In May this year, a pickpocket on the Guangzhou Metro was identified by passengers and, in retaliation, sprayed pepper spray indiscriminately into the crowd. The incident put Chinese metro screening back in the spotlight: how had the spray passed through inspection undetected? It is a routine reminder that the system’s filtering capacity is partial, uneven, and well understood as such by the public — but it does not, by itself, tell us what the system is for. The more important fact is that the screening system itself has become the most capillary, material expression of China’s “stability maintenance” (维稳, weiwen) regime.
Weiwen has been the core operating logic of the Chinese Party-state for the past fifteen years: it treats social stability as a hard political necessity and converts that necessity into fiscal line items, staffing quotas, and technology deployments. Security screening is how this logic materializes at every routine point of public access.
About 90 million people pass through a security check in China each day. This floor estimate is derived by averaging annual transport figures across the year, including civil aviation’s 752 million departing and transit passenger movements, urban rail’s 33.38 billion trips, and the railway system’s 4.255 billion trips. An ordinary Chinese citizen may be scanned two to five times in a single day — at the metro, a high-speed rail station, an airport, a hospital outpatient or emergency ward, a court, a government service hall, a shopping mall, or a concert venue. In Xinjiang, the number rises by an order of magnitude.
Sustaining this daily scanning requires roughly 1.1 to 1.3 million “narrow-sense screeners” on duty nationwide at any one time, alongside 77,464 civil aviation screeners, around 130,000 metro screeners, and 150,000 railway screeners. Annual industry output is on the order of RMB 120 billion.
This system did not emerge as a passive response to social need. It is a central component of the government’s stability-maintenance strategy over the past fifteen years — a deliberately constructed and continuously expanding “stability-maintenance industrial chain.” It converts the political demand for public security into procurable contracts, employable labor, depreciable equipment, and exportable projects. It turns “governance” into “industry.”
Understanding this requires breaking with two common readings. The first treats screening as wasteful spending — the price society pays for authoritarian rule. The second treats it as legitimate public-safety investment, since airports, metros, and hospitals do face real risks. Neither is sufficient. The system is both of these things, but it has a third character as well: it is a RMB 120 billion micro-industry that runs from listed companies and equipment manufacturing through labor outsourcing, local fiscal subsidies, foreign aid, and cross-border project delivery. The chain is designed politically, not economically.
In public-safety discourse, it is “the guarantor of the people’s sense of security.” In labor-market discourse, it is “a sub-sector of private security services.” In going-global discourse, it is “Chinese high-end equipment going out.” Only when these three identities are stacked together does the full political-economy picture come into view: a stability-maintenance industrial chain built jointly out of fiscal money, low-wage labor, equipment exports, and aid-financed projects. It is not the largest sector of the Chinese economy, but it is the clearest case study of how the system converts governance demand into material production.
A clarification is in order before proceeding. The argument that follows does not require the screening system to fail at producing public safety. China’s recorded violent-crime rates are low by middle-income-country standards, and screening may well contribute to that outcome — though attribution is hard given the absence of counterfactual data and the much larger role of broader policing and surveillance infrastructure. The political-economy reading offered here runs alongside the public-safety reading, not in place of it. The question is what the system reveals when read through the additional lens of fiscal structure, labor markets, industrial policy, and overseas diffusion — a lens largely absent from English-language accounts of Chinese “stability maintenance.”
This article addresses the political economy of that chain: how its fiscal accounts work, why this apparent “wasted money” has become an economic engine, what it carries with it when it goes abroad, and what it reveals about the form of the Chinese state.
I. Four procurement models and the “shadow fiscal” architecture
To see how the chain works, start with where the money comes from. On the surface, security screening is a public service and should be paid for out of public finance. In practice, funding sources are fragmented, hidden, and different in every setting.
The fragmentation appears as four parallel procurement models. Shanghai uses an “enterprise responsibility” model: the metro company absorbs the cost itself, which surfaces in the tension between fares and public subsidies. Shenzhen uses a “direct PSB procurement” model, in which the municipal Public Security Bureau (PSB) draws directly from local fiscal funds. Only in 2024 did the Shenzhen PSB’s audit report disclose the line item “metro screening and security guard subsidy funds” for the first time, implying an annual citywide subsidy of roughly RMB 1.2 to 2.5 billion — money that does not appear on the metro company’s books at all. Beijing uses a “hybrid procurement” model: the four operating subsidiaries, Beijing Infrastructure Investment, and the Rail Transit Command Office co-finance the function, making it hard to isolate the figure from any single annual report. Second- and third-tier cities such as Fuzhou, Hefei, and Qingdao use a “full fiscal” model, with local governments paying outright.
The railway system is the most opaque of all. Beijing-Shanghai HSR, the only listed pure-line railway operator, folds screening costs into its “entrusted transport management fee” and provides no separate disclosure. The 18 regional railway group companies each operate on their own terms, with no unified disclosure standard equivalent to the Civil Aviation Administration’s. Public spaces are even more fragmented. Court screening runs on judicial-police personnel budgets; hospital screening is bundled into “Ping’an Hospital” (Safe Hospital) integrated security programs; office-building screening sits across categories such as “property services” and “agency operations.” The same “gate money” can land in a property contract, an agency operating budget, an outsourced security agreement, an equipment tender, a dedicated fiscal grant, or any line on a corporate P&L.
This is the “shadow fiscal” architecture. Its dispersed structure is not a technical design problem but a political design choice. It keeps screening spending inconspicuous under any single category. There is no accounting heading anywhere that can tell you how much a city spends each year simply to put people through gates.
This is the first structural feature of the stability-maintenance system: it is not a flagged, named, openly budgeted “political stability” line. It is a spending bundle deliberately dispersed, hidden, and broken up to the point where it cannot be added together. The invisibility is itself the political point.
II. The RMB 120 billion ledger — measured against pensions and social insurance
So what is the order of magnitude of RMB 120 billion?
Against GDP, it is about 0.09 percent — seemingly negligible. But the meaning of a ratio depends less on the decimal places than on what it is compared with.
The most direct comparison is the rural basic pension. In 2025, the central government’s subsidy under the Urban-Rural Residents’ Basic Pension scheme stands at a floor of RMB 143 per recipient per month (RMB 160-180 in select provinces), with total central subsidies to the 170 million recipients nationwide reaching roughly RMB 290 billion. Put differently, the money China spends each year to give 170 million elderly rural residents a “minimum survival pension” is only a little over twice what it spends to put their grandchildren through security gates each day. Adding the RMB 120 billion in “gate money” to the rural pension envelope would raise the pension by roughly 50 percent, or lift the monthly per-capita subsidy from RMB 143 to RMB 203.
Parallel comparisons run through medical insurance and education. In 2024, the per-capita fiscal subsidy under Urban-Rural Residents’ Medical Insurance was RMB 670, covering 960 million people, for a total fiscal contribution of about RMB 640 billion. Central transfers to local governments for education in 2024 totaled around RMB 408.5 billion. RMB 120 billion is equivalent to raising the per-capita medical subsidy from RMB 670 to RMB 795, or raising the national per-student funding base for compulsory education by 20 percent.
These comparisons are not an argument that the money should be reallocated from screening to pensions — that is a politically impossible counterfactual. They simply make one fact visible: in fiscal terms, the stability-maintenance industrial chain is a spending choice on the same order of magnitude as basic social welfare. Behind every scanned passenger, every CT machine, and every screener pulling a 16-hour shift sits a distributional decision.
From a welfare-economics standpoint, screening produces no accumulating public good in the conventional sense: it builds no schools or bridges, trains no doctors, and leaves no asset for the next generation. Its marginal contribution to public safety may be real, but it is unmeasured and is in any case incommensurable with the foregone alternatives in pensions, healthcare, and education set out above. The point of the comparison is not to claim that screening produces no safety, but to make the distributional choice visible. What the spending also produces, alongside whatever safety it provides, is the central government’s visibility over local public order and local officials’ political credit for “zero major incidents.” These are consumption expenditures, not investment.
But the intuition has a blind spot. While welfare-economically the chain is pure consumption, in industrial terms it generates real employment, equipment, manufacturing, and export revenue. Of the RMB 120 billion, RMB 94 billion flows to labor services, RMB 19.5 billion to equipment, and RMB 6.5 billion overseas. Viewed as welfare it is anti-welfare; viewed as industrial policy, it is industrial policy.
III. Why “wasted money” became an economic engine
Switching from a welfare lens to an industrial lens changes the picture entirely.
The RMB 94 billion in services translates into 1.1 to 1.3 million labor positions at private security firms. Median monthly pay is around RMB 4,500 — roughly RMB 5,500-9,000 at first-tier-city airports and RMB 2,800-4,000 in third- and fourth-tier county locations — with annual turnover of 60-80 percent. That means close to 2 million people cycle through these roles each year. This is a high-throughput “human-wave checkpoint,” not a stable employment pool. China’s registered private security workforce grew from 5.22 million in 2020 to 6.76 million by the end of 2023; of the 1.54 million net additions, screening-related positions probably account for 400,000 to 500,000.
The RMB 19.5 billion in equipment spending shows up as orders for firms such as Nuctech (同方威视), whose nuclear-technology applications revenue reached RMB 6.376 billion in 2024 (up 12.57 percent), and Shengxun (声迅股份), a transit-screening specialist. It sustains a technology ladder — X-ray machines, CT-based screening, millimeter-wave body scanners, trace explosives detection, AI-assisted image review, centralized image review, and “screening large models.” Each new rung triggers another procurement cycle.
2024 was a policy-cluster year. The Civil Aviation Administration issued six new screening standards. The Administrative Measures on Railway Passenger Security Screening came into force. And in July 2024, after a doctor named Li Sheng was killed inside a hospital in Wenzhou, hospitals nationwide rushed to install screening equipment. The combined effect pushed 2024-2027 into a peak replacement cycle for China’s security-screening hardware.
The dynamic becomes clearer when viewed at the level of local employment and fixed-asset investment. A prefecture-level city’s “Safe Hospital construction” project typically bundles screening equipment, fire safety, video surveillance, anti-riot gear, in-hospital police posts, and a smart-screening platform — running anywhere from a few million to tens of millions of yuan per project. Local resellers, equipment manufacturers, and security firms all see revenue; the municipal health commission and the PSB book political credit. It is a four-way win on the fixed-investment ledger — except that the “asset” produced is not hospital beds, medical devices, or community services, but entry gates, X-ray machines, screener positions, and a “Safe Hospital” plaque on the wall.
This is the real mechanism by which the chain functions as an “economic engine.” It is not a macroeconomic engine — its GDP contribution is under 0.1 percent. It is a local-level, narrowly sectoral, labor-intensive micro-engine that converts fiscal spending into visible jobs, procurable equipment, and accumulating political capital.
But the engine is losing momentum. In 2025, the number of civil aviation screeners fell by 1,132 — the first year-on-year decline in the history of Chinese aviation security. The Guangzhou Metro has already cut its image-review staff from over 2,400 to fewer than 70. Technological substitution is underway, and the policy window will close around 2027, after which the equipment market will settle into a steady state of RMB 10-13 billion. The engine was never sustainable: it has always relied on political campaigns to generate orders.
IV. Scanner diplomacy — capacity and governance logic moving abroad together
If China’s security-screening industry stayed at home, the story would end here. But it has a third lane: going overseas. This branch carries the same logic — fiscal fragmentation, human-wave checkpoints, equipment refresh cycles — and exports it, in a different form, to 65 countries.
A February 2025 report by the AidData research lab at William & Mary maps the diffusion clearly. Between 2000 and 2022, Chinese state-linked entities provided scanner equipment to 65 low- and middle-income countries through 210 grants and loans. Of these, 108 involved customs-inspection equipment: 41 for seaports, 25 for land border crossings, 24 for airports, and 4 for railway stations. Of those 108 projects, 57 (52.8 percent) were supplied by Nuctech (同方威视).
Nuctech itself is the pivotal case. It originated as a spin-off from Tsinghua University’s Institute of Nuclear and New Energy Technology; its controlling shareholder was once Tsinghua Holdings; CNNC (the China National Nuclear Corporation) holds 21 percent; and its former chairman, Hu Haifeng, is the son of former CCP General Secretary and PRC President Hu Jintao, and has since moved into senior central-government posts. Nuctech is not an ordinary commercial firm. It is a “governance-infrastructure supplier” assembled out of state technical institutions, central-SOE equity, and political-family lineage.
“Scanner diplomacy” is the practice of moving Nuctech’s X-ray machines, CT-based screening systems, and millimeter-wave body scanners abroad through Chinese foreign-aid channels — China Eximbank concessional loans, China Development Bank policy loans, and Silk Road Fund project financing — to Serbia, Cambodia, Tajikistan, Morocco, Papua New Guinea, and others. This is not a commercial export. In 2024, China’s total global exports under HS code 902219 (non-medical X-ray equipment) were USD 448 million, far below the USD 1.67 billion in installations recorded by AidData. The gap is the non-commercial diffusion underwritten by aid and lending channels.
There is an economic logic to the strategy. As one analysis puts it, “donations and zero-interest loans are a deliberate commercial strategy for Nuctech”: once the equipment is in place, the next 8 to 15 years of maintenance, software updates, spare parts, and training are locked in with the Chinese supplier. What looks like a loss up front becomes a lock-in downstream. The piece hardest to replicate is that no competitor has a national-level aid channel of its own.
But the real political significance of scanner diplomacy is not the commercial returns; it is what travels with the equipment. A security scanner is not a neutral safety device. It is a “governance interface” that defines how a country’s customs, airports, borders, and rail stations process flows of people, goods, and information. Its algorithms, rules, classifications, and alert logics all carry the supplier’s prior judgments about what counts as dangerous, what is contraband, and what merits re-inspection. When Senegal’s customs adopts a Nuctech smart-customs solution, when a Malaysian airport rolls out “one face” biometric self-screening, when Astana’s light rail in Kazakhstan integrates Chinese urban-rail standards, those countries are running their critical public-space chokepoints on governance logic designed in China.
The U.S. and EU response targets precisely this. In April 2024, the European Commission carried out unannounced inspections at Nuctech’s offices in Warsaw and Rotterdam — the first such inspection ever conducted under the EU’s Foreign Subsidies Regulation (FSR). In December 2025, the case advanced to an “in-depth investigation”, the first FSR case to reach the second phase. Most recently, in May 2026, China’s Ministry of Justice escalated the standoff by formally instructing Chinese companies and citizens not to cooperate with European authorities in the ongoing investigation.
The United States added Nuctech to its Entity List in December 2020. The U.S. Maritime Administration has issued five separate warnings since 2023, repeatedly noting that Nuctech equipment can access “biometric data, personally identifiable information, behavioral patterns, cargo information, and geolocation metadata.” In December 2024, Warsaw Chopin Airport announced it would replace four large Nuctech hold-baggage scanners.
The Western reaction is not really about narrow technical security; it is about governance sovereignty. What Western officials see is not a machine but a chain that runs from Chinese state institutions — Tsinghua University, CNNC, the China Atomic Energy Authority’s nuclear-technology security research center, and MOFCOM’s foreign-aid department — into their own customs, airports, and ports.
Geopolitical tension has, paradoxically, accelerated the Belt-and-Road end of the business. As Western markets close, Nuctech’s order book in the Middle East, Southeast Asia, Central Asia, Africa, and Latin America is being refilled. A two-track structure has emerged: the West tightens, the Belt and Road opens. Capacity is going out, and so is the governance logic.
V. The paradoxes of the rentable state
When the first four sections are taken together, what emerges is not an industry but a form of state.
China’s security-screening regime is not a “government department.” It is a loose alliance — local treasuries, the public-security system, metro corporations, airport groups, railway bureaus, the health commission system, the education ministry, security companies, equipment manufacturers, labor-dispatch firms, local agents, and foreign-aid channels. Its responsibilities are dispersed, its finances opaque, its labor force fluid, its technology policy-driven, and its overseas presence treated as state-strategic. The chain converts governance authority from “the state acting directly” into “the market absorbing the function” — every gate rests on a contract, a job, an invoice, and an entry in someone’s annual report.
This is the “rentable state” — the working form of China’s all-purpose government inside a market economy. It does not outsource governance to private actors: security firms, equipment manufacturers, and labor-dispatch companies are mostly controlled by state capital or local governments in practice. Instead, it industrializes governance, makes it contract-based, and renders it procurable. What used to be a state monopoly — public security — becomes a service market that can be divided, priced, tendered, scored, and fined. This is the distinctive way the Chinese government “domesticates the market,” and what sets it apart from a Soviet-style command economy: rather than the market serving the plan, the plan operates through the market as its instrument.
The model has market efficiency. It makes 90 million daily passes through gates feasible (even if screening quality is uneven), gives 1.1 million workers contracted access to the labor market, sustains a stable RMB 19.5 billion equipment market, and turns RMB 6.5 billion in overseas projects into hard geopolitical assets. It also has political utility. It allows the state to extend governance into any setting it chooses without expanding civil-service headcount, adding ministry layers, or triggering disputes over budget transparency. This is the Chinese version of what Foucault called “capillary discipline” — state power extended through marketization, contracting, and KPIs.
But the form has three internal paradoxes, all pressing on it from different directions.
The first is political blowback at the labor end. The chain sustains a pool of 1.1 million workers through low wages, high turnover, and human-wave tactics — high-speed-rail screeners paid RMB 1,750 a month, logistics-sorting screeners with 100 percent annual turnover, county-level screeners who pay labor agencies RMB 28,000 in fees just to be placed in a job. These workers are the chain’s operational endpoint, and they are also the lowest-income population with the most direct daily contact with the system. Their grievances toward the “China model” are likely sharper than those of any outside observer.
The second is technological compression of the chain’s own economic function. AI image review, centralized review, CT-based replacement, and millimeter-wave replacement are all eroding the human-wave checkpoint. Once algorithmic gates replace human gates, the chain’s political utility as a local job pool weakens. The low-end jobs disappear, but the low-end population does not. There is no mechanism to absorb the political pressure of that transition.
The third is the imbalance of the two-track overseas strategy. The EU FSR, the U.S. Entity List, and decisions by Canada, Lithuania, and Poland have effectively redefined Chinese screening equipment as “hostile foreign technology.” While the chain expands across 65 low- and middle-income markets, the high-end Western markets — where global technical standards are written — are closing fast. The repayment, sovereign-debt, and political-stability risks in Central Asia, Africa, Southeast Asia, and Latin America are of a different character from the compliance risks in the West, but both push in the same direction. The two-track structure looks balanced; in substance, the overseas track is a lower-quality, higher-risk road.
Taken together, the three paradoxes describe a textbook “introverted leviathan.” All its parts mesh inward: fiscal money sustains local jobs, local jobs produce political credit, political credit pulls in more fiscal money. It directs only two things outward — foraging (fiscal capture) and biting (stability maintenance). The stability-maintenance industrial chain is the most concrete expression of that leviathan: it turns governance demand into industry, industry into employment, employment into political stability, and political stability back into a rationale for more governance demand. The chain’s distinctive feature, beyond stability maintenance proper, is that it has become a self-reproducing mechanism — and this is what makes it rigid even as its three paradoxes intensify.
Even with all three paradoxes in play, the chain itself will not disappear. It has grown too large, too deep, too dispersed, with too many people depending on it. The gates will keep opening and closing. The scanners will keep running. The 77,464 civil aviation screeners will keep showing up for their shifts. The Chinese equipment in 65 countries will keep flashing its green and red status lights.
This is the Chinese gate in 2026. It is not public safety alone; it is not an economic engine alone; it is not political governance alone. It is all three layered together, in a way no single category can contain. It consumes RMB 120 billion and produces RMB 120 billion: what it consumes is the possibility of public goods; what it produces is the material foundation of the stability-maintenance system.