Democracy may demand a splintered internet

Democracy may demand a splintered internet

The Trump administration’s attack on TikTok, which may simply be motivated by the president’s “America First” instincts, nevertheless raises an important question of principle. Can cutting off market access to a foreign digital company because of how it may handle the data it collects be a legitimate act, based on principled policy preferences? Or is it necessarily an act of protectionism, serving only to tilt the playing field in one’s own company’s advantage, and which free traders ought to oppose?

The US itself is quick to shout “protectionism” if others — in particular the EU — put constraints on American internet companies’ market access. This is illustrated by former US trade negotiator Charlene Barshefsky, who berates Europe’s digital sovereignty agenda in an FT op-ed.

But double standards aside, the basic question remains, and I find myself siding with Trump against TikTok. Surely governments can legitimately put in place rules for how their citizens’ data are handled. Citizens have an obvious and profound interest in this, given how data can be used to spy on, monitor and manipulate both their knowledge and their behaviour, through algorithmic filters of what they encounter on the internet. Given the reasonable expectation that data gathered by TikTok are in the last instance accessible by its Chinese parent and China’s government, Trump’s determination to cut off this access is defensible.

A different perspective is that from my colleague Yuan Yang, who in a recent column nudges readers away from the idea of breaking ties with China or its companies. She rightly points out that “if you’re worried about Beijing getting its hands on your data, you should be concerned about giving it to American groups too”. 

But how far does this analogy take us? After all, public opinion in many countries is in fact concerned about how American companies use their data; this why the EU puts conditions on transferring personal data elsewhere, and why its court has just ruled that Privacy Shield, the existing system for data transfers to the US, is not good enough.

A second contrast between China and liberal democracies is that the governments of the latter are accountable (if sometimes imperfectly) to their citizens. There is a way for citizens’ preferences over data handling to be expressed in the rules that govern them. But Beijing is not accountable to its own people, let alone users of Chinese digital services in other countries.

Restrictions on what companies can do with data their services collect primarily have to do not with trade but with democracy. Europe’s particularly consumer-friendly data rules reflect the preferences of Europeans, who seem on average more privacy-conscious than Americans. EU law is much more restrictive than US law when it comes to harvesting location data from mobile users, for instance.

That instance is a good illustration why data restrictions — and excluding companies and services that violate them — are not inherently protectionist. Many apps made and serviced by US companies are available to European consumers, but they may not legally collect and sell location data outside strict limits. That is not protectionist; it simply requires US digital services providers to abide by the same standards as European ones. It is similar to how exporters of industrial goods to Europe must abide by the EU product standards required for “CE” marking. While that can be used for protectionism if intentionally designed to discriminate against foreign production, it is not inherently protectionist.

There is a difference, however. Physical goods can be kept out if non-compliant and let in when they conform with the rules. But data, once collected, are potentially there forever — at least if they have been sent to another jurisdiction — and there is no way to undo the collection if the rules are broken later. Which means that if we take seriously the rules and the public preferences that ideally underpin them, we may have to restrict data from being transferred across jurisdictions in the first place. Hence the Privacy Shield challenge; hence, too, data rules emerging as significant, if often unnoticed, elements of the Brexit negotiations.

As an FT editorial this week acknowledges, this could lead to a “splinternet” — separate online arenas that allow data to flow within but not (or not as much) between jurisdictions. I have argued before why, for Europe at least, this is nothing to fear. But supporters of free trade should wish for greater agreement between countries on what the rules governing data should be, in order to facilitate more global data flows. This is, of course, what the EU has achieved inside its own bloc, where common rules underpin fully free flows of data. More generally, it illustrates the important principle that today, to be in favour of trade liberalisation is to promote common rule-making.

But common rule-making presupposes some fundamental common ground on what the rules need to do. That exists (just) among EU members. It would be good if it could be established between the EU and the US, as my colleague Rana Foroohar hopes for, because it would give them greater leverage with China. The reality, however, is that both sides try to encourage third parties to follow their own regulations; the new US-Mexico-Canada trade deal, for example, includes language mirroring the US’s own “Section 230” that affords platforms legal immunity for the content they host.

For the foreseeable future, we should learn to love the splinternet.

Coronanomics readables

  • In my latest FT column, I argue that fiscal orthodoxy is an important casualty of the Covid-19 crisis.

  • Another, more depressing, casualty is British artisanal cheesemaking.

  • If you think the coronavirus downturn itself is bad, wait until you see how it has messed with the minds of consumers and managers. A new paper by Julian Kozlowski, Laura Veldkamp and Venky Venkateswaran suggests that the long-term costs of “scarred beliefs” — people now thinking disastrous outcomes are more likely than they thought before — will be many times greater than the short-term losses from the downturn itself.

Other readables

  • After Libor and other benchmark interest rates were revealed to be scandalously liable to manipulation, central banks and regulators have worked on establishing new reference interest rates to replace them. But Michael Troege warns that the replacements, too, can be manipulated.

Numbers news

  • The expiry of the US’s federal supplement to state unemployment insurance could cause a collapse in aggregate demand. A new study finds the demand effect of more generous unemployment insurance is particularly strong, because the money reaches those most likely to spend it. “The decline in consumption from ‘no supplement’ is greater than the entire peak-to-trough decline of the Great Recession,” says one of the authors, Peter Ganong, in a Twitter thread on the research. Meanwhile, the Washington Post’s Catherine Rampell rounds up five studies disproving the worry that the US’s temporarily generous unemployment insurance, which leaves some workers better off than while working, might undermine efforts to bring down unemployment.