‘Network Effects’ Multiply a Viral Threat

 This isn’t merely a bad flu season. Covid-19 is spreading far faster than most Americans realize.

‘The coronavirus panic is dumb.” I hesitate to disagree with Elon Musk, but here goes.

The wrong way to think about the rapid spread around the world of the novel coronavirus, 2019-nCoV, and the disease it causes, Covid-19, is to say—as another smart and wealthy man put it to me last Monday—“Remember the H1N1-A virus of 2009? Neither do I. It infected a significant chunk of the globe, killed 20,000 U.S. citizens and we got over it pretty quickly.” He might have added that 20,000 is less than half the number of Americans who died of influenza and pneumonia in 2017.

H1N1, also known as swine flu, was a form of influenza. The reproductive number—the number of people a carrier typically infected, R0 for short—was 1.75. In the U.S., the CDC estimates that H1N1 infected 60.8 million people and killed 12,469, for a mortality rate of 0.02%

This new coronavirus—which is not influenza—appears to have a higher R0 and a much higher mortality rate. That rate is almost certainly lower than the World Health Organization suggested last week (3.4%), but it is still much higher than for H1N1. South Korea, which probably has the most accurate data given its aggressive testing regime, reports 50 deaths from 7,313 infections, a mortality rate of 0.68%. If as many Americans catch Covid-19 as caught swine flu, the death toll could exceed 440,000.

In short, Covid-19 has the potential to make 2020 much more than a bad flu season. To understand why, we need to apply more sophisticated frameworks than are being employed by most lay commentators, billionaires included.

“In Wuhan there also seems to be a new outbreak of pneumonia that’s bad.” That was the first mention in my email of the coming pandemic, on Jan. 4. Just over two weeks later, I noted that the new virus had “already caused three deaths in the city of Wuhan” but warned that it could spread rapidly, like SARS in 2003, as the outbreak coincided with the Chinese New Year celebrations. Ten days after that, on Jan. 30, the number of confirmed cases world-wide was 9,776 and the total deaths 213. As of March 8, there are more than 107,000 confirmed cases and close to 3,700 deaths.

At first, the number of cases outside China did not grow exponentially. But that changed in February. Three weeks ago, the number was doubling every eight days. Now it is doubling every five days.

Standard epidemiological models tend to understate the threat posed by a virus such as 2019-nCoV, because they don’t take account of the topology of the social networks that transmit it. Thanks to the work of network scientists such as Romualdo Pastor-Satorras and Alessandro Vespignani, we now understand the extraordinary power of modern transportation networks combined with the social-network hubs known as “superspreaders.”

To quote László Barabási, Mr. Vespignani’s colleague at Northeastern University, “When it comes to the spreading of a pathogen, the epidemic parameters are of secondary importance. The most important factor is the structure of the mobility network. . . . An influenza virus moves through a continent with the speed of a sports car or of a smaller airplane.”

That means travel restrictions tend to be imposed too late to stop the spread of a contagious virus along the routes between the world’s 3,000 busiest airports. What matters is not geographic distance but “effective distance” in journey time.

Then there are all the local networks on the ground, where airports, shopping malls, supermarkets and schools act as hubs connected to countless homes and offices.

Finally, and crucially, there are social networks, which have the same “scale-free” character as the transportation networks: A relatively small number of nodes have an amazingly high number of edges. In good times, these are the frequent flyers, the gregarious networkers, the sexual Lotharios. In times of contagion, they are the superspreaders, like Gaëtan Dugas, the Canadian flight attendant who claimed to have had 2,500 sexual partners and came to be known as “patient zero” in the early histories of AIDS, or Liu Jianlun, the physician from Guangdong Province who brought SARS to Hong Kong when he checked into the Metropole Hotel on Feb. 21, 2003.

Nicholas Christakis has shown that there was a similar pattern at Harvard when H1N1 came to campus. “The speed with which people acquired the flu during the epidemic, depended on various aspects of their social network position,” writes Mr. Christakis, who is now at Yale. “Those with more friends, those who were more central in the network, and those whose friends did not know each other got it sooner.”

Similar processes have caused Covid-19 to spread with startling speed around the world, then outward from transport and social hubs. A British businessman who went from a conference in Singapore to a ski trip in the French Alps was one of the first superspreaders to be identified in this epidemic. The virus reached Switzerland via a group of tourists returning from Italy.

For all these reasons, the number of known cases in the U.S. (436) must be off by at least one order of magnitude and more likely two, simply because of the disastrous shortage of test kits. According to Messrs. Pastor-Satorras and Vespignani’s Global Epidemic and Mobility model, the United States is the fifth-likeliest country to import Covid-19 from abroad—after Thailand, Japan, Taiwan and South Korea. If the U.S. turns out to have proportionately as many cases as South Korea, it will soon have some 46,000 cases and more than 300 deaths—or 1,200 deaths if the U.S. mortality rate is as high as Italy’s.

Network effects are the reason it is anything but dumb to worry about the novel coronavirus. Not only is it spreading much faster than most Americans realize; it is also disrupting global manufacturing supply chains as well as all the economic activities that depend on travel and proximity. It could set off a cascade of defaults in the corporate bond market, disrupting the global financial network.

Finally, cable news and online social networks can be relied upon to disseminate alarmist and downright fake stories about the pandemic—like the widely circulated map of global air routes that, according to one Australian website, depicted the “flight data of an estimated five million Wuhan residents who fled during the critical two weeks before the outbreak city was placed under lockdown.”

That aspect of the panic is indeed dumb. But that doesn’t make it smart to underestimate the scale of the Covid-19 pandemic—a perfect illustration of the vulnerability and fragility of our networked world.

Mr. Ferguson is writer and host of “Niall Ferguson’s Networld,” based on his 2018 book, “The Square and the Tower,” which will air March 17 on PBS.


The Wall Street Journal
  • Show All
  • Newsweek/Daily Beast
  • The Washington Post
  • The Australian
  • Daily Mail
  • Huffington Post
  • Vanity Fair
  • FORA.tv
  • The Telegraph
  • Time Magazine
  • Foreign Affairs
  • The Sunday Times
  • London Evening Standard
  • The Spectator
  • The Atlantic
  • The Globe and Mail
  • Politico Magazine
  • The Times Literary Supplement
  • The Wall Street Journal
  • Bloomberg
33 Article Results

Turning Points