Privatize the Spectrum

The proposed merger between AT&T and T-Mobile is getting a lot of attention with most of the focus being on whether consumers will pay higher prices. The answer is maybe. Prices per minute have been falling and this was true even following the two big mergers in 2004-05 (Cingular/AT&T Wireless, and Sprint/Nextel). Quality-adjusted prices, i.e. taking into account the post-merger buildout of 3G networks, have fallen even further.  On the other hand, although there are competitors in many large local markets and potential competitors (the Cable companies own a chunk of spectrum not yet in use) a merger could increase market power. But even if consumer prices did rise the merger is probably still a good idea. It’s long been known that even small cost savings can outweigh losses to consumers from a price increase (Nobelist Oliver Williamson was one of the first to drive this point home.)

The big issue, however, is not the merger. The big issue is reallocating spectrum from low to high value uses. My colleague Thomas Hazlett argues that spectrum currently being used for low-value over-the-air broadcast of television could, if it were reallocated to high-value uses like wireless, increase consumer welfare by over a trillion dollars. Moreover, for a price of about 3 billion we could switch almost all of the tv-viewers to cable or satellite.  President Obama has pledged to move a big chunk of spectrum, about 500 mhz, to wireless but the process is slow and highly politicized. What really needs to be done is to auction off as much spectrum as possible with as few restrictions on it use as possible. Let the market allocate spectrum across all uses, allowing value maximizing trades. More spectrum would not only be good in itself it would alleviate any concerns about the merger.

Comments

Prices per minute have been falling

Prices per minute isn't really a good measure. Prices have been falling because people are losing interest in voice calls and are switching to text and data services instead. That said, reallocating spectrum to high-value uses is a good idea. Mobile data services are much more valuable than broadcast TV.

I'm not sure auctions are the way to handle spectrum all the way down. We do very well with the commons on which most non-cellphone wireless devices--networks, Bluetooth, microphones, garage door openers, etc.--operate, for example, and it would be hard to replicate this feature with private property because either we'd have to make auctions far too granular or we'd lock out developers because they'd have to pay pure rents to spectrum owners.

"President Obama has pledged to move a big chunk of spectrum, about 500 mhz, to wireless but the process is slow and highly politicized. What really needs to be done is to auction off as much spectrum as possible with as few restrictions on it use as possible."

Alex, what does the first quoted sentence imply for your recommendation in the second quoted sentence?

Best.

Dan

The falling price per minute statistic can be misleading. For many low-to-mid volume users, the given price per minute of usage is increasing. These users are essentially offered take-or-leave it plans where the lowest-priced plan includes many times as many minutes as is desired. The minimum-priced plans have gone up in price considerably in recent years. Yes, the number of minutes in the minimum-priced plans has also gone up, but the price per minute of usage (for these users) has gone up. As you mention, though, there may be off-setting reasons that make this merger beneficial.

So when I ditched cable to go back to over-the-air broadcasts, I became a free rider? And the only way to fix this is submit the over-the-air broadcast spectrum to auctions that will likely result in me losing most of my 'free with adverts' television programming?

How depressing.

"It’s long been known that even small cost savings can outweigh losses to consumers from a price increase"

Why should we be happy if the merger reduces their "consumer surplus" but causes a larger increase in the "producer surplus"? The Williamson model seems to treat those two quantities identically. I doubt that would be a convincing argument to most people.

Secondly, aren't the two issues separate: (1) post-merger price increases. (2) spectrum shift from TV to other uses

Or is there a link here that I don't see?

Wasn't the president just chastising the carriers for not using the spectrum they've already got?

There's value to the nation holding onto this spectrum to see what comes next if the carriers don't need it currently. We might end up wanting to sell it to a disruptive technology that's far more interesting and efficient than phones.

Are these spectrum licenses for perpetuity or is there a specific time in years after which they revert back to the government?

How much is NPR's spectrum worth? I wonder this every time they claim they don't get a lot of subsidy. I think it is worth a lot in a roundabout way. By having that notch on the dial, they get to set up a tournament for syndicated shows and also a national news organization. It is an economic moat and why they are able to have such high production quality, which translates into public support and donations.

NPR doesn't own spectrum. It produces programs and sells them to broadcasters.

There is a slice of FM reserved for non-commercial broadcasters of all types, which includes NPR affiliates, college stations, and religious organizations.

First, judging by the mess that was redoing just the UHF fraction of frequencies dedicated to TV, this won't be so easy. Also, the analog-to-digital broadcast transition cost over $2 billion ($1.5B originally + $650M in the stimulus) just to subsidize the purchase of digital converter boxes at about $50 a piece. Converting to cable or satellite would be more expensive, and that's not accounting for the ongoing cost of programming or the impossibility of doing so shortly after putting the country through years of drama associated with the digital broadcast transition.

Second, about 10-year licenses for bits of 90 MHZ of the old UHF spectrum were auctioned off in 2008. This was a competitive bidding process with reserve prices high enough that one block did not receive bids above its reserve price. The total haul was $20B. Can $20B of spectrum be leveraged into hundreds of billions of consumer welfare?

Lastly, shouldn't radio be on the chopping block before broadcast TV? In terms of the public value of emergency broadcasting, I'd guess that more people are capable of receiving over-the-air television (all new TVs are required to have a digital broadcast tuner) than radio in their homes at this point. Probably don't want folks going to their cars to tune in the emergency FM broadcast in the midst of radioactive fallout or whatever.

"Lastly, shouldn’t radio be on the chopping block before broadcast TV?"

Not going to happen while 90% of people don't have satellite radio in their cars

Some of this has already happened/happening in the uk and elsewhere
e.g. analogue TV is being switched off shortly

http://www.bbc.co.uk/news/business-12811122

"The actual parts of the spectrum being sold - at the 800MHz and 2.6GHz bandwidths - will include parts of the wireless spectrum historically used by analogue TV, which is being switched off as digital is rolled out."

First, I'm not sure the argument that wireless offers more value than TV broadcasts is valid. Second, I'll point out again that the reason the U.S. has less broadband than other nations is the Department of Defense hogs a lot. I'm surprised somebody so libertarian is focusing on private industry instead of the government.

Lastly, shouldn’t radio be on the chopping block before broadcast TV? In terms of the public value of emergency broadcasting, I’d guess that more people are capable of receiving over-the-air television (all new TVs are required to have a digital broadcast tuner)

But a digital tuner with no antenna doesn't do anything for you, and how many people with cable or satellite (which is most people) have an antenna connected to any of their TVs? I know we don't. But we have lots of things that can pick up radio (including even cell-phones).

Maybe the AM-Radio could be chopped and the FM might be left alone? Isn't FM quality better anyways?

AM's only allocated to ~1MHZ of fairly low-frequency spectrum that's not very useful for data transmission (although it can be used much more efficiently than it is now by broadcasting digitally). FM's in the middle of the VHF frequencies.

Count me as one of those people -- only in my shed where no coax runs, though. I'm not sure what the other "85% of households" do. But I suspect not many do. Whether we can infer that the other 15% of households watch over the air, I'll leave it up to you.
http://www.businesswire.com/news/home/20110322006178/en/Research-Markets-USA---Digital-TV-Market

Although I enjoy the free-ness of over-the-air, on-demand TV seems to be the future paradigm. At one point in time, a colleague asserted that we'd see more channels going to over-the-air. Does anyone know why we haven't seen this? Links appreciated.

"But a digital tuner with no antenna doesn’t do anything for you, and how many people with cable or satellite (which is most people) have an antenna connected to any of their TVs?"

Fair enough. In many markets you can find a channel without an antenna and you can make your own easily but it's unreasonable to expect people know this.

I think that Comcast, in a roundabout way, has shown us the way forward.

Comcast has digitized its cable system in the Chicago area. They have taken almost all their analog channels and made them digital, requiring a cable box to receive most stations, even "basic cable" ones.

The only stations left that you can tune in with an old analog tuner TV are the broadcast ones (which are ironically now digitally broadcasted). This has done wonders for Comcast, forcing all the people who were stealing cable to subscribe to get any of the stations beyond the broadcast ones.

But on the other hand, if they opened the system to anyone who wants to connect to it for free, anyone doing so would get the broadcast stations without an antenna. Pretty much everyone in Comcast's area, which is every single Chicago suburb, would be covered.

It would be a win-win, Comcast would have the opportunity to upsell people on its service and the broadcast frequencies would be freed up. And it would cost the government nothing.

I think that Comcast, in a roundabout way, has shown us the way forward.

I'd say, actually, that YouTube and Netflix (streaming) have shown us the real future. Deliver everything as digital data, with the resolution scaled to network bandwidth and device (low res on small devices and over wireless, high res for large TVs over wired broadband). Dedicate nearly all the spectrum to wireless data, and run video and audio services over that.

Just want to say that this is a very interesting usage of the word "privatize"

From reading even this short post it's clear you haven't read TIm Wu's indispensable book "The Master Switch."

Please read it. Seriously.

The relationship between high-value use of spectrum and "the market" isn't nearly as obvious as one would imagine. "Let the market allocate spectrum" sounds good in theory; in practice, things don't always work out so well.

Probably, but government allocation isn't perfect either.

I'm not sure what your concern is, kent. Wu warns about the rise of monopolies, but the key monopolies in his book arose precisely because government gave them the ability to constrict potential competition, whether by over-the-air regulations on spectrum, or by government-created monopoly (AT & T).

I agree with your post that more spectrum should come on the market. I think you are wrong to imply, however, that this administration has been the cause of delay; rather, it has been the source of speed up, overcoming objections of incumbent owners of current spectrum.

The timing of release of spectrum for auctions has a great deal of potential for creating new products and creating new competitors. What you want are auction markets in which a new entrant can cover sufficient geography to enter nationally--Gooogle or Apple as your phone carrier in the future, anyone? So, it is not just making spectrum available, but how you make it available and what auction format you choose that is important.

I disagree, however, with the post if it implies that all the spectrum should be opened up immediately, rather than being released over time. First, current technology companies would pick up the spectrum--how would you have liked it if 30 years ago the spectrum had all been put up for sale at once and ATT (the old line monopolist) had acquired it all to block the emergence or control the emergence of cell phones? Second, the government, as a monopolist of the spectrum, just as any owner, has interests in maximizing revenue; but, as a public entity, also has to balance that against promoting entry. As a monopolist, it might want to withhold spectrum; but as a public entity it might want to release it more generously to promote new uses of spectrum. Finally, the government is not the only source of spectrum. Perhaps we need to create auction mechanisms that would facilitate private spectrum sales with the government being willing to permit incumbents to reallocate existing spectrum to other uses. So, if the fire departments motorola system can be replaced with a digital system, the fire department should be able to sell that spectrum.

Bill you wrote

"Perhaps we need to create auction mechanisms that would facilitate private spectrum sales with the government being willing to permit incumbents to reallocate existing spectrum to other uses. So, if the fire departments motorola system can be replaced with a digital system, the fire department should be able to sell that spectrum."

Exactly. This is what I mean by auctioning off with few restrictions and letting the market allocate.

I don't read the post as implying anything. However, it's just funny that Obama could know that the spectrum was going to go to wireless, but he's still compelled to allocated it to wireless. It's like these guys can't help themselves.

Point being, Alex is the one not saying he knows how to run the world because he's not allocating anything except to the market. But we'll always have people (\/ \/ \/ \/ ) to remind us that that amounts to allocating it to the robber barons.

Perhaps you didn't read this: "President Obama has pledged to move a big chunk of spectrum, about 500 mhz, to wireless but the process is slow and highly politicized.'

Williamson trade-off model is exactly why people hate economists.

Not once in history of the universe has there been anything benefiting the rich and powerful, for which some economists didn't make a model showing it's really in everybody's best interest.

Cost savings are bullshit, price increases are real.

I thought the thread of comments started off poorly, but then became more interesting until Tomasz uploaded his wisdom.

I wonder how spectrum allocation would have proceeded if the FCC had never gotten involved, other than to define boundaries to prevent interference and to establish the validity of ownership. I think Alex's auction idea is an attempt to get a fresh start on a more rational basis.

There are legitimate critiques of the Williamson model. e.g. (wikipedia):

The "trade-off" in the Williamson model involves a gain in producer's (firms') surplus and a loss in consumers' surplus. Thus, in focusing the analysis on total surplus, it neglects distributional issues and treats changes in both consumers and producers welfare symmetricaly. However, anti-trust policy as actually practiced in many countries (Europe, Canada and US) appears to have the goal of maximizing consumer surplus (or this goal is stated explicitly).[11] In that sense, as long as the post-merger market price is higher than pre-merger, the fact that producer surplus and firm profits rise is immaterial from the point of view of the regulators. In that case, only those mergers in which the fall in unit cost is sufficiently large to ensure a lower price after the merger should be permitted.[5] For this reason the Williamson model is not applicable in European Community competition law and is controversial in Canada.[12]

A big reason you see consumer outrage over this is that T mobile is the one major company that affers both good deals for people who use unlocked phones (like those that travel to Europe often), and offers affordable pay as you go pricing that makes a lot of sense for those who barely use their phones at all.

The 1000 minutes lasting a year for $100 prepaid package, for instance, beats the pants out of any ATT offering for those who use a phone sparingly. In the equivalent plans for ATT, minutes expire after 3 months.

So the issue for consumers is that pricing systems designed for some usage patterns are going the way of the dodo bird, as the way the industry is moving is to give you near infinite amounts of minutes for a high premium every month: Infinite supply that many people would never use.

In contrast, see what those companies do in regard to wired internet options: Caps on usage designed to slow down the growth of independent streaming video services, just so that the TV packages that they sell on the same line still appear like a value proposition. The fact that all of this services are tied to each other at the hip hinders competition, and makes us pay more for less.

This is my issue, exactly. I'm on T-Mobile's pay-as-you-go plan, and I end up paying between $30-$50 a year (depending on how much I actually use the phone) for service. AT&T hates this kind of pricing, so I expect this plan will die as soon as they get the buyout finalized, and I'll be stuck having to pay that much per month instead of per year, while getting no more value.
Complicating things on my end is that I was just about ready to completely drop my AT&T landline when this news hit. So now I'm stuck between getting screwed by their landline division or getting screwed by someone's wireless. Joy.

Given modern digital technology, we should be working on eliminating allocated bandwidth itself.

If the government created a framework for acceptable use, and actually enforced it, there is no reason why radio stations (for example), couldn't all share common bandwidth and use some sort of digital packet system, where by it would be possible to let anyone who wants to start a digital radio station.

The reason we keep the old system, is because the "owners" of bandwidth want to keep the ability to broadcast scarce... and the government would find it hard to control the media if anyone could start their own radio/TV station.

That's an interesting thought. If we see wide swathes of the country covered by internet access a lot of the other providers and media might become redundant. Essentially all content and communication could hop over IP networks. We could have radio and TV channel content hopping over IP.That'd free a lot of bandwidth. Many local radio stations already have a streaming service the bottleneck's getting internet access in a moving car. Not in the near future, but in the longer run cellphone providers could themselves become redundant if VOIP calls can be made on the move.

The technical problem is the adoption of WiMax or some other wireless protocol giving good bandwidth and large area coverage. Current offerings seem too stingy on bandwidth and heavy on cost. A challenge is the conflict of interest. Cellphone providers don't have much incentive for internet access provision if those will kill their voice-communication cash-cows

"Not in the near future, but in the longer run cellphone providers could themselves become redundant if VOIP calls can be made on the move."

Not in the near future? What about Google Voice, Skype Mobile or Vonage? VOIP works fine on the move, even on a 3G connection if its reliable. Every major metropolitan area in the US will have WiMax and/or LTE service through a telecom provider within two years. The conflict of interest you point out is exactly what is holding back this more efficient use of bandwidth.

The idea of abolishing exclusive individual rights to pieces of spectrum ("supercommons") is not new, but the reality is that there are enormous technical obstacles that aren't anywhere near being overcome.

I would also note that simply using digital packet-switched radio access networks does not eliminate or alleviate physical phenomena like interference or bandwith congestion. In fact there are fundamental limits to the amount of information that can be transmitted over a given bandwidth. Consequently, technology cannot fully solve the problem of excessive demand to scarce resource. I expect spectrum regulation to be with us for the foreseeable future.

Even if we can't have a true commons, we should be able to radically increase the number of stations available on the FM bandwidth (as an example) using digital technology. There is no reason 10,000 stations shouldn't be able to fit in the FM bandwidth. That would be a significant improvement.

There is a reason why 10 000 radio stations (I suppose you mean simultaneous sound broadcast transmitters) with reasonable sound quality can't fit into 20.5 MHz of spectrum. That reason is the laws of physics coupled with the state of current technology.

Even if a new digital packet-switched broadcasting system was established, you would still need someone to set up that access network (i.e. a licensee). Once the network was set up the frequencies would have to be exclusively occupied. This is due to the fact that any unlicensed use in that band would be detrimental to the S/N ratio and thus to the reception quality of the broadcast signal. So you haven't eliminated exclusive individual rights at all, you have created another monopoly (which incidentally may be a good idea...)

How does the Bandwidth and theoretical information carrying density of, say, a single FM channel radio band compare with, say, optical fiber cable. I was thinking more along the lines of efficiency: Is having hundreds of separate FM channels, TV channels, cellphone provides using their exclusive slices of spectrum less or more efficient than consolidating it all into one (or more, if monopoly is a concern) identical-protocol packet switched networks and then having the flexibility to use it all as per usage patterns. Maybe right now 90% information-density is cellular traffic, tonight 90% might be local-TV (video) and so on.....

That way the protocol becomes agnostic to the actual content; all that matters is information density.

Rahul -

Yes the situation you describe would very probably be a more efficient use of spectrum (from a technical and economical standpoint) than what we have now.

However even in such a world, allocation of bandwidth and assignment of licenses would still be with us. It's important to realise that wireless spectrum is a scarce resource and will remain so for the foreseeable future.

In the case of optical fibre cable between two points technology has reached a stage where the available bandwidth is practically unlimited (limits are set by other pieces of equipment in the network). This is not true when it comes to wireless transmission of signals.

Xeeleee-

You are right about the technology restrictions for long range packet switched networks. As a side-note though, licensing of a spectrum is not always necessary in the wireless commons. An example is the bluetooth and WiFi device-protocols. Given a good industry standard, communication over a common medium can happen by token passing. So long as all devices play fair and dont talk while others are speaking.

Funny story I hear from a while ago was that Cisco was complaining that Meru routers were trying to selfishly hog the network by using some loophole in the standards so that Meru routers got better bandwidths and latencies. Not sure how that dispute resolved.

A simple spectrum auction is not adequate relief for an ATT-TMobile acquisition.

First, spectrum is not the only item you need to enter the cellular market. You need towers.

Second, You also need cellular capacity in specific areas: having bandwith in Wyoming doesn't do you any good if you want to compete in NYC. So, if there is a bandwith auction, unless you get combinations of locations, you are not a national competitor. Can you hear me now?

Third, carriers lock in customers with long term contracts. This is also a barrier to entry even by someone who builds towers, gets spectrum in the right places, etc.

Bottom line: if there is a divestiture package, expect it to be more than spectrum, and don't expect spectrum to be the only piece.

What is a car without tires, or tires without a car?

I have XM in my cars, but get far more interesting music stations through Rhapsody and Pandora, and with a Mobile wifi hotspot ($30 a month for 3 Gig) I can really see this making the XM superfluous. More bandwidth could really start to undermine the cable companies with their local monopolies.

While we could privatise spectrum and make whole the losing side of the political process, we won't. "trillion" to private interests will not result in billions to "public", because it never has.
Immigration hurts poor people, even if it is better for the nation as a whole. We could compensate poor people for those lost jobs, but we don't.
Imports hurt poor people, even if it is better for the nation as a whole. We could compensate poor people for those lost jobs, but we don't.
Rinse, lather, repeat.

I'm not sure that the auction of the spectrum, there is a treatment along the road. We are not encyclopedia most cell phones, wireless networks, Bluetooth, microphones, openers garage door, and so on, for example, works very well, and it is difficult to replicate this facility with private property, either we will have to a "far too thin D for the auction, or we lock out the window to the developers because the owners of the spectrum they must pay the rent.

This is backwards. The most efficiently used spectrum in the world is that used for WiFi, which was only made available because it was 'junk spectrum' with water resonances in it.
Now the carriers with their expensively bought licenses are pleading with cellphone users to offload bandwidth onto WiFi instead.

Releasing the old TV bands under similar rules to the existing WiFI bands is the way to create maxim efficient utilization, not some medieval monopoly system.

The trouble with auctioning off more spectrum to counter the effects of an AT&T/T-Mobile merger is that building a nationwide network of cell phone towers is expensive and time consuming. Case in point: a few years ago Canada auctioned some spectrum to new wireless carriers nationwide but so far these new carriers operate mostly in the big cities only. So the proposed merger will reduce competition because it will take several years for new entrants to build a nationwide tower network. Now it will be interesting to see what happens to the smaller national carriers (MetroPCS, Leap, US Cellular), I wonder if they will merge or be bought out and fill the vacuum left by T-Mobile.

Comments for this post are closed