Spectrum, Scarcity, and Centralized Control

by on September 5, 2006 · 28 comments

I was generally impressed with Tim Wu’s paper on Hayekian analysis of intellectual property, but I did want to note one place where his analysis goes off the rails:

A second example is broadcast spectrum reform, which has been under consideration for about a decade in the United States. The question is whether broadcasting at certain frequencies should be propertized. In other words, the question is whether some firm should own the alienable rights to broadcast between frequencies X and Y. The impact of the government’s decision whether to grant property rights or not will have important decisional consequences. Granting no rights will create decentralized market entry for spectrum-dependent projects or technologies. Any entity willing to make the investment may develop a project that depends on access to spectrum, albeit at the cost of many failed projects. Granting government-specified licenses or property rights, conversely, makes some kind of hierarchical decision structure possible in the first place. That is, we should expect to see greater screening of spectrum-dependent projects or technologies before they are launched.

Which is better is slightly ambigious. For some uses of spectrum there may be good arguments for a hierarchical, centralized authority who decides what the spectrum will be used for, perhaps to ensure public safety. But otherwise, whether we want propertized spectrum depends on whether there is any argument that spectrum-dependent projects be carefully screened. Absent risk the public, the answer must sometimes be no.

This strikes me as rather misguided. As Jerry has explained in this space before, the difference between spectrum and ideas is that spectrum is rivalrous and scarce, whereas ideas are not. Complete decontrol is never an option–somebody has to pick the rules governing how the resource will be consumed, and the only question is who will make the rules.


So it’s nonsense to claim that only propertized spectrum is subject to a “hierarchical decision structure.” Unlicensed spectrum is subject to the “hierarchical decision structure” of the FCC, which makes various rules about how unlicensed spectrum may be used. So far, most unlicensed spectrum has been subject to limits on the power of transmitters, making unlicensed spectrum extremely useful for short-range transmission technologies like WiFi and cordless phones, but not especially useful for longer-range applications. Conceivably, the government could establish different, more complex rules for other unlicensed spectrum, thereby enabling certain long-range applications to operate in an unlicensed environment.

But if that’s the policy that Wu is proposing, then our choice is not between “hierarchical decisions” made by private companies and decentralization. The choice is simply a matter of who the hierarchical decision maker will be: a private company or the state?

Now, I do think there’s something to be said for having some portion of the spectrum set aside for low-power usage, as has been done for the band used by WiFi. It may be that there are many low-power uses that are not individually lucrative enough to be worth anyone’s trouble to obtain a license, but that collectively are quite valuable. In particular, wouldn’t want to propertize the bands already dedicated to low-power uses like WiFi, which have proven extremely useful.

But this doesn’t work so well for higher-powered transmissions. Hence, the primary reason for advocating that “spectrum-dependent projects be carefully screened” isn’t public safety, it’s interference among uses. There can only be so many transmitters using a particular band, at a particular power level for a particular purpose. Barring a major technological breakthrough, there will be more demand for spectrum than supply for the foreseeable future. Someone is going to have to decide among competing uses. For a variety of reasons that Jerry ably lays out here, the best way for society to make that decision is through the price mechanism. When it comes to scarce, rivalrous resources, a “commons” is just another word for keeping the government in charge.

There’s an interesting symmetry here. Many policy analysts on the right are so used to defending property rights and markets that they go overboard advocating the propertization of everything in sight, including non-scarce intellectual products (such as software patents) that might be better left in the public domain. The left, in contrast, is so infatuated with their responses to this overreach by the right that they go too far in the opposite direction, wrongly asserting that we can dispense with market mechanisms not only for non-scarce ideas, for scarce resources like spectrum too. Both are making the same fundamental error by failing to appreciate the fundamental difference between rivalrous and non-rivalrous resources. Property rights are essential to ensuring the efficient allocation of the former among competing uses. In contrast, property rights in the latter are at best a necessary evil, because the most efficient use of a non-rivalrous resource is to make it as widely available as possible.

  • http://techdirt.com/ Mike Masnick

    Just to throw a wrench into this discussion… there are a number of fairly knowledgable folks who claim that spectrum can be made to be non-rivalrous. I’m not yet convinced of this (because, despite those claims, interference is still an issue). However, I do wonder *if* this does prove true and equipment can be made where interference no longer is a problem… how will that change the policy impact of what you are saying?

    And, if so, is it worth considering it at this time, or do we just try to unravel the mess we created with property rights that weren’t actually needed at a later date?

    Again… I’m not yet convinced that those who claim it are right. However, it does seem like an issue worth thinking about now, rather than later.

  • http://techdirt.com/ Mike Masnick

    Just to throw a wrench into this discussion… there are a number of fairly knowledgable folks who claim that spectrum can be made to be non-rivalrous. I’m not yet convinced of this (because, despite those claims, interference is still an issue). However, I do wonder *if* this does prove true and equipment can be made where interference no longer is a problem… how will that change the policy impact of what you are saying?

    And, if so, is it worth considering it at this time, or do we just try to unravel the mess we created with property rights that weren’t actually needed at a later date?

    Again… I’m not yet convinced that those who claim it are right. However, it does seem like an issue worth thinking about now, rather than later.

  • Steve R.

    Mike: With an ANALOG signal, one user occupies the bandwidth. However, with a DIGITAL signal many users can occupy the same bandwidth. Nevertheless given enough users and data transmission load, I assume (not being an engineer), that a band could eventually be filled-up since capacity is not infinite. At this point in time, I have no idea on whether access to a band would or would not be competitive (based on usage rather than leased for resale).

    Also, as an aside, higher frequencies tend to be in demand since they can carry more information than the lower frequencies.
    BR>

  • http://www.techliberation.com Tim

    Mike,

    I have a feeling that engineers and economists mean slightly different things when they talk about non-rivalrousness. As I understand it, the proposals for non-rivalrous uses of spectrum involve all devices cooperating with each other in particular ways: sharing the bandwidth by time-slices, forming themselves into a mesh network where devices talk to their neighbors, etc. I don’t think it will ever be true that anybody will be able to transmit anything they like at certain frequencies–getting “non-rivalrous” spectrum uses to work will always require everyone following fairly complex rules.

    If that’s true, it means that for the foreseeable future, somebody’s going to have to make the rules. I would rather have a lot of different private companies making those decisions than putting all the decisions in the hands of the FCC’s bureaucracy.

    If, on the other hand, the engineers figure out ways to have genuinely non-rivalrous uses of spectrum, such that the new uses don’t interfere at all with the older uses at all, then that’ll be a whole new ballgame. But in that case, all that will need to happen is for the FCC to exempt the new technologies from the traditional spectrum rules. It might still make sense to have a property-based system to govern the older, rivalrous uses of the spectrum. As I understand it, though, those technologies are still on the drawing board. It would be silly to set aside spectrum for uses that don’t exist yet.

  • Steve R.

    Tim: I agree with your analysis that Wu is off the mark on analyzing the issue of “spectrum, scarcity, and centralized control. On the question of whether “certain frequencies should be propertized” I would like to add that they are already propertized, the radio spectrum is owned by the government on behalf of the citizens.

    Simply because a piece of property is in the public domain does not mean that the government shouldn’t operate it as a business. The government, like any business, can and should use pricing mechanisms as a means of determining who can use scarce rivalrous resources.

  • http://www2.blogger.com/profile/14380731108416527657 Steve R.

    Mike: With an ANALOG signal, one user occupies the bandwidth. However, with a DIGITAL signal many users can occupy the same bandwidth. Nevertheless given enough users and data transmission load, I assume (not being an engineer), that a band could eventually be filled-up since capacity is not infinite. At this point in time, I have no idea on whether access to a band would or would not be competitive (based on usage rather than leased for resale).

    Also, as an aside, higher frequencies tend to be in demand since they can carry more information than the lower frequencies.
    BR>

  • http://www.techliberation.com Tim

    Mike,

    I have a feeling that engineers and economists mean slightly different things when they talk about non-rivalrousness. As I understand it, the proposals for non-rivalrous uses of spectrum involve all devices cooperating with each other in particular ways: sharing the bandwidth by time-slices, forming themselves into a mesh network where devices talk to their neighbors, etc. I don’t think it will ever be true that anybody will be able to transmit anything they like at certain frequencies–getting “non-rivalrous” spectrum uses to work will always require everyone following fairly complex rules.

    If that’s true, it means that for the foreseeable future, somebody’s going to have to make the rules. I would rather have a lot of different private companies making those decisions than putting all the decisions in the hands of the FCC’s bureaucracy.

    If, on the other hand, the engineers figure out ways to have genuinely non-rivalrous uses of spectrum, such that the new uses don’t interfere at all with the older uses at all, then that’ll be a whole new ballgame. But in that case, all that will need to happen is for the FCC to exempt the new technologies from the traditional spectrum rules. It might still make sense to have a property-based system to govern the older, rivalrous uses of the spectrum. As I understand it, though, those technologies are still on the drawing board. It would be silly to set aside spectrum for uses that don’t exist yet.

  • http://www2.blogger.com/profile/14380731108416527657 Steve R.

    Tim: I agree with your analysis that Wu is off the mark on analyzing the issue of “spectrum, scarcity, and centralized control. On the question of whether “certain frequencies should be propertized” I would like to add that they are already propertized, the radio spectrum is owned by the government on behalf of the citizens.

    Simply because a piece of property is in the public domain does not mean that the government shouldn’t operate it as a business. The government, like any business, can and should use pricing mechanisms as a means of determining who can use scarce rivalrous resources.

  • David Evans

    I wanted to post about a company called Cyren Call. It has a different approach to public safety interoperability. Its proposal has gained support from associations like NENA (National Emergency Numbers Association) and APCO (Association of Public-Safety Communications Officials-International).

    Cyren Call’s website says the following about its proposal: “The 30 MHz of spectrum in the 700 MHz band that is currently allocated for commercial use only and which is scheduled to be auctioned in 2008 is ideally suited for the creation of the nationwide, next-generation network for public safety because of the unique physical properties of the spectrum.” Cyren is currently working to create a trust to oversee the 30MHz and use it for public safety use.

    The website is http://www.cyrencall.com. It goes into detail about why the 30 MHz is so ideal for first responders and public safety.

    Do you think this plan could work? What are your thoughts? Have you heard any pros or cons with regard to their proposal?

  • David Evans

    I wanted to post about a company called Cyren Call. It has a different approach to public safety interoperability. Its proposal has gained support from associations like NENA (National Emergency Numbers Association) and APCO (Association of Public-Safety Communications Officials-International).

    Cyren Call’s website says the following about its proposal: “The 30 MHz of spectrum in the 700 MHz band that is currently allocated for commercial use only and which is scheduled to be auctioned in 2008 is ideally suited for the creation of the nationwide, next-generation network for public safety because of the unique physical properties of the spectrum.” Cyren is currently working to create a trust to oversee the 30MHz and use it for public safety use.

    The website is http://www.cyrencall.com. It goes into detail about why the 30 MHz is so ideal for first responders and public safety.

    Do you think this plan could work? What are your thoughts? Have you heard any pros or cons with regard to their proposal?

  • http://bennett.com/blog Richard Bennett

    However, with a DIGITAL signal many users can occupy the same bandwidth.

    Not true, you just get digital interference instead of analog interference. It works the same way – somebody else’s message gets combined with your message and the result is garbage – and it does so because bits are converted to analog before transmission.

    Radio waves are always analog, whether they’re carrying a digital payload or not.

    higher frequencies tend to be in demand since they can carry more information than the lower frequencies.

    Also not true, as higher frequencies suffer more attenuaton with distance.

  • http://bennett.com/blog Richard Bennett

    However, with a DIGITAL signal many users can occupy the same bandwidth.

    Not true, you just get digital interference instead of analog interference. It works the same way – somebody else’s message gets combined with your message and the result is garbage – and it does so because bits are converted to analog before transmission.

    Radio waves are always analog, whether they’re carrying a digital payload or not.

    higher frequencies tend to be in demand since they can carry more information than the lower frequencies.

    Also not true, as higher frequencies suffer more attenuaton with distance.

  • Steve R.

    Richard: Since I am not an engineer, I can’t argue the science, but I disagree.

    1. DIGITAL: Computers can use simultaneous wireless digital connections to access the internet, and as far as I know the wireless connections operate on the same frequency. Also think of how our computers connect to the interet. Our signal is carried by either a copper wire and/or fiber optic cable. Because the signal is digital (packetized), they don’t interfere. In the old days, phones were analog and each phone required its own copper wire to avoid interference.

    2. ANALOG: This one takes a bit of thought. In an overt sense you are correct, but you miss how the information is carried. Take AM radio for example. With a true analog signal the value of the signal is dependent on the detected voltage. With a digital signal the value of the signal is carried by a digitized bit so the receiver does not actually measure the voltage but translates the value of the bit (packet) back into a sound. Additionally, like an ethernet packet, you can add other values to the bit (packet) so that radio “A” receives one signal while radio “B” receives another signal.

    3.HIGHER FREQUENCIES: You are correct to say that higher frequencies suffer from greater attenuation with distance, however that is NOT a reflection of the amount of data it can carry. It simply means that the data can only be carried a short distance. An example of this are microwave relay sites which are used to “refresh” the high capacity signal. Also look at the evolution of cordless phones and cell phones, they have moved up in radio frequency because of better sound quality. Radio has also evolved from low frequency AM radio to the higer frequency FM radio because of better signal capacity.

  • Steve R.

    Richard: I took a quick peek at your blog and was amused by The Original Blog. Periodically, I visit Jerry Pournelle’s website and he has an asertion that his site may be one of the originals. I have no idea of who would be correct. It may make for an interesting story for the forum.

    You guys may very well know each other. His website is http://www.jerrypournelle.com/.

  • http://www2.blogger.com/profile/14380731108416527657 Steve R.

    Richard: Since I am not an engineer, I can’t argue the science, but I disagree.

    1. DIGITAL: Computers can use simultaneous wireless digital connections to access the internet, and as far as I know the wireless connections operate on the same frequency. Also think of how our computers connect to the interet. Our signal is carried by either a copper wire and/or fiber optic cable. Because the signal is digital (packetized), they don’t interfere. In the old days, phones were analog and each phone required its own copper wire to avoid interference.

    2. ANALOG: This one takes a bit of thought. In an overt sense you are correct, but you miss how the information is carried. Take AM radio for example. With a true analog signal the value of the signal is dependent on the detected voltage. With a digital signal the value of the signal is carried by a digitized bit so the receiver does not actually measure the voltage but translates the value of the bit (packet) back into a sound. Additionally, like an ethernet packet, you can add other values to the bit (packet) so that radio “A” receives one signal while radio “B” receives another signal.

    3.HIGHER FREQUENCIES: You are correct to say that higher frequencies suffer from greater attenuation with distance, however that is NOT a reflection of the amount of data it can carry. It simply means that the data can only be carried a short distance. An example of this are microwave relay sites which are used to “refresh” the high capacity signal. Also look at the evolution of cordless phones and cell phones, they have moved up in radio frequency because of better sound quality. Radio has also evolved from low frequency AM radio to the higer frequency FM radio because of better signal capacity.

  • http://www2.blogger.com/profile/14380731108416527657 Steve R.

    Richard: I took a quick peek at your blog and was amused by The Original Blog. Periodically, I visit Jerry Pournelle’s website and he has an asertion that his site may be one of the originals. I have no idea of who would be correct. It may make for an interesting story for the forum.

    You guys may very well know each other. His website is http://www.jerrypournelle.com/.

  • http://bennett.com/blog Richard Bennett

    Steve, you’re confusing “concurrent” use of a common medium with “simultaneous” use. If two stations transmit at the same time on Bob Metcalfe’s Ethernet, their messages collide and neither is successful, and the same thing happens on a WiFi network. But if they transmit their messages concurrently, or one after the other, they can both use the medium successfully.

    The messages on these networks carry addresses, and by convention we don’t look at other people’s addresses. So we can share the medium in an orderly fashion as long as we use the same conventions for access, addressing, and delivery. That’s not “simultaneous” access, it just looks that way.

    And the amount of information an analog signal can carry is a function of the size of the slice of spectrum it uses – bandwidth – not the frequency at which it operates. So the key point here is “large” bandwidth, not “high” bandwidth. This is a common misunderstanding.

    I don’t really have the original blog blog, just the original activist blog; I started one for a fathers’ rights organization before any other activist group was using the web. As that lead in some sense to Daily Kos, it was apparently quite a serious mistake, one for which I will certainly burn in hell for all eternity.

  • http://bennett.com/blog Richard Bennett

    Steve, you’re confusing “concurrent” use of a common medium with “simultaneous” use. If two stations transmit at the same time on Bob Metcalfe’s Ethernet, their messages collide and neither is successful, and the same thing happens on a WiFi network. But if they transmit their messages concurrently, or one after the other, they can both use the medium successfully.

    The messages on these networks carry addresses, and by convention we don’t look at other people’s addresses. So we can share the medium in an orderly fashion as long as we use the same conventions for access, addressing, and delivery. That’s not “simultaneous” access, it just looks that way.

    And the amount of information an analog signal can carry is a function of the size of the slice of spectrum it uses – bandwidth – not the frequency at which it operates. So the key point here is “large” bandwidth, not “high” bandwidth. This is a common misunderstanding.

    I don’t really have the original blog blog, just the original activist blog; I started one for a fathers’ rights organization before any other activist group was using the web. As that lead in some sense to Daily Kos, it was apparently quite a serious mistake, one for which I will certainly burn in hell for all eternity.

  • Steve R.

    Richard: Thanks for responding. I am sure that we will have much to mull over in the future. It’s good that you are supporting father’s rights.

  • http://www2.blogger.com/profile/14380731108416527657 Steve R.

    Richard: Thanks for responding. I am sure that we will have much to mull over in the future. It’s good that you are supporting father’s rights.

  • http://bennett.com/blog Richard Bennett

    Somebody has to.

  • http://bennett.com/blog Richard Bennett

    Somebody has to.

  • Tim Wu

    The point made by Tim Lee is decent. It is certainly true that the FCC would have to state some kind of standard to make possible permissionless entry into the spectrum market (as it does for the garage band used by 802.11b). In addition, private actors could, if they wanted, similarly allow permissionless use of spectrum. The question is why they would want to.

    In general I cannot understand the strength of Jerry’s and others’ objection to the substance of rules that would create permissionless market entry into the spectrum market. In my view, reflected in that paper, permissionless market entry is one of the holy grails of an effective market system.

    But that’s an ongoing debate,

    Tim Wu

  • Tim Wu

    The point made by Tim Lee is decent. It is certainly true that the FCC would have to state some kind of standard to make possible permissionless entry into the spectrum market (as it does for the garage band used by 802.11b). In addition, private actors could, if they wanted, similarly allow permissionless use of spectrum. The question is why they would want to.

    In general I cannot understand the strength of Jerry’s and others’ objection to the substance of rules that would create permissionless market entry into the spectrum market. In my view, reflected in that paper, permissionless market entry is one of the holy grails of an effective market system.

    But that’s an ongoing debate,

    Tim Wu

  • Steve R.

    Tim: Based on the postings and my own views, I believe we actually have two concepts at work here.

    The first concept is that the manufacturing of physical products such as a car’s fuel pump should be a permissionless market.

    The second concept is the use of a resource, such as water, air, and the radio spectrum. In low population/usage situations the use of these resources could be in the form of permissionless market. However, as demand grows there is a need to regulate the use of the resource. The regulation is not simply in terms of the capcity of the resource but also in terms of how the users use that resource. (For example, in terms of usage, radio waves can travel world wide. Two users on the same frequency could easily interfere with each other even though other frequencies are available.) Clearly, I am advocating managing the commons and that these resources should be in the public domain and should be managed by the government for the benefit of the public. Though these resources belong in the public domain, the government – through the auction/lease process – can make some of these resources available for commercial uses. The government (to a degree) must act as a business.

  • http://www2.blogger.com/profile/14380731108416527657 Steve R.

    Tim: Based on the postings and my own views, I believe we actually have two concepts at work here.

    The first concept is that the manufacturing of physical products such as a car’s fuel pump should be a permissionless market.

    The second concept is the use of a resource, such as water, air, and the radio spectrum. In low population/usage situations the use of these resources could be in the form of permissionless market. However, as demand grows there is a need to regulate the use of the resource. The regulation is not simply in terms of the capcity of the resource but also in terms of how the users use that resource. (For example, in terms of usage, radio waves can travel world wide. Two users on the same frequency could easily interfere with each other even though other frequencies are available.) Clearly, I am advocating managing the commons and that these resources should be in the public domain and should be managed by the government for the benefit of the public. Though these resources belong in the public domain, the government – through the auction/lease process – can make some of these resources available for commercial uses. The government (to a degree) must act as a business.

  • http://bennett.com/blog Richard Bennett

    I think the issue with spectrum management comes about because of a failure of technical understanding. The people who argue that spectrum doesn’t need to be managed by regulation and auction believe, incorrectly in my opinion, that interference is not a problem for modern packet radio systems.

    This wrong idea has severe consequences. There is a concern, somewhat legitimate, that limited consumer choice in wireline broadband access to the Internet demands some sort of regulation on Internet access plans and services.

    The market solution to this problem is simply to expand consumer choice, and the most straightforward way to do this is through wireless technologys like EVDO and WiMax. But these services depend on reliable wireless channels, and with the radio technologies of today (and the foreseeable future) reliable radio channels depend on “permission” and limited access.

    So the very people who complain about a dearth of choices in broadband Internet access argue for a regulatory status quo for wireless that effectively prevents it from ever becoming a useful alternative to wireline access. It’s easy to become a conspiracy nut when you see things like this, but it’s actually a matter of technical ignorance rather than malice (in most cases; some net neutrality freaks actually want the government to own broadband Internet access networks.)

    The Internet neutrality access regulators have got the problem upside down: the laws of networking suggest choice is enhanced by the de-regulation of wireline services and the regulation of wireless ones, and it’s all because of interference and scarcity. We can always add bandwidth to a wireline system by adding more cable, but with wireless, once its gone its gone.

    They aren’t making any more spectrum, you see.

  • http://bennett.com/blog Richard Bennett

    I think the issue with spectrum management comes about because of a failure of technical understanding. The people who argue that spectrum doesn’t need to be managed by regulation and auction believe, incorrectly in my opinion, that interference is not a problem for modern packet radio systems.

    This wrong idea has severe consequences. There is a concern, somewhat legitimate, that limited consumer choice in wireline broadband access to the Internet demands some sort of regulation on Internet access plans and services.

    The market solution to this problem is simply to expand consumer choice, and the most straightforward way to do this is through wireless technologys like EVDO and WiMax. But these services depend on reliable wireless channels, and with the radio technologies of today (and the foreseeable future) reliable radio channels depend on “permission” and limited access.

    So the very people who complain about a dearth of choices in broadband Internet access argue for a regulatory status quo for wireless that effectively prevents it from ever becoming a useful alternative to wireline access. It’s easy to become a conspiracy nut when you see things like this, but it’s actually a matter of technical ignorance rather than malice (in most cases; some net neutrality freaks actually want the government to own broadband Internet access networks.)

    The Internet neutrality access regulators have got the problem upside down: the laws of networking suggest choice is enhanced by the de-regulation of wireline services and the regulation of wireless ones, and it’s all because of interference and scarcity. We can always add bandwidth to a wireline system by adding more cable, but with wireless, once its gone its gone.

    They aren’t making any more spectrum, you see.

Previous post:

Next post: