A new study finds that incumbent parties lose votes after their citizens get online
THE EARLY days of the internet were full of predictions about access to information unleashing a wave of democratisation. More recently, views on the internet’s impact have soured, as states use it to spy on dissidents and influence foreign elections.
Opinions on this topic are abundant, but hard data are scarce. No one knows whether the Arab spring could have occurred without the internet, or whether Russia’s online efforts to boost President Donald Trump’s campaign had any effect. Nonetheless, scholars can sometimes find natural experiments to substitute for such counter-factual scenarios. A recently revised study by the economists Sergei Guriev, Nikita Melnikov and Ekaterina Zhuravskaya, which is now undergoing peer review, uses the growth of mobile broadband to reveal a link between internet access and scepticism of government.
Most of the 4.1bn people now online got connected after 2010. To measure how new users’ views changed as a result, the authors combined two datasets. First, for each year in 2007-18, they estimated the share of people in each of 2,232 regions (such as states or provinces), spread across 116 countries, that could access at least 3G-level mobile internet. Then they used surveys by Gallup, a pollster, to measure how faith in government, courts and elections changed during this period in each area.
In general, people’s confidence in their leaders declined after getting 3G. However, the size of this effect varied. It was smaller in countries that allow a free press than in ones where traditional media are muzzled, and bigger in countries with unlimited web browsing than in ones that censor the internet. This implies that people are most likely to turn against their governments when they are exposed to online criticism that is not present offline. The decline was also larger in rural areas than in cities.
A similar pattern emerged at the ballot box. Among 102 elections in 33 European countries, incumbent parties’ vote-share fell by an average of 4.7 percentage points once 3G arrived. The biggest beneficiaries were parties classified as populist—though this may simply have been because they happened to be in opposition when voters turned against parties in power, rather than because of their ideology.
A central (and disconcerting) implication is that governments that censor offline media could maintain public trust better if they restricted the internet too. But effective digital censorship requires technical expertise that many regimes lack. In Belarus, where the government tries to control media both online and off, an opposition news channel on Telegram, an encrypted mobile app, has 2m subscribers—one-fifth of the country’s population.