State of Surveillance (Budgeting for Surveillance)

By Mareike Ohlberg and Jessica Batke

This article is part of a larger investigation of the PRC’s surveillance apparatus, including a feature article, a number of sidebar articles, and supplemental materials, all originally appearing on ChinaFile. The text below does not include illustrations and charts that are part of the full article.


The work of buying and building out surveillance infrastructure in China largely falls to local governments. Numerous localities and agencies make purchases as they implement a range of centrally-mandated surveillance projects and campaigns. Given the number of bureaucracies and initiatives taking part—as well as the Party-state’s tendency to withhold information from the public—it is difficult to precisely reconstruct a full account of China’s spending on surveillance-related products and services. According to China Digital Times, the central government distributed 3.1 billion renminbi (U.S.$463,000) to Sharp Eyes projects between its launch in 2015 and the end of 2017. The central government’s five-year plan for video surveillance projected that “from 2016 to 2020, 250 model cities [would] allocate 10 billion renminbi to ‘Project Sharp Eyes.’”

ChinaFile’s review of some 76,000 government procurement notices offers a far more detailed accounting of just how much money central and local governments are committing to the surveillance push, particularly within the ambit of Sharp Eyes. Even with our likely incomplete dataset, which includes notices from June 29, 2004 through May 19, 2020, we tallied up more than 14.3 billion renminbi spent on surveillance acquisitions that referenced Sharp Eyes. Notices specifically for facial recognition technology across the whole country (including both notices mentioning and not mentioning Sharp Eyes) accounted for spending of some 8.6 billion renminbi.

How does this expenditure compare to what any given city might spend on, say, education, or other spheres of governance? In order to get a sense for the relative cost of surveillance purchases around the country, we zeroed in on seven cities to look at their budgets more in depth. These were localities for which our computer program was able to extract price information for the majority of that locality’s procurement notices. We then identified publicly-available budget information from local government bureaus of these places. (We could not find annual budgets for every year, for every bureau, in every location.) Comparing surveillance spending in our data with these local bureau budgets, we were able to see that surveillance-related purchases represented larger or smaller proportions of overall government spending in different regions. (Note that we cannot confirm how much surveillance spending in our dataset is represented in local public security budgets, so we have treated them as separate in the chart below.)

In one example, Henan’s Party and government offices in Zhoukou city—home to some 11 million people—ponied up 270 million renminbi for surveillance purchases in 2018, almost equaling that year’s education budget and outstripping by 2.5 times the amount the city allocated to environmental protection. By contrast, in Beijing’s Fengtai district (population 2 million, including more than 600,000 migrants) our data shows 127 million renminbi spent on surveillance-related products and services in 2018. Surveillance-related spending, while not insignificant, didn’t represent a massive outlay for the district: The Fengtai Public Security Bureau spent nearly 2 billion renminbi that same year, while the Education Commission spent nearly 6 billion renminbi. There are a number of potential explanations for the fact that Fengtai district, in China’s well-surveilled capital city, spent a relatively small sum on surveillance in 2018. One is that, because it is in the capital, it already had an extensive surveillance structure in place and did not need to spend much to expand it. Or, perhaps local officials were actually spending more on surveillance equipment than it appears in the data, but, being uniquely attuned to what the central government might consider “sensitive,” opted not to upload notices on certain topics. Of course, Fengtai may have simply spent less than Zhoukou because it has less than a fifth of the population. In the end, these are all educated guesses, not definitive explanations.

Even beyond the gaps in our data, all of these figures almost certainly understate actual spending: Our tally includes a number of procurement notices that referenced funding through public-private partnerships. A notice for Shawan county in Xinjiang, for example, listed a budget of 0 renminbi, while supplemental materials indicated that a public-private partnership would pay the estimated 105.1 million renminbi cost of the planned project.

While some “model cities” report receiving subsidies from the central government (for example, a procurement notice from Xiamen in Fujian province explains the city received 23 million renminbi from the central government to set up and upgrade several city-wide video-sharing platforms), surveillance budgets are not infinite, particularly for local governments working to implement underfunded central mandates. To this end, some local governments work to ensure their surveillance equipment can serve multiple purposes. For example, a feasibility study from Shawan county, in the Xinjiang Uighur Autonomous Region, concluded that street lights could also be installed on camera posts in under-serviced areas in order to save money. Further, the study noted, LED screens could be attached to these camera posts to run advertisements.

China’s spending on surveillance demonstrates the depth of concern about “social stability” animating the country’s leaders. But it also underscores real-world constraints on how much money can be thrown at the problem, and hints at the geographical disparities in available funding, local priorities, and existing surveillance infrastructure upon which to build. It is still an open question how much of a burden surveillance spending places on local budgets. How easy is it for local officials to get funding for these projects? Who benefits the most from this spending—local officials, who can report to their superiors that they are prioritizing “social stability”? The contractors who actually install and manage the equipment? Technology manufacturers who sell the equipment to contractors? These questions, among others, require further research to better understand how surveillance-related purchases fit into larger government spending priorities and how they affect political and corporate ambitions.

—Jessica Batke and Mareike Ohlberg