The Long Arm of the CCP: Tracing Patterns of Influence in U.S. Higher Education, Municipal Politics, and Corporate Finance

The argument proceeds across four levels that must remain analytically distinct. At the individual level, migration selection explains why exit does not identify regime opposition. At the network level, cross-system brokerage, family ties, status, and coercive connectivity can preserve incentives for homeland engagement. At the institutional level, Chinatown associations and universities can develop durable PRC-facing relationships for reasons that include constituency representation, internationalization, revenue exposure, and organizational path dependence. At the symbolic level, flags and public events reveal institutional choices and mobilization but do not identify hidden ideology.

The 2018–20 rupture then supplies the temporal test: if mass opinion turns downward faster than established organizational relationships, the resulting divergence is evidence of institutional persistence, not proof that every participant shares the same political motive. The legal analysis enters only after this separation, asking when protected advocacy becomes foreign-directed activity and when ordinary scrutiny becomes unsupported attribution or discriminatory suspicion.

A common intuition, particularly prevalent among U.S. observers skeptical of the PRC, runs as follows: if a Chinese immigrant favored the CCP or the PRC government, that person would have stayed; the fact of settling permanently in the United States is therefore itself evidence of opposition to the regime left behind. This inference is a selection-effect fallacy — it treats a single binary fact (departure) as though it screens the population on the one variable it is being used to prove, when Pew's own data show education and family reunification, not political opposition, as the dominant stated reasons for emigration.

This paper's contribution is not merely to note that the inference is invalid, but to supply the positive mechanism that explains the pattern the fallacy fails to predict: a durable propensity toward PRC-facing institutional and economic engagement despite physical relocation to the United States. The cross-system brokerage rent hypothesis developed in Section 07 does this work. Actors positioned between the United States and a partially integrated but institutionally divergent PRC can earn rents unavailable to actors on either side alone — from information asymmetry, bilingual and bicultural network access, regulatory arbitrage, and status portability. For a large fraction of the population whose institutional visibility this paper documents in San Francisco, the economically rational strategy was never "oppose the regime and cut ties" or "support the regime out of belief" — it was to remain positioned to extract value from the asymmetry between the two systems, which requires maintaining PRC-facing relationships, fluency, and standing regardless of private political conviction. Section 06's M3 (economic pragmatism) and M4 (institutional path dependence) mechanisms, and Section 08's account of why institutional relationships outlive shifts in mass opinion, describe the same underlying logic from different angles. None of this requires, or claims, ideological CCP support; it requires only that PRC-facing access remained more valuable than Taiwan-facing access for many occupations during the high-integration era, which is a testable economic proposition, not a claim about anyone's private loyalty.

The two national flags at issue are not peer contemporary symbols. The flag commonly called the "Taiwan flag" is the national flag of the Republic of China (ROC), adopted by law in 1928 from a design by Lu Hao-tung (陸皓東, 1895) that Sun Yat-sen carried as the revolutionary emblem. From 1928 until the United States switched diplomatic recognition to the People's Republic in 1979, this was the flag of the U.S.-recognized government of China. The PRC flag was adopted in 1949. A San Francisco Chinatown association displaying the ROC flag before 1979 was displaying the flag of the diplomatically recognized Chinese state and of the republican tradition continuous from the 1911 revolution. A post-1949 shift to the PRC flag is therefore the substitution of a specific communist-state symbol for the older republican-Chinese one, not the exchange of two peer contemporary options. (The earlier 1912–1928 Republic used the Five-Coloured Flag [五色旗]; the current ROC design was a 1928 national-flag adoption, not a continuous 1912 pedigree.)

The ROC flag's iconography repays close reading. Its central emblem is a sun in which a central disk is separated from twelve radiating triangular rays by an intervening blue ring — an eclipse of the sun, in which the corona of rays still shines around a ringed core. The twelve rays carry a double resonance: twelve tribes — recalling the twelve tribes of Israel as the constitutive scriptural polity — and twelve knights around the round table, a chivalric fellowship united in service to a common sovereign — the Arthurian round-table figure. Read together, the emblem carries republican, scriptural, and chivalric layers; a reading as "just the Taiwan flag" flattens all three.

My own sense during those eight years was of an environment saturated with Beijing-aligned institutional visibility — flags, consular events, association leadership, chamber-of-commerce ties — concentrated in a traditional Chinatown establishment that the measured record above independently shows was genuinely PRC-majority in exactly that period. That is an observation about institutional density, a different quantity from individual belief. The separately sourced measure of individual sentiment is the national survey trend above: Chinese-American favorability toward China declining from roughly 68% in 2016 to 41% in 2022–23, with no survey in the record supporting an estimate near 75%. The two measurements answer different questions — how saturated an institutional environment felt to a resident, and what individuals told pollsters they believed — and this paper reports them as two separate data points rather than treating either as a corrected version of the other.

I was, during this period, publicly and openly anti-CCP, and my own experience navigating that institutional environment on those terms — particularly across 2019–2021 — was that it felt unwelcoming rather than neutral. I want to be careful about what that supports: I am not claiming personal targeting rising to the level of the individually documented transnational-repression cases cited elsewhere in this paper (Section 08, Section 09), and I have no independent record comparable to those cases. What I can say is that my subjective experience aligns with the independently documented pattern: independent reporting shows organized pro-Beijing counter-mobilization confronting Hong Kong-solidarity demonstrators across multiple U.S. cities that same August,^[95] and the Washington Post's and Hong Kong Democracy Council's independently documented pattern of intimidation against anti-CCP demonstrators recurred as late as the November 2023 APEC summit in this same city.^{[24, 25]} Taken together, that record makes an unfriendly climate for an openly anti-CCP resident of this specific institutional environment during this specific window a documented, not merely felt, possibility — my own account is one data point consistent with a pattern independently reported by others, not a freestanding claim resting on personal impression alone.

A flag census requires three separate statistics. Conflating them can reverse the apparent result.^{[38, 39]}

The official San Francisco Chinatown Area Plan encompasses roughly thirty blocks and provides a reproducible geographic denominator. Dated street imagery, Mapillary, and geotagged photographs can support a normal-day census distinct from October 1, October 10, Lunar New Year, and major diplomatic visits.

The Flag Alignment Index can be computed separately for institutions and physical flag density. A value of −1 indicates an entirely Taiwan-flag sample; +1 indicates an entirely PRC-flag sample.

An association's flag choice is not a private opinion collected by a pollster; it is a discrete, standing decision made by that association's leadership and displayed continuously in shared civic space. Treated that way, the historical sequence in the table below reads less like a survey trend and more like a recorded institutional vote: 3–3–1 (Taiwan–PRC–neither) among the seven CCBA constituent associations in 2004, moving to roughly 70%–30% PRC–Taiwan by the mid-2010s, and to 2 of 7 associations still displaying a Taiwan flag by 2024. Each of those numbers is a count of governing bodies, not people, and it says nothing about how any individual member of those associations privately votes in an actual election — that distinction is the same institution/population separation this section already establishes elsewhere.

What the "vote" framing adds is a different — and, for anyone navigating this specific civic infrastructure, a more consequential — quantity: not what the Chinese-American population as a whole believes, but what an individual actually encounters when working through that infrastructure. A Taiwan-aligned or anti-CCP-aligned resident dealing with Chinatown's traditional association system in 2004 was engaging a body split down the middle. By the late 2010s, the same resident was engaging a system where PRC-oriented leadership, programming, and symbolism had become the numerically dominant environment across the institutions historically organized to represent Chinese San Franciscans collectively. That shift is a real structural fact about which doors are open, which associations offer support or standing, and which public events, resources, and leadership networks a person aligned against Beijing can expect to access inside that particular institutional space — independent of, and not resolved by, the separate national survey finding that individual Chinese-American opinion moved in the opposite direction over the same years. The resident's-eye account earlier in this section describes the lived texture of exactly that asymmetry; the association-level vote count is the structural fact behind it.

The closest online approximation to a ground census is a dated, address-level inventory covering every publicly visible frontage inside the official planning boundary. Google Street View, Mapillary, recent Wikimedia photographs, news photography, and organization websites can be cross-checked rather than treated as interchangeable snapshots.

Mapillary is especially useful because its current platform distinguishes Best Captures from All Captures, with Best Captures favoring higher-quality, accurately positioned, and newer imagery where available. Its mobile views can also color captures by recency, allowing older observations to be separated from recent ones.

A separate mechanism concerns the university rather than the student. Private universities operate with substantial fixed and quasi-fixed costs: buildings, debt service, tenure and faculty lines, administrative infrastructure, libraries, laboratories, information systems, and student services cannot be reduced in proportion to enrollment on short notice. A country-specific enrollment source can therefore matter far more than its raw percentage of total headcount suggests. The relevant quantity is not simply share of students; it is the contribution of that market to net tuition and contribution margin, its substitutability, and the speed with which costs can adjust if the market disappears.

USF provides unusually strong evidence of material China-market exposure. In September 2016, President Paul J. Fitzgerald stated at a USF China Business Studies Initiative forum that 10% of USF students came from different provinces throughout China. That is a direct contemporary university statement about students from China, not an estimate of Chinese ancestry. Earlier reporting in 2012 counted 781 Chinese nationals among 10,017 students and reported then-full tuition of $36,000, implying a gross sticker-price scale of about $28.1 million before aid and program differences. By Fall 2020, USF reported 525 students from China among 1,272 degree-seeking international students, or about 41.3% of that international population.

The most probative school-level evidence comes from USF's FY2027 Three-Year Financial Plan. The School of Management enrolled 640 Chinese undergraduates and 107 Chinese graduate students in Fall 2015, or 747 students. By Fall 2025, those figures had fallen to 31 and 12, respectively: a decline of 704 students, or approximately 94%. Over the same decade, total School of Management undergraduate enrollment fell from 2,291 to 994 and graduate enrollment from 732 to 411. USF states that the school's enrollment contraction was driven overwhelmingly by the collapse in international enrollment, alongside decline in the legacy College of Professional Studies.

The institutional response was not cosmetic. The same plan reports that four graduate programs were sunset, the remaining graduate portfolio was redesigned, full-time faculty fell from 86 to 47, and professional staff fell from 72 to 37. The loss of Chinese students alone numerically equals about 43.5% of the School of Management's total headcount decline over the period. That ratio is descriptive, not a causal allocation: it does not mean China caused 43.5% of every cut. It does show that the exposure was large enough to be structurally consequential.

This fixed-cost structure explains why an exposure of “only 10%” can be strategically large. If a national cohort supplies roughly one-tenth of enrollment or tuition while most costs cannot fall by one-tenth immediately, the revenue shock is absorbed first by the university's much smaller flexible margin. The result can require disproportionately large adjustments in hiring, sections, program portfolios, services, capital projects, or tuition discounting.

A useful concept from the companion financial analysis is a revenue-shock tipping threshold, \(\lambda^{*}\): the smallest revenue loss that pushes recurring operations below break-even before management can fully resize costs. The concept is analytically valuable because it explains why a 5–10% source-market exposure can matter disproportionately in a high-fixed-cost university. The companion manuscript's specific numerical calibration of a roughly 5.4% tipping point, however, depends on assumed fixed/variable cost classifications, expense elasticity, spillovers, donor responses, and timing. It should be treated as a stress-test scenario, not an audited USF threshold, until those inputs are independently reconstructed.

This is a feedback mechanism, not an inevitable death spiral. Universities can offset a country shock through domestic recruitment, replacement international markets, philanthropy, price changes, new programs, reserves, endowment support, or cost restructuring. Rankings are likewise an indirect channel: closing a program does not mechanically lower a U.S. News score, but sustained resource, faculty, retention, outcome, or reputation changes can affect the admissions market and ranking inputs over time.^[66]

For the symbolic-accommodation hypothesis, the key consequence is an incentive rather than a demonstrated motive. A university with hundreds of students tied to one foreign market has rational reasons to invest in country-specific recruitment, alumni networks, orientation, language-specific communications, student organizations, cultural programming, and inexpensive symbols of belonging. USF also built institutional China-facing infrastructure during the high-dependence period, including the China Business Studies Initiative and high-level U.S.–China programming. Whether the PRC flag itself was adopted or retained for recruitment, diplomacy, generic internationalism, or some combination cannot be inferred without the flag-policy decision record.^{[54, 58, 59]}

Here A_j denotes accommodation toward national market j. The hypothesis predicts more country-specific recruitment, services, networks, and symbolic inclusion as material exposure rises, controlling for a university's general level of internationalization. It does not predict that administrators consciously trade political principles for each tuition dollar, nor that accommodation is inherently improper.

The previously modeled estimate of approximately $40.9 million in “Chinese-heritage” tuition is not used as an observed revenue measure because it combines a 2016 statement about students from China, a 2020 international headcount, a later total-enrollment denominator, and 2025–26 tuition. Those years and populations are not commensurable, and “students from China” cannot be redefined as all domestic students of Chinese heritage. The directly supported evidence is narrower: China-origin enrollment reached roughly 10% of the university in 2016; the School of Management contained an unusually large Chinese cohort in 2015; and the later contraction coincided with substantial program and staffing restructuring. The separate $394,000 FY2024 operating-deficit figure is likewise omitted from the quantitative baseline because the cited audited statements report different accounting measures—a $15.873 million decrease in net assets without donor restrictions and a $22.929 million total increase in net assets—without identifying the $394,000 figure as the same accounting concept.

Observed financial exposure, recruitment infrastructure, enrollment concentration, and organizational contraction establish an incentive environment but not the motive for a particular symbolic decision. Administrative friction avoidance, revenue-motivated symbolic accommodation, and reduced comparative political instruction remain causal hypotheses unless contemporaneous policy records, communications, minutes, or testimony connect the incentive to the decision.

The financial mechanism just quantified for USF is not a theoretical possibility invented for this paper. Human Rights Watch has documented the pattern at the level of the U.S. higher-education sector as a whole: declining state funding for public universities and growing reliance on international tuition and Chinese-government-linked partnerships has made universities “susceptible to Chinese government influence,” producing censorship and self-censorship by administrators and academics who did not want to “irk Chinese authorities.” A GAO review found Chinese government entities provided individual U.S. universities between $1.5 million and over $15 million through Confucius Institute partnerships and related arrangements — funding relationships that came with documented content and hiring conditions.^{[96, 97]}

HRW’s interviews document the mechanism operating exactly as the friction-avoidance hypothesis below predicts: one scholar reported that a senior university administrator asked him “as a personal favor” to decline media interviews critical of Xi Jinping during a presidential visit, citing concern that criticism could damage the university’s standing with Chinese partners and recruitment pipeline. HRW’s research also documents Chinese-government complaints to university officials specifically about hosting speakers PRC diplomats considered “sensitive,” and instances of course material being altered or removed after objections framed in terms of protecting the institution’s China-facing relationships. Separately, in March 2019 — within the same Bay Area region as USF, though at a different institution (UC Berkeley) — mainland-student attendees attempted to shout down speakers at a public event addressing the Xinjiang human-rights crisis, an example of the peer-pressure channel operating alongside the administrative one.^{[98, 99]}

The mechanism HRW documents nationally is structurally identical to the one this section quantifies for USF: an institution financially exposed to a national market has a rational incentive to avoid programming, speech, or symbolism that the government of that market’s students finds objectionable, independent of any individual administrator’s personal views on China. That incentive does not require a conspiracy or explicit instruction from Beijing — it operates through ordinary institutional risk-aversion once the revenue exposure quantified above exists.

Campus-flag evidence requires separate coding for outdoor display, recurring display, administrative prominence, China-based operations, and ceremonial use. Several materially different U.S.-campus cases are already documented:

International flags plus Chinese enrollment do not establish the presence of a current PRC flag. A campus flag claim therefore requires a verified flag identity together with its date, location, display rule, and duration.

The meaning of a national flag depends partly on who is displaying it and what that institution exists to do. Some Chinatown organizations are clan, district, hometown, benevolent, business, cultural, or explicitly cross-Strait political organizations whose histories are directly bound up with Chinese communal identity. For such organizations, a PRC or Taiwan flag can have a comparatively direct representational, historical, or constituency function. Even there, the flag does not prove CCP ideology or foreign direction; it may express nationality, ancestral identification, diplomacy, association politics, or a particular cross-Strait position.

The University of San Francisco occupies a different institutional category. USF is a private American university whose published mission emphasizes education, diversity, critical engagement, and learning. It is not a PRC diplomatic mission, a PRC state institution, or a Chinese community association whose constitutive purpose requires representation of the PRC as a sovereign. A standing PRC flag on USF property is therefore more accurately treated as a discretionary internationalization and inclusion choice, not an institutionally necessary expression of USF's own national identity.

That discretionary-inclusion rationale is testable against USF's own conduct: USF recognizes a Taiwanese Student Association alongside its Chinese Students and Scholars Association (Section 04.5), so a generic policy of representing the university's actual student-nationality mix predicts a Taiwan flag at the same installation. The reported PRC-without-Taiwan asymmetry at Lone Mountain (Section 04.3 row; Section 11.C census pending) is evidence against the neutral-internationalization account and toward the institutional-accommodation account this section develops: a truly nationality-representative policy has no reason to include one Chinese-speaking polity's flag while excluding the other's.

This creates a useful Institutional-Role Necessity Gradient. A PRC embassy or consulate has an obvious mission-derived reason to display the PRC flag. A community organization explicitly organized around PRC national representation may also have a strong internal rationale. A broadly Chinese hometown or benevolent association has a contextual but contestable rationale. An American university has the weakest intrinsic nationality-based necessity: its rationale must instead come from a policy of international representation, student inclusion, diplomacy, recruitment, or another university-specific purpose.

The Necessity Gradient identifies the structural condition — USF has weak intrinsic reason for national flag display, so a discretionary rationale is required. Which rationale operates is what H10 tests: whether USF's flag-selection rule, administrative location, stated purpose, recruitment environment, and decision process point toward accommodation of the material tuition exposure established in subsection 1. Under the neutral-internationalization account the record should also carry a Taiwan flag at the same installation; under H10 it should not. The reported Lone Mountain asymmetry, once the Section 11.C census confirms it, is the observable prediction on which the two accounts differ.

The institutional-cue hypothesis is sharpened when a national symbol is not encountered only once. USF currently recognizes a Chinese Students and Scholars Association within its culturally focused student-organization structure, alongside a Taiwanese Student Association. USF also describes CSSA leadership as organizing orientation sessions tailored to Chinese international students, managing a dedicated WeChat group, and serving as a bridge between the university and Chinese international students. These functions are ordinary student-support activities, but they make the organization a consequential peer-network node.

A separate firsthand account from the 2019–2021 period describes CSSA participants marching on campus with an approximately thirty-foot-wide PRC flag. No independent archived university or news source has yet been located that establishes the dimensions or dates of those marches. The observation is therefore coded as a firsthand archival lead, not as a verified historical finding. If dated photographs, video, CSSA posts, or university records are recovered, the event can be upgraded to a directly observed mobilization variable.

The causal proposition is an interaction rather than a direct flag effect. A student who arrives with pre-existing political priors may receive repeated cues from co-national peers and from the university as an authority environment. The prediction is not that either cue creates CCP support. It is that the combination can reduce the likelihood that inherited assumptions are independently re-examined when comparative civic education and competing Chinese-language political narratives are weak.

Established: USF carried material, quantifiable financial exposure to China-origin enrollment during the 2012–2018 period (subsection 1), built dedicated China-facing recruitment infrastructure including a China Business Studies Initiative and a China-specific admissions office and WeChat recruitment channel, and displays a PRC flag under administrative authority in a high-salience location (subsection 3). USF also experienced a severe, well-documented contraction in Chinese enrollment and associated program and staffing cuts after 2018 (subsection 1), consistent with the national pattern of universities suffering real financial harm from reduced Chinese enrollment — which is itself further evidence that the underlying incentive to avoid China-market friction was, and for any remaining exposure still is, financially real rather than speculative.

H10 as the specific testable claim: the USF-specific causal question — whether USF administration has cancelled, discouraged, or disadvantaged specific anti-CCP or pro-Taiwan speakers, events, courses, or student organizations because of China-market concerns — is registered as H10 and addressed by the design in Section 11, not asserted here as a discovered incident. This is a deliberate posture, not a research failure. HRW's national research finds that self-censorship of this kind typically leaves no documentary trace, since it takes the form of an event never being proposed, a course topic never being assigned, or an invitation quietly not extended. A search for a specific public cancellation is therefore the wrong instrument: the design instrument in Section 11 — organization-year records, faculty and program-level testimony, comparative programming density across high- and low-China-exposure institutions — is the right one. What this section establishes is the incentive structure and the general national mechanism; H10 is what Section 11 decides.

A separate reported episode concerns motive attribution during a Lone Mountain visit undertaken to check the university's PRC-flag display. According to the firsthand account, ordinary campus photography was interpreted as possible interest in a distant Bay Area security or military facility. The author reports that a vehicle — a blue Tesla consistent with a campus public-safety fleet — sounded its horn and stopped suddenly and closely near him while he was photographing the flag installation, and that the vehicle's occupant subsequently proceeded to the campus public-safety (CASA) building. No one approached the author directly or stated a reason for the stop at the time. The author did not learn of any "hostile reconnaissance" characterization through a direct conversation; his understanding of how the incident was interpreted is inferential, based on the sequence of events rather than on anything communicated to him.

The evidentiary standard here is symmetrical: just as PRC flag ≠ CCP agent, the inference visitor + camera near a flag ≠ evidence of intelligence activity requires its own independent evidentiary bridge, which is absent on the reported facts. The episode is not independently adjudicated and cannot establish institutional policy, discriminatory motive, retaliation, or coordination. Its value for the paper's argument is precisely that it is an ordinary, undramatic sequence — a horn, a close stop, a drive to a security office, no conversation — which is the same low-specificity evidentiary basis this section argues is insufficient to support a hostile-intent inference in either direction. If the university's own PRC flag is displayed as an ordinary inclusion symbol, an interested visitor's photography of that same flag should not, on this evidentiary record, independently trigger security scrutiny.

The proper inferential sequence is Bayesian rather than accusatory. A routine visitor/flag-documentation explanation and a hostile-reconnaissance explanation are competing hypotheses. Evidence should update between them according to how specifically it is predicted by each. A camera aimed at the campus flag is strongly expected under the first hypothesis. The hostile-reconnaissance hypothesis becomes materially stronger only with behavior such as persistent targeting of a sensitive facility, unusual magnification or pattern collection, access circumvention, concealed tasking, or a foreign-principal relationship.

Security personnel are not barred from asking neutral questions when conduct creates a genuine site-specific concern. The methodological and civil-liberties boundary is crossed when a low-specificity fact—appearance, ancestry, language, or ordinary public photography—is converted into a high-specificity accusation without the missing evidentiary bridge. The same evidentiary principle applies throughout: observable symbols and identities are not substitutes for evidence of hidden motive.

A direct causal test combines three complementary designs to measure the perceptual and legitimacy effect Section 04.5 predicts:

The friction-avoidance hypothesis targets institutional incentives facing university administration. The remedy that follows is administration-level: the transparency and symmetric-enforcement toolkit specified in Section 09 — disclosed funding relationships, viewpoint-neutral event and speaker policy, and published flag-and-symbol criteria. Individual Chinese or Chinese-American students are outside its scope; their presence on campus is not, per Section 00, evidence of any political position. The experimental record supports this allocation directly: Fan, Pan, Shao, and Xu find that anti-Chinese discrimination among Chinese undergraduates in the United States reduces support for political reform and increases support for authoritarian rule, while non-racist criticism of the Chinese government produces no such backlash. Student-level suspicion would therefore be both wrong in principle and counterproductive in effect; institutional-level accountability is the mechanism that actually works.

In 2016, 68% of surveyed Chinese Americans characterized Chinese investment in the United States as driven by financial opportunity rather than an attempt to increase China’s influence; 51% expected such investment to help the U.S. economy and 44% expected it to improve U.S.–China relations. These figures support a high-engagement environment without establishing ideological support for the CCP.

Per the source-weighting rule in Section 11.A, PRC state-media outlets (Xinhua, PRC MFA, Consulate General) are not treated as neutral measurements of crowd size or motive. The August–October 2019 peak claim does not depend on them. It rests on the following independently sourced facts, any one of which is sufficient to establish organized institutional mobilization; taken together, they establish a coordinated, cross-city wave.

The peak claim therefore survives the strongest available provenance objection. State-media coverage adds volume estimates and celebratory framing that the paper does not credit as neutral measurement; it does not supply the occurrence facts, the organizer identities, the named venues, or the U.S. official participation. Removing every PRC-sourced sentence from the record above leaves the coordinated August–October 2019 pattern intact, and Section 11.F specifies the falsification standard: only consistent organization-year evidence placing higher PRC-oriented mobilization outside 2019 — once event size and source bias are normalized — would displace the peak.

Treating May 2019 as the plausible onset, rather than October as the plausible close, follows from ordinary causal reasoning about how organizational confidence is built and sustained rather than switched on for a single autumn. Three chained facts support this: (1) the 3–14 May rhetorical escalation supplied the ideological vocabulary — "people's war" — that reframed routine trade friction as civilizational confrontation, which is the kind of framing that mobilizes organizations rather than individuals reconsidering a survey answer; (2) the 17–18 August cross-city counter-mobilization shows that vocabulary had already been operationalized into coordinated street presence within three months, well before the September/October San Francisco events this section documents; and (3) Hong Kong's National Security Law was not enacted until June 2020, meaning the underlying grievance driving PRC-aligned institutional mobilization did not resolve or de-escalate at the end of 2019 — it continued, and the pandemic-era rise in U.S. anti-Chinese sentiment supplied a second, independent motive for PRC-aligned institutions to close ranks defensively through 2020–21. Deductively, an institutional-confidence variable driven by an unresolved ideological trigger, a demonstrated capacity for rapid mobilization, and a compounding defensive motive should be modeled as elevated across a multi-year window, not as a single-season spike that resets by November 2019. The October 2019 close date in the table above is a documentary boundary — it is where this paper's discrete, dated SF events happen to end — not a claim that institutional confidence itself declined immediately afterward.

This inference is falsifiable in the same way the rest of this paper's claims are: it predicts that an organization-year panel of the kind specified in Section 10 should show San Francisco PRC-oriented institutional mobilization remaining elevated, not declining, across 2020–21 relative to the pre-2019 baseline, and declining only where an independent shock (such as pandemic-era program contraction) reduces institutional capacity rather than institutional will. That prediction has not yet been tested against a reproducible panel; it is presented here as the paper's best current inference, not as an established finding.

If \(\alpha\) is large, accumulated relationships generate persistence. A network built through consular contacts → provincial and municipal relationships → delegations → business introductions → community banquets → donors → media relationships → board interlocks can remain valuable even after mass China favorability falls.

This mechanism explains how 68% national Chinese-American China favorability in 2016 fell to 41% by 2022–23 while traditional Chinatown institutions remained substantially more PRC-oriented.

Here \(F\) denotes family exposure, \(A\) assets, \(B\) business exposure, and \(P\) professional exposure inside the homeland state. Physical migration can increase legal protection while leaving these channels of leverage intact. Documented PRC transnational-repression cases include pressure through family members. The existence of the mechanism is established; its prevalence among ordinary Chinese Americans is not.

Xi Jinping’s November 2023 visit to San Francisco should not be interpreted as proof that broad Chinese-American opinion returned to earlier levels. It is more useful as a mobilization shock revealing organizational capacity that may otherwise remain latent. Contemporary reporting described hundreds of pro-Xi supporters, some arriving by tour bus, and messages on university networks offering transportation, food, or accommodation. Large Tibetan, Hong Kong, Uyghur, Chinese-dissident, and other anti-CCP demonstrations occurred simultaneously.

The Washington Post later reconstructed the confrontations from more than 2,000 photographs and videos and over 21 hours of footage. A separate Hong Kong Democracy Council / Students for a Free Tibet report catalogued 34 alleged incidents of harassment, intimidation, or assault. These sources concern particular events and actors, not the political preferences of Chinese Americans as a population.

In September 2025, PRC Ministry of Foreign Affairs reporting stated that the Chinese Consulate General and San Francisco city government jointly held a PRC National Day flag-raising ceremony at City Hall, with Mayor Daniel Lurie participating and proclaiming October 1 “China–US Friendship and Heritage Day.” Whatever the normative interpretation, the event demonstrates continuing official and institutional connectivity years after Chinese-American national China favorability had fallen sharply.

Foreign-influence law does not create a general offense of being sympathetic to a foreign country. FARA is principally a disclosure regime for specified activities performed at the order, request, direction, or control of a foreign principal, subject to exemptions. Section 951 is a distinct criminal notification statute for persons acting in the United States as agents of foreign governments. Neither statute is established by favorable views of China, a PRC flag, Chinese-language activity, consular attendance at a public event, or ordinary academic/business engagement standing alone.^{[31, 67, 81]}

The FBI defines transnational repression as foreign-government conduct reaching into the United States to intimidate, silence, coerce, harass, or harm diaspora and exile communities. Reported tactics include stalking, threats, family pressure abroad, cyber activity, coercive return efforts, assault, and attempted kidnapping. The category identifies a threat pattern; it does not create a presumption that ordinary diaspora organizations, students, or political advocates act for a foreign government.^{[29, 71]}

Where a concrete incident involves repeated following, a credible threat, threatened injury, coerced transfer of property, interference with a legally protected right, or disclosure of protected institutional records, the appropriate legal analysis begins with the elements of the relevant statute. The nationality of the alleged actor is not a substitute for those elements.^{[74, 75, 76, 77, 78, 73, 69]}

The legal implications follow from the same distinction that structures the empirical analysis: regulation is most defensible when it targets undisclosed foreign-government direction, coercion, discriminatory enforcement, misuse of protected records, and opaque financial relationships rather than nationality, lawful advocacy, cultural affiliation, or political opinion. A rule aimed at suppressing “pro-PRC” speech as such would collapse viewpoint into agency and reproduce the attribution error identified by the evidence.

Australia’s Foreign Influence Transparency Scheme and the United Kingdom’s Foreign Influence Registration Scheme provide comparative models centered on specified relationships and influence activity rather than ethnic identity. Their details differ from U.S. law, but both reinforce the same structural principle: disclosure duties should attach to defined foreign-principal or foreign-power arrangements, not to a person’s ancestry or political preference.^{[32, 33]}

Section 03 argued that the Chinatown flag ratio functions as something closer to an institutional vote than an opinion poll, and that a Taiwan-aligned or anti-CCP resident occupying the minority side of that vote faces a real, structural disadvantage independent of national survey trends. The forward-looking question is what a numerical minority inside that specific institutional ecosystem can do about it — and the answer, consistent with everything argued in this section, is the same lawful, viewpoint-neutral toolkit available to any minority faction inside a voluntary association system, not an exception to it. Section 06's M4 (institutional path dependence) mechanism explains how the current PRC-oriented majority was built: sustained participation in leadership, event programming, and organizational continuity over decades. Rebalancing it requires the same mechanism run in the opposite direction, not a shortcut around it.

None of this requires, and this paper does not endorse, treating PRC-sympathetic Chinese Americans as a class to be excluded, monitored, or displaced. The strategy above targets institutional positions and voluntary association membership — conduct available to any political faction — not the ancestry, nationality, or private belief of anyone inside or outside that faction. A 5–10 year horizon is realistic precisely because Section 08 shows institutional alignment moves at the speed of leadership turnover and organizational renewal, not at the speed of a single election cycle or a single dramatic event; that is also why an unbounded or extralegal shortcut would not actually work faster — it would simply violate the same conduct-not-identity standard this section spends its length defending.

China operates two confirmed overseas military bases. The PLA Support Base in Djibouti opened in 2017. The Ream Naval Base joint logistics and training center in Cambodia was inaugurated in April 2025 and includes a 650-meter pier able to berth aircraft-carrier-sized vessels — capacity Cambodia's own navy does not use.^[100] The Department of Defense's 2025 report to Congress states that Beijing is weighing additional overseas logistics sites across some combination of 21 countries in the Indian Ocean, Africa, and the Western Pacific, and that this basing effort advances alongside a six-carrier expansion target by 2035.^[100] RAND's site-selection research, which evaluated 108 candidate host nations, finds the PLA's basing criteria driven primarily by economic and logistical utility rather than by preparation to strike U.S. forces; a companion RAND study of Chinese military writings concludes the PLA currently has neither the intent nor the capability to conduct offensive operations from its overseas bases through at least 2030, though it recommends the U.S. government continue monitoring for a shift.^[101] This distinction matters: the basing network is a documented and expanding fact; its wartime offensive utility is a separately assessed and currently negative finding, not an open question this paper resolves.

BRI-financed ports show a documented dual-use pattern independent of any Chinese base agreement. C4ADS's analysis of fifteen BRI-funded port projects across the Indo-Pacific and CSIS's research on Chinese military access in the Indian Ocean both find that Chinese-financed port infrastructure is regularly built to specifications exceeding ordinary commercial need — deepwater berths capable of accommodating naval vessels, reinforced access roads, and fuel and cold-storage capacity — that match PLA operational requirements and support Chinese naval refueling and resupply without requiring a return to Chinese territory.^{[102, 103]} Beijing's public framing of the Belt and Road Initiative as a purely economic program does not account for this pattern, which multiple independent research organizations have documented using open-source and satellite-imagery methods.

If the Chi Haotian document is genuine, its claims describe CCP ambitions reaching past regional hegemony toward the long-term displacement of the American population and the opening of North America to Chinese settlement and resource extraction. On that reading — and only on that reading — the documented PLA basing network and BRI dual-use ports described in Subsection A would function not as ordinary power projection but as forward staging for that outcome, and the institutional footholds traced in Sections 03–08 would represent the erosion of the target population's capacity to recognize or resist it. A CCP biological attack, in this same conditional scenario, would be harder to attribute and to counter if the party's influence reached into American epidemiology, law enforcement, telecommunications, and government. And a public exposed long enough to United Front elite cooptation and public-opinion operations might, on this reading, come to treat a gradual takeover as unavoidable, or even preferable to a fractious domestic politics.

Every sentence in this subsection depends on crediting a document this paper has not been able to authenticate. It is included as the clearest illustration of what would be at stake if the document is genuine, not as a finding this paper makes.

Primary statutes, official institutional records, audited financial statements, and original survey reports receive the greatest evidentiary weight. Government and advocacy sources are used for attributed claims or event occurrence unless independently corroborated; state-media crowd estimates and causal characterizations are not treated as neutral measurements. Survey findings retain their original question wording, sampling frame, and methodological limits. This weighting rule separates evidence of occurrence from evidence of motive, prevalence, or causation.

The primary unit of observation is the organization-year. A 100–300-organization panel can separate durable institutional alignment from one-off events and from individual public opinion.

The official approximately thirty-block Chinatown planning boundary supplies the denominator. Each visible address is coded by organization, PRC large/small flags, Taiwan large/small flags, U.S. flags, no national flag, image date, source, and confidence. Normal-day observations are estimated separately from October 1, October 10, Lunar New Year, and major diplomatic visits.

A nationally representative experiment can randomize why a hypothetical China-born adult migrated to the United States: graduate education, economic opportunity, family reunification, political persecution, or an unspecified reason. Respondents then estimate the person’s likely views of the PRC government, CCP, Taiwan, U.S.–China restrictions, and continuing China-linked professional relationships. Party identification, ideology, education, age, military background, and China knowledge enter as moderators.

Candidate discontinuities include the 2018 trade confrontation, May 2019 nationalist escalation, June–November 2019 Hong Kong protests, the 2020 pandemic rupture, the June 2020 Hong Kong National Security Law, 2021–22 strategic competition, and the November 2023 APEC summit.

Separate outcomes include PRC flags, Taiwan flags, pro-PRC events, anti-CCP events, institutional statements, attendance, leadership-network density, media tone, and commercial-exposure proxies.

China's population peaked at 1.4126 billion in 2021 and has fallen for four consecutive years, to roughly 1.405 billion at the end of 2025.^[104] The UN's 2024 World Population Prospects projects China will record the largest absolute population loss of any country in the 2024–2054 window (−204 million) and, if current fertility persists, the largest through 2100 (−786 million, returning to a population size last seen in the late 1950s).^[105] Births fell to a record low of 7.92 million in 2025, down 17 percent year-on-year, against a total fertility rate near 1.0 — well below the 2.1 replacement level.^[104]

The aging half of that trend is what converts a population problem into a fiscal one. China's 65-and-over population stood at 209.78 million (14.9 percent of the total) in 2022; official projections put it above 300 million by 2033 and above 400 million by 2050, peaking near 436 million around 2057 — an increase of roughly 100 million elderly people within about a decade.^[106] The working-age population (15–64), by contrast, peaked in 2015 and has been declining since.^[107] Two standard demographic ratios quantify the resulting strain:

For comparison, the United States is projected to retain more than 2.5 working-age adults per person 65 and older in 2050 — a materially larger support base for the same-aged pension and health-care obligations.^[107]

A simple growth-accounting identity connects this demographic arithmetic to the GDP projections below. Using the standard Solow decomposition, output growth is approximately the sum of labor-force growth, capital's contribution, and productivity growth:

The realized numbers track that arithmetic. China's National Bureau of Statistics reported 5.0 percent GDP growth in 2025 (independent estimates put the figure closer to 4.0 percent); the IMF projects 4.4 percent for 2026 and 4.2–4.3 percent for 2027, with its longer-range review flagging a slide toward roughly 3.3 percent by 2029 on aging and slowing productivity growth.^{[109, 110]} The fiscal side compounds the same pressure. A standard debt-dynamics identity gives the mechanism:

China's official 2026 budget deficit target is 4.0 percent of GDP; the IMF's augmented measure, which folds in local-government financing-vehicle borrowing, puts the true public-sector deficit closer to 14 percent of GDP, a level the Fund has flagged as unsustainable absent local-government debt restructuring and reduced industrial subsidies.^[111] The IMF's 2026 review names a "deeper-than-expected" property-sector downturn as the primary domestic risk to the entire trajectory.^[111]