How Scary is Wireless Duopoly?

by on March 15, 2012 · 19 comments

I’m puzzled by Harold Feld’s latest post. Last month, Harold laid out his argument that society faces a tradeoff between allocating the wireless spectrum to the highest bidder and maintaining competition in wireless services. I responded by showing geometrically that in most cases, efficient production is much more important than maximal competition; in economics, trapezoids are bigger than triangles.

Harold seemed to get it. My argument was only an intuitive one, but he admitted that he needed to be more rigorous to show that competition in the wireless industry is more important than lowering the cost of producing wireless services.



Today, Harold reprised nearly the same post, without additional economic rigor, and at greater length. So maybe you can see why I’m puzzled.

I continue to believe that trapezoids are bigger than triangles. But for this post, let’s put that issue to the side to focus on duopoly, of which Harold seems to have a visceral fear. He’s probably not alone.

Duopoly sounds really bad, because it’s 1 away from monopoly. But modern economists tend not to place too much emphasis on the number of competitors in an industry, at least not in isolation from other factors. For instance, it is well understood that under what is known as a Bertrand duopoly, economic profits to the firms are zero, and the price to consumers is as low as it is when there are many firms. That does not sound too scary, does it?

Duopoly can be consistent with vigorous competition. For instance, most of the bestselling digital SLR cameras on Amazon.com are made by either Nikon or Canon. Sure, a few other companies have a toehold in the industry, but it is as much a duopoly as the wireless industry is. Despite the dominance of these two firms, the price of DSLRs falls every year, and quality continuously goes up.

So what clues can we look at to see if AT&T and Verizon constitute a harmful or a harmless duopoly? The first one is profits. Now, this is tricky, because we want to look at economic profits, which are not the same as accounting profits. Economic profits are usually less than accounting profits, since we need to subtract out the ordinary return to the physical capital owned by the firm. However, even if we look at accounting profits, it turns out that both AT&T and Verizon lost money last quarter. This is not the kind of thing that nasty duopolies tend to do. Now, it could be that both firms had a bad quarter, and they will be back in the black soon. In fact, that is what I expect. But even so, neither firm is likely to rake in profits hand over fist any time soon—and if you think they are, you should put your money where your mouth is and buy their stock.

Another clue we can look at is fixed costs. In an industry with high fixed costs of being in business, lots of companies is neither a stable equilibrium nor a desirable outcome. Suppose there is an industry in which fixed costs are $100 billion. This would mean that for there to be 10 firms, society would have to pay $1 trillion in fixed costs. By reducing the number of firms from 10 to 2, society saves $800 billion. This situation roughly exists in the market for large, commercial jet aircraft, where the only competitors are Boeing and Airbus. Would we really want there to be more commercial jet producers? There would be a whole lot of duplication of costs, and the price of jetliners and air travel would increase, not decrease. We’re better off with a duopoly, and in fact we get duopoly precisely because vigorous competition between the jumbojet giants keeps everyone else out.

Wireless service also has high fixed costs, or, more accurately, declining average costs. Harold notes that if AT&T had been allowed to buy T-Mobile, “it would enjoy ‘synergies.’ Combining with T-Mo would make it bigger, better able to extract the value (by laying off ‘redundant’ T-Mo staff and other cost cutting measures).” I’m not sure why he puts sneering quotation marks around “redundant;” those redundancies represent exactly the socially-beneficial fixed cost-cutting I am talking about. If the industry were not to need those employees any longer, then they could be employed elsewhere, which is what is best for society. If Harold has some other normative criterion, he should come out with it.

The bottom line is that appeals to competition based solely on the number of competitors are just no longer persuasive to anyone up on the economics of firms and industries, also known as industrial organization. As both an avid consumer of wireless service and someone with more than a little fondness for IO, I’m not scared of wireless duopoly at all.

  • http://tinyurl.com/CowboyBooksBlog fgoodwin

    I’d like to see a study of how many bloggers and economists are optimal?  Would society be better off if the excess were employed elesewhere?

  • Blake Johnson

    Most bloggers do so unpaid and of their own volition, so it is kind of a consumption good for the bloggers. So unless someone were willing to offer them money to stop blogging it seems unlikely that social welfare could be improved by moving them to other endeavors unless perhaps they were moved to some position with large positive externalities.

  • Harold Feld

    Tsk. Youth has no patience, especially with us pedantic and long winded geezers. As I noted in the last pargraph, there are bunch of counter arguments which I hope to reach in Part III. But I figured  that clocking in at over 5000 words was already enough for my section on “why secondary markets experience the same dynamic (assuming my intuition is correct).”

    As I say, I am hoping to get to your criticisms (and those of others) in Part III (once I get a chance to write it, you may note it took about a month between installments). 

  • http://twitter.com/binarybits Timothy B. Lee

    I think it’s a mistake to focus too heavily on price. Wireless connectivity is a (direct or indirect) input to a lot of other products–phones and tablets, apps, text messages, mobile audio and video services, etc. The primary danger of too much concentration in the wireless market, in my view, is that the duopolists will vertically integrate into these adjacent markets and freeze out other competitors. The more wireless carriers there are, the less leverage any one of them has against would-be partners (like device makers, content companies, etc.).

    Also, the comparison to the camera market is inapt because while the market might currently be highly concentrated, there are relatively few legal barriers to entry in the camera market. Canon and Nikon are disciplined not only by their current competitors, but also by the knowledge that new firms (Apple, or some new startup) could enter the market with a disruptive new product. The threat of new entrants, or of existing competitors significantly increasing their market share, prevents the incumbents from resting on their laurels.

    In contrast, if AT&T and Verizon were allowed to, between them, buy more than 2/3 of the spectrum suitable for mobile services, then in practice it would mean that other firms are operating at a permanent, legally-guaranteed disadvantage. No matter how bad AT&T and Verizon’s service gets, it might be technically impossible for other firms to compete with them because there isn’t enough spectrum to go around.

    To be clear, I’m not saying we’ve reached that point, or that we’re even close to it. The point is just that a legally-enforced duopoly is a very different thing than a market like cameras or sodas where the two firms are able to maintain their dominance only if they can keep customers happy. In principle, at least, I think there’s clearly a legitimate basis for regulators to be concerned about the possibility that any two firms could obtain such a large fraction fo the available spectrum that it becomes infeasible for a third or fourth firm to enter the market.

  • http://twitter.com/binarybits Timothy B. Lee

    I think it’s a mistake to focus too heavily on price. Wireless connectivity is a (direct or indirect) input to a lot of other products–phones and tablets, apps, text messages, mobile audio and video services, etc. The primary danger of too much concentration in the wireless market, in my view, is that the duopolists will vertically integrate into these adjacent markets and freeze out other competitors. The more wireless carriers there are, the less leverage any one of them has against would-be partners (like device makers, content companies, etc.).

    Also, the comparison to the camera market is inapt because while the market might currently be highly concentrated, there are relatively few legal barriers to entry in the camera market. Canon and Nikon are disciplined not only by their current competitors, but also by the knowledge that new firms (Apple, or some new startup) could enter the market with a disruptive new product. The threat of new entrants, or of existing competitors significantly increasing their market share, prevents the incumbents from resting on their laurels.

    In contrast, if AT&T and Verizon were allowed to, between them, buy more than 2/3 of the spectrum suitable for mobile services, then in practice it would mean that other firms are operating at a permanent, legally-guaranteed disadvantage. No matter how bad AT&T and Verizon’s service gets, it might be technically impossible for other firms to compete with them because there isn’t enough spectrum to go around.

    To be clear, I’m not saying we’ve reached that point, or that we’re even close to it. The point is just that a legally-enforced duopoly is a very different thing than a market like cameras or sodas where the two firms are able to maintain their dominance only if they can keep customers happy. In principle, at least, I think there’s clearly a legitimate basis for regulators to be concerned about the possibility that any two firms could obtain such a large fraction fo the available spectrum that it becomes infeasible for a third or fourth firm to enter the market.

  • http://elidourado.com/ Eli Dourado

    I’d be curious to know what exactly your model of vertical integration is. There are a lot of cases in which vertical integration is socially desirable, as well as some in which it is harmful. Broadly speaking, if you have market power in one industry, it does not benefit you to acquire market power in a competitive upstream or downstream industry. Vertical relations between firms is a really fun IO topic, but it’s also very counterintuitive at times, so I’d caution against claims that just assume vertical integration is bad.

    Also, the whole point of making spectrum allocations tradable is to make incumbents bear the opportunity cost of holding the spectrum. Even if AT&T and Verizon owned all the good spectrum, if they were doing a poor job, then an entrant could offer to buy them out of some or all of the business. If we assume that the entrant can create more value, then by definition there is a Pareto-improving trade between them. The fact that the amount of spectrum is limited does not constitute the kind of legal barrier to competition that economists worry about.

  • Ryan Radia

    What about the countervailing power of mobile OS vendors (Apple, Google, Apple, RIM, etc)? I sense that users tend to have stronger preferences with respect to the operating system (“ecosystem”) that runs on their mobile device than their wireless network. If VZW were to attempt to vertically integrate the wireless market, Apple could buy the whole network in cash! 

  • http://twitter.com/binarybits Timothy B. Lee

    Broadly speaking, if you have market power in one industry, it does not benefit you to acquire market power in a competitive upstream or downstream industry.
    I don’t think I can sketch out a coherent theory of vertical integration in a blog post, but I’ll just mention that the closest example we have–AT&T’s pre-1984 monopoly–tells a different story. AT&T used its local phone monopolies to vertically integrate into a number of adjacent industries–long distance, handsets–and would have vertically integrated into many others–like computing devices–if the government hadn’t prevented them from doing so. It’s possible that AT&T was behaving irrationally, but it’s also possible that the real world is more complex than the economic model behind your claim above. For example, vertical integration allows kinds of price discrimination in downstream markets that wouldn’t emerge from a competitive market.
    Also, I think it’s unrealistic to assume firms always behave rationally. Firms are run by people with big egos, who often over-estimate how good they are at their jobs. So even if the profit-maximizing course of action is to stay out of a particular downstream market and charge the firms in that market monopoly rents, the guy running the firm with market power may not realize this. See, for example, Microsoft’s perpetual, failing efforts to create successful web services and digital media devices.

  • http://twitter.com/binarybits Timothy B. Lee

    Smartphone vendors’ power is a direct function of the relative competitiveness of the smartphone and network markets. A big reason AT&T was interested in the iPhone was that it would allow them to significantly increase their market share. The larger a share of the market AT&T already has, the less incentive they have to relinquish power to a third party like Apple.

    The point is a general one. If there are two networks and four major OS vendors, the network providers are going to be a much stronger bargaining position than the OS vendors. If there are 2 OSes and 4 networks, the opposite is true.

  • Pingback: Em breve…puzzles. Por agora, uma breve notícia interessante sobre duopólios « Microeconomia, Econometria e muito mais

  • http://elidourado.com/ Eli Dourado

    I think the old AT&T should be viewed as effectively a branch of the government (albeit with shareholders), rather than analogous to AT&T and Verizon today.

    When I ask for a model of vertical integration, I’m not looking for a comprehensive theory, just an abstract statement of what exactly you’re worried about. Which you provide with your sentence on price discrimination, so thanks. Again, I’m not too worried. It’s relatively easy to price discriminate on wireless service and relatively difficult to do so on, say, handsets. Also, in declining average cost industries, price discrimination is a necessary feature of the competitive outcome, not a sign of market power. See a very readable paper, one of my favorites, by Michael Levine for this point.

    I think your concerns would be extremely valid for, say, the wheat industry, where there are low fixed costs and flat or rising average costs. Industries with lots of infrastructure and network externalities are just different.

    I take your point about overconfidence among executives. Nevertheless, I think we should stipulate that the demand curve for irrationality slopes down. Other things equal, people are less irrational when they have a lot of money on the line. There is a pretty solid consensus among economists about improvements to corporate governance that could be made, but I don’t think that in the absence of those improvements we need to throw up our hands and predict that executives will behave totally unpredictably. And I think that’s also a pretty flimsy basis for distorting the spectrum market.

  • Pingback: Wetmachine » Tales of the Sausage Factory » My Insanely Long Field Guide To The Verizon/SpectrumCo/Cox Deal.

  • Pingback: wRDwgC1HqM wRDwgC1HqM

  • Pingback: the glades

  • Pingback: management consulting

  • Pingback: lainat

  • Pingback: book of ra deluxe

  • Pingback: book of ra deluxe

  • Pingback: Maid Services

Previous post:

Next post: