China’s Covid lockdowns are likely costing the country at least $46 billion a month, or 3.1% of GDP, in lost economic output, and the impact could double if more cities tighten restrictions.

That’s a minimum estimate from an economist at the Chinese University of Hong Kong based on the assumption that cities generating about 20% of China’s gross domestic product are currently imposing targeted lockdowns. That cost would double if those areas had to follow Shanghai and impose stricter policies requiring most residents to remain at home. 

China’s tough measures mean “the economic cost of lockdown is clearly bigger than we see in other countries,” said Zheng Michael Song, a CUHK professor who is part of a research team using real-time data to measure the impact of lockdowns. The 3.1% estimate—equivalent to 295 billion yuan ($46.3 billion)—is “conservative,” he added, as it doesn’t include the effect on incomes through inflation.

Song and his team use data on the location of nearly 2 million trucks which crisscross China, and whose movements are highly correlated with local economic activity. 

By including the impact on nationwide inflation and spillover effects from supply chains, the effect on the economy would be much worse. A strict lockdown in Shanghai alone could reduce China’s real GDP by 4%, Song and his co-authors estimate. If China’s four largest cities all underwent a strict lockdown together, national inflation-adjusted GDP would fall 12% for the duration of the shutdowns.

A worst case scenario would be a lockdown of all cities for one month, which would cut national GDP by 53% over that period.

China has been fighting a wave of the more-infectious omicron variant of coronavirus since the beginning of the month, with daily confirmed cases rising above 6,000 this week. Goldman Sachs Group Inc. estimates areas with mid to high-risk outbreaks under some form of lockdown restriction account for about 33% of GDP.

The key to the prospects for China’s economy will depend on whether the highly infectious omicron variant will require more intense or longer lockdowns. Shanghai initially tried a targeted approach, but that didn’t bring down case numbers and it said this week it would require half of the city to stay at home in turn over a total of eight days while mass-testing and isolation are carried out. 

In Shanghai, which began introducing coronavirus restrictions earlier this month, trucking data indicates economic activity in the city fell 40% below normal even before the half-city lockdown began on Monday. Activity could fall further still if the city follows the path of Changchun, the capital of Jilin province in the northeast, which imposed a lockdown over omicron cases on March 11. It’s seen economic activity levels plunge by more than 66% from normal levels since then, according to the trucking data.

Shenzhen appears to represent a best case scenario of what China’s lockdowns can achieve against omicron. The city’s week-long lockdown, during which residents were tested three times, appears to have been relatively light—it only reduced economic activity by about 34%, data tracked by Song and his team suggests. The shorter duration also reduced the economic impact, although truck activity was still 20% below normal a week after the lockdown ended.

The team previously estimated that China’s coronavirus lockdowns in the first quarter of 2020 led to a 19.4% decline in real GDP, compared to the official figure of a 6.9% decline. Economists at Nomura Holdings Inc. wrote this weekend that China’s economy faces its worst outlook since that quarter due to the spreading virus outbreaks. 

Song forecasts a better outcome than 2020 if Shenzhen serves as a model.

“The one week full lockdown in Shenzhen seems to have been more effective,” he said. If other cities can achieve a similar result “the overall performance of the economy will be much better even if a third to half of cities see Covid cases, as they did in early 2020.”

It’s still possible for China’s economy as whole to grow even if lockdowns reduce economic output, if other regions not under restrictions increase their output. But that becomes less likely the larger the lockdown impact is.

Lockdowns have become more targeted since 2020, with restrictions imposed in specific rather than entire cities, and so tend to reduce local activity by around 30% compared with more than 60% for earlier lockdowns like those seen in Wuhan in early 2020, according to Song. The shutdowns have also become shorter, lasting 24 days on average before the current omicron wave compared with Wuhan’s two and a half months. 

“That’s an indication that the government learned from past experience and has updated their policies,” he said.