An assessment of the transnational repression apparatus the Chinese Communist Party runs through the United Front Work Department, what the public record actually shows, and what comes next as the police-station modality gets sacrificed.
Headline assessment
Confirmed at scale, ongoing despite exposure, and structurally larger than the headlines suggest. The Chinese Communist Party operates a single integrated transnational repression apparatus through the United Front Work Department (UFWD), with at least 102 overseas police service stations documented across 53 countries as one of four parallel modalities. The other three (Confucius Institutes, Chinese Students and Scholars Associations, and a coordinated Chinese-language diaspora media environment) extend the same operational logic to different segments of the diaspora and to host-country institutions. The system is documented in PRC government records published by provincial Public Security Bureaus, in independent investigations across more than two dozen jurisdictions, in US Department of Justice prosecutions including a conviction handed down on 13 May 2026, and in the public inquiry findings of three G7 nations. High confidence on architecture, scale, and continuing operational function. Medium-high confidence on the active operational reach of any individual station.
The mainstream framing as “Chinese police stations abroad” reads the visible artifact and misses the system around it. Stations have been closed in the Netherlands, Ireland, Canada, and the UK; formal investigations are open across most of Europe, North America, and parts of Asia. The closures are real. The function is not gone. It has migrated to WeChat, to hometown associations operating under their original names, and to other modalities that sit in legal gray zones where host-state response options are weaker. This is the Confucius Institute pattern from 2020, where physical closure produced institutional continuity, replaying on a different vector. The police-station modality is the part Beijing was always going to concede.
United Front Work Department, the 2018 consolidation, and the four modalities that sit under one chain of command.
The mainstream framing treats “Chinese police stations abroad” as a discrete scandal that broke in September 2022, when the Madrid-based human rights NGO Safeguard Defenders published 110 Overseas — China’s Transnational Policing Gone Wild, followed in December by Patrol and Persuade. The first report identified 54 stations in 30 countries. The second brought the total to at least 102 stations across 53 countries on five continents. Both reports were built from publicly available PRC government documentation, including notices and press releases from the provincial Public Security Bureaus running the operation.
That framing misses what the stations are. They are one surface of an integrated influence-and-repression apparatus operated by the Chinese Communist Party against its diaspora, against its dissidents, and against the institutions of the host countries those people live in.
The connective tissue is the United Front Work Department (UFWD) of the CCP Central Committee. Through a March 2018 institutional reorganization, UFWD absorbed the State Council’s Overseas Chinese Affairs Office (OCAO) and the State Administration for Religious Affairs as internal bureaus, and took effective control of the overseas-Chinese work of Chinese diplomatic missions from the Ministry of Foreign Affairs. The Jamestown Foundation’s 2018 documentation of this reorganization remains the cleanest public account of how the party concentrated diaspora work under one chain of command: three of UFWD’s twelve bureaus and two of its eight vice ministers are now dedicated to overseas Chinese affairs. The doctrine running through it is qiaowu, or overseas Chinese affairs work, which the political scientist James To has characterized as a comprehensive program that ostensibly serves diaspora interests but in practice works to legitimize and protect the CCP’s hold on power, uphold China’s international image, and retain influence over channels of access to social, economic, and political resources both inside China and abroad. A 2022 Canadian Federal Court ruling described the same office as engaged in “covert and surreptitious intelligence gathering.”
Four modalities sit under that umbrella, each addressed to a different segment of the diaspora and a different access point to host society. The overseas police stations are run by four provincial Ministry of Public Security jurisdictions: Fuzhou in Fujian, Wenzhou and Qingtian in Zhejiang, and Nantong in Jiangsu. They are physically embedded in pre-existing hometown associations. The Confucius Institutes operate on host-country university campuses, now under the rebranded Centre for Language Education and Cooperation (CLEC) and the Chinese International Education Foundation (CIEF). The Chinese Students and Scholars Associations (CSSAs) run on roughly 150 US campuses and equivalents across the UK, Canada, Australia, Germany, and beyond. Chinese-language diaspora media and a curated WeChat environment carry the information layer, with funding flows traceable to OCAO and consular networks.
These are not parallel programs. They report into overlapping party structures, share funding pathways through consulates and the Ministry of Education, share personnel networks, and run the same response-to-exposure playbook: deny, rebrand, retain personnel, move the function online.
The MPS did not build new infrastructure. It walked into networks that had been operating for a century.
The selection of Fuzhou, Wenzhou, Qingtian, and Nantong as the four PSB jurisdictions running the overseas stations is not random. These are the historical sending regions for the modern Chinese diaspora into Europe and the Americas. Fuzhou (Fujian province) is the origin of much of post-1980s Chinatown New York and a substantial fraction of the Chinese diaspora in Ireland and continental Europe. Qingtian and Wenzhou (Zhejiang province) feed the Chinese communities of Italy (Prato, Milan), Spain, Portugal, France, and the Western Balkans. Nantong (Jiangsu province) has heavy outflows into Japan and Southeast Asia.
Hometown associations, known in Chinese as qiaoxiang organizations, were already dense in those diaspora hubs. They predate the 110 program by decades and in many cases by more than a century. They were founded during the mass emigration of the late Qing and early Republican periods and have continued to provide genuine community functions ever since: paperwork help, business networking, funeral and wedding logistics, holiday celebrations, language schools for the children of immigrants. Many remain active and useful institutions to their membership. The Overseas Chinese Affairs Office (now under UFWD) has formally designated a subset of them as Overseas Chinese Service Organizations and incorporated them into the official qiaowu network.
When MPS decided to project policing capacity abroad, it did not need to build new infrastructure. It walked into networks that already existed. Office space, community legitimacy, multilingual personnel, contact lists, recurring foot traffic from diaspora members handling routine business: all already present. The MPS station was added as a function inside the existing association. Sometimes a separate signed office. Sometimes a desk inside the larger association space. Often the same individuals served both functions.
The Fuzhou PSB stations carry public signage that reads “Fuzhou Police Overseas Service Station, [City Name].” The Dublin station’s signage was photographed and reported by the Irish Times in September 2022, and it sat openly on Capel Street. The Manhattan station, on the third floor of a building at 107 East Broadway shared with the America ChangLe Association, displayed a banner reading “Fuzhou Police Overseas Service Station, New York USA,” which federal prosecutors entered into evidence at the Lu Jianwang trial. The stations were not designed to be clandestine. They operate openly under the assumption that host states will not push back. Where bilateral police-cooperation agreements with the MPS were in place, host states did not push back for years.
The repression operation is rarely dramatic. The mechanism is graduated, mostly legal-adjacent pressure designed to produce a “voluntary” return.
The repression operations themselves are rarely dramatic. Most coverage skips this part. The mechanism is graduated, mostly legal-adjacent pressure designed to make the target “voluntarily” return to China. Safeguard Defenders’ classification across Involuntary Returns (2022) and Chasing Fox Hunt (2024) maps three modes.
By far the most common. The state detains, harasses, or threatens the target’s family members back in China. Children are pulled from school. Parents lose pensions. Siblings are interrogated and instructed to make contact. A video meeting is then arranged. The target, sitting in Madrid or Belgrade or Rotterdam, opens WeChat and sees their mother seated beside a “family representative” placard at a Procuratorate office, with a Public Security officer in frame. The overseas hometown association running the local police station sets up the video link and applies the local social pressure. The Liu case from Qingtian, reported in detail by Safeguard Defenders, ran exactly this pattern: the Qingtian Public Security Bureau, the County Procuratorate, the County Overseas Chinese Federation, the Qingtian Hometown Association of Spain, and the Madrid overseas station all participated in the same video meeting that produced Liu’s “voluntary” return in January 2020.
MPS officers travel to the host country on tourist or business visas. They locate the target, photograph, follow, and sometimes approach host-country business partners or family-friends to apply additional social pressure. In some cases they lure the target across a border to a less protective jurisdiction (Thailand, the UAE, Serbia) where Chinese extradition cooperation is closer or where Interpol Red Notices land more easily. The 34 MPS officers indicted in absentia by the US Department of Justice on 17 April 2023 fall into this operational mode: the indictments describe them as having used fake social media accounts to harass and intimidate Chinese nationals inside the United States, with the intent of suppressing dissident speech.
Rare, with the highest diplomatic cost if it surfaces. Reserved for high-value targets where the political calculation favors the risk.
The hometown association and police station is the middleware across all three modes. It identifies and locates targets, hosts the video pressure sessions, sometimes provides on-the-ground harassment, occasionally rents the safe house or arranges transport. It rarely conducts the kidnapping itself. That work falls to traveling MPS officers.
The political-dissident cohort within those numbers is smaller than the headline figures suggest and harder to count precisely. The 230,000 figure is dominated by lower-level Southeast Asian telecom-fraud cases. The 12,000 forced-returns figure better approximates the higher-value Fox Hunt and Sky Net political and economic-crime targets. Well-documented political cases on the public record include the Wang Jingyu harassment campaign run from the Rotterdam station against a Dutch-resident dissident; the Liu environmental-pollution case from Madrid; the Xia theft case from Belgrade; and dozens of US-based cases that surfaced in the April 2023 DOJ filings.
Why Beijing’s “administrative service center” framing is partly true, and why it does not change the assessment.
Beijing’s standing line is that the overseas service centers are administrative service centers, handling driver’s license renewals, document notarization, and COVID-era logistical help for diaspora members locked out of China by border closures. The defense at the Lu Jianwang trial in Brooklyn leaned hard on this framing. Lu’s attorney told jurors the Manhattan outpost was a community space where people renewed Chinese driver’s licenses remotely during COVID lockdowns and met to play ping-pong and mahjong.
The argument works partly because it is partly true. Many of these centers do help diaspora members with paperwork. The dual-use structure is the operational point. Most foot traffic is innocuous, which gives host-state prosecutors a hard time isolating the bad acts from the routine ones. The architecture is the same as a money-laundering front: legitimate volume launders the illegitimate function.
The most rigorous version of the skeptical reading comes from Jeremy Daum at Yale Law School’s China Center and Matt Schrader, then writing for the Jamestown Foundation. Their position is that many of the centers described in the Safeguard Defenders reports predate the MPS 110 program by years, were founded as genuine community-serving institutions tied to UFWD’s qiaowu outreach, and were opportunistically tagged or absorbed by MPS later in their existence. By that reading, calling all 102 documented stations “police stations” overstates the case. The accurate framing is that a subset of these centers was operationalized for active policing functions, often without the broader hometown association membership fully knowing.
That reading does not change the structural conclusion. It locates the problem one institutional layer up, at UFWD itself, which built and maintains the underlying network into which MPS plugged its policing capacity. The hometown association is the foundation. The police station is the second story.
The police stations are one vector. The Confucius Institute trajectory is the precedent for what comes next.
The four modalities of the apparatus do not share branding or organizational charts. They share the operational logic that makes each work: a real, partly legitimate front function layered over a political one; embedding inside a pre-existing local institution that provides legitimacy and physical cover; party-state funding and coordination delivered through deniable channels; and a response-to-exposure pattern that rebrands rather than retreats.
| Modality | Target segment | Host institution | Legal exposure |
|---|---|---|---|
| Police stations / 110 Overseas | Working- and business-class diaspora, fraud and dissident targets | Hometown associations | High Vienna Convention; FARA |
| Confucius Institutes | Host-country academic and policy elites | Universities | Medium Foreign-funding disclosure |
| CSSAs | Chinese students abroad and overseas Chinese researchers | Student association charter | Low Student-association autonomy |
| Chinese-language media + WeChat | The Chinese-speaking diaspora information environment | Independent media, platform-as-service | Very low Free-press protections |
The asymmetry in legal exposure across the rows is what determines Beijing’s pattern of response when any individual modality is challenged. The police-station modality is the most exposed: it claims policing authority on foreign soil, which the Vienna Convention on Consular Relations treats as a sovereignty violation absent host-state consent, and which US FARA-style foreign-agent registration statutes can prosecute when the operating personnel are local residents. The Confucius Institute modality is moderately exposed: foreign-funding disclosure laws and university policies allow refusal or non-renewal, but no individual is prosecuted for hosting a CI. The CSSA modality is barely exposed at all: student associations are protected by the right of free association, and even consular funding flows that bypass university financial controls are not formally illegal in most host states. The information layer is effectively beyond reach of host-state response, protected by free-press doctrine and platform-as-service architectures.
The Confucius Institute trajectory from 2017 to the present is the cleanest precedent for what comes next with the police stations. US CI counts fell from a peak of around 120 in 2017 to single digits by 2024 under federal pressure and state legislation. In July 2020, in response to closures, the Chinese government dissolved Hanban (the Office of Chinese Language Council International, which had administered the CIs) and reconstituted its function as the Centre for Language Education and Cooperation (CLEC), then spun off a nominally non-governmental foundation, the Chinese International Education Foundation (CIEF), to fund and oversee CI replacements. Research by the Foundation for Defense of Democracies (FDD) documented that several US universities that closed their CIs maintained funding flows and personnel relationships with their former Chinese partner universities; the University of Michigan, for example, continued to receive Hanban-source funding after its formal CI closure. The University of Central Arkansas case is similar. The pattern is consistent: physical closure of the named entity, institutional continuity of the underlying function under a new label.
The CSSA modality has not been challenged at scale in the same way. There are roughly 150 CSSA chapters on US campuses, with equivalents at virtually every major Anglophone, European, and Australian university. The US State Department’s 2020 fact sheet on the CCP on campus states plainly that CSSAs were created in the late 1970s to monitor Chinese students and mobilize them against views that dissent from the CCP’s stance, and are collectively overseen by UFWD with funding and direction flowing through Chinese embassies and consulates. Foreign Policy’s 2018 investigation documented consular funding deposited directly into individual CSSA officers’ personal bank accounts to bypass university financial oversight. A China Service Centre for Scholarly Exchange smartphone application, documented in UK reporting from 2024 to 2026, awards financial points to Chinese students abroad for logging their daily location and consulate-event attendance, and deducts points for “extreme political or separatist rhetoric” and “words or deeds that undermine national security and ethnic unity.” China’s 2017 National Intelligence Law obligates all Chinese citizens to cooperate with state intelligence work regardless of where in the world they reside. That law is the structural condition under which all of this operates.
Liberal democracies with active diaspora dissident communities have moved. Most other host states have not.
The host-country response to the Safeguard Defenders reports clusters into three tiers, with the dividing lines mapping cleanly onto the strength of rule-of-law institutions and the political weight of dissident diaspora communities inside each country.
FBI raided the Manhattan station in October 2022. DOJ unsealed indictments against Lu Jianwang, Chen Jinping, and 34 MPS officers in absentia in April 2023. Chen pleaded guilty in December 2024. Lu was convicted in Brooklyn federal court on 13 May 2026. The Netherlands formally declared the stations illegal in November 2022 and ordered closures. Ireland ordered the Dublin stations closed in October 2022. Canada moved publicly with uniformed RCMP presence at four sites in March 2023, but the Montreal investigations closed in September 2025 without charges, and the targeted community groups are now suing the RCMP for $2.5 million in damages.
The UK reviewed stations in Glasgow, London-Hendon, and Croydon without bringing prosecutions. Germany’s RTL/ntv investigation in November 2024 found at least four stations still operating in Berlin, Hamburg, Düsseldorf, and Munich, plus one in the Ruhr area. France’s DGSI in June 2025 suspended the deportation of a Chinese businessman it suspected of running a station from within a Fujian hometown association in Paris. Spain identified nine stations, including the Madrid station tied to the Liu case. Sweden’s Säpo investigation continues. Tokyo police raided a station in February 2024 in connection with a COVID-fraud investigation.
Italy ran joint police patrols with uniformed Chinese MPS officers walking Italian streets until December 2022. ProPublica reporting documented organized-crime ties in the leadership of the Prato station. Joint policing agreements between China and Croatia, Serbia, Romania, Hungary, and South Africa gave Beijing physical legitimacy for the station network as a normalized partner activity. The Belgrade station was the platform for the Xia “persuasion to return” case in 2018. South Africa’s Cape Town consulate publicly endorsed the framework. Most Belt-and-Road African states with Chinese workforce presence host stations: Angola, Lesotho, Nigeria, Tanzania.
The Italy / Croatia / Serbia / Romania pattern is the strategically important one. Pre-existing bilateral policing agreements between these host countries and the MPS gave Beijing physical legitimacy and helped pilot the wider network expansion before Safeguard Defenders made it visible. Italy’s joint patrols placed uniformed Chinese officers on European Union streets with host-country police authority. That precedent of Chinese policing as a normalized partner activity, rather than as a sovereignty violation, is what made the station network politically viable across continental Europe in the first place. Italy ended the joint patrols in December 2022; the diplomatic ground was already prepared.
FARA gets convictions. Conspiracy does not. The template for future US prosecutions is now visible.
The Brooklyn jury convicted Lu Jianwang on 13 May 2026, on one count of acting as an unregistered agent of a foreign government and one count of obstruction of justice for deleting WeChat messages with his MPS handler. The jury acquitted him on a third count, conspiracy to act as an agent of a foreign government, after sending the judge several notes during deliberation asking for clarification on what “conspiracy” meant in the charging instructions. Lu faces up to 30 years at sentencing. Chen Jinping, the co-defendant, had already pleaded guilty in December 2024 to a related conspiracy count.
The split verdict matters for what comes next.
The conviction validates the prosecution model. A FARA-style failure-to-register theory combined with obstruction for digital evidence destruction does not require proving the substantive content of any surveillance work. It requires showing that the defendant was acting at the direction of a foreign government and did not notify the Attorney General. Lu attended the January 2022 Fuzhou ceremony where the MPS publicly announced the global station program. He photographed himself there holding a sign that read “Fuzhou Public Security Bureau, Overseas 110 Report to Police Service Station.” He established the Manhattan outpost the following month. He communicated with handlers over WeChat. He deleted those communications when the FBI got close, then admitted as much to FBI agents during the post-raid interview. That sequence was enough.
The acquittal on conspiracy signals where the case is harder. Substantive coordination theories, where multiple actors are alleged to have conspired to surveil and harass dissidents, require juries to absorb a more complex story than a paperwork-failure theory does. The defense narrative, that Lu was a well-meaning community leader handling driver’s license renewals and mahjong games who tripped over a registration requirement, carries enough sympathetic weight to peel the conspiracy charge off even when the technical FARA violation is unambiguous.
The template for future US prosecutions is now clear: single defendant, FARA plus obstruction, well-documented digital trail, ideally an admission during the post-raid interview. The 34 MPS officers indicted in April 2023 will never see a US courtroom. Those indictments function as diplomatic signaling, supporting visa restrictions, travel limitations, and named-and-shamed personnel files for Five Eyes intelligence sharing. They are not endgame prosecutions.
Six steelmanned counters, with leans.
The Daum and Schrader reading is that many of the documented locations are genuine community-serving hometown association offices that predate the 110 program, were folded into MPS branding opportunistically, and have no operational policing function. Safeguard Defenders’ methodology relied heavily on open-source PRC government documents; some of those documents reflected aspiration or rhetorical inclusion rather than active operations.
LeanValid on the facts, does not change the structural assessment. A subset of the 102 stations is the operationally active set; the rest are institutional cover. The total count is less informative than the question of how many active stations there are, which is unknown and probably smaller. The architecture point survives intact: even a smaller operational set sits inside a UFWD-coordinated hometown-association network that provides the same scaffolding for the modalities that are not police stations.
A real subset of targets are genuine criminal suspects who would in principle be subject to extradition under proper mutual legal assistance treaties (MLATs) with PRC authorities. The political-dissident overlap is real but is a minority of total cases. Treating the entire campaign as a dissident-targeting operation misreads the operational composition.
LeanTrue on the composition, does not justify the methodology. The Liu environmental-pollution case from Qingtian to Madrid is the clean instance: the underlying offense was real, but the persuasion methodology violated Spanish sovereignty. The legal problem is not whether the target is guilty under PRC law. The legal problem is that the persuasion-to-return apparatus circumvents host-state extradition processes, threatens family members back in China as collateral, and operates outside any framework the host state has consented to. The proper channel exists and is sometimes available. Beijing routes around it.
The Montreal community groups suing the RCMP for $2.5 million in damages represent a real political and social cost of enforcement actions that do not result in convictions. Hometown association leadership tends older, business-class, and either pro-CCP-aligned or transactionally engaged with consular networks; younger, professional, dissident-aware diaspora members tend to align with the Safeguard Defenders framing. Aggressive enforcement creates intra-community fractures that are exploited politically.
LeanReal cost, not a reason to disengage. The Lu Jianwang prosecution is the cleanest case in point: the conviction was on the technical FARA violation with documented WeChat traffic to an MPS handler, not on a profiled assumption about Chinese-American community structures. The Montreal RCMP case is the worst case in point: enforcement action without prosecutable charges generates the political backlash without the legal payoff. The implication for prosecutors is to wait for the digital evidence trail before raiding, and to charge narrowly when they do.
CIs are nominally academic and have ongoing legitimate language-teaching functions. CSSAs are nominally student organizations with ongoing legitimate social functions. Bundling them with active police stations as one apparatus inflates the threat and risks chilling legitimate community and academic activity.
LeanEquivalence is the wrong word; common operational logic is the right frame. The four modalities are not equivalent in their primary function or in their legal exposure (the comparison table in Section 05 makes this explicit). They share UFWD coordination, share funding pathways through consulates and the Ministry of Education, share personnel networks, and share the rebrand-don’t-retreat response pattern. The Hanban-to-CLEC-to-CIEF rebrand of 2020 is the documented case. The CSSA app surveillance described in UK reporting is the documented operational vector. The argument is structural, not equivalent.
Bureaucratic alignment under UFWD is not the same as operational integration. The four PSBs running stations do not coordinate with each other. Individual CSSA chapters operate with varying degrees of consular involvement. The information layer is genuinely decentralized. Treating this as a single apparatus imposes more coherence than the evidence supports.
LeanThe integration is structural, not operational. The four PSBs do not need to coordinate with each other for the system effect to hold; they all report into the same political logic under UFWD’s overseas-Chinese affairs bureaus, draw on the same qiaowu doctrine, and operate inside a shared hometown-association network that UFWD maintains. The CCP is not a monolithic actor at the tactical level. It is a coordinated actor at the strategic level. The 2017 National Intelligence Law and the 2018 UFWD reorganization are the strategic-coordination mechanisms. Tactical decentralization is a feature, not evidence against the architecture.
The G7 has elevated transnational repression as a priority during Canada’s 2025 presidency. The Hogue Report named the PRC as the most active perpetrator of state-based foreign interference against Canadian democratic institutions. EU member states have stiffened their public posture. The cost of maintaining the apparatus has gone up. Beijing has rational reasons to retreat.
LeanPossible but not the dominant trajectory. The 2018 institutional reorganization and the 2017 National Intelligence Law both predate the current diplomatic environment and signal a strategic commitment that is unlikely to be reversed under Xi. The pattern Beijing has demonstrated is concession of the most exposed vector (police stations) and continued investment in the less exposed vectors (CIs under CLEC/CIEF, CSSAs, the WeChat information environment). That is not retreat. That is portfolio rebalancing toward the more defensible positions.
The architecture, the trajectory, and what the next five years look like.
The PRC operates a coordinated transnational repression apparatus under the United Front Work Department of the CCP Central Committee. The apparatus has at least four operational modalities, of which the overseas police stations are the most legally exposed and therefore the most visible. The architecture and the scale of the police-station modality are confirmed by PRC government documents, by US Department of Justice prosecutions, by the public inquiry findings of three G7 nations, and by independent investigations across more than two dozen jurisdictions. None of the underlying facts is in serious dispute. The disagreements that exist are about composition (how many of the 102 stations are operationally active versus institutionally inactive), about scope (how to count political-dissident targets within the larger fraud-suspect population), and about framing (whether to analyze the modalities together or separately).
The trajectory is more informative than the headline numbers. The Confucius Institute precedent from 2017 to the present is the cleanest signal of how Beijing responds to host-country pushback against any individual modality: it concedes the exposed entity, rebrands the parent institution, and migrates the function to less visible channels. CIs were closed; CLEC and CIEF replaced them; some of the underlying university relationships and funding flows continued. The police stations are now on the same trajectory. Stations have been closed in the Netherlands, Ireland, and the UK. The function migrated to WeChat groups, to hometown associations operating under their original names, and to undeclared online persuasion-to-return operations that do not need a physical office. Where stations remain operational (in Germany, in parts of France, across the Tier 3 host countries), the political cost of maintaining them has not yet exceeded Beijing’s tolerance for continued operation.
The CSSA modality and the diaspora information environment are the durable surfaces of the apparatus. They face the weakest host-state response options. The State Department’s 2020 fact sheet, Foreign Policy’s 2018 investigation into consular funding flows, and the UK reporting on the China Service Centre app all document the operational mechanics. The 2017 National Intelligence Law provides the structural condition under which they continue to function. Host-state response in these domains is constrained by free-association doctrine, by academic freedom, and by free-press protections that protect the surface activities even when their underlying coordination is established.
The active operational question for the next five years is not whether Beijing operates a transnational repression apparatus against its diaspora. That is settled. The active question is how much of the apparatus continues to function in plain sight, and what fraction of educated Western opinion learns to identify the modalities that are still operating after the police stations finish being closed.
Sources