Odlyzko on Net Neutrality, Price Discrimination, PrivacyFail, Search & Cloud Neutrality

by on November 12, 2009 · 20 comments

by Berin Szoka & Adam Thierer

The latest call for “search neutrality” and “cloud neutrality” comes from Andrew Odlyzko of the University of Minnesota’s School of Mathematics & Digital Technology Center—and probably among the top ten most influential academics in Internet policy. In his latest Review of Network Economics article “Network Neutrality, Search Neutrality, and the Never-ending Conflict between Efficiency and Fairness in Markets,” Odlyzko shows (discussed by Ars) just how slippery the slippery slope of Net neutrality regulation will be—exactly as we predict in our recent paper Net Neutrality, Slippery Slopes & High-Tech Mutually Assured Destruction.” Odlyzko concludes:

for pervasive infrastructure services that are crucial for the functioning of society, rules about allowable degrees of discrimination have traditionally applied, and are likely to be demanded for the Internet in the future. Those rules have often been set by governments, and are likely to be set by them in the future as well. For telecommunications, given current trends in demand and in rate and sources of innovation, it appears to be better for society not to tilt towards the operators, and instead to stimulate innovation on the network by others by enforcing net neutrality. But this would likely open the way for other players, such as Google, that emerge from that open and competitive arena as big winners, to become choke points. So it would be wise to prepare to monitor what happens, and be ready to intervene by imposing neutrality rules on them when necessary.

Odlyzko identifies search and cloud computing as the next most likely targets of “neutralization” and explains how calls for regulating these virtual “networks” would flow logically from the current arguments for neutrality mandates at the infrastructure layer:

The net neutrality debate is often pictured as a contest between the two most prominent corporate champions of the opposing sides, AT&T and Google. But the underlying issue predates both companies by centuries. It was never resolved completely, since it arises from a conflict between society’s drives for economic efficiency and for fairness. There is no reason to expect that this conflict will lessen, and instead there are arguments that suggest it will intensify. Should something like net neutrality prevail, the conflict would likely move to a different level. That level might become search neutrality. (And allegations about discriminatory behavior of a web search provider have surfaced recently in China, Tschang (2009).) Or, to take another currently popular concept, if “cloud computing” does become as significant as its enthusiasts claims, it could lead to dominance of a single service provider. The effective monopoly of that dominant player could then become perceived as far more insidious than any of the “walled gardens” or “intelligent network” that telcos would like to build.

There is, of course, an entirely different approach to the issue that does not involve the sort of across-the-board cyber-meddling that Odlyzko suggests: Freedom for all players at all layers of the Net to invest and innovate in the “networks” or “platforms” that offer content, connectivity and services. Such an approach rests on a recognition that high-tech markets are driven by dynamic competition: Network effects often create markets with a limited number of players, so from a purely static perspective, such markets seem “monopolistic” or “oligopolistic,” but over the perspective of just a few years, competition among technological titans tends to be highly rivalrous, with today’s leaders often becoming tomorrow’s laggards. Smaller entrepreneurial “furry mammals” quickly run circles around entrenched “dinosaurs” because established players necessarily struggle to keep pace with “disruptive innovation.”

Alleged Search Bias towards Advertisers

Odlyzko begins by citing the famous passage (“advertising funded search engines will be inherently biased towards the advertisers and away from the needs of the consumers”) from the 1998 paper written by Sergey Brin and Larry Page when Google was just a cool Stanford research project. Despite putting the term “Search neutrality” the title of this paper, Odlyzko offers little elaboration on this point.  But in principle, he appears to concur Seton Hall law professor Frank Pasquale’s statement of the “problem.” From Pasquale’s 2008 congressional testimony calling for establishment of a Federal Search Commission:

Many webmasters live in fear of the “Google Death Penalty”—relegation to the bottom of results for a “gray” search engine optimization tactic. The thin and ever-shifting line between “black hat” and “white hat” search engine optimization raises serious questions about arbitrariness. More ominously, search engines can openly profit from opacity here. If there is no clear route to the top of “organic results” for a given term, the only way to assure one’s association with it is to buy “paid results” from search engines’ themselves. Just as search engines worry that cable and telecommunications carriers may deliberately impair quality of service in order to force application providers to pay for a “fast lane,” content providers may legitimately worry that dominant search engines “churn” organic results in order to make paid ads the only guaranteed method of reaching customers.

Wharton law professor Eric Clemons has similarly asserted that “Google is abusing its monopoly position by overcharging corporations for access to consumers [through advertisements].”

Capturing the Value Created by Price Discrimination

Odlyzko doesn’t address such arguments directly but seems to endorse them, before moving on to his central claim, that Google’s “huge stock market valuation” is likely based in large part on the prediction by investors that Google will be able to reap huge profits in the future by facilitating price discrimination for advertisers—and engaging in price discrimination of its own:you

Serving up the standard online ads has turned out to be extremely profitable… And there are plenty more ads that can migrate online. But there are limits, since the advertising field is not all that large. (The total amount spent on advertising is lower than spending on voice telephony.) But if “behavioral targeting” can be exploited, extensive price discrimination would open up entirely new, and much larger, revenue sources. (Enterprises already spend on direct marketing and other promotions amounts comparable to what they spend on ads, and effective differential pricing could replace and augment that.) And the company in control of the process, the one with the information about customers, and in control of the delivery of the offers, could potentially end up in control of product pricing, essentially relegating the good and service providers to a commodity role. The profit potential of such a role is gigantic.

In other words, Odlyzko argues, as much as Widget, Inc. might value “behavioral” data about consumers’ interests (derived from “tracking” users’ browsing) in order to focus limited advertising budgets on consumers who are actually likely to be interested in widgets, even more valuable is the ability to charge personalized prices based on each consumer’s demand elasticity. In short, Odlyzko argues that what will most drive the expansion of neutrality mandates is popular fears about increased “price discrimination” facilitated by what he sees as the deterioration of privacy online: with greater information about individual users, online service providers and retailers could charge prices unique to each user according to their demand elasticity. He concludes, dramatically, that, “whether AT&T or Google wins the net neutrality battle, the outcome at a high level may be similar, namely society exposed to the prospect of an unprecedented degree of discrimination.”

Odlyzko emphasizes that, “from the standard economic point of view, there is nothing nefarious about attempting to introduce differential charging,” especially because it allows “service providers [to] obtain more funding, which may be used to build out networks.” In other words, “differential charging” or “price discrimination” can certainly increase efficiency, ultimately benefiting the consumers who enjoy the improved products and services increased funding makes possible, but notes that “social acceptability of such practices is at wide variance with the recommendations of the conventional economic models.” Translation: “gosh darn it, it’s just not fair!” Odlyzko appears to be resigned to the conclusion that this essentially emotional reaction will ultimately drive the expansion of neutrality mandates, but doesn’t clearly distinguish between this descriptive prediction and a normative judgment as to whether such an expansion would actually be good for consumers.

Odlyzko wouldn’t be as widely respected if he weren’t the sort of academic who takes such nuanced, if ambiguous (and perhaps even deliberately opaque), positions. But he could have followed in the footsteps of others who have mounted an unequivocal defense of “price discrimination.” Among the best at explaining why we’d all be better off with less “fairness” is former FCC economist Rafi Mohammed, whose 2005 masterpiece The Art of Pricing: How to Find the Hidden Profits to Grow Your Business, never uses a single variant of the word “discriminate.” Mohammed realizes that a rational conversation is hard enough when economic theory conflicts with what appears to be hard-wired emotional responses—but probably impossible when the words used for the economic theory is one as fraught with nasty historical connotations as “discrimination.” So, to dispel any subconscious associations with racist thugs like Bull Connor, Mohammed simply—and brilliantly—introduces the term “the multi-price mindset” (id. at 86-91).

Mohammed isn’t just read branding an impolitic term: The problem  racial “discrimination” (what we usually think of when we hear the word “discrimination”) is that it fails to discriminate among individuals of the same race, whereas Odlyzko’s worries that pricing will be too finely focused on the individual consumer rather than the group. Moreover, Mohammed groups three related forms of “price discrimination” together under the term “multi-price mindset”: differential pricing, versioning, and segment-based pricing.  He explains brilliantly why this mindset is good for business:

A multi-price mindset is a series of strategies that enables you to profit from each customers unique product valuation. Remember those customers you thought would pay more? The multi-price mindset charges them higher prices, which they willingly pay. The multi-price mindset tips its hat to the law of demand by recognizing that lower prices will draw in new customers. Remember those shoppers who demurred because of price? The multi-price mindset discreetly offers them discounts. The multi-price mindset is the key to pricing for profits and growth for the following three reasons. First, price is aligned with value. If you aren’t thinking about price in this manner, you are undoubtedly missing out on profits. Second, you’ll make higher profits from the customers who value your product the most. Third, working in conjunction with the law of demand, the multi- price mindset increases your customer base through lower prices and offering new selling strategies. While you will earn fewer profits from these customers, you will nonetheless be earning additional, previously hidden, profits.

So what’s in it for consumers? Odlyzko notes that increased profits mean improved networks. In the Schumpeterian model of competition among Titans whose reigns as market leaders come and go with changes in technologies, the profits made possible by a multi-price mindset are precisely what creates every Goliath—but also makes the challenge worthwhile for every new David. By contrast, the more a market approximates the hypothetical ideal of “perfect competition,” the less “juice” they will be in the system ongoing innovation and investment in new technologies.

Of course, differential pricing means that many customers will pay less—or will buy products whose “one-size-fits-all” price simply would have exceeded their valuation of the product. Nowhere are benefits to consumers from multi-price mindset more apparent than in the realm of “Free.” Wired editor Chris Anderson discusses three varieties of the “free” business model.

  • In the first two, cross subsidies and advertising support, the amount of funding available for “free” content, products and services depends directly on how profitable the related business—advertising or otherwise—is.
  • The third, “freemium,” is itself nothing but “price discrimination”: “a few people subsidize everyone else.”

To be blunt: the sustainability of all these models requires that somebody has pricing power. If the government takes that away because differential pricing seems “unfair,” what will fund the “free” content, services and products of the future? The counter-intuitive and politically unpopular reality is that greater “fairness” may actually make consumers worse off by impoverishing their digital lives.

Cloud Computing

Odlyzko’s final argument for inevitability of broad neutrality mandates lies in fears about monopolization of “cloud computing.” As the FCC recently asked in its Wireless Innovation and Investment Notice of Inquiry, “can a dominant cloud computing position raise the same competitive issues that are now being discussed in the context of network neutrality?” Odlyzko argues:

Yet another way that Google could justify its high valuation would be if the current hype about “cloud computing” becomes reality. If most of our data and computing migrate into a distributed network of processors and storage, entirely new opportunities would arise. Given the economics and the technology, with the “first mover advantage”, “economies of scale,” and similar factors, cloud computing could easily become dominated by one or a handful of players, who might then have huge power, and a corresponding profit potential. And Google is ideally positioned to grab such a role, since its core expertise is in deploying a distributed information technology infrastructure.

Odlyzko approaches “cloud computing” as “an extreme form of vertical integration, just carried out by other companies than the telecom service providers, and at higher levels of the protocol stack.” But unlike others who have expressed such concerns, Odlyzko notes that market forces would themselves constrain such “dominance:”

In either scenario, though, Google (or Yahoo! or Microsoft, or Facebook, or some other entity that might reach a position to exploit its position in the ways outlined here) would have to tread very carefully, since the public could easily decide its practices amounted to “being evil”.

Odlyzko also distinguishes himself from other Chicken Littles who claim the “Cloud is Falling!”, like Jonathan Zittrain.  While  Zittrain abhors the fact that not everyone wants to “tinker” with highly “generative” devices and applications, Odlyzko rightly recognizes (as Adam Thierer has) that vertical integration benefits the majority of users who just don’t want to bother:

There are attractions to end-users in some levels of vertical integration. The concept of a do-it-yourself end-to-end network is attractive, but few users have the skills and patience to make it a reality. So we will likely see extensive vertical integration, the only question is where, and what dangers will it produce.

So give these market constraints on dominance of cloud computing and the benefits of vertical cloud integration to consumers… why would we need broad “cloud neutrality” mandates ? Or, why are antitrust laws (such as PFF proposed with its DACA project) inadequate? Odlyzko doesn’t answer these critical questions—but we can be sure that others will attempt to do so in the future. Again, Odlyzko’s argument seems to be an almost deterministic one about the inevitability of regulatory creep than a direct endorsement of broad regulation.

A Final Word about Mutually Assured Destruction

Just as Milton Friedman lamented the “Business Community’s Suicidal Impulse,” we’ve warned that Silicon Valley seems to be driving the expansion of neutrality mandates in a regulatory “war of all against all.” Odlyzko suggests an economic explanation for why some large online service providers might favor neutrality mandates for the infrastructure layer:

For Google (or some other entity) to fully exploit the potential of differential pricing or of cloud computing (and the latter would surely also involve extensive price discrimination), it is necessary to have a communications infrastructure that either that entity controls, or that is approximately net neutral. Otherwise, the communications network could grab the profits by charging differential fees to Google that would absorb most of the benefits, or could deploy its own competing “cloud” which might be less efficient, but could exclude competitors.

This kind of thinking probably does play a role in some corporate strategizing, but if businesses were really able to think this ahead, they should also be able to see that this isn’t a game they can ultimately win, because the same basic argument will ultimately be used to impose neutrality mandates on them once they rise to “the top of the heap.” As we concluded paper about Net Neutrality, Slippery Slopes & High-Tech Mutually Assured Destruction:

Unless we find a way to achieve “Digital Détente,” the consequences of this increasing regulatory brinkmanship will be “mutually assured destruction” (MAD) for industry and consumers…  “Digital Détente” would require that all parties concede something and work constructively toward a more “peaceful” (i.e., less regulatory) resolution. And yet, no Internet company wants to disarm unilaterally, foreswearing politics as a continuation of competition by other means. Only through multilateral disarmament could they break out of the current cycle of regulatory one-upmanship: If the companies in the Internet ecosystem could form a united front against increased government regulation and in favor of removing existing regulatory obstacles to competition, they could all return to their core competencies of creativity and innovation.

  • Pingback: Oh Farts! The Droid, the iPhone & the Lessig-Zittrain Thesis — Technology Liberation Front

  • mwendy

    Some very smart people wrote the FCC's Net Neut NPRM. Yet, it reveals a pacuity of factual underpinning to guide its conclusion for more regulation. Madison River + Comcast + some vague concerns about choking off free speech by gatekeepers on the Internet = prophylaxis regulation to save us from ourselves.

    Regulations represent the commons into which one can easily litter without fear of having to internalize those costs. Smart people, being who they are, think that they can always stay ahead of the slippery slope. It rarely comes to pass. Exceptions, loopholes, new mandates continue almost unabated – the CFR grows like a weed.

    Detente demands that the regulators stop, too.

  • http://uis.blogspot.com/ Aakash

    Wow – Thanks for writing an in-depth and detailed piece, on this issue.

    Americans for Tax Reform has been discussing this issue recently; this is a particularly interesting piece.

    I wish I had more time to look into, and write about, this important issue. Thanks again for your coverage.

  • friend of NN

    Odlyzko is right that “cloud neutrality” might be called for next, but as someone who supports net neutrality, I wouldn't support that.

    I support net neutrality in the last mile because I think that broadband access in the home is not competitive, and as fiber rolls out the natural monopoly tendencies are going to become ever more pronounced. I have a choice of two broadband providers, and I'm lucky. Some people only have one.

    I'm talking only about last-mile stuff here. No one that I know of has ever called for the regulation of the core network, which is highly competitive.

    You may disagree how and if natural monopolies should be regulated, and even if they exist (e.g. you may believe, as I don't, that wireless is a viable competitor to fiber in the relevant ways), but that's my basis for believing in NN. Laying fiber to houses is like laying sewers. (And yes, overbuilding in some areas will lessen the need for regulation in those areas.)

    But Google has only lock-in, no real natural monopoly. Consumers will jump ship if something better comes along, and along with that the fear of a Google death sentence. It's just totally different.

  • http://techliberation.com/author/berinszoka/ Berin Szoka

    So, how competitive is “competitive enough?” And what makes you so sure that policymakers in the future will continue to draw the line, as you have, at what you would consider “natural monopolies?”

  • friend of nn

    I don't think that you should avoid good regulations for fear of later bad regulations. Food safety regulations can go too far, blocking GMOs or irradiation, for example, but I still support them.

    I know some Cato types oppose even food and aeronautical safety regulation, and I'm not read to go down that rabbit hole.

    What's “competitive enough”? Maybe an HHI of less than 3000? But there needs to be a time factor, as well. A market may be very concentrated but still subject to competitive entry, Schumpetarian or otherwise, if the current actors get too cocky. And again, many sectors of last mile bb are very competitive (in urban cores).

    The fact that line-drawing is hard doesn't mean that, in life and in law, you don't need to draw lines sometimes. I'd rather we picked a bright line rule than a balancing kind of test to determine what's the market concentration point that triggers regulation, because in my view the advantages of hard-and-fast rules outweigh the occasional unjust results.

    Part of my belief in NN also comes from my prediction that fiber to the home is the end of the line for the foreseeable future, for communications infrastructure built to the home. Intermodal competition is a historical fluke. There is nothing on the horizon technologically that can compete with fiber. Not satellite, not wireless. (Maybe docsis 3.0, but that's a likely fiber substitute–I think we'll not see docsis 3 and fiber in the same market for a while)

  • http://techliberation.com/author/berinszoka/ Berin Szoka

    Of course, it's true that if you define the market narrowly enough, it will seem less competitive. But, In all seriousness, what makes you think fiber is a unique market unto itself? You're certainly right that for power users, there's nothing like fiber. But for the vast majority of users, even the relatively slow and high-latency connections of 3G wireless is perfectly adequate for 95% (if not more) of what they do.

  • mwendy

    I have three b-band pipes available to my house. I can switch. That's risky stuff – if even it were two. Last mile represents some of the most expensive investment, and one which should have a high reward.

    What gets me is that some provider who pays simple access to a NAP or POP in some far away location – let's say San Jose – gets access to me in Alexandria, VA. And they don't have to worry about any of the research or risk it took to wire me up. They pay once (or relatively little in re value) and then get to “fly” to some highly desirable / valuable locations (like my zip) with little additional investment or risk. What a deal!

    I don't advocate an access charge regime like with POTs, but one should have sympathy and some understanding for those that are doing some heavy lifting here at the last mile. You don't have a network without that risk.

  • friend of NN

    Odlyzko is right that “cloud neutrality” might be called for next, but as someone who supports net neutrality, I wouldn't support that.

    I support net neutrality in the last mile because I think that broadband access in the home is not competitive, and as fiber rolls out the natural monopoly tendencies are going to become ever more pronounced. I have a choice of two broadband providers, and I'm lucky. Some people only have one.

    I'm talking only about last-mile stuff here. No one that I know of has ever called for the regulation of the core network, which is highly competitive.

    You may disagree how and if natural monopolies should be regulated, and even if they exist (e.g. you may believe, as I don't, that wireless is a viable competitor to fiber in the relevant ways), but that's my basis for believing in NN. Laying fiber to houses is like laying sewers. (And yes, overbuilding in some areas will lessen the need for regulation in those areas.)

    But Google has only lock-in, no real natural monopoly. Consumers will jump ship if something better comes along, and along with that the fear of a Google death sentence. It's just totally different.

  • http://techliberation.com/author/berinszoka/ Berin Szoka

    So, how competitive is “competitive enough?” And what makes you so sure that policymakers in the future will continue to draw the line, as you have, at what you would consider “natural monopolies?”

  • friend of nn

    I don't think that you should avoid good regulations for fear of later bad regulations. Food safety regulations can go too far, blocking GMOs or irradiation, for example, but I still support them.

    I know some Cato types oppose even food and aeronautical safety regulation, and I'm not read to go down that rabbit hole.

    What's “competitive enough”? Maybe an HHI of less than 3000? But there needs to be a time factor, as well. A market may be very concentrated but still subject to competitive entry, Schumpetarian or otherwise, if the current actors get too cocky. And again, many sectors of last mile bb are very competitive (in urban cores).

    The fact that line-drawing is hard doesn't mean that, in life and in law, you don't need to draw lines sometimes. I'd rather we picked a bright line rule than a balancing kind of test to determine what's the market concentration point that triggers regulation, because in my view the advantages of hard-and-fast rules outweigh the occasional unjust results.

    Part of my belief in NN also comes from my prediction that fiber to the home is the end of the line for the foreseeable future, for communications infrastructure built to the home. Intermodal competition is a historical fluke. There is nothing on the horizon technologically that can compete with fiber. Not satellite, not wireless. (Maybe docsis 3.0, but that's a likely fiber substitute–I think we'll not see docsis 3 and fiber in the same market for a while)

  • http://techliberation.com/author/berinszoka/ Berin Szoka

    Of course, it's true that if you define the market narrowly enough, it will seem less competitive. But, In all seriousness, what makes you think fiber is a unique market unto itself? You're certainly right that for power users, there's nothing like fiber. But for the vast majority of users, even the relatively slow and high-latency connections of 3G wireless is perfectly adequate for 95% (if not more) of what they do.

  • mwendy

    I have three b-band pipes available to my house. I can switch. That's risky stuff – if even it were two. Last mile represents some of the most expensive investment, and one which should have a high reward.

    What gets me is that some provider who pays simple access to a NAP or POP in some far away location – let's say San Jose – gets access to me in Alexandria, VA. And they don't have to worry about any of the research or risk it took to wire me up. They pay once (or relatively little in re value) and then get to “fly” to some highly desirable / valuable locations (like my zip) with little additional investment or risk. What a deal!

    I don't advocate an access charge regime like with POTs, but one should have sympathy and some understanding for those that are doing some heavy lifting here at the last mile. You don't have a network without that risk.

  • Pingback: https://www.youtube.com/watch?v=tYE3oe7ddbQ

  • Pingback: How to Astral Project

  • Pingback: IMP SOURCE

  • Pingback: Dave Pascht

  • Pingback: topsail rentals

  • Pingback: Website

  • Pingback: Ibiza

Previous post:

Next post: