An economist’s perspective on ‘gain-of-function’ virus research

There has recently been a great deal of focus on whether COVID-19 was created in a lab, or whether it crossed over to humans from an animal host. While this is an important question to resolve, there is a larger one that also needs to be addressed: Do some types of lab research on viruses inherently involve unacceptable risk to the broader public? 

Even if COVID-19 did not escape from a lab, this pandemic has important implications for the future path of “gain-of-function” research, where viruses are genetically modified to become more virulent or easily transmitted, or both. To see why, we need to consider three facts. 

First, it turns out that lab accidents are actually quite common, even in supposedly safe facilities. Here’s Zeynep Tufekci in the New York Times:

“Nearly every SARS case since the original epidemic has been due to lab leaks — six incidents in three countries, including twice in a single month from a lab in Beijing. In one instance, the mother of a lab worker died.

“In 2007, foot-and-mouth disease, which can devastate livestock and caused a massive crisis in Britain in 2001, escaped from a drainage pipe leak at an English lab with the highest biosafety rating, BSL-4.

“Even the last known person who died of smallpox was someone infected because of a lab incident in Britain in 1978.”

In addition, it is highly likely that the global H1N1 flu epidemic of 1977, which killed hundreds of thousands of people, was caused by a lab leak.

Second, while this research can help us to better understand potential new viruses, Harvard epidemiologist Marc Lipsitch has argued that gain-of-function research provides only very small benefits relative to the risks associated with potential lab leaks. Partly as a result of these concerns, in 2014 the Obama administration put a moratorium on the funding of gain-of-function research (although some exceptions were allowed). This restriction was lifted by the Trump administration in 2017.

Third, the consequences of a lab accident could be horrific. Whether or not COVID-19 was caused by a lab leak, the mere possibility that a human accident has killed at least 3.9 million people (with millions of deaths in Russia, Latin America, South Asia and Africa having likely gone unrecorded) is jarring. COVID has also devastated the global economy. Future gain-of-function research might well produce far more dangerous viruses, as the infection fatality rate for COVID-19 is believed to be below 1 percent.

As an economist, I worry about whether decisionmakers face the appropriate incentives, and whether there are important external costs of their decisions being imposed on others. Do lab workers underestimate the probability of a lab leak? Are they 1 million times more careful when dealing with viruses that could kill a million people than when dealing with viruses that might kill one infected worker? Given what we know about cognitive biases, there is reason to worry that lab workers – even professional scientists – might not fully appreciate the risks being taken in their research.

Misaligned incentives pose an additional problem. Richard Enright observed that working in (lower-level) BSL-2 labs “effectively increas[ed] rates of progress [in research], all else being equal, by a factor of 10 to 20.” In other words, researchers with laxer standards have a huge time advantage in the race for scientific breakthroughs.

The United States is in an especially influential position in this area of research. David Wallace-Wells argues that we dominate this area of science at almost all levels, including the training of scientists, the development of theories and techniques and even the funding of labs such as the Wuhan Institute of Virology. He suggests that scientists tend to do research that can get published in top American science journals, and if those journals stop publishing inappropriate gain-of-function research, then most scientists would stop doing these experiments.

The debate over the origin of COVID-19 has now become highly politicized. In both the West and in China, the lab leak hypothesis is widely seen as being much more embarrassing for China than the animal market hypothesis, although it is far from obvious why this should be the case. Western countries also do gain-of-function work on viruses and occasionally have lab leaks. 

There is a danger that the debate over the Wuhan lab could divert attention from the much more important question: Would a continuation of widespread gain-of-function research pose unacceptable risks to the public? In my view, the answer after a full weighing of the facts is likely to be yes, even if COVID-19 did not come from a lab.

Scott Sumner is the Ralph G. Hawtrey Chair of Monetary Policy with the Mercatus Center at George Mason University and a professor emeritus at Bentley University.