- China Leadership Monitor
CLM Insights Interview with Minxin Pei
Minxin Pei. The Sentinel State: Surveillance and the Survival of Dictatorship in China. Cambridge, MA. Harvard University Press, 2024. 336 pp. ISBN-10: 0674257839; ISBN-13: 978-0674257832.
Insights Interview
What motivated you to write this book? What were the main challenges in collecting the data?
About eight years ago, as the economy continued to perform poorly, I thought that the Chinese Communist Party (CCP) would need to rely more on repression to maintain control, I was going to write a book on China’s repressive capacity But the topic is too broad, so I wanted to find a focus. Then I read a lot of press reports about China’s hi-tech surveillance state and how this system monitors activities. So the Chinese surveillance state seemed to be the perfect focus because it was a hot topic and very little comprehensive research had been done on the subject because of the difficulties in gaining access to official sources. Indeed, initially, collecting data was a real challenge. Most of the data on the surveillance system, such as the scope of surveillance (the number and kind of people targeted) are difficult to find. But over time, I was able to use local data (mainly yearbooks and gazetteers) to piece together the puzzle. China has nearly 3,000 county-level jurisdictions, most of which publish their own yearbooks that provide brief reports on their accomplishments in various areas, including law enforcement. Some sensitive information is bound to leak. The more data I found, the more I learned what else I needed to find and where to look.
What are the most important findings of the book?
Overall, China has a far more sophisticated and capable surveillance system than any other dictatorship either in history or today. Perhaps it is unfair to compare today’s CCP with the Communist Party of the Soviet Union or the Communist regime in the former East Germany because the latter two did not have advanced technologies. But even if we exclude the technology factor, the Chinese system is better organized and more economical. There are several institutional innovations that allow the CCP to control the secret police and coordinate domestic security, in particular surveillance of potential targets, more effectively. One is the establishment of a specialized party bureaucracy, the political-legal committees at all levels of the state, to oversee domestic security. None of the former Communist regimes in the Soviet bloc had a similar bureaucracy. Another is the ability to distribute the tasks of surveillance to various security agencies and state-affiliated actors, such as universities, neighborhood committees, and state-owned enterprises. This allows China to maintain a relatively small secret police force. If China had the same ratio of secret police per population as did the former East Germany, the headcounts of the Ministry of State Security and of the Political Security Protection units of the Ministry of Public Security (MPS) (the main domestic secret police) would total some 8.5 million personnel. But today the total number of uniformed police in China comes to about 2 million. Finally, the Chinese Leninist party-state has an extensive and deep reach into Chinese society and the economy, thus facilitating the recruitment of a large number of informants to perform surveillance tasks, mostly without compensation. My rough estimate based on disclosures in local government yearbooks is that xinxiyuan, the most generic category of informants, likely account for 0.7–1.1 percent of the population (or 9.8–15.4 million). This figure excludes the more specialized informants who are recruited by the police.
Contrary to the prevailing view that modern technologies, such as facial recognition and big data, constitute the pillars of the Chinese surveillance state, the backbone of the system actually is labor and organization. To be sure, modern technologies have greatly enhanced the party’s surveillance capacity. But surveillance technology has many blind spots, so labor is still needed. Most surveillance tasks, such as identifying and verifying the physical whereabouts of so-called Key Individuals, a category that includes repeat petitioners, practitioners of cults, ex-PLA soldiers, and ethnic minorities, must be performed by local officials and informants.
Because by design China has a system of “distributed surveillance,” coordinating surveillance requires effective bureaucratic organization and mobilization. This is where human labor, not technology, performs a far better job. At the beginning of each year, the central Political-Legal Committee convenes an annual conference that sets the domestic security agenda, which is then communicated to the lower levels of the government through a series of meetings convened by the local political-legal committees. Operationally, the party typically steps up surveillance around so-called “sensitive dates,” such the annual sessions of the National People’s Congress. Extra security measures taken by local authorities, such as the assignment of teams of five people to watch one key target, are very labor-intensive.
Last but not the least, despite the media hype on the subject, we really do not know much about how China’s hi-tech surveillance system actually works. Nearly all of the coverage of this subject focuses on the acquisition of hardware, based on examinations of government procurement records, which are sometimes accessible online. But it is actually quite difficult to access the nitty-gritty technical details of the system because these pieces of information are highly classified. Aside from what kind of hardware China has installed, there is a lot of additional work to do for scholars who want to understand how the hi-tech piece of the system works and how it is integrated into the labor-intensive piece.
Is the Chinese surveillance state exportable?
The tech piece is definitely exportable, but not the labor and organization-intensive piece. Countries or autocracies importing China-made hi-tech gear will likely increase their surveillance capabilities, but they will not acquire the same level of surveillance as that which now exists in China. This is because they cannot import the CCP. Their regimes do not have the same reach and penetration of society and the economy as the CCP. If they lack the same leverage over ordinary people that the CCP has, then they are unlikely to be able to recruit and maintain a vast network of mostly volunteer informants. As hi-tech surveillance can primarily spy on what people do in public and its targets are able to take simple evasive measures, such as wearing masks or not carrying their phones, it has many blind spots. This does not mean that autocracies will not benefit from China-made hardware. But with hardware alone they will not acquire the same degree of social control as the CCP.
Will surveillance be able to keep the party in power in the potentially difficult times ahead?
The CCP has done very well in the post-Tiananmen era, but its sophisticated surveillance apparatus should be given only partial credit. A booming economy obviously produced fewer potential enemies as the overall level of satisfaction among the population has been high. But if the economy goes into a sustained slump, high unemployment, falling income, and other economic ills could increase the level of social tensions and produce more potential threats. Chinese surveillance does a relatively effective job when the number of people under watch is small. Based on my analysis of the local yearbooks, the total number of people blacklisted on two mass surveillance programs, Key Populations (run by the MPS) and Key Individuals (most likely under the direction of the local political-legal committees) is between 7.3 and 12.7 million people, or slightly under 1 percent of the total population. If economic stagnation drives another 1 percent of the population to engage in protests or other anti-government or anti-social activities, the current surveillance system could become overwhelmed.
Another potential problem China will face during economic stagnation is how to pay for the maintenance and upgrading of the hi-tech piece of its surveillance state. Local governments, which cover about 85 percent of China’s spending on domestic security, footed the upfront costs to build the system when the economy was doing much better than it is today. As the hi-tech gears require intensive maintenance (facial recognition cameras must be cleaned constantly for them to remain effective and integrated data networks are prone to glitches) and regular updates are essential to improve their security and performance, cash-strapped local governments may be hard pressed to shoulder such expenses in the future.
Protracted economic stagnation will most likely lead to a less effective surveillance state. But that does not mean this factor alone will threaten regime survival. The party can always fall back on harsher methods of repression to maintain control. In the overall context of political and economic malaise, perhaps a deteriorating surveillance system is not the first thing that keeps Xi Jinping up at night. There are many other items on that list, in particular his authority within the party because only regime insiders pose a plausible threat to his rule today.
