May 22, 2012

"Centralization can lead to price controls, censorship without due process, lack of reader privacy, and resistance to innovators" (*). Can one best this blast against centralization, the enemy of capitalism? Most pernicious, as it kills the invisible hand which Adams Smith credited with wealth creation. Most insidious, because the progressive disappearance of what is invisible to begin with is bound to escape notice.

Brewster Kahle and Rick Prelinger's warning focuses on current attempts to build a universal digital library. The authors condemn the lot as being monopolistic in principle or, at least, in practice. They call instead for "a robust, distributed publishing and library system" with "lots of publishers, booksellers, authors, and readers - and lots of libraries". Transpose the warning to your own domain of interest, you will not impair its validity.

If centralization is dangerous though, does it follow that decentralization is necessarily good?

The two periods preceding and following the Xth century have both been honored with the term "Renaissance". One can hardly take this as a compliment for the intervening time during which our Western civilization experimented with decentralization on a grand scale.

One should not confuse decentralization with anarchy. Xth century Europe structured itself according to feudalism, which recognized a sovereign authority, be it that of the Carolingian emperor or his successors.

The former should look familiar to us today as a typical case of overextended big government, unable to muster sufficient economic resources to back its implicit promises of empire wide protection. Hugh Capet (1), one of the latter, would appeal on the contrary to those who dream of small government. While his family had an excellent record of local resistance against Viking marauders, he was elected king precisely because his relative lack of resources reassured his more powerful vassals.

Judged on the economic record of the Xth century, decentralization is not a panacea by itself. Nor is small government.

A thousand years later, the same conclusion still applies. Take insider trading. Left unchecked, it organizes an invisible, hence powerful transfer mechanism which taxes the many to benefit the few.

Gretchen Morgenson makes the case that, in the United States, the successful prosecution of individual villains like Raj Rajaratnam, former head of Galleon, "glosses over risks that insider trading can and does occur regularly at many Wall Street firms" (**). In what she dubs its "institutionalized" form, "Wall Street firms alert favored clients to analyst research before the investing public". R. Foster Winans showed how rewarding this can be.

Whether or not this is true and could be proven in court, it illustrates this apparent paradox. To work, decentralization depends on a strong government. Financial markets are but a sham if run for insiders and the SEC doesn't make them disgorge all illegal profits.

This is why the current campaign in the United States for small government is so dangerous. Beyond the need to reassess the necessary mix of capitalism and solidarity which powers society, it attacks the right of the government to rule over the markets.

Railing against taxes levied by the government while fighting to protect one's right to levy invisible taxes is a movement as medieval as it is modern. The comparison is even more accurate if one takes into account the predominant source of wealth. Back in the Food Age, feudalism was based on the right to land revenues. Now in the Information Age, the new lords demand the right to data revenues.

When ready money is rare, it is in fact a very reasonable request to ask for what is abundant, whether land a thousand years ago or data today.

In his column about online education (***), Thomas L. Friedman lets "Daphne Koller, a Stanford computer science professor who [co]founded Coursera.org" explain that her company "will also be working with employers to connect students - only with their consent - with job opportunities that are appropriate to their newly acquired skills". Professor Koller nicely nods to student privacy. But I would be surprised if students refused this way of looking for a job, recruiters did not have to pay to be matched and any of this revenue was distributed back to participating students.

Coursera.org is hardly a case of egregious abuse. According to Nick Wingfield, Microsoft could also claim it is sensitive to users' concerns when it decided that "sharing Internet searches in an identifiable way with Facebook friends seem[ed] to cross some kind of line" and pulled the feature out of its latest search engine (****). The real world though is not without rogues and in such a world self-regulation is an oxymoron. My point is that both companies mentioned operate in the absence of robust eprivacy rules established and enforced by the United States government.

Instead said government finds itself busy defending itself. Two years ago, it was the target of the Digital Due Process Coalition. Today "the Republican-led House voted to eliminate the [American Community Survey] altogether, on the grounds that the government should not be putting its nose into Americans' homes" (*****).

I fully appreciate that governments the world over, democratic or not, are wont to consider eprivacy hinders their missions. If for instance the American Community Survey were to ask respondents for their consent, Catherine Rampell reports it would cost money. Indeed consent against money is one aspect of what eprivacy is about.

But the purpose of the data lords is not to protect our privacy, it is to make sure nothing undercuts their own right to appropriate our personal data for free, with or without the fig leaf of a coerced consent. In its present form, the American Community Survey is a hidden tax, and data lords want all such taxes to flow into their coffers. By competing with them, governments play into their hands and weaken themselves into paralysis.

The data lords recognize even weak governments can make laws. Read Somini Sengupta's account of Facebook lobbying in Washington (******). When connected with campaign contributions, this allows companies to buy laws. The originality of Facebook is to pitch the use of its service by US Congress members as a valuable campaign contribution in itself. And no limit ever applies since the service is free!

Meanwhile Facebook bills itself the universal people's directory, Google wants to provide all the information one will ever need.

Forget decentralization. We now witness the rise of competing centers of power, the very spectres feared by Brewster Kahle and Rick Prelinger.

Philippe Coueignoux

• (*) ........... Many libraries, by Brewster Kahle and Rick Prelinger (Technology Review) - May/June, 2012
• (**) ......... Is Insider Trading Part of the Fabric, by Gretchen Morgenson (New York Times) - May 20, 2012
• (***) ....... Come the Revolution, by Thomas L. Friedman (New York Times) - May 16, 2012
• (****) ..... A Feature Bing Left Out, by Nick Wingfield (New York Times) - May 14, 2012
• (*****) ... The Beginning of the End of the Census, by Catherine Rampell (New York Times) - May 20, 2012
• (******) . Facebook Builds Network of Friends in Washington, by Somini Sengupta (New York Times) - May 19, 2012
• (1) for more information, see Hugh Capet in the wikipedia