties Interference Elasticity Substitutability Substitutability Substitutability Substitutability Interference Elasticity Interference Elasticity Interference Elasticity LICENSING COMMONS Cournot Equilibria when native SNR is 50dB Figure C.18: Cournot Equilibria when Native SNR is 50dB 102 Consumer Consumer Surplus 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 Surplus 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 1.9 4.9E+8 6.4E+8 8.2E+8 10.4E+8 13.0E+8 15.9E+8 19.3E+8 23.2E+8 27.5E+8 32.3E+8 1.9 30.7E+8 30.7E+8 30.7E+8 30.7E+8 30.7E+8 30.7E+8 30.7E+8 30.7E+8 30.7E+8 30.7E+8 1.7 5.0E+8 6.6E+8 8.7E+8 11.1E+8 14.0E+8 17.3E+8 21.1E+8 25.5E+8 30.4E+8 35.9E+8 1.7 33.2E+8 33.2E+8 33.2E+8 33.2E+8 33.2E+8 33.2E+8 33.2E+8 33.2E+8 33.2E+8 33.2E+8 1.5 5.1E+8 6.9E+8 9.2E+8 11.9E+8 15.2E+8 19.0E+8 23.4E+8 28.4E+8 34.1E+8 40.5E+8 1.5 36.2E+8 36.2E+8 36.2E+8 36.2E+8 36.2E+8 36.2E+8 36.2E+8 36.2E+8 36.2E+8 36.2E+8 1.3 5.2E+8 7.3E+8 9.8E+8 12.9E+8 16.7E+8 21.1E+8 26.3E+8 32.2E+8 38.9E+8 46.5E+8 1.3 39.7E+8 39.7E+8 39.7E+8 39.7E+8 39.7E+8 39.7E+8 39.7E+8 39.7E+8 39.7E+8 39.7E+8 1.1 5.4E+8 7.6E+8 10.6E+8 14.2E+8 18.6E+8 23.9E+8 30.1E+8 37.3E+8 45.4E+8 54.6E+8 1.1 44.0E+8 44.0E+8 44.0E+8 44.0E+8 44.0E+8 44.0E+8 44.0E+8 44.0E+8 44.0E+8 44.0E+8 0.9 5.6E+8 8.1E+8 11.6E+8 15.9E+8 21.3E+8 27.8E+8 35.5E+8 44.4E+8 54.7E+8 66.3E+8 0.9 49.3E+8 49.3E+8 49.3E+8 49.3E+8 49.3E+8 49.3E+8 49.3E+8 49.3E+8 49.3E+8 49.3E+8 0.7 5.7E+8 8.8E+8 12.9E+8 18.4E+8 25.2E+8 33.6E+8 43.6E+8 55.4E+8 69.0E+8 84.5E+8 0.7 55.9E+8 55.9E+8 55.9E+8 55.9E+8 55.9E+8 55.9E+8 55.9E+8 55.9E+8 55.9E+8 55.9E+8 0.5 5.8E+8 9.5E+8 14.9E+8 22.1E+8 31.5E+8 43.2E+8 57.5E+8 74.5E+8 94.2E+8 1.2E+10 0.5 64.3E+8 64.3E+8 64.3E+8 64.3E+8 64.3E+8 64.3E+8 64.3E+8 64.3E+8 64.3E+8 64.3E+8 0.3 5.7E+8 10.5E+8 17.9E+8 28.9E+8 43.8E+8 63.1E+8 87.3E+8 1.2E+10 1.5E+10 1.9E+10 0.3 74.9E+8 74.9E+8 74.9E+8 74.9E+8 74.9E+8 74.9E+8 74.9E+8 74.9E+8 74.9E+8 74.9E+8 0.1 4.4E+8 10.6E+8 23.5E+8 47.2E+8 84.9E+8 1.4E+10 2.1E+10 3.0E+10 4.2E+10 5.5E+10 0.1 86.4E+8 86.4E+8 86.4E+8 86.4E+8 86.4E+8 86.4E+8 86.4E+8 86.4E+8 86.4E+8 86.4E+8 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 1.9 12.7E-4 11.6E-4 10.6E-4 9.8E-4 9.1E-4 8.5E-4 8.0E-4 7.5E-4 7.1E-4 6.8E-4 1.9 0.0948 0.0948 0.0948 0.0948 0.0948 0.0948 0.0948 0.0948 0.0948 0.0948 1.7 12.6E-4 11.5E-4 10.5E-4 9.7E-4 9.0E-4 8.4E-4 7.8E-4 7.4E-4 6.9E-4 6.6E-4 1.7 0.1036 0.1036 0.1036 0.1036 0.1036 0.1036 0.1036 0.1036 0.1036 0.1036 1.5 12.7E-4 11.5E-4 10.5E-4 9.6E-4 8.9E-4 8.2E-4 7.7E-4 7.2E-4 6.8E-4 6.4E-4 1.5 0.1143 0.1143 0.1143 0.1143 0.1143 0.1143 0.1143 0.1143 0.1143 0.1143 1.3 12.9E-4 11.6E-4 10.5E-4 9.5E-4 8.8E-4 8.1E-4 7.5E-4 7.1E-4 6.6E-4 6.2E-4 1.3 0.1274 0.1274 0.1274 0.1274 0.1274 0.1274 0.1274 0.1274 0.1274 0.1274 1.1 13.2E-4 11.7E-4 10.5E-4 9.5E-4 8.7E-4 8.0E-4 7.4E-4 6.9E-4 6.5E-4 6.1E-4 1.1 0.1439 0.1439 0.1439 0.1439 0.1439 0.1439 0.1439 0.1439 0.1439 0.1439 0.9 13.6E-4 12.0E-4 10.7E-4 9.6E-4 8.7E-4 8.0E-4 7.4E-4 6.8E-4 6.4E-4 6.0E-4 0.9 0.1653 0.1653 0.1653 0.1653 0.1653 0.1653 0.1653 0.1653 0.1653 0.1653 0.7 14.4E-4 12.5E-4 11.0E-4 9.8E-4 8.8E-4 8.0E-4 7.3E-4 6.8E-4 6.3E-4 5.9E-4 0.7 0.1942 0.1942 0.1942 0.1942 0.1942 0.1942 0.1942 0.1942 0.1942 0.1942 0.5 15.6E-4 13.3E-4 11.5E-4 10.1E-4 8.9E-4 8.1E-4 7.3E-4 6.7E-4 6.2E-4 5.8E-4 0.5 0.2353 0.2353 0.2353 0.2353 0.2353 0.2353 0.2353 0.2353 0.2353 0.2353 0.3 18.3E-4 15.0E-4 12.5E-4 10.7E-4 9.3E-4 8.3E-4 7.4E-4 6.7E-4 6.2E-4 5.7E-4 0.3 0.2985 0.2985 0.2985 0.2985 0.2985 0.2985 0.2985 0.2985 0.2985 0.2985 0.1 28.1E-4 21.1E-4 16.2E-4 12.9E-4 10.6E-4 9.0E-4 7.8E-4 6.9E-4 6.2E-4 5.7E-4 0.1 0.4082 0.4082 0.4082 0.4082 0.4082 0.4082 0.4082 0.4082 0.4082 0.4082 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 1.9 9.8E+8 12.8E+8 16.4E+8 20.8E+8 26.0E+8 31.9E+8 38.7E+8 46.4E+8 55.0E+8 64.6E+8 1.9 6.8E+8 6.8E+8 6.8E+8 6.8E+8 6.8E+8 6.8E+8 6.8E+8 6.8E+8 6.8E+8 6.8E+8 1.7 10.0E+8 13.3E+8 17.4E+8 22.2E+8 28.0E+8 34.6E+8 42.3E+8 51.1E+8 60.9E+8 71.9E+8 1.7 7.4E+8 7.4E+8 7.4E+8 7.4E+8 7.4E+8 7.4E+8 7.4E+8 7.4E+8 7.4E+8 7.4E+8 1.5 10.3E+8 13.9E+8 18.3E+8 23.9E+8 30.3E+8 38.0E+8 46.8E+8 56.9E+8 68.3E+8 81.0E+8 1.5 8.2E+8 8.2E+8 8.2E+8 8.2E+8 8.2E+8 8.2E+8 8.2E+8 8.2E+8 8.2E+8 8.2E+8 1.3 10.5E+8 14.6E+8 19.6E+8 25.8E+8 33.4E+8 42.3E+8 52.6E+8 64.5E+8 77.9E+8 93.0E+8 1.3 9.1E+8 9.1E+8 9.1E+8 9.1E+8 9.1E+8 9.1E+8 9.1E+8 9.1E+8 9.1E+8 9.1E+8 1.1 10.8E+8 15.3E+8 21.2E+8 28.4E+8 37.3E+8 47.9E+8 60.3E+8 74.6E+8 90.9E+8 1.1E+10 1.1 10.3E+8 10.3E+8 10.3E+8 10.3E+8 10.3E+8 10.3E+8 10.3E+8 10.3E+8 10.3E+8 10.3E+8 0.9 11.1E+8 16.3E+8 23.2E+8 31.9E+8 42.7E+8 55.7E+8 71.1E+8 88.9E+8 1.1E+10 1.3E+10 0.9 11.8E+8 11.8E+8 11.8E+8 11.8E+8 11.8E+8 11.8E+8 11.8E+8 11.8E+8 11.8E+8 11.8E+8 0.7 11.4E+8 17.6E+8 25.9E+8 36.8E+8 50.5E+8 67.3E+8 87.4E+8 1.1E+10 1.4E+10 1.7E+10 0.7 13.9E+8 13.9E+8 13.9E+8 13.9E+8 13.9E+8 13.9E+8 13.9E+8 13.9E+8 13.9E+8 13.9E+8 0.5 11.7E+8 19.1E+8 29.7E+8 44.2E+8 63.1E+8 86.5E+8 1.2E+10 1.5E+10 1.9E+10 2.3E+10 0.5 16.8E+8 16.8E+8 16.8E+8 16.8E+8 16.8E+8 16.8E+8 16.8E+8 16.8E+8 16.8E+8 16.8E+8 0.3 11.5E+8 20.9E+8 35.9E+8 57.8E+8 87.7E+8 1.3E+10 1.7E+10 2.3E+10 3.0E+10 3.8E+10 0.3 21.3E+8 21.3E+8 21.3E+8 21.3E+8 21.3E+8 21.3E+8 21.3E+8 21.3E+8 21.3E+8 21.3E+8 0.1 8.9E+8 21.2E+8 47.1E+8 94.5E+8 1.7E+10 2.8E+10 4.2E+10 6.1E+10 8.4E+10 ###### 0.1 29.2E+8 29.2E+8 29.2E+8 29.2E+8 29.2E+8 29.2E+8 29.2E+8 29.2E+8 29.2E+8 29.2E+8 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 1.9 43120 49304 55930 62934 70271 77900 85793 93924 102272 110819 1.9 119586 119586 119586 119586 119586 119586 119586 119586 119586 119586 1.7 41164 47531 54346 61444 68996 76778 84853 93224 101787 110567 1.7 119586 119586 119586 119586 119586 119586 119586 119586 119586 119586 1.5 39246 45669 52484 59850 67508 75513 83863 92435 101271 110305 1.5 119586 119586 119586 119586 119586 119586 119586 119586 119586 119586 1.3 36941 43532 50556 57983 65906 74177 82754 91580 100685 110012 1.3 119586 119586 119586 119586 119586 119586 119586 119586 119586 119586 1.1 34541 41048 48293 55977 64111 72628 81478 90624 100044 109684 1.1 119586 119586 119586 119586 119586 119586 119586 119586 119586 119586 0.9 31702 38318 45738 53612 62035 70834 80023 89519 99277 109296 0.9 119586 119586 119586 119586 119586 119586 119586 119586 119586 119586 0.7 28336 35108 42599 50778 59508 68668 78246 88144 98355 108830 0.7 119586 119586 119586 119586 119586 119586 119586 119586 119586 119586 0.5 24169 30927 38598 47068 56192 65822 75909 86367 97153 108218 0.5 119587 119587 119587 119587 119587 119587 119587 119587 119587 119587 0.3 18583 25098 32868 41679 51328 61621 72458 83723 95357 107308 0.3 119587 119587 119587 119587 119587 119587 119587 119587 119587 119587 0.1 9468.9 14593 21735 30777 41281 52856 65212 78161 91584 105394 0.1 119587 119587 119587 119587 119587 119587 119587 119587 119587 119587 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 1.9 831 910 990 1070 1150 1232 1314 1395 1477 1558 1.9 10 10 10 10 10 10 10 10 10 10 1.7 930 1022 1117 1212 1309 1406 1502 1599 1696 1792 1.7 10 10 10 10 10 10 10 10 10 10 1.5 1048 1159 1272 1388 1503 1620 1737 1853 1969 2085 1.5 10 10 10 10 10 10 10 10 10 10 1.3 1192 1329 1468 1610 1752 1895 2038 2180 2322 2463 1.3 10 10 10 10 10 10 10 10 10 10 1.1 1379 1550 1724 1902 2081 2261 2441 2620 2799 2976 1.1 10 10 10 10 10 10 10 10 10 10 0.9 1628 1850 2078 2310 2545 2780 3015 3248 3480 3710 0.9 10 10 10 10 10 10 10 10 10 10 0.7 1984 2287 2601 2924 3249 3574 3899 4220 4540 4856 0.7 10 10 10 10 10 10 10 10 10 10 0.5 2550 3002 3480 3970 4465 4961 5455 5944 6429 6907 0.5 10 10 10 10 10 10 10 10 10 10 0.3 3632 4435 5307 6215 7138 8061 8979 9886 10780 11662 0.3 10 10 10 10 10 10 10 10 10 10 0.1 7085 9443 12310 15508 18853 22223 25558 28832 32040 35181 0.1 10 10 10 10 10 10 10 10 10 10 Colors indicate whether the value is or than its counterpart under the alternative management regime LOWER HIGHER Substitutability Prices Substitutability Quantities Substitutability Qualities Substitutability Varieties Substitutability Interference Elasticity Prices Interference Elasticity Interference Elasticity Quantities Interference Elasticity Substitutability Interference Elasticity Qualities Interference Elasticity Varieties Interference Elasticity Substitutability Substitutability Substitutability Substitutability Interference Elasticity Interference Elasticity Interference Elasticity LICENSING COMMONS Cournot Equilibria when native SNR is 60dB Figure C.19: Cournot Equilibria when Native SNR is 60dB 103 Consumer Consumer Surplus 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 Surplus 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 1.9 9.3E+8 11.5E+8 14.2E+8 17.2E+8 20.6E+8 24.5E+8 28.9E+8 33.7E+8 39.0E+8 44.8E+8 1.9 41.8E+8 41.8E+8 41.8E+8 41.8E+8 41.8E+8 41.8E+8 41.8E+8 41.8E+8 41.8E+8 41.8E+8 1.7 9.7E+8 12.1E+8 15.1E+8 18.5E+8 22.3E+8 26.8E+8 31.7E+8 37.2E+8 43.2E+8 49.9E+8 1.7 45.2E+8 45.2E+8 45.2E+8 45.2E+8 45.2E+8 45.2E+8 45.2E+8 45.2E+8 45.2E+8 45.2E+8 1.5 10.1E+8 12.9E+8 16.2E+8 20.0E+8 24.5E+8 29.5E+8 35.2E+8 41.5E+8 48.6E+8 56.3E+8 1.5 49.3E+8 49.3E+8 49.3E+8 49.3E+8 49.3E+8 49.3E+8 49.3E+8 49.3E+8 49.3E+8 49.3E+8 1.3 10.6E+8 13.8E+8 17.6E+8 22.0E+8 27.2E+8 33.1E+8 39.7E+8 47.2E+8 55.5E+8 64.7E+8 1.3 54.1E+8 54.1E+8 54.1E+8 54.1E+8 54.1E+8 54.1E+8 54.1E+8 54.1E+8 54.1E+8 54.1E+8 1.1 11.3E+8 15.0E+8 19.4E+8 24.6E+8 30.8E+8 37.8E+8 45.8E+8 54.8E+8 64.9E+8 76.0E+8 1.1 59.9E+8 59.9E+8 59.9E+8 59.9E+8 59.9E+8 59.9E+8 59.9E+8 59.9E+8 59.9E+8 59.9E+8 0.9 12.2E+8 16.5E+8 21.8E+8 28.1E+8 35.7E+8 44.4E+8 54.3E+8 65.6E+8 78.3E+8 92.4E+8 0.9 67.1E+8 67.1E+8 67.1E+8 67.1E+8 67.1E+8 67.1E+8 67.1E+8 67.1E+8 67.1E+8 67.1E+8 0.7 13.3E+8 18.5E+8 25.2E+8 33.2E+8 42.9E+8 54.2E+8 67.3E+8 82.3E+8 99.1E+8 1.2E+10 0.7 76.1E+8 76.1E+8 76.1E+8 76.1E+8 76.1E+8 76.1E+8 76.1E+8 76.1E+8 76.1E+8 76.1E+8 0.5 14.7E+8 21.5E+8 30.3E+8 41.4E+8 54.9E+8 71.0E+8 89.7E+8 1.1E+10 1.4E+10 1.6E+10 0.5 87.6E+8 87.6E+8 87.6E+8 87.6E+8 87.6E+8 87.6E+8 87.6E+8 87.6E+8 87.6E+8 87.6E+8 0.3 16.6E+8 26.3E+8 39.8E+8 57.4E+8 79.6E+8 1.1E+10 1.4E+10 1.8E+10 2.2E+10 2.7E+10 0.3 1.0E+10 1.0E+10 1.0E+10 1.0E+10 1.0E+10 1.0E+10 1.0E+10 1.0E+10 1.0E+10 1.0E+10 0.1 17.8E+8 35.9E+8 65.8E+8 1.1E+10 1.7E+10 2.5E+10 3.5E+10 4.7E+10 6.2E+10 7.8E+10 0.1 1.2E+10 1.2E+10 1.2E+10 1.2E+10 1.2E+10 1.2E+10 1.2E+10 1.2E+10 1.2E+10 1.2E+10 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 1.9 10.2E-4 9.5E-4 8.9E-4 8.3E-4 7.8E-4 7.4E-4 7.0E-4 6.7E-4 6.3E-4 6.1E-4 1.9 0.0948 0.0948 0.0948 0.0948 0.0948 0.0948 0.0948 0.0948 0.0948 0.0948 1.7 10.1E-4 9.4E-4 8.8E-4 8.2E-4 7.7E-4 7.2E-4 6.8E-4 6.5E-4 6.2E-4 5.9E-4 1.7 0.1036 0.1036 0.1036 0.1036 0.1036 0.1036 0.1036 0.1036 0.1036 0.1036 1.5 10.1E-4 9.3E-4 8.7E-4 8.1E-4 7.6E-4 7.1E-4 6.7E-4 6.3E-4 6.0E-4 5.7E-4 1.5 0.1143 0.1143 0.1143 0.1143 0.1143 0.1143 0.1143 0.1143 0.1143 0.1143 1.3 10.2E-4 9.3E-4 8.6E-4 8.0E-4 7.5E-4 7.0E-4 6.6E-4 6.2E-4 5.9E-4 5.6E-4 1.3 0.1274 0.1274 0.1274 0.1274 0.1274 0.1274 0.1274 0.1274 0.1274 0.1274 1.1 10.3E-4 9.4E-4 8.6E-4 8.0E-4 7.4E-4 6.9E-4 6.5E-4 6.1E-4 5.8E-4 5.5E-4 1.1 0.1439 0.1439 0.1439 0.1439 0.1439 0.1439 0.1439 0.1439 0.1439 0.1439 0.9 10.5E-4 9.5E-4 8.7E-4 8.0E-4 7.4E-4 6.8E-4 6.4E-4 6.0E-4 5.7E-4 5.4E-4 0.9 0.1653 0.1653 0.1653 0.1653 0.1653 0.1653 0.1653 0.1653 0.1653 0.1653 0.7 10.9E-4 9.7E-4 8.8E-4 8.0E-4 7.4E-4 6.8E-4 6.3E-4 5.9E-4 5.6E-4 5.3E-4 0.7 0.1942 0.1942 0.1942 0.1942 0.1942 0.1942 0.1942 0.1942 0.1942 0.1942 0.5 11.5E-4 10.1E-4 9.1E-4 8.2E-4 7.4E-4 6.8E-4 6.3E-4 5.9E-4 5.5E-4 5.2E-4 0.5 0.2353 0.2353 0.2353 0.2353 0.2353 0.2353 0.2353 0.2353 0.2353 0.2353 0.3 12.9E-4 11.0E-4 9.6E-4 8.5E-4 7.6E-4 6.9E-4 6.4E-4 5.9E-4 5.5E-4 5.1E-4 0.3 0.2985 0.2985 0.2985 0.2985 0.2985 0.2985 0.2985 0.2985 0.2985 0.2985 0.1 17.8E-4 14.1E-4 11.5E-4 9.7E-4 8.4E-4 7.4E-4 6.6E-4 6.0E-4 5.5E-4 5.1E-4 0.1 0.4082 0.4082 0.4082 0.4082 0.4082 0.4082 0.4082 0.4082 0.4082 0.4082 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 1.9 18.6E+8 23.1E+8 28.3E+8 34.4E+8 41.3E+8 49.1E+8 57.7E+8 67.4E+8 78.0E+8 89.7E+8 1.9 9.2E+8 9.2E+8 9.2E+8 9.2E+8 9.2E+8 9.2E+8 9.2E+8 9.2E+8 9.2E+8 9.2E+8 1.7 19.3E+8 24.3E+8 30.2E+8 37.0E+8 44.7E+8 53.5E+8 63.4E+8 74.4E+8 86.5E+8 99.8E+8 1.7 10.1E+8 10.1E+8 10.1E+8 10.1E+8 10.1E+8 10.1E+8 10.1E+8 10.1E+8 10.1E+8 10.1E+8 1.5 20.2E+8 25.8E+8 32.5E+8 40.1E+8 49.0E+8 59.1E+8 70.4E+8 83.1E+8 97.2E+8 1.1E+10 1.5 11.1E+8 11.1E+8 11.1E+8 11.1E+8 11.1E+8 11.1E+8 11.1E+8 11.1E+8 11.1E+8 11.1E+8 1.3 21.3E+8 27.7E+8 35.2E+8 44.1E+8 54.4E+8 66.2E+8 79.5E+8 94.5E+8 1.1E+10 1.3E+10 1.3 12.4E+8 12.4E+8 12.4E+8 12.4E+8 12.4E+8 12.4E+8 12.4E+8 12.4E+8 12.4E+8 12.4E+8 1.1 22.7E+8 30.0E+8 38.8E+8 49.3E+8 61.6E+8 75.6E+8 91.6E+8 1.1E+10 1.3E+10 1.5E+10 1.1 14.0E+8 14.0E+8 14.0E+8 14.0E+8 14.0E+8 14.0E+8 14.0E+8 14.0E+8 14.0E+8 14.0E+8 0.9 24.4E+8 33.0E+8 43.6E+8 56.3E+8 71.4E+8 88.8E+8 1.1E+10 1.3E+10 1.6E+10 1.8E+10 0.9 16.1E+8 16.1E+8 16.1E+8 16.1E+8 16.1E+8 16.1E+8 16.1E+8 16.1E+8 16.1E+8 16.1E+8 0.7 26.5E+8 37.1E+8 50.4E+8 66.5E+8 85.9E+8 1.1E+10 1.3E+10 1.6E+10 2.0E+10 2.4E+10 0.7 18.9E+8 18.9E+8 18.9E+8 18.9E+8 18.9E+8 18.9E+8 18.9E+8 18.9E+8 18.9E+8 18.9E+8 0.5 29.4E+8 43.1E+8 60.8E+8 82.9E+8 1.1E+10 1.4E+10 1.8E+10 2.2E+10 2.7E+10 3.3E+10 0.5 22.9E+8 22.9E+8 22.9E+8 22.9E+8 22.9E+8 22.9E+8 22.9E+8 22.9E+8 22.9E+8 22.9E+8 0.3 33.2E+8 52.7E+8 79.7E+8 1.1E+10 1.6E+10 2.1E+10 2.8E+10 3.5E+10 4.4E+10 5.4E+10 0.3 29.1E+8 29.1E+8 29.1E+8 29.1E+8 29.1E+8 29.1E+8 29.1E+8 29.1E+8 29.1E+8 29.1E+8 0.1 35.7E+8 71.8E+8 1.3E+10 2.2E+10 3.4E+10 5.0E+10 7.0E+10 9.5E+10 ###### ###### 0.1 39.7E+8 39.7E+8 39.7E+8 39.7E+8 39.7E+8 39.7E+8 39.7E+8 39.7E+8 39.7E+8 39.7E+8 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 1.9 59429 66253 73415 80877 88609 96585 104784 113188 121801 130558 1.9 139517 139517 139517 139517 139517 139517 139517 139517 139517 139517 1.7 57353 64261 71652 79341 87229 95442 103874 112479 121310 130313 1.7 139517 139517 139517 139517 139517 139517 139517 139517 139517 139517 1.5 55100 62274 69815 77606 85773 94198 102822 111705 120784 130045 1.5 139518 139518 139518 139518 139518 139518 139518 139518 139518 139518 1.3 52643 60030 67718 75768 84136 92789 101728 110860 120206 129748 1.3 139518 139518 139518 139518 139518 139518 139518 139518 139518 139518 1.1 49997 57494 65362 73684 82334 91242 100453 109901 119553 129419 1.1 139518 139518 139518 139518 139518 139518 139518 139518 139518 139518 0.9 46868 54563 62685 71237 80194 89435 98966 108771 118793 129035 0.9 139518 139518 139518 139518 139518 139518 139518 139518 139518 139518 0.7 43140 51000 59424 68283 77582 87207 97164 107393 117867 128566 0.7 139518 139518 139518 139518 139518 139518 139518 139518 139518 139518 0.5 38369 46441 55159 64437 74179 84317 94810 105600 116655 127952 0.5 139518 139518 139518 139518 139518 139518 139518 139518 139518 139518 0.3 31576 39818 48955 58767 69168 80038 91313 102935 114856 127042 0.3 139518 139518 139518 139518 139518 139518 139518 139518 139518 139518 0.1 18932 26842 36331 47061 58709 71060 83968 97330 111067 125125 0.1 139519 139519 139519 139519 139519 139519 139519 139519 139519 139519 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 1.9 1029 1106 1184 1263 1342 1421 1501 1580 1659 1737 1.9 10 10 10 10 10 10 10 10 10 10 1.7 1158 1250 1342 1437 1530 1625 1719 1812 1906 1999 1.7 10 10 10 10 10 10 10 10 10 10 1.5 1313 1425 1537 1650 1763 1877 1990 2102 2214 2326 1.5 10 10 10 10 10 10 10 10 10 10 1.3 1510 1646 1784 1922 2061 2200 2338 2476 2613 2750 1.3 10 10 10 10 10 10 10 10 10 10 1.1 1764 1936 2110 2283 2458 2632 2806 2979 3151 3322 1.1 10 10 10 10 10 10 10 10 10 10 0.9 2112 2336 2562 2790 3019 3246 3472 3698 3921 4143 0.9 10 10 10 10 10 10 10 10 10 10 0.7 2623 2932 3246 3561 3877 4190 4503 4813 5121 5426 0.7 10 10 10 10 10 10 10 10 10 10 0.5 3464 3934 4411 4891 5372 5850 6325 6796 7261 7722 0.5 10 10 10 10 10 10 10 10 10 10 0.3 5167 6030 6917 7811 8706 9594 10474 11343 12202 13049 0.3 10 10 10 10 10 10 10 10 10 10 0.1 11227 14159 17320 20571 23832 27061 30241 33365 36430 39439 0.1 10 10 10 10 10 10 10 10 10 10 Colors indicate whether the value is or than its counterpart under the alternative management regime LOWER HIGHER Substitutability Prices Substitutability Quantities Substitutability Qualities Substitutability Varieties Substitutability Interference Elasticity Prices Interference Elasticity Interference Elasticity Quantities Interference Elasticity Substitutability Interference Elasticity Qualities Interference Elasticity Varieties Interference Elasticity Substitutability Substitutability Substitutability Substitutability Interference Elasticity Interference Elasticity Interference Elasticity LICENSING COMMONS Cournot Equilibria when native SNR is 70dB Figure C.20: Cournot Equilibria when Native SNR is 70dB 104 Consumer Consumer Surplus 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 Surplus 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 1.9 15.3E+8 18.4E+8 21.9E+8 25.9E+8 30.2E+8 35.0E+8 40.4E+8 46.2E+8 52.5E+8 59.4E+8 1.9 54.5E+8 54.5E+8 54.5E+8 54.5E+8 54.5E+8 54.5E+8 54.5E+8 54.5E+8 54.5E+8 54.5E+8 1.7 16.2E+8 19.6E+8 23.6E+8 28.0E+8 32.9E+8 38.4E+8 44.4E+8 51.1E+8 58.3E+8 66.2E+8 1.7 59.0E+8 59.0E+8 59.0E+8 59.0E+8 59.0E+8 59.0E+8 59.0E+8 59.0E+8 59.0E+8 59.0E+8 1.5 17.2E+8 21.1E+8 25.6E+8 30.6E+8 36.3E+8 42.6E+8 49.5E+8 57.2E+8 65.6E+8 74.7E+8 1.5 64.3E+8 64.3E+8 64.3E+8 64.3E+8 64.3E+8 64.3E+8 64.3E+8 64.3E+8 64.3E+8 64.3E+8 1.3 18.5E+8 23.0E+8 28.1E+8 33.9E+8 40.5E+8 47.9E+8 56.1E+8 65.1E+8 75.1E+8 85.9E+8 1.3 70.7E+8 70.7E+8 70.7E+8 70.7E+8 70.7E+8 70.7E+8 70.7E+8 70.7E+8 70.7E+8 70.7E+8 1.1 20.1E+8 25.3E+8 31.3E+8 38.3E+8 46.1E+8 55.0E+8 64.9E+8 75.9E+8 87.9E+8 1.0E+10 1.1 78.3E+8 78.3E+8 78.3E+8 78.3E+8 78.3E+8 78.3E+8 78.3E+8 78.3E+8 78.3E+8 78.3E+8 0.9 22.2E+8 28.4E+8 35.8E+8 44.3E+8 54.0E+8 65.0E+8 77.4E+8 91.1E+8 1.1E+10 1.2E+10 0.9 87.7E+8 87.7E+8 87.7E+8 87.7E+8 87.7E+8 87.7E+8 87.7E+8 87.7E+8 87.7E+8 87.7E+8 0.7 25.1E+8 32.9E+8 42.2E+8 53.1E+8 65.8E+8 80.1E+8 96.4E+8 1.1E+10 1.3E+10 1.6E+10 0.7 99.5E+8 99.5E+8 99.5E+8 99.5E+8 99.5E+8 99.5E+8 99.5E+8 99.5E+8 99.5E+8 99.5E+8 0.5 29.3E+8 39.8E+8 52.5E+8 67.7E+8 85.6E+8 1.1E+10 1.3E+10 1.6E+10 1.9E+10 2.2E+10 0.5 1.1E+10 1.1E+10 1.1E+10 1.1E+10 1.1E+10 1.1E+10 1.1E+10 1.1E+10 1.1E+10 1.1E+10 0.3 36.3E+8 52.4E+8 72.6E+8 97.5E+8 1.3E+10 1.6E+10 2.0E+10 2.5E+10 3.0E+10 3.6E+10 0.3 1.3E+10 1.3E+10 1.3E+10 1.3E+10 1.3E+10 1.3E+10 1.3E+10 1.3E+10 1.3E+10 1.3E+10 0.1 51.0E+8 87.2E+8 1.4E+10 2.1E+10 2.9E+10 4.0E+10 5.3E+10 6.8E+10 8.5E+10 ###### 0.1 1.5E+10 1.5E+10 1.5E+10 1.5E+10 1.5E+10 1.5E+10 1.5E+10 1.5E+10 1.5E+10 1.5E+10 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 1.9 8.6E-4 8.1E-4 7.7E-4 7.3E-4 6.9E-4 6.6E-4 6.3E-4 6.0E-4 5.7E-4 5.5E-4 1.9 0.0948 0.0948 0.0948 0.0948 0.0948 0.0948 0.0948 0.0948 0.0948 0.0948 1.7 8.5E-4 8.0E-4 7.5E-4 7.1E-4 6.8E-4 6.4E-4 6.1E-4 5.8E-4 5.6E-4 5.4E-4 1.7 0.1036 0.1036 0.1036 0.1036 0.1036 0.1036 0.1036 0.1036 0.1036 0.1036 1.5 8.5E-4 7.9E-4 7.4E-4 7.0E-4 6.6E-4 6.3E-4 6.0E-4 5.7E-4 5.4E-4 5.2E-4 1.5 0.1143 0.1143 0.1143 0.1143 0.1143 0.1143 0.1143 0.1143 0.1143 0.1143 1.3 8.5E-4 7.9E-4 7.4E-4 6.9E-4 6.5E-4 6.2E-4 5.9E-4 5.6E-4 5.3E-4 5.1E-4 1.3 0.1274 0.1274 0.1274 0.1274 0.1274 0.1274 0.1274 0.1274 0.1274 0.1274 1.1 8.5E-4 7.9E-4 7.3E-4 6.9E-4 6.5E-4 6.1E-4 5.8E-4 5.5E-4 5.2E-4 5.0E-4 1.1 0.1439 0.1439 0.1439 0.1439 0.1439 0.1439 0.1439 0.1439 0.1439 0.1439 0.9 8.6E-4 7.9E-4 7.3E-4 6.8E-4 6.4E-4 6.0E-4 5.7E-4 5.4E-4 5.1E-4 4.9E-4 0.9 0.1653 0.1653 0.1653 0.1653 0.1653 0.1653 0.1653 0.1653 0.1653 0.1653 0.7 8.8E-4 8.0E-4 7.4E-4 6.9E-4 6.4E-4 6.0E-4 5.6E-4 5.3E-4 5.0E-4 4.8E-4 0.7 0.1942 0.1942 0.1942 0.1942 0.1942 0.1942 0.1942 0.1942 0.1942 0.1942 0.5 9.2E-4 8.3E-4 7.5E-4 6.9E-4 6.4E-4 6.0E-4 5.6E-4 5.3E-4 5.0E-4 4.7E-4 0.5 0.2353 0.2353 0.2353 0.2353 0.2353 0.2353 0.2353 0.2353 0.2353 0.2353 0.3 9.9E-4 8.8E-4 7.9E-4 7.1E-4 6.5E-4 6.0E-4 5.6E-4 5.2E-4 4.9E-4 4.6E-4 0.3 0.2985 0.2985 0.2985 0.2985 0.2985 0.2985 0.2985 0.2985 0.2985 0.2985 0.1 12.5E-4 10.5E-4 9.0E-4 7.9E-4 7.0E-4 6.3E-4 5.8E-4 5.3E-4 4.9E-4 4.6E-4 0.1 0.4082 0.4082 0.4082 0.4082 0.4082 0.4082 0.4082 0.4082 0.4082 0.4082 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 1.9 30.7E+8 36.8E+8 43.8E+8 51.7E+8 60.5E+8 70.1E+8 80.8E+8 92.5E+8 1.1E+10 1.2E+10 1.9 12.0E+8 12.0E+8 12.0E+8 12.0E+8 12.0E+8 12.0E+8 12.0E+8 12.0E+8 12.0E+8 12.0E+8 1.7 32.4E+8 39.3E+8 47.1E+8 56.0E+8 65.8E+8 76.8E+8 88.9E+8 1.0E+10 1.2E+10 1.3E+10 1.7 13.2E+8 13.2E+8 13.2E+8 13.2E+8 13.2E+8 13.2E+8 13.2E+8 13.2E+8 13.2E+8 13.2E+8 1.5 34.5E+8 42.3E+8 51.2E+8 61.2E+8 72.6E+8 85.2E+8 99.1E+8 1.1E+10 1.3E+10 1.5E+10 1.5 14.5E+8 14.5E+8 14.5E+8 14.5E+8 14.5E+8 14.5E+8 14.5E+8 14.5E+8 14.5E+8 14.5E+8 1.3 37.0E+8 46.0E+8 56.2E+8 67.9E+8 81.1E+8 95.9E+8 1.1E+10 1.3E+10 1.5E+10 1.7E+10 1.3 16.2E+8 16.2E+8 16.2E+8 16.2E+8 16.2E+8 16.2E+8 16.2E+8 16.2E+8 16.2E+8 16.2E+8 1.1 40.3E+8 50.6E+8 62.7E+8 76.6E+8 92.4E+8 1.1E+10 1.3E+10 1.5E+10 1.8E+10 2.0E+10 1.1 18.3E+8 18.3E+8 18.3E+8 18.3E+8 18.3E+8 18.3E+8 18.3E+8 18.3E+8 18.3E+8 18.3E+8 0.9 44.4E+8 56.9E+8 71.6E+8 88.6E+8 1.1E+10 1.3E+10 1.5E+10 1.8E+10 2.1E+10 2.5E+10 0.9 21.0E+8 21.0E+8 21.0E+8 21.0E+8 21.0E+8 21.0E+8 21.0E+8 21.0E+8 21.0E+8 21.0E+8 0.7 50.2E+8 65.8E+8 84.5E+8 1.1E+10 1.3E+10 1.6E+10 1.9E+10 2.3E+10 2.7E+10 3.1E+10 0.7 24.7E+8 24.7E+8 24.7E+8 24.7E+8 24.7E+8 24.7E+8 24.7E+8 24.7E+8 24.7E+8 24.7E+8 0.5 58.7E+8 79.7E+8 1.1E+10 1.4E+10 1.7E+10 2.1E+10 2.6E+10 3.1E+10 3.7E+10 4.4E+10 0.5 29.9E+8 29.9E+8 29.9E+8 29.9E+8 29.9E+8 29.9E+8 29.9E+8 29.9E+8 29.9E+8 29.9E+8 0.3 72.8E+8 1.0E+10 1.5E+10 2.0E+10 2.5E+10 3.2E+10 4.1E+10 5.0E+10 6.0E+10 7.2E+10 0.3 37.9E+8 37.9E+8 37.9E+8 37.9E+8 37.9E+8 37.9E+8 37.9E+8 37.9E+8 37.9E+8 37.9E+8 0.1 1.0E+10 1.7E+10 2.8E+10 4.1E+10 5.9E+10 8.0E+10 ###### ###### ###### ###### 0.1 51.9E+8 51.9E+8 51.9E+8 51.9E+8 51.9E+8 51.9E+8 51.9E+8 51.9E+8 51.9E+8 51.9E+8 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 1.9 76393 83696 91275 99187 107228 115481 123970 132586 141372 150325 1.9 159449 159449 159449 159449 159449 159449 159449 159449 159449 159449 1.7 74241 81765 89563 97609 105822 114316 123001 131864 140893 150076 1.7 159449 159449 159449 159449 159449 159449 159449 159449 159449 159449 1.5 71991 79670 87645 95888 104373 113081 121994 131095 140371 149810 1.5 159449 159449 159449 159449 159449 159449 159449 159449 159449 159449 1.3 69417 77333 85486 93990 102696 111681 120844 130232 139794 149515 1.3 159449 159449 159449 159449 159449 159449 159449 159449 159449 159449 1.1 66599 74680 83104 91834 100837 110089 119566 129267 139137 149184 1.1 159449 159449 159449 159449 159449 159449 159449 159449 159449 159449 0.9 63281 71614 80319 89350 98674 108260 118085 128127 138370 148800 0.9 159449 159449 159449 159449 159449 159449 159449 159449 159449 159449 0.7 59340 67917 76955 86320 96035 106013 116266 126749 137442 148331 0.7 159450 159450 159450 159450 159450 159450 159450 159450 159450 159450 0.5 54207 63153 72554 82376 92579 103096 113894 124943 136226 147716 0.5 159450 159450 159450 159450 159450 159450 159450 159450 159450 159450 0.3 46766 56127 66086 76557 87471 98754 110366 122267 134424 146804 0.3 159450 159450 159450 159450 159450 159450 159450 159450 159450 159450 0.1 32001 41831 52721 64423 76772 89649 102957 116632 130622 144884 0.1 159450 159450 159450 159450 159450 159450 159450 159450 159450 159450 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 1.9 1216 1292 1370 1447 1524 1601 1678 1755 1832 1908 1.9 10 10 10 10 10 10 10 10 10 10 1.7 1375 1466 1557 1649 1740 1832 1924 2015 2106 2196 1.7 10 10 10 10 10 10 10 10 10 10 1.5 1569 1679 1789 1899 2009 2119 2229 2339 2447 2556 1.5 10 10 10 10 10 10 10 10 10 10 1.3 1813 1948 2083 2218 2353 2488 2622 2756 2889 3022 1.3 10 10 10 10 10 10 10 10 10 10 1.1 2135 2304 2474 2644 2814 2983 3151 3319 3486 3652 1.1 10 10 10 10 10 10 10 10 10 10 0.9 2580 2800 3022 3244 3466 3686 3906 4124 4341 4556 0.9 10 10 10 10 10 10 10 10 10 10 0.7 3243 3550 3856 4162 4468 4772 5075 5375 5673 5968 0.7 10 10 10 10 10 10 10 10 10 10 0.5 4360 4826 5293 5760 6226 6688 7147 7601 8051 8497 0.5 10 10 10 10 10 10 10 10 10 10 0.3 6710 7576 8446 9314 10178 11034 11882 12720 13549 14368 0.3 10 10 10 10 10 10 10 10 10 10 0.1 15916 19020 22185 25351 28490 31588 34637 37636 40584 43484 0.1 10 10 10 10 10 10 10 10 10 10 Colors indicate whether the value is or than its counterpart under the alternative management regime LOWER HIGHER Substitutability Prices Substitutability Quantities Substitutability Qualities Substitutability Varieties Substitutability Interference Elasticity Prices Interference Elasticity Interference Elasticity Quantities Interference Elasticity Substitutability Interference Elasticity Qualities Interference Elasticity Varieties Interference Elasticity Substitutability Substitutability Substitutability Substitutability Interference Elasticity Interference Elasticity Interference Elasticity LICENSING COMMONS Cournot Equilibria when native SNR is 80dB Figure C.21: Cournot Equilibria when Native SNR is 80dB 105 Consumer Consumer Surplus 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 Surplus 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 1.9 23.2E+8 27.1E+8 31.6E+8 36.4E+8 41.8E+8 47.6E+8 53.9E+8 60.8E+8 68.1E+8 76.1E+8 1.9 69.0E+8 69.0E+8 69.0E+8 69.0E+8 69.0E+8 69.0E+8 69.0E+8 69.0E+8 69.0E+8 69.0E+8 1.7 24.7E+8 29.1E+8 34.1E+8 39.6E+8 45.6E+8 52.2E+8 59.5E+8 67.3E+8 75.7E+8 84.8E+8 1.7 74.7E+8 74.7E+8 74.7E+8 74.7E+8 74.7E+8 74.7E+8 74.7E+8 74.7E+8 74.7E+8 74.7E+8 1.5 26.6E+8 31.6E+8 37.3E+8 43.5E+8 50.5E+8 58.1E+8 66.4E+8 75.4E+8 85.2E+8 95.8E+8 1.5 81.4E+8 81.4E+8 81.4E+8 81.4E+8 81.4E+8 81.4E+8 81.4E+8 81.4E+8 81.4E+8 81.4E+8 1.3 28.9E+8 34.7E+8 41.3E+8 48.5E+8 56.7E+8 65.6E+8 75.4E+8 86.1E+8 97.6E+8 1.1E+10 1.3 89.4E+8 89.4E+8 89.4E+8 89.4E+8 89.4E+8 89.4E+8 89.4E+8 89.4E+8 89.4E+8 89.4E+8 1.1 31.9E+8 38.7E+8 46.5E+8 55.2E+8 64.9E+8 75.7E+8 87.5E+8 1.0E+10 1.1E+10 1.3E+10 1.1 99.1E+8 99.1E+8 99.1E+8 99.1E+8 99.1E+8 99.1E+8 99.1E+8 99.1E+8 99.1E+8 99.1E+8 0.9 35.9E+8 44.2E+8 53.7E+8 64.4E+8 76.4E+8 89.8E+8 1.0E+10 1.2E+10 1.4E+10 1.6E+10 0.9 1.1E+10 1.1E+10 1.1E+10 1.1E+10 1.1E+10 1.1E+10 1.1E+10 1.1E+10 1.1E+10 1.1E+10 0.7 41.6E+8 52.1E+8 64.3E+8 78.2E+8 93.9E+8 1.1E+10 1.3E+10 1.5E+10 1.8E+10 2.0E+10 0.7 1.3E+10 1.3E+10 1.3E+10 1.3E+10 1.3E+10 1.3E+10 1.3E+10 1.3E+10 1.3E+10 1.3E+10 0.5 50.4E+8 64.7E+8 81.6E+8 1.0E+10 1.2E+10 1.5E+10 1.8E+10 2.1E+10 2.4E+10 2.8E+10 0.5 1.4E+10 1.4E+10 1.4E+10 1.4E+10 1.4E+10 1.4E+10 1.4E+10 1.4E+10 1.4E+10 1.4E+10 0.3 66.4E+8 89.3E+8 1.2E+10 1.5E+10 1.9E+10 2.3E+10 2.8E+10 3.3E+10 3.9E+10 4.6E+10 0.3 1.7E+10 1.7E+10 1.7E+10 1.7E+10 1.7E+10 1.7E+10 1.7E+10 1.7E+10 1.7E+10 1.7E+10 0.1 1.1E+10 1.7E+10 2.4E+10 3.4E+10 4.5E+10 5.9E+10 7.4E+10 9.2E+10 ###### ###### 0.1 1.9E+10 1.9E+10 1.9E+10 1.9E+10 1.9E+10 1.9E+10 1.9E+10 1.9E+10 1.9E+10 1.9E+10 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 1.9 7.5E-4 7.2E-4 6.8E-4 6.5E-4 6.2E-4 5.9E-4 5.7E-4 5.5E-4 5.3E-4 5.1E-4 1.9 0.0948 0.0948 0.0948 0.0948 0.0948 0.0948 0.0948 0.0948 0.0948 0.0948 1.7 7.4E-4 7.0E-4 6.7E-4 6.4E-4 6.1E-4 5.8E-4 5.5E-4 5.3E-4 5.1E-4 4.9E-4 1.7 0.1036 0.1036 0.1036 0.1036 0.1036 0.1036 0.1036 0.1036 0.1036 0.1036 1.5 7.4E-4 6.9E-4 6.6E-4 6.2E-4 5.9E-4 5.7E-4 5.4E-4 5.2E-4 5.0E-4 4.8E-4 1.5 0.1143 0.1143 0.1143 0.1143 0.1143 0.1143 0.1143 0.1143 0.1143 0.1143 1.3 7.3E-4 6.9E-4 6.5E-4 6.1E-4 5.8E-4 5.6E-4 5.3E-4 5.1E-4 4.9E-4 4.7E-4 1.3 0.1274 0.1274 0.1274 0.1274 0.1274 0.1274 0.1274 0.1274 0.1274 0.1274 1.1 7.3E-4 6.8E-4 6.4E-4 6.1E-4 5.8E-4 5.5E-4 5.2E-4 5.0E-4 4.8E-4 4.6E-4 1.1 0.1439 0.1439 0.1439 0.1439 0.1439 0.1439 0.1439 0.1439 0.1439 0.1439 0.9 7.3E-4 6.8E-4 6.4E-4 6.0E-4 5.7E-4 5.4E-4 5.1E-4 4.9E-4 4.7E-4 4.5E-4 0.9 0.1653 0.1653 0.1653 0.1653 0.1653 0.1653 0.1653 0.1653 0.1653 0.1653 0.7 7.4E-4 6.9E-4 6.4E-4 6.0E-4 5.7E-4 5.4E-4 5.1E-4 4.8E-4 4.6E-4 4.4E-4 0.7 0.1942 0.1942 0.1942 0.1942 0.1942 0.1942 0.1942 0.1942 0.1942 0.1942 0.5 7.7E-4 7.0E-4 6.5E-4 6.1E-4 5.7E-4 5.3E-4 5.0E-4 4.8E-4 4.5E-4 4.3E-4 0.5 0.2353 0.2353 0.2353 0.2353 0.2353 0.2353 0.2353 0.2353 0.2353 0.2353 0.3 8.1E-4 7.4E-4 6.7E-4 6.2E-4 5.8E-4 5.4E-4 5.0E-4 4.7E-4 4.5E-4 4.3E-4 0.3 0.2985 0.2985 0.2985 0.2985 0.2985 0.2985 0.2985 0.2985 0.2985 0.2985 0.1 9.7E-4 8.4E-4 7.5E-4 6.7E-4 6.1E-4 5.6E-4 5.1E-4 4.8E-4 4.5E-4 4.2E-4 0.1 0.4082 0.4082 0.4082 0.4082 0.4082 0.4082 0.4082 0.4082 0.4082 0.4082 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 1.9 46.4E+8 54.3E+8 63.2E+8 72.8E+8 83.6E+8 95.2E+8 1.1E+10 1.2E+10 1.4E+10 1.5E+10 1.9 15.2E+8 15.2E+8 15.2E+8 15.2E+8 15.2E+8 15.2E+8 15.2E+8 15.2E+8 15.2E+8 15.2E+8 1.7 49.4E+8 58.3E+8 68.2E+8 79.2E+8 91.3E+8 1.0E+10 1.2E+10 1.3E+10 1.5E+10 1.7E+10 1.7 16.7E+8 16.7E+8 16.7E+8 16.7E+8 16.7E+8 16.7E+8 16.7E+8 16.7E+8 16.7E+8 16.7E+8 1.5 53.2E+8 63.2E+8 74.6E+8 87.1E+8 1.0E+10 1.2E+10 1.3E+10 1.5E+10 1.7E+10 1.9E+10 1.5 18.4E+8 18.4E+8 18.4E+8 18.4E+8 18.4E+8 18.4E+8 18.4E+8 18.4E+8 18.4E+8 18.4E+8 1.3 57.9E+8 69.5E+8 82.6E+8 97.1E+8 1.1E+10 1.3E+10 1.5E+10 1.7E+10 2.0E+10 2.2E+10 1.3 20.5E+8 20.5E+8 20.5E+8 20.5E+8 20.5E+8 20.5E+8 20.5E+8 20.5E+8 20.5E+8 20.5E+8 1.1 63.9E+8 77.5E+8 93.0E+8 1.1E+10 1.3E+10 1.5E+10 1.8E+10 2.0E+10 2.3E+10 2.6E+10 1.1 23.1E+8 23.1E+8 23.1E+8 23.1E+8 23.1E+8 23.1E+8 23.1E+8 23.1E+8 23.1E+8 23.1E+8 0.9 71.9E+8 88.4E+8 1.1E+10 1.3E+10 1.5E+10 1.8E+10 2.1E+10 2.4E+10 2.8E+10 3.2E+10 0.9 26.6E+8 26.6E+8 26.6E+8 26.6E+8 26.6E+8 26.6E+8 26.6E+8 26.6E+8 26.6E+8 26.6E+8 0.7 83.2E+8 1.0E+10 1.3E+10 1.6E+10 1.9E+10 2.2E+10 2.6E+10 3.1E+10 3.5E+10 4.0E+10 0.7 31.2E+8 31.2E+8 31.2E+8 31.2E+8 31.2E+8 31.2E+8 31.2E+8 31.2E+8 31.2E+8 31.2E+8 0.5 1.0E+10 1.3E+10 1.6E+10 2.0E+10 2.5E+10 3.0E+10 3.5E+10 4.2E+10 4.9E+10 5.6E+10 0.5 37.9E+8 37.9E+8 37.9E+8 37.9E+8 37.9E+8 37.9E+8 37.9E+8 37.9E+8 37.9E+8 37.9E+8 0.3 1.3E+10 1.8E+10 2.3E+10 3.0E+10 3.7E+10 4.6E+10 5.6E+10 6.7E+10 7.9E+10 9.2E+10 0.3 48.0E+8 48.0E+8 48.0E+8 48.0E+8 48.0E+8 48.0E+8 48.0E+8 48.0E+8 48.0E+8 48.0E+8 0.1 2.2E+10 3.4E+10 4.9E+10 6.8E+10 9.1E+10 ###### ###### ###### ###### ###### 0.1 65.7E+8 65.7E+8 65.7E+8 65.7E+8 65.7E+8 65.7E+8 65.7E+8 65.7E+8 65.7E+8 65.7E+8 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 1.9 93920 101589 109584 117689 126057 134536 143228 152034 161012 170106 1.9 179381 179381 179381 179381 179381 179381 179381 179381 179381 179381 1.7 91700 99600 107732 116078 124621 133348 142277 151325 160523 169862 1.7 179381 179381 179381 179381 179381 179381 179381 179381 179381 179381 1.5 89345 97424 105822 114368 123125 132078 141240 150535 159988 169597 1.5 179381 179381 179381 179381 179381 179381 179381 179381 179381 179381 1.3 86771 95090 103665 112425 121456 130684 140094 149674 159413 169301 1.3 179381 179381 179381 179381 179381 179381 179381 179381 179381 179381 1.1 83890 92400 101197 110294 119580 129109 138805 148702 158761 168973 1.1 179381 179381 179381 179381 179381 179381 179381 179381 179381 179381 0.9 80478 89266 98403 107769 117386 127254 137322 147564 157994 168584 0.9 179381 179381 179381 179381 179381 179381 179381 179381 179381 179381 0.7 76364 85484 94938 104692 114714 124997 135494 146180 157062 168115 0.7 179381 179381 179381 179381 179381 179381 179381 179381 179381 179381 0.5 71037 80545 90439 100690 111235 122046 133104 144377 155847 167500 0.5 179381 179381 179381 179381 179381 179381 179381 179381 179381 179381 0.3 63210 73278 83814 94763 106059 117669 129555 141686 154038 166586 0.3 179382 179382 179382 179382 179382 179382 179382 179382 179382 179382 0.1 47258 58264 70016 82365 95210 108476 122098 136029 150228 164665 0.1 179382 179382 179382 179382 179382 179382 179382 179382 179382 179382 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 1.9 1395 1470 1546 1621 1697 1772 1848 1923 1998 2072 1.9 10 10 10 10 10 10 10 10 10 10 1.7 1582 1671 1761 1851 1940 2030 2119 2208 2297 2385 1.7 10 10 10 10 10 10 10 10 10 10 1.5 1812 1920 2028 2135 2243 2350 2458 2564 2670 2776 1.5 10 10 10 10 10 10 10 10 10 10 1.3 2103 2236 2368 2500 2631 2763 2894 3024 3153 3282 1.3 10 10 10 10 10 10 10 10 10 10 1.1 2489 2655 2820 2987 3152 3317 3481 3644 3806 3968 1.1 10 10 10 10 10 10 10 10 10 10 0.9 3026 3242 3459 3675 3890 4105 4319 4531 4741 4951 0.9 10 10 10 10 10 10 10 10 10 10 0.7 3836 4135 4434 4733 5030 5325 5619 5910 6200 6487 0.7 10 10 10 10 10 10 10 10 10 10 0.5 5219 5675 6130 6584 7035 7483 7928 8369 8806 9239 0.5 10 10 10 10 10 10 10 10 10 10 0.3 8199 9047 9893 10735 11571 12400 13220 14032 14835 15630 0.3 10 10 10 10 10 10 10 10 10 10 0.1 20628 23712 26796 29855 32880 35862 38801 41695 44545 47352 0.1 10 10 10 10 10 10 10 10 10 10 Colors indicate whether the value is or than its counterpart under the alternative management regime LOWER HIGHER Substitutability Prices Substitutability Quantities Substitutability Qualities Substitutability Varieties Substitutability Interference Elasticity Prices Interference Elasticity Interference Elasticity Quantities Interference Elasticity Substitutability Interference Elasticity Qualities Interference Elasticity Varieties Interference Elasticity Substitutability Substitutability Substitutability Substitutability Interference Elasticity Interference Elasticity Interference Elasticity LICENSING COMMONS Cournot Equilibria when native SNR is 90dB Figure C.22: Cournot Equilibria when Native SNR is 90dB 106
Abstract (if available)
Abstract
Radio spectrum is a fundamental input in communications technology and the allocation of spectrum has gained ever increasing importance with technological advances. In this dissertation we start by recoupling the accurate scientific definition and progression of spectrum with the economic and policy analysis of spectrum. We then identify and provide insights to the current valuation, allocation and management problems in spectrum equipped with this improved understanding. We go on to provide consumer surplus estimates for unlicensed spectrum that had been overlooked in the literature and show that the consumer surplus from the utilization of unlicensed spectrum is commensurate with that of licensed spectrum found in previous studies. To assess the possible negative effects of excessive entry under unlicensed allocations, we continue with a model of a communications device market whereby the allocation of spectrum determines the technological environment, the competitive environment and consequently the quality and the variety of the devices in the market. We then go on to simulate the welfare consequences of instituting the alternative management regimes. We conclude that the regime choice has to be thoroughly informed by the preference and technology structure and the optimal allocation must be an interior solution with a finely tuned mix of the two regimes.
Linked assets
University of Southern California Dissertations and Theses
Conceptually similar
PDF
Essays on the economics of non-profit institutions
PDF
Essays on bundling and discounts
PDF
Case studies of national cultural protection in the era of globalization
PDF
Essays on pricing and contracting
PDF
Essays in retail management
PDF
Empirical essays on industrial organization
PDF
Three essays on supply chain networks and R&D investments
PDF
Essays on the properties of financial analysts' forecasts
PDF
Legal institutions of the market economy: private commercial arbitration as transaction costs reducer in Colombia
PDF
Essays on the econometrics of program evaluation
PDF
Essays on the econometrics of program evaluation
PDF
Essays in international economics
PDF
Essays on the economics of subjective well-being in transition countries
PDF
Optimal resource allocation and cross-layer control in cognitive and cooperative wireless networks
PDF
Essays on labor and development economics
PDF
Adaptation, assets, and aspiration. Three essays on the economics of subjective well-being
PDF
Two essays on the impact of exchange rate regime changes in Asia: examples from Thailand and Japan
PDF
Essays on the turning points of the product life cycle
PDF
Essays on urban and real estate economics
PDF
Essays on competition between multiproduct firms
Asset Metadata
Creator
Bayrak, Ergin
(author)
Core Title
Essays on the economics of radio spectrum
School
College of Letters, Arts and Sciences
Degree
Doctor of Philosophy
Degree Program
Economics
Publication Date
11/17/2011
Defense Date
10/28/2009
Publisher
University of Southern California
(original),
University of Southern California. Libraries
(digital)
Tag
differentiated goods,OAI-PMH Harvest,oligopoly,radio spectrum,spectrum allocation,spectrum management
Language
English
Contributor
Electronically uploaded by the author
(provenance)
Advisor
Wilkie, Simon J. (
committee chair
), Aronson, Jonathan (
committee member
), Cheng, Harrison (
committee member
)
Creator Email
ebayrak@usc.edu,erginbayrak@gmail.com
Permanent Link (DOI)
https://doi.org/10.25549/usctheses-m2741
Unique identifier
UC1444453
Identifier
etd-Bayrak-3391 (filename),usctheses-m40 (legacy collection record id),usctheses-c127-278914 (legacy record id),usctheses-m2741 (legacy record id)
Legacy Identifier
etd-Bayrak-3391.pdf
Dmrecord
278914
Document Type
Dissertation
Rights
Bayrak, Ergin
Type
texts
Source
University of Southern California
(contributing entity),
University of Southern California Dissertations and Theses
(collection)
Repository Name
Libraries, University of Southern California
Repository Location
Los Angeles, California
Repository Email
cisadmin@lib.usc.edu
Tags
differentiated goods
oligopoly
radio spectrum
spectrum allocation
spectrum management