Adam Thierer's Blog, page 134

Information Control, Market Concentration, and the AT&T/T-Mobile Deal

Like Milton, I'm very worried about the political vulnerabilities that might arise if the wireless sector grows more concentrated. Still, I think it's a big mistake to legitimize one repressive incarnation of coercive state power (antitrust intervention) to reduce the likelihood that another incarnation (information control) will intensify. This approach is not only defeatist, as Hance argues, but it also requires a tactical assessment that rests on several dubious assumptions.

First, Milton overestimates the marginal risk that the AT&T – T-Mobile deal will pave the way for an information control regime. The wireless market isn't static; the disappearance of T-Mobile as an independent entity (which may well occur regardless of whether this deal goes through) hardly means we're forever "doomed" to live with 3 nationwide wireless players. With major spectrum auctions likely on the horizon, and the possibility of existing spectrum holdings being combined in creative ways, the eventual emergence of one or more nationwide wireless competitors is quite possible — especially if, as skeptics of the AT&T – T-Mobile deal often argue, the wireless market underperforms in the years following the acquisition.

More importantly, network operators, like almost all Internet gatekeepers, face mounting pressure from their users not to facilitate censorship, surveillance, and repression. Case in point: AT&T is a leading member of the Digital Due Process coalition (to which I also belong) that's urging Congress to substantially strengthen the 1986 federal statute that governs law enforcement access to private electronic communications. Consider that AT&T's position on this major issue is officially at odds with the official position of the same Justice Department that's currently reviewing the AT&T – T-Mobile deal. Would a docile, subservient network operator challenge its state overseers so publicly?

Or take Google. Arguably, it's an enormously important gatekeeper — in many respects, it's an even greater "chokepoint" than any single network operator — but the firm has held strong against substantial pressures from the U.S. government to facilitate censorship and surveillance. (See, e.g.: Google's successful 2006 challenge to a Justice Department request seeking search query logs; Google's recent refusal to remove DUI apps from its Android market; Google's widely noted hesitance toward censoring search results absent a lawfully issued takedown request; Google's 2010 joint amicus brief urging a federal district court to reject a Justice Department subpoena seeking to compel Yahoo! to disclose the contents of a user's opened emails).

I could go on. The point is that large network operators are often willing to vigorously resist — both in private and in public — governmental demands that they facilitate information control. Working in cahoots with unpopular governmental actors is terrible PR; some major players seem to think it's bad for business, too.

It's often overlooked that antitrust intervention deprives us of beneficial competitive reactions to business deals — even deals that, viewed in isolation, appear to be "harmful." The consequence of the AT&T – T-Mobile deal won't simply be the two companies operating as a single entity; the deal will also force rivals to respond in unforeseeable ways that will tend to benefit consumers and fuel innovation. As Hance reminds us, this virtuous cycle of Schumpeterian creative destruction is fundamental to the long-term evolution of markets. When government blocks proposed business arrangements, it contributes to stasis — and static markets tend to be much easier to regulate and control than relatively dynamic markets.

To be sure, if the combination of AT&T and T-Mobile exacerbates political pressures for imposing a network information control regime, we cyber-libertarians should fight back vigorously. Turning to antitrust intervention to keep markets relatively unconcentrated — and, hence, more difficult to regulate — is a mistake.

If you welcome the growing pressures for regulating business arrangements in the high-tech sector, an emboldened antitrust regime is just what the doctor ordered.

Bitcoin, intermediaries, and information control

I'm gratified that my recent writing on the Bitcoin virtual currency project has stirred much conversation and I thought I'd take a moment to continue that conversation.

Tim Lee has written two posts critiquing the viability of Bitcoin from the supply and demand side. Dan Rothschild has responded in part. Tyler Cower also weighed in.

To address Tim I'll simply say this: Do I think Bitcoin will replace the dollar? No. Might Bitcoin have certain systemic design flaws that might impede its success? Quite possibly. Will Bitcoin become the de facto, manipulation-proof currency of the internet? Who knows. Tim's posts are a somewhat technical critique of Bitcoin's long-term feasibility. It's a great contribution, but since I'm neither a gold bug nor a Bitcoin booster per se, I don't find it especially interesting.

That all said, what I do think is revolutionary about Bitcoin is that its developers have solved, without the use of a middleman, the double-spending problem faced by virtual currencies. That gives us license to realistically imagine a world without regulable financial intermediaries online.

While Tim overlooks what makes Bitcoin radical, Tom Sydnor groks it viscerally. Writing in a lengthy comment on my post, Tom expresses dismay at what Bitcoin represents and offers what I would, with apologies, characterize as the cyber-conservative response.

In his work looking at who should govern the internet and who should regulate activities on it, Milton Mueller has made the case for a system of "denationalized liberalism," which "favors a universal right to receive and impart information regardless of frontiers, and sees freedom to communicate and exchange information as fundamental and primary elements of human choice and political and social activity." For reasons I won't go into here, I like the idea of denationalized liberalism. However, short of asking us to show up at ICANN meetings to fight back government encroachments, Milton has not really given us any practical paths to such a world. The reason is that it is very difficult. As Jack Goldsmith and Tim Wu have pointed out, governments ultimately have control over the persons (and servers) under their physical jurisdiction.

In "A History of Online Gatakeeping," (PDF) Jonathan Zittrain catalogs how intermediaries serve as the obvious targets of regulation for governments seeking to control information. Think of Napster, PayPal, and DNS registrars. Ever improving peer-to-peer technologies, like Bitcoin, remove this layer of intermediation by avoiding centralized servers that can be regulated or shut down. Now here is a path to denationalized liberalism: decentralized the infrastructure so that it is less susceptible to bordered control. Bitcoin show us a way to decentralize money transfers, and others are working on P2P DNS systems. It's this path I want to explore in my research.

Now, it isn't a clear path. As Zittrain writes:

The loss of these natural points of control will cause those with challenged interests to foreground a new and less palatable set of intermediaries: software authors. These authors may be asked to write their software in such a way that it can be recalled or modified after it has been obtained by a user and then put to an undesirable purpose. They may even be asked to program their software to disable the installed software of others. Control over software — and the ability of PC users to run it — rather than control over the network, will be a future battleground for Internet regulation, a battleground primed by an independently-motivated movement by consumers away from open, generative PCs and toward more highly regulable endpoint platforms.

I'll stay away from the question of "generative" vs. "sterile" devices for now and just say that he's right that the battle will move to software. Tom Sydnor's response to my comment highlights how those who oppose an unregulable internet for order and stability reasons will react.

First, Tom says that getting people to run the Bitcoin software (thus creating the Bitcoin P2P network) is fraud because consumers would not engage in what might be legally questionable behavior unless they were "tricked" into doing so. The problem with this logic is that it's not clear who is doing the tricking or defrauding. There is no Bitcoin company inducing circumvention of financial regulations. That's the point of decentralization.

Second, and more important, Tom says that the reason running Bitcoin is legally questionable is that an argument can be made that 1) the predominant use of the Bitcoin network is trafficking in illicit goods and services, and 2) running the software should therefore carry vicarious or contributory liability for those offenses. This is in effect an argument for regulation of software, and I suspect this is the type of argument that will be advanced in any efforts to ban software like Bitcoin.

I'll skip the practical feasibility of banning software short of also banning "generative" PCs, much less doing so globally, and stick to addressing the legal questions. Are Bitcoin's predominant uses illicit? Well Dude, we just don't know. But there might certainly be "substantial non-infringing uses," to borrow a phrase. You can imagine contributing to WikiLeaks and other unpopular organizations, buying banned content under repressive regimes, and preserving personal privacy, say if you're buying something legal but potentially embarrassing.

More to the point, Bitcoin is less like Grokster, Inc., and more like BitTorrent the protocol. Again, there is no third party inducing bad behavior. Bitcoin is a tool that can be used for good or ill. (A lot like paper cash.) Hopefully governments will prosecute those who do ill, and not simply seek to ban the technology. If they do, that's a battle we'll have to be prepared to fight. As for vicarious or contributory liability from simply joining the Bitcoin P2P network, last I checked doing math was not yet a crime. When you run Skype on your computer, you help create the Skype P2P network over which no doubt illegal activities transpire all the time. Should Skype users be held vicariously liable?

I'm looking forward to discussing this further. What other pieces of the online infrastructure can be decentralized? How far can we decentralize? What will the reactions from governments be? What are the implications for crime and security? What are the implications for personal liberty?

An Internet Eraser Button to Protect Privacy? Unwise & Probably Impossible

In my latest "Technologies of Freedom" column for Forbes, I take a closer look at the idea of an "Internet eraser button" as one method of protecting privacy or safeguarding reputation online. The child safety group Common Sense Media has suggested it is needed to help kids and others wipe out embarrassing facts we've place online but later come to regret. The Eraser Button idea is similar to "the right to be forgotten" proposal currently being hotly debated in Europe.

In my column, I argue that "it is unlikely that such a mechanism could be implemented, and even if it could, it would have troubling ramifications for freedom of speech, digital commerce, and Internet governance more generally." I dwell a bit on the free speech issues and note that "What we are talking about here is the destruction of history, otherwise known as censorship. Few would have suggested that burning books was a smart way to protect privacy in the past. Is burning binary bits of information any wiser?" But the point seems moot in light of the significant enforcement challenges the notion faces, including the question: Who actually owns the data collected by online sites and services?

Anyway, read the rest of the essay over at Forbes. And here are a few other pieces we've run here at the TLF on the issue: 1, 2, 3.

Open minded on the AT&T/T-Mobile merger

Is it "insane" for free market oriented thinkers to support the AT&T/T-Mobile merger?  Although AT&T says there are five choices of wireless providers to choose from in 18 of 20 major markets, Milton Mueller argues that 93 percent of wireless subscribers prefer a seamless, nationwide provider.  If the merger is approved, there would only be three such providers.

A market dominated by three major providers is neither competitive nor noncompetitive as a definitional matter.  Factual analysis is necessary to determine competitiveness.

And it may be premature to conclude that there is no competitive significance either to the fact there are over a hundred providers currently delivering nationwide service on the basis of voluntary roaming agreements that are common in the industry, or to assume that the possibility the FCC will double the amount of spectrum available for wireless services will not impact the structure of the industry.

The trouble with antitrust generally is the possibility that government will choose to protect weak or inefficient competitors, thus preventing meaningful competition that attracts private investment which leads to innovation, better services and lower prices.  Antitrust is supposed to protect consumers, not politically influential producers.  Although this sounds simple in theory, it can get confusing in practice.  As free market oriented thinkers, we do not want government picking winners and losers.

As Larry Downes correctly points out, if concentration leads to higher prices there will be more profits for competitors and more competition.  Schumpeter teaches that short-run monopolies come and go, while long-run monopolies are exceedingly rare – unless "buttressed by public authority."

Mueller makes a couple other important points that reflect the thinking of a lot of people.  One is AT&T's customer satisfaction ratings, prices and network performance.   This merger would provide extra spectrum the carrier needs to address each of these issues, however.

Although there are other ways to increase network capacity, they won't be enough.  The National Broadband Plan notes that data traffic on AT&T's mobile network increased 5,000% between 2006 and 2009.  According to the chairman of the FCC, "We need to tackle the looming spectrum crunch by dramatically increasing the new spectrum available for mobile broadband, and the efficiency of its use."  That's why the broadband plan is proposing to double the amount of spectrum available for wireless, a process that could take several years.  Historically, it has taken 6-13 years to reallocate spectrum.

Mueller also points out that AT&T cited network capacity issues when it opposed microwave-based competition in long-distance (in the 1960s).  AT&T has been criticized for the advocacy it used in those days.  The real problem the company faced then was being saddled with a pricing structure that was incompatible with competition, but policymakers were not about to change that.  Regulation mandated that AT&T charge high prices for long-distance to subsidize "affordable" (i.e., at-or-below cost) local service.  That set up an attractive cream-skimming opportunity for potential competitors such as MCI.  The new entrant's business plan was to compete for the fat long-distance profits and not get stuck with money-losing local service obligations.  That was a real problem.

Finally, Mueller suggests limited government intervention as a preventative for excessive government intervention.

If you support a competitive industry where one can reasonably expect the public and legislators to rely on market forces as the primary industry regulator, this merger has to be stopped. On the other hand, if you welcome the growing pressures for regulating carriers and making them the policemen and chokepoints for network control, a bigger AT&T is just what the doctor ordered.

A similar argument was made in the early 1980s with regard to local and long-distance service.  Regulate local so long-distance can be free.  This is so defeatist.  As free market thinkers, we need to stand for principle.  Unwarranted government intervention in the free market should be opposed.

If the FTC, Department of Justice and/or FCC do find that such a merger would substantially lessen competition, they could also impose specific conditions to ameliorate the impact – making it unnecessary to block a merger outright.  Minimum intervention ought to be the first choice for free market thinkers.

Gavin Andresen on Bitcoin

On the podcast this week, Gavin Andresen, project lead of the open source, decentralized, and anonymous virtual currency project Bitcoin, talks about the project. Andresen explains how the peer-to-peer currency functions and talks about what allows Bitcoin to operate without a central bank, why it doesn't have to rely on intermediaries, and how it overcomes the double-spending problem. He also discusses the project's implications for government regulation, what attracted him to the project, and Bitcoin inventor Satoshi Nakomoto's motivation for creating the currency.

Related Links


"Online Cash Bitcoin Could Challenge Governments, Banks," by Jerry Brito
Bitcoin: A Peer-to-Peer Electronic Cash System (pdf), by Satoshi Nakomoto
ClearCoin, Andresen's startup Bitcoin escrow service
"How to Start Your Own Private Currency," The Atlantic

To keep the conversation around this episode in one place, we'd like to ask you to comment at the web page for this episode on Surprisingly Free. Also, why not subscribe to the podcast on iTunes?

Why I fear the AT&T-T-Mobile merger

I was surprised to read a defense of the AT&T-T-Mobile merger on this blog, but I suppose I should have expected it. The stakes are high and AT&T's spin machine is formidable, extending even into free market blogs.

Let's begin at the beginning and ask why this merger is happening. It's not as if AT&T is gaining dominance the way Google gained it in search and advertising, or the way Intel did in chips: i.e., through low prices, superior products and customer loyalty. No, last time I looked AT&T was the carrier with the lowest customer satisfaction ratings, some of the highest prices and one of the weakest network performance metrics. In my opinion there is no reason for this merger to take place other than to make life easier for AT&T by reducing competitive pressures on it. AT&T seems to be driven by the following calculus. It can either grow its services and its network under the harsh constraints of market pricing and competition, or it can attempt to reduce the field to an oligopoly with tacit price controls by using its size and financial bulk to eliminate a pest who keeps downward pressure on pricing and service requirements. I think it is rational for AT&T to try to get away with the latter. I think it's insane for free market oriented thinkers to support it.

Larry Downes can't argue with the extremely high level of market concentration and the scary HHI measurements that the merger would produce. So he plays the game that clever antitrust advocates always play: shift the market definition. Downes argues that "both Justice and the FCC have consistently concluded that wireless markets are essentially local." I see no citation to any specific document in Downe's claim, but if FTC and FCC have concluded that "local" means "my metropolitan area" they are wrong.

Let's reacquaint everyone with a very basic but pertinent fact: 93% of the wireless users in the U.S. are served by the national carriers. This number (the proportion served by national as opposed to regional providers) has generally increased over the past decade, driven by both demand-side requirements, mergers, and supply-side efficiencies. The choices of consumers have rendered a decisive verdict negating Downes's claim. Whether it's voice or data, people expect and want seamless national service; a small but significant segment wants transnational compatibility as well.

Increases in the scope of service will intensify as we move from a primarily voice-driven market to a data-driven market. Carriers who have to impose roaming charges and interconnection fees on their users will not be competitive. Nor will they be able to attract the interest of the cutting-edge handset manufacturers and service developers. Can you imagine Apple signing an iPhone exclusivity deal with Cricket?

It is no accident that the dominant mobile network operators have national brands and national footprints. Most Americans travel outside their metro areas at least once a month, and go places further away than that at least once or twice a year. The 93% who choose a national carrier are rationally calculating that it pays to not have to guess the service area limits of their provider. Of course, a highly budget-constrained segment of the market will accept limited local service for a lower price. To say that those smaller providers are in the same market as a T-Mobile or AT&T is not plausible. They occupy a niche. And if one allows a major merger like this on the grounds that these tiny players constitute a competitive alternative to the likes of AT&T, what are you going to say as the last of these local providers is gobbled up?

How about that "spectrum efficiency" argument? Downes, like the AT&T Corp. he is probably consulting for, makes the same claim that the old AT&T made when it said there should be no microwave-based competition in long distance. As a matter of pure engineering efficiency, it is of course true that a single, optimizing planner can make better use of limited spectrum bands than multiple, competitive providers. But then, that argument applies to any and all carriers (an AT&T-Verizon merger, for example) and to any resource – that's why it was used by the socialists of the 19th century to claim that capitalism was inherently wasteful and inefficient. Dynamic efficiencies of competition typically benefit the public more than a few allocative efficiencies. And there are plenty of ways for AT&T to expand network capacity without merging.

But there is an interesting twist to this line of reasoning. Notice how the "market is local" claim suddenly disappears. AT&T needs to take over a smaller national rival, according to Downes, so it can "accelerate deployment of nationwide mobile broadband using LTE technology, including expansion into rural areas." Voila! Once we start talking about spectrum efficiencies and the promotion of universal service we take a nationwide perspective, not a local one. Doesn't this obvious contradiction make anyone suspicious?

Notice also the ominous historical overtones of AT&T's claim that it will be able to promote universal broadband service in rural areas if it has a stronger monopoly er, if it gains consolidation efficiencies. Hey, rural areas don't have congested spectrum, do they? What's stopping them from doing that now? If they need help to do it, where are the subsidies going to come from? Would more market power make that possible? One cannot help but ask: Is AT&T doing this to get more spectrum or is it trying to pull a neo-Theodore Vail, and promise the government that it will subsidize rural access if it has more market power?

Bottom line: this is one step too far back to the days of a single telephone company. If you support a competitive industry where one can reasonably expect the public and legislators to rely on market forces as the primary industry regulator, this merger has to be stopped. On the other hand, if you welcome the growing pressures for regulating carriers and making them the policemen and chokepoints for network control, a bigger AT&T is just what the doctor ordered.

Balancing Risk and Regulation

So a few weeks ago I hit up Adam Thierer, who has done and is continuing to do great work on all things regulation, on some materials for a project I was working on regarding the precautionary principle in the digital space. Turns out Adam was in the middle of his own Digital Precautionary Principle piece as well. I'll take our simpatico as a sign that this phenomenon may actually be taking place and that I'm not paranoid. (If you haven't read his earlier piece on TLF, please do so).

While my piece on DPP is coming, hopefully this week, I'll start things off with my article in today's RealClearMarkets.com on regulations and risk and how regulating agencies are engaging in traditional "risk aversion behavior" to the detriment of the risk takers (aka entrepreneurs) in the private market. A smarter approach to regulating would incorporate both benefits and risks of NOT regulating. So many times the discussion is geared towards the notion that something has to be done, so how can we minimize the negative impacts, rather than, should we be doing anything at all or should we encourage the trial and error mechanisms that markets utilize?

While the piece isn't targeted directly at the technology industry, I think it can apply there just as much as any other industry.

 

The AT&T – T-Mobile Merger: Beyond the Arithmetic

Following AT&T's announcement last month of its planned acquisition of T-Mobile USA, pundits and other oddsmakers have settled in for a long tour of duty. Speculation, much of it uninformed, is already clogging the media about the chances the $39 billion deal—larger even than last year's merger of Comcast and NBC Universal—will be approved.

Both the size of the deal and previous consolidation in the communications industry lead some analysts and advocates to doubt the transaction will or ought to survive the regulatory process.

Though the complex review process could take a year or perhaps even longer, I'm confident that the deal will go through—as it should. To see why, one need only look to previous merger reviews by the Department of Justice and the Federal Communications Commission, both of which must approve the AT&T deal.

Critics of the deal argue principally that a reduction from four national wireless carriers to three will create grave risks to competition. But Justice and the FCC have consistently rejected such simple-minded analysis. Instead, as consent decrees for several wireless mergers over the last decade–under Democratic and Republican administrations—make clear, both agencies approach the unique economic features of mobile communications with more subtle tools.

For example, both Justice and the FCC have consistently concluded that wireless markets are essentially local. Their competitive analysis—the key in reviewing horizontal mergers of this type—therefore focuses on the choices available to consumers where they buy wireless service; typically where they live, work or shop.

In today's dynamic mobile industry, some national wireless carriers are strong in some cities or rural areas and weak or absent from others. Beyond the national carriers, lower-priced providers including MetroPCS and Cricket, as well as established regional companies including US Cellular, are strong in local markets. The Justice-FCC market analysis will consider market structure at the local level, counting all providers who are genuinely competitive.

The discussion so far about market concentration levels—measured by the Herfindahl-Hirschman Index, or HHI—ignores the more detailed analysis that the DOJ and the FCC have performed for past mergers including Sprint-Clearwire and Verizon-Alltel (both in 2008).

HHIs are commonly used as starting screens to identify markets where anti-competitive effects might result. But the two agencies have historically concluded that anti-competitive effects will occur only where concentration is especially high–at levels of the HHI well above the published estimates for the AT&T-T-Mobile deal in most markets. In particular, the focus has historically been on local markets where the merger would result in too few remaining competitors. In those markets, local divestitures have been required.

On that analysis, AT&T probably will be required to divest some consumers in some local markets, but fewer than would result from strict application of the high-level HHI screens.

The FCC must also consider potential benefits of the deal that improve the ''public interest."  Here, the agency will take into account serious capacity constraints both AT&T and T-Mobile are already experiencing on their networks. AT&T argues the merger will allow it to optimize scarce spectrum, improve network performance and quality, and accelerate deployment of nationwide mobile broadband using LTE technology, including expansion into rural areas.

The FCC and DOJ will require evidence to support these claims, of course. But assuming AT&T can back them up, they constitute strong public interest benefits.  After all, these are all goals the FCC itself established as part of last year's National Broadband Plan. Likewise, as part of its evaluation, the DOJ will consider these and other claimed synergies as pro-competitive efficiencies produced by the merger.

Finally, that the deal is being vigorously opposed by some competitors—notably Sprint—actually helps AT&T's case. Antitrust enforcers are understandably skeptical of claims by competitors that a merger will hurt them. Why? If a merger leads to higher prices for consumers, competitors such as Sprint would actually benefit. So when a competitor complains a merger is anticompetitive, the agencies take that as evidence the deal will in fact produce a stronger rival. And a stronger rival is good for consumers.

In the end, Justice and the FCC will have to weigh the competitive risks of further consolidation against the benefits for American consumers of improved service and accelerated deployment of nationwide mobile broadband. If history is any guide, the merger will ultimately be approved with specific conditions, including divestitures, to ensure that local competition is preserved even as national benefits are achieved.

That, of course, assumes the agencies follow their own best practices and not naysayers who can only count down from four to three. Let's hope they do.

Reproduced with permission from Daily Report for Executives, 69 DER B-1 (Apr. 11, 2011). Copyright 2011 by The Bureau of National Affairs, Inc. (800-372-1033)

The Epsilon Breach: Inference and Exaggeration

News about the Epsilon breach has spread relatively slowly. The breach of data held by an email service provider is bad—no question—but it's not terribly consequential. Emails aren't generally kept private.

But the Epsilon story may soon heat up. The presence of an email address on a list creates inferences about aspects of a person's life that may be sensitive. So it is with GlaxoSmithKline's lists related to prescriptions. As the Coalition Against Unsolicited Commercial Email points out, correlation between email addresses and interest in particular drugs makes spear-phishing attacks more potent. Fraudulent email that is tailored to a medication a person takes will have a higher uptake than average, and could be used to defraud people on matters relating to their health.

But is it helpful to exaggerate this serious threat? CAUCE titles its post: "Criminals Now Know What Prescriptions You Take." Thought leaders like Jules Polenetsky have picked up that meme and run with it.

For people who are not data-literate, a likely implication of "criminals know what prescriptions you take" is that criminals have access to lists of the prescriptions they take. A person on ten different medications might think that criminals know each and every prescription he or she takes. That's more frightening than knowing that an association between one or two prescriptions and an email address is available to criminals. (It's possible that people have signed up for email relating to each of their prescriptions, all of which are from drug companies who use Epsilon as their email service provider, but I think it is unlikely and rare enough to treat as an irrelevant outlier.)

What criminals know is that people are on lists related to prescriptions. Many do take that prescription. Some used to take that prescription. Some have a loved one who takes it, some sell it, some prescribe it, and so on.

What's the point of this observation? Not much. But under the rule of media and politics—"if it bleeds, it leads"—we may soon see a media and policy stampede. That stampede will treat an important security issue that deserves careful attention as a techno-cyber-apocalypse that demands immediate overreaction.

Bitcoin: Imagine a net without intermediaries

Yesterday the FBI effectively shut down three of the largest gambling sites online and indicted their executives. From a tech policy perspective, these events highlight how central intermediary control is to the regulation of the internet.

Department of Justice lawyers were able to take down the sites using the same tools we've seen DHS use against alleged pirate and child porn sites: they seize the domain names. Because the sites are hosted overseas (where online gambling is legal), the feds can't physically shut down the servers, so they do the next best thing. They get a seizure warrant for the domain names that point to the servers and to point them instead to a government IP address, such as 50.17.223.71. The most popular TLDs, including .com, .net, .org, and .info, have registrars that are American companies within U.S. jurisdiction.

Another intermediary point of control for the federal government are payment processors. The indictments revealed yesterday relate to violations of the Unlawful Internet Gambling Enforcement Act, which makes it illegal for banks and processors like Visa, MasterCard and PayPal to let consenting adults use their money to gamble online. According to the DOJ, in order to let them bet, the poker sites "arranged for the money received from U.S. gamblers to be disguised as payments to hundreds of non-existent online merchants purporting to sell merchandise such as jewelry and golf balls." (PDF)

Now, imagine if there were no intermediaries.

In my TIME.com Techland column today, I write about Bitcoin, a completely decentralized and anonymous virtual currency that I think will be revolutionary.

Because Bitcoin is an open-source project, and because the database exists only in the distributed peer-to-peer network created by its users, there is no Bitcoin company to raid, subpoena or shut down. Even if the Bitcoin.org site were taken offline and the Sourceforge project removed, the currency would be unaffected. Like BitTorrent, taking down any of the individual computers that make up the peer-to-peer system would have little effect on the rest of the network. And because the currency is truly anonymous, there are no identities to trace.

And if a P2P currency can make it so that there is no fiscal intermediary to regulate, how about a distributed DNS system so that there are no registrars to coerce? This is something Peter Sunde of Pirate Bay fame has been working on. These ideas may sound radical and far-fetched, but if we truly want to see an online regime of "denationalized liberalism," as Milton Mueller puts it, then getting rid of the intermediaries in the net's infrastructure might be the best path forward.

Again, check out my piece in TIME for a thorough explanation of Bitcoin and its implications. I plan to be writing about it a lot more and devote some of my research time to it.