Everyone – or at least everyone who reads the Cayman Financial Review – knows about tax havens, but what about data havens? The idea is simple.
Put some servers in a country with strong privacy laws and no other laws to speak of, run some fast network cables to them, and open for business. Data havens are designed to serve customers afraid of badge-bearing censors, copyright rent-a-cops, onerous discovery requests, narrow-minded gaming authorities and every other kind of governmental intrusion into their bits’ private business.
Such, at least, was the dream of HavenCo, a data haven start-up launched in 2000 by a pair of American libertarian computer geeks. Sean Hastings and Ryan Lackey went far beyond the usual island suspects, though, when it came time to pick their offshore home. Its founders met on Anguilla, but HavenCo’s chosen jurisdiction was the Principality of Sealand. And if HavenCo was unlikely, then Sealand is truly off the wall.
Sealand, more a nanonation than a micronation, claims as its territory a dilapidated anti-aircraft platform in the North Sea, seven miles offshore just past the mouth of the Thames Estuary. The platform, built in World War II for coastal defence, was abandoned by the British government after the war, then reoccupied in the 1960s by pirate radio broadcasters. Its offshore location, safely beyond the three-mile limit of England’s territorial waters, allowed it to fit into a temporary jurisdictional lacuna in the United Kingdom’s broadcasting laws.
By the time a 1967 statute finally fit the noose around offshore broadcasters, Sealand was in the hands of Roy Bates, a charismatic fisherman-slash-radio entrepreneur with a genius for public relations, a taste for adventure and a penchant for occasional violence. He fought off the staff of Radio Caroline, a competing pirate broadcaster, with air rifles and Molotov cocktails. And then, in September of 1967, he declared his own country, and named himself Prince.
The British government, for obvious reasons, disputed Bates’ claims. But, with classic bureaucratic ineffectuality, it found itself unable to do anything about it. Bates spurned offers to buy him out, and when the government started making plans to occupy or demolish the platform, he embarrassed it with strategic leaks to the press, who leapt at the chance to help this colourful David poke fun at a stodgy Goliath.
In 1968, Bates’ son Michael forced the issue by taking potshots at a passing survey ship. Father and son were indicted on firearms charges. But in court, Bates prevailed. Whatever kind of thing an anti-aircraft platform was (ship? island?), it was outside the three-mile limit of territorial jurisdiction, and Parliament hadn’t extended admiralty jurisdiction to cover firearms offenses. In the hands of Sealand’s advocates, this limited ruling would become a basis for claims of sovereignty.
So Sealand was its own country, or at least enough of one for a plausible business plan. In 1999 and 2000, Hastings and Lackey made contact with the Bates family and negotiated a deal with Michael Bates, now running Sealand day-to-day under the name of Prince Regent Michael.
The royal family would take a part ownership stake in HavenCo in exchange for extending its legal protection to the company. No child pornography, no spam, but a friendly home for dissidents, Page Three pornography, and other content in popular demand but guaranteed to offend the politicos and prudes. It was the pirate radio dream reborn.
And, for the first few years, everything seemed to be going well. The press coverage was wall-to-wall; HavenCo’s principals shared their hosts’ intuitive flair for drama. Online gambling sites rented servers, as did the makers of unauthorised accelerators for IBM servers. Ryan Lackey spoke grandly of machine-gun-toting guards, server rooms in the platform’s concrete legs and ambitious technical enhancements.
But when outside reality caught up with the HavenCo dream, it wasn’t pretty. The dot-com crash exposed the economic impracticality of shlepping servers seven miles offshore, only to pipe the data back, slowly and at great expense, to network access points back on dry land. Customers were few; there never had been huge racks of servers. (It wasn’t security concerns that kept reporters out of the legs of the tower.)
And then, in the summer and fall of 2002, HavenCo and its hosts found themselves falling out. HavenCo, completely unsurprisingly, had a high tolerance for copyright risk: it found a customer eager to set up a streaming-video service similar to one shut down by a federal judge in the summer of 2011. But Sealand’s advisors, hoping for broader recognition from the great nations of the world, were wary of anything that might offend the United Kingdom or the United States. They nixed the deal and also started making demands on the company that Lackey, by now in sole charge of HavenCo, saw as incompetent intrusions into his technical area of expertise.
What started as an attempt at an amicable separation turned into an ugly divorce. Lackey left the platform, never to return. He accused the Sealand government of nationalising HavenCo and expropriating his personal computers. Sealand said that he was no longer an employee of HavenCo and that he had no valid visa even to set foot on its concrete-and-steel soil.
What went wrong?
For one thing, in the early 2000s, governments were discovering that they weren’t nearly so powerless before the Internet as they had feared. It didn’t matter if locals were offshoring their servers; they couldn’t offshore themselves or their customers.
People actually present in a jurisdiction could still be arrested and served. And as for those located abroad, well, a combination of intermediary liability and outright Internet blocking provided substantial leverage. It wasn’t perfect, to be sure, and still isn’t but the idea that simply bouncing one’s bits off the right data haven would solve one’s problems with the authorities was becoming more obviously false.
What of Sealand’s sovereignty? HavenCo’s “product” was formal legal compliance without substance. Its services were subject to Sealand law, but the entire point of being subject to Sealand law was that there wasn’t any. But without a navy, diplomatic clout or any foreign trade worthy of note, Sealand didn’t have much success in convincing the major nations of the world that Sealand law was worthy of any deference whatsoever. No one was interested in negotiating a tax treaty with Sealand, much less an Internet freedom treaty.
All of this made HavenCo’s services an unappealing prospect for law-abiding businesses. For law-breakers, there wasn’t much point either. This was the age of Napster and the birth of peer-to-peer networks. Why pay for expensive colocation service when free sharing works just as well for samizdat? And for the true criminals, even formal legal compliance was pointless: better just to rent server space using stolen credit card numbers or laundered money.
And HavenCo’s fall also illustrates a dark irony about law. HavenCo was dependent on the protection of Sealand’s government; otherwise it would just have been the world’s most impractical data centre. But Sealand’s family-run government was a constitutional monarchy vesting near-absolute executive, legislative and judicial power in the Prince.
When Sealand took over HavenCo and exiled Lackey, where could he go for redress? Not to Sealand courts run by his erstwhile business partner. And not to any other court in the world, not without conceding the very jurisdictional principle on which HavenCo was founded: Sealand’s absolute sovereignty.
HavenCo’s play was based on legal avoidance: it tried to monetise the absence of Internet law. And in one sense, its failure demonstrates the resilience of law and national authority. In another, HavenCo failed because it didn’t have enough law. Having moved to the near anarchy of Sealand, HavenCo was helpless when Sealand law proved inadequate to protect it. HavenCo tried to undermine law, but in the end, law was the one thing it needed most.