How Zero Covid failed again in Chengdu

NEW YORK – Chengdu’s 21 million people went under Covid lockdown on September 1, among 65 million Chinese in 33 cities now held in place under the country’s “Zero Covid” policy, including Guiyang, the capital of Guizhou province.

Western media reported that local officials in Chengdu enforced the lockdown even as tremors hit the city from an earthquake in Sichuan that measured 6.8 on the Richter scale.

Shenzhen had closed its city center and ordered most residents to remain in place last weekend after a spurt of Covid cases. Shenzhen reported 71 new cases on Monday.

Thousands of Chinese factories are now working in a so-called closed loop, in which employees are required to sleep at their factories.

The damage will not be as great as what resulted from the month-long shutdown of Shanghai, with its 26 million people and its leading role in Chinese manufacturing, finance and transport, but nonetheless is extremely serious.

Chengdu emerged as a center of China’s aerospace and information technology industries. Its relationship to Shenzhen, the nerve center of Chinese information technology, is something like that of Austin, Texas, to Palo Alto.

Once again, we observe a bouncing-ball pattern of Covid transmission from province to province.

After the fact, statistical evidence is overwhelming that a small number of cases in a distant province – in the most recent case Xinjiang – can ignite a Covid wave in major population centers.

The twin cities of Chengdu and Chongqing, the former in Sichuan province, the latter a provincial-level municipality under the direct administration of the central government, form China’s largest population center with a combined population of 52 million. A major outbreak in the dual megalopolis would have serious consequences for China’s economy.

But, before the fact, it’s impossible to lock down the highly transmittable Omicron strain of Covid-19 fast enough to avoid further outbreaks – except by shutting down enormous swaths of China’s economy.

China’s complacency after its 2020 success in containing Covid-19 led Beijing to neglect measures that would have made the tragic Shanghai lockdown avoidable. Once the highly-contagious new Covid strains hit Shanghai in early April, China had no choice but to lock the city down, with a cost in death and suffering that still remains to be tallied.

But the disaster could have been mitigated by the mass application of better vaccines and by improvement in the AI/Big Data capacity that China applied with success in 2020.

How China’s Covid controls broke down: a quantitative inquiry

Statistical evidence is overwhelming that a small number of cases in one province can lead rapidly to a much larger outbreak in another province. The Omicron strain transmits so rapidly that China’s systems cannot contain it.

In the charts below, we disaggregate the national picture into pairs of provinces to illustrate the speed of transmission.

Shandong’s largest city, Qingdao, is 750 kilometers by road from Shanghai on the coast of the Yellow Sea. But the transmissibility of the new Covid strains is so high that a few travelers on an intercity bus might be sufficient to bring the pandemic to China’s economic capital.

By the time the authorities decided to lock down Shanghai, there probably was no other choice. Cases had begun to appear in Beijing, despite quarantine procedures for Chinese as well as international travelers to China’s capital. The Shanghai lockdown gave the government time to control the Beijing outbreak, which never exceeded a few hundred cases a week.

But now we observe the same pattern repeated between Xinjiang and Sichuan. Xinjiang’s capital Urumqi is nearly 3,000 kilometers from Chengdu, but there are 18 scheduled commercial flights a day between the two cities, as well as rail and bus connections.

The cross-correlograms below show the correlation between leads and lags of Covid cases (using a 5-day change in the case volume) between Shandong and Shanghai, and Xinjiang and Sichuan, respectively.

Econometric analysis confirms the strong predictive relationship between lagged values for Xinjiang Covid cases and current values of Sichuan cases.

Anecdotal evidence suggests that public opinion in China, especially among younger Chinese, has turned against “Zero Covid.” This contrasts markedly with public support for the government during the initial wave of the pandemic in 2020. An opinion poll conducted by the University of California San Diego in March 2020 showed strong confidence in China’s Covid policies:

Results from three of our most recent surveys – which asked respondents about their trust in the central and local governments in China, support of China’s political system, and opinions toward the United States – show remarkable growth in favorable opinions of the Chinese government, and declines in favorable opinion of the US.

The average levels of trust in both the central and local governments in China have steadily increased during the past year. On a scale of 1 to 10, the average level of trust in the central government – already high – increased from 8.23 in June 2019, to 8.65 in Feb 2020, and to 8.87 in May 2020. There was a similar upward trend for the average level of trust in municipal governments.

No credible polling data is available, but a sign of dissent came last week from a Chinese think tank, the Anbound Research Center, which titled a new report, “It’s Time for China to Adjust Its Virus Control and Prevention Policies.” According to Western news services, the report was posted on social media but deleted the following day. It gave no details of proposed changes.

It is unlikely that China will revise its Covid policy before the 20th Party Congress of the Communist Party in mid-October. President Xi Jinping has staked a lot of credibility on Zero Covid, and would be reluctant in the extreme to admit error prior to his presumed re-election. After the Congress, a more rational approach is likely. This probably would include:

  1. Enhanced monitoring of vital signs of a large sample of the population through smartphones and wearables (temperature, blood oxygen, pulse) to provide a more fine-grained “heat map” of Covid’s spread;
  2. The use of mRNA vaccines used successfully in Hong Kong to control the spread and severity of the pandemic earlier this year; and
  3. Shorter and more localized lockdowns rather than all-city shutdowns.

The chart above shows the number of mRNA vaccine doses administered in Hong Kong prior to the winter 2022 outbreak. The spike in cases was brief and rapidly controlled, in significant measure due to the superior vaccine.

Follow David P Goldman on Twitter at @davidpgoldman