Telecoms regulation considered harmful

Doc Searls asked me to put the argument for total telecoms deregulation into a nutshell, then blog it so he could point at it. Here it is.

Telecoms regulation, to the extent it was ever justified, was justified on the basis of preventing or remedying market failures — such as, in particular, lack of market incentives to provide universal coverage.

The market failures in telecoms all derive from the high fixed-capital costs of conventional wirelines. These have two major effects: (1) incentives to provide service in rural areas are weak, because the amount of time required to amortize large fixed costs makes for poor discounted ROI; and (2) in higher-density areas, the last mile of wire is a natural monopoly/oligopoly.

New technologies are directly attacking this problem. Wi-Fi, wireless mesh networks, IP over powerlines, and cheap fenceline cable dramatically lower the fixed capital costs of last-mile service. The main things holding these technologies back are regulatory barriers (including, notably, not enough spectrum allocated to WiFi and UWB).

The right answer: deregulate everything, free the new technologies to go head-to-head against the wired last mile, and let the market sort it all out.


  1. Thanks Eric, succinct and to the point. I am certainly open to arguments on the merits on net neutrality, but I am much less open to achieving those ends via regulation. There is no net neutrality legislation that I am aware of, so why punish a crime that does not exist? It presumes the carriers guilty until proven innocent, and adds rules to a place (hat tip to Doc) that is flourishing without rules. It also asks a government to predict and categorize a market. It’s like asking a dinosaur to do astronomy.

    More here:

  2. Anti-free-marketeers have for decades pointed at specific flaws that supposedly came about as a result of communications deregulation (e.g. the rise of Clear Channel), when in fact the problems are the result of *not enough* deregulation. Drop the barriers to entry on low-power FM and Clear Channel suddenly has thousands of new competitors that each know their local market better than CC’s database. Drop the barriers to entry on the “last mile” and you can tell the Baby Bells where to shove themselves.

  3. >in order to write laws that stifle competition

    Let me draw you a picture: when the telecoms companies buy politicians to write laws to stifle the competition, it means they fear the competition and fear market pressure.

    I’m for deregulating everything so the market pressure they fear will destroy the telecoms monopolies.

    Is that clear enough?

    For your next exercise, Mark, learn why monopolists love market regulation and antitrust law. The hilarious part (if you have a really cynical sense of humor) is that they sell the laws that protect them as the laws that keep them in check.

  4. Internet is a network of computers, nothing less and nothing more. Internet network bandwith (pipes) will be free, any other scenario does not fit elementary game theory (see prosoners’ dilemma). It is content that has value. Content owners will win. As if this were not obvious.

  5. If you just changed the way new subdivisions are developed so that the demarc point between the “last mile monopoly” was at the subdivision concentrator, you’d find plenty of competition for the homeowners’ business. Negotiate a wire maintenance agreement for the subdivision and a contract for bandwidth for the entire subdivision and suddenly you have enough economic power to draw plenty of people to compete for your business.

    This doesn’t actually take any legislation, just farsighted builders who simply arrange a better way to arrange telecom/datacom for the people who are going to be living in the structures they are building.

  6. The truly ironic part about the Net Neutrality advocated by Lessig and Co. is that it essentially turns broadband Internet access provided over DSL pipes into a heavily regulated common carrier service.

    With the FCC deciding what I get and what I don’t.

    And the Bells are arguing for NOT regulating it.

    A complete inversion.

  7. Subdivision demarc: I’m not aware of any subdivision or other detached housing development that pays any attention to this. Unfortunately the only related trend, in apartment/condo buildings, that I’ve witnessed is quite the opposite. Whether kickbacks are involved or not, when anything telecoms-related is offered in these scenarios it tends to be from a “competitive” provider who is handed a captive pool of customers and usually offers service even worse/more expensive than the prevailing monopolist in the area.

  8. Dead on commentary. Don’t let unknowledgable commentaries detract – I was surprised to see someone who actually understands the issues.

    I work in the rural broadband industry and indeed WiFi spectrum issues is the constraint on growth. In communities of 200 households, we service more than 40% typically, but need FCC progression on WiFi frequencies (ala 3.6 GHz which was permitted but then locked up by satellite providers who can’t understand their extreme latency product has limitations based on laws of physics that preclude it being useful to rural telework subscribers, for instance).

    LMDS and MMDS were frequency auction nightmares paid for by dot-com action constructs (many of which defaulted – and those that didn’t default were usually snapped up by incumbants to PREVENT any introduction of competitive broadband service) – we don’t need repeats of this, and from my understanding, most in the FCC realize it was penny smart, pound foolish.

    Our own capital structure has led to our taking a holding position due to frequency issues – we’ve achieved 40-50% market penetration in about 1/3 of a state, but cannot proceed further due to regulatory/interference issues that have no regulatory response and cause financial constraints that make new projects unviable.

    If rural broadband matters, FCC, then you’re gonna have to tell those who have impeded progress that their dollars are less important than a critical national goal.

  9. My fear is that total deregulation will lead us back to the AT&T monopoly days. SBC just bought AT&T — without *some* rules in place, what’s to stop them from buying up Verizon, Cingular, &c? A free market means that companies are able to buy each other up. A free market doesn’t automatically mean there’s competition.

  10. The AT&T monopoly days were brought to you by government regulation: after the end of the initial Bell patent (= first government monopoly), the former monopoly was losing market shares rapidly. After 10 years of lobbying, and 50% of market loss, it got a law enacted whereby there could be only one operator on a given territory, and it was always in the best position to be granted this monopoly in every place. There was NEVER a “natural monopoly”. Monopolies were always government interventions. The market always provided diversity.

    Sure, with deregulation, SBC can by Verizon, Cingular & c. And the day after they are suspected to rise prices, a competitor will emerge who will take half their market. The only way they can get a “de facto” monopoly and keep it is by keeping quality high and prices low, and even then there will always be marginal competitors, unless the government intervenes. See my article about monopolies.

  11. The BBC had a nice story a few months back about how Somalia, a near-anarchy, had easily the best telecom system in the region. (Somalia may raise other issues for Libertarians, but in this aspect it’s a real poster child for the free market.)

  12. >There was NEVER a “natural monopoly”.

    Much as I’d like to agree with this on libertarian grounds, I don’t think it’s true. Natural monopolies do happen. For example, there is essentially only one place on earth you can mine helium. That’s a permanent natural monopoly (or, at best, oligopoly) at least until H fusion gets cheap enough that we can synthesize the stuff in bulk :-). The only way you beat a monopoly like that is by finding goods to substitute for the product.

    My point in the topic post is that the ‘last-mile’ natural monopoly is being undermined by technological change. Usually such phenomena are quite temporary.

  13. I think the idea of regulation being predicated on the idea of scarcity is right on the money. Now that the cost of bandwidth and the machinery to use that bandwidth is asymptoting to zero, it’s am anachronism.

    Even among the technically-savvy, there’s little knowledge of just how much bandwidth will be available as technologies mature. The low-absorption window in fibre between 1430 and 1580nm is 20 TeraHertz wide. When I was working on wireless LANs back in the 90’s, the blue-sky frequency was in the 60 GHz oxygen absorption band. The absorption gives you frequency re-use, and there’s so much bandwidth up there that putting a Gigabit Ethernet link – or five – over WiFi is easy. There are still challenges with the hardware (it’s not easy doing RF engineering in the EHF band – in fact you’ve got edges on signals that are heading into far infra-red, and doing DSP at that speed is non-trivial) but I know a lot of people who are working to make it happen. Ten years ago, a wireless LAN working in the 802.11 bands (2.4 GHz or 5.15 Ghz) occupied thirty feet of bench space and cost a half-million dollars. Now they’re PCMCIA-sized and cost sixty bucks. 60 GHz is unlicensed at the moment – long may it remain so.

  14. 1) WiFi’s “problem” isn’t a lack of spectrum, and we covered “WiFi can’t scale” (an is inappropriate for mesh networks in particular) several posts back, thanks.

    2) David Gillies is correct about 60 GHz (which is also license free, and allows you to build full-duplex Gpbs links over about a mile), though nobody runs at DSP (or even ADCs) up at 60GHz (just calculate the Nyquist sampling rate to see why). I/F is a good thing, David. (And its unlikely that the 802.11MAC is the appropriate structure for a full duplex link, but I digress.

    As a side note, “10 years ago” 802.11 wasn’t a ratified IEEE standard (that didn’t happen until Sep 97), but wireless LAN cards were available for $600-700 each, and “access points” for 2-3X that amount. (The market was quite small.) NCR’s wireless LAN stuff was available in the very early 90s in PCMCIA form-factor for less than $1000 each. (Obviously I take issue with the “30′ of bench space and 0.5 million.)

    3) Fenceline networking (thank you for the plug) isn’t regulated now, at least, not until you cross the street. (See #2… or UWB)

    4) IP over powerline will fuck Amateur Radio *forever*. This is bad.

    In economics, (Where is Russ Nelson?), a natural monopoly occurs when, due to the economies of scale of a particular industry, the maximum efficiency of production and distribution is realized through a single supplier. You don’t have to have a lock on the sole source of Helium to become a natural monopoly in Helium sales.

    Its also not clear that wires (or fiber) on poles (or in the ground) constitute a “natural monopoly”. You can always string more wire, though it costs. (In Texas, and other states there is an “open ditch” law where *anyone* can throw *anything* into an open ditch. Thus, when someone trenches for fiber, *all* the providers run after them putting fiber in the ground.)

    In the end:

    Spouting off about “deregulate everything and the market will sort it out”: hopelessly naive.
    Building the network “bottom up”: Good.

  15. Mark Musante: “That’s rather a naive position to take. When you’ve got politicians being bought by the telecom industry in order to write laws that stifle competition, it doesn’t take much to figure out that some regulation is required in order to promote that ‘free’ market.”

    That’s fairly naive on your part Mark, and demonstrates a clear misunderstanding of WHY the telecom industry is trying to buy politicians.

  16. Mr. Cowperthwaite,

    Are you by any chance related to Sir John Cowperthwaite, the former laissez-faire Financial Secretary of British Hong Kong? Regardless of your ancestry, you appear to have learned his lessons well.

  17. Somewhere back in the family tree Sir John and I are related. The families diverged in the late 19th century, as close as I can tell.

    Not sure if you meant that as a compliment, or not, but as a “pragmatic neo-classicist”, economically speaking, I’ll take it as one. :-)

  18. Mr Cowperthwaite: My understanding of the situation is that the telecoms are buying politicians so that they can get regulations written which stifle competition. My understanding of ‘free market’ is that there are no laws in place which dictate how the market can be divvied up. My conclusion, therefore, was that one company would eventually become the dominant, and then the only, company in control of telecommunications. Is this wrong?

  19. >Is this wrong?

    Yes. Monopoly is not the natural outcome of a free market. With only rare and transient exceptions, monopolies require government market-rigging to survive. Um…Bell Tel was monopoly created by government fiat, you know.

  20. Monopoly is difficult, to impossible, in an unregulated market. A perfect free market is one where all players have perfect information, which leads to nothing being sold for more than its marginal cost. If all players in the market have perfect information, monopoly becomes impossible, except in unique cases where there is only one source of a resource in the entire world and one person owns that resource. Regulation, by its nature, reduces information available to the market entities, thus moving the market away from free.

    For example, if the government mandates that a potential employee must work for no less than $7.00/hour, then we will never know how many employees would be willing to work for less than that, nor will we know how many employers might be willing to pay more than that for unskilled, entry level labor. You might argue that the minimum wage laws are a net good because they ensure that no one is being exploited in their desire to get a job, but you cannot argue that it doesn’t prevent all the people involved in the employment market from gaining full information about the market. Each incremental reduction in information makes it more likely that a monopoly could come into existence. I think it is interesting to note that employers like McDonald’s, that employ large numbers of unskilled workers, are not lobbying for a change to, or removal of, minimum wage laws. We should be suspicious when regulations that are supposed to protect us from being exploited are not opposed, or even lobbied for, by the very market players that they are supposed to be restraining. Just think about the implications of McDonald’s favoring minimum wage laws. Is that because they love people and want to be forced to pay them a higher wage? Or is it because it is good for McDonald’s?

  21. > Monopoly is not the natural outcome of a free market. With only rare and transient exceptions, monopolies require government
    > market-rigging to survive.

    Damnit Eric…

    Your statement is only true when the competing entities have equal access to the commodity that is for-sale.

    Show me the government ‘market-rigging’ for OPEC, or the DeBeers cartel.

    Or… read this:

  22. I said “rare and transient exceptions”. Monopolies based on privileged access to a resource do arise from time to time, but this is a phenomenon with a short half-life. They get done in by substitution effects.

    DeBeers is a screwy and interesting case. To understand why their monopoly has been so persistent, you need to know two things: (1) they only monopolize gem-quality diamonds; the market for industrial diamonds is normally competitive. (2) gem diamonds are a ‘positional’ good, that is, they have effectively no value other than as a social signal of conspicuous consumption. Thus, as it turns out, nobody has any sustained interest in lowering their price!

  23. > Just think about the implications of McDonald’s favoring minimum wage laws.

    Following that train of logic leads me to conclude that minimum wage laws prevent some small companies from starting, because they can’t afford to pay employees the minimum wage. If that’s true, minimum wage laws are good for the individual with a job (he gets more than he would have gotten), but bad for the individual without a job (a potential employer of his can’t hire him).

  24. >Following that train of logic leads me to conclude that minimum wage laws prevent some small companies from starting, because they can’t afford to pay employees the minimum wage.

    That’s right. It turns out that when you add the deadweight loss (look it up) the effect of minimum-wage laws is always negative.

  25. If the equilibrium wage being paid is higher then the new minimum wage, wages will remain unchanged. Theoretically, price floors below equilibrium have no effect on prices. Note that under these conditions, an increase in the minimum wage would not increase unemployment – even if the demand for labor is reduced.

    Under these conditions there is no negative effect due to minimum wage laws, and esr’s assertion is false. Your “potential employer” wouldn’t be able to find anyone to work at the “minimum wage”.

    (“Always”, Eric? You know better! The “invisible hand-job”, to be sure.)

  26. > Monopolies based on privileged access to a resource do arise from time to time, but this is a phenomenon with a short half-life.
    > They get done in by substitution effects.

    This assumes that the good (or service) has available substitutions which can be made available at a lower cost.

    This is not always the case. (examples abound)

  27. >Under these conditions there is no negative effect due to minimum wage laws,

    No negative effect on employment rates, no, not when the equilibrium wage is higher than the mandated floor. But deadweight loss still gets you. Among other ways, even useless laws have a cost of enforcement.

  28. If the equilibrium wage is higher than the minimum wage, what economically-sound bureaucracy would bother “enforcing” a minimum wage law?

    “Useless” laws go un-enforced all the time. In Morrisville, PA, women need a permit to wear cosmetics. In Harrisburg, Pennsylvania it is illegal to have sex with a truck driver inside a toll booth. It is against the law to whale hunt in Oklahoma. In Breton, Alabama, there is a law on the town’s books against riding down the street in a motorboat. In Houston, TX, it is illegal to sell Limburger cheese on Sunday.

    Jim — who remembers Taco Bell and other “fast food” places desperate to hire shift workers at $10.50/hr with a $500 “sign-on” bonus in Austin, TX, circa 1999.

  29. Jim, since we have a significant “undocumented workforce” willing to work for less than minimum wage, clearly the equilibrium wage is lower than the minimum wage. Even if we assume that having a minimum wage is fine so long as it is below equilibrium, that is clearly not the case in the USA today. Setting a minimum wage higher than equilibrium drives higher unemployment and inflation, at a minimum. It also deeply contributes to a social problem of illegal immigrants who are essentially slaves (there’s other contributing factors, but that is one of them). Any valid economic analysis of the minimum wage law should conclude that it is a net negative.

    Mark, you worked out the logic correctly. Except for one thing. Why does McDonald’s want a minimum wage law?

  30. Eric (Cowperthwaite),

    I’m actually “mostly against” the minimum wage laws. My issue (and the point of my postings) is that ESR (and others like him) make definitive statements (which aren’t backed by evidence). My purpose here has been (mostly) been one of “lets tell the truth”.

    I have seen times (and markets) where no “legal” worker needed to accept minimum wage. Most recently Austin, TX, where the percentage of “undocumented” workers can be quite high.

  31. Jim, I think you’re missing the point that the undocumented workers drastically distort the employment market, in much the same way that a blackmarket of goods will arise in a marketplace that has high taxes and tarrifs.

  32. We have laws (that esr is against) to prevent “undocumented workers” from being employed at Taco Bell (or elsewhere except on a cash basis.)

    Are you really claiming that “undocumented workers” are “blackmarket labor”?

Leave a comment

Your email address will not be published. Required fields are marked *