The Anxiety Machine

There is a pattern running through 125 years of Australian political life that the event-response model of foreign policy cannot explain. It is simple enough to state: Parliament fills with China-threat language not when China does threatening things, but when Australian workers are being squeezed. The discourse tracks domestic stress, not external behaviour. This is not a metaphor, a conjecture, or an ideological position. It is a measurement.

The measurement comes from my 105-year quantitative study combining computational text analysis of the National Library of Australia's digitised Trove newspaper corpus (1900–2025), Hansard parliamentary records (2006–2025), and annual CAMS institutional health scores for both Australia and China. The full project is documented at neuralnations.org/epiphenomenon. What it finds is structurally so regular — so resistant to the particular actors, vocabularies, and crises that fill each era — that it begins to acquire the character of a physical law rather than a historical observation. The threat discourse is not produced by the threatening behaviour of the named threat. It is produced by the institutional conditions of the society doing the naming.

To understand why Parliament fills with China-threat language when working people are squeezed, it is first necessary to understand a prior condition: why it stops doing so when the squeeze becomes genuinely catastrophic.

My CAMS Managed Threat Ecology model identifies two discourse registers that compete for the same political and emotional bandwidth in any society: class discourse (wages, strikes, unemployment, housing, the cost of living) and foreign-threat discourse (the external enemy who explains and justifies national anxiety). These two registers are not additive. They are rivals. When one floods the available space, the other contracts.

The Depression years demonstrate this at maximum intensity. As Australian institutional health collapsed through the 1930s — mass unemployment, wage cuts, bread queues, evictions — the Yellow Peril virtually disappeared from the press. Class language surged to its all-time corpus peak. The correlation between class discourse and Australian institutional health across fifty-five years runs at r = −0.700 (p < 0.001) — the strongest single correlation in the entire Trove analysis. The inverse relationship is foundational: strong institutions suppress class discourse; weak institutions amplify it. And crucially, when class discourse surges under genuine crisis, foreign-threat discourse retreats. You cannot manufacture an existential enemy abroad when the existential enemy is the empty larder at home.

This crowding-out mechanism is the positive control — the methodological ground truth against which everything else in my analysis is calibrated. The question the essay addresses is what happens when the machinery of threat management becomes sophisticated enough to suppress the crowding-out mechanism itself.

The pattern long precedes data. The Yellow Peril fever that produced the White Australia Policy in 1901 emerged at Federation — Australia at its most institutionally coherent, not its most stressed. In the Trove data this shows as yellow_peril discourse correlating positively with institutional health at r = +0.272: a prosperous, coherent society performing its own boundaries from a position of strength. Threat language, in this foundational regime, is a luxury of coherence, not a symptom of fear.

The Depression inverts this. My CAMS Regime 2 (1914–1939) shows institutional health at its lowest mean value of any of the three identified regimes, yellow_peril at its lowest mean (111.5 articles per million, collapsed from the 329.9 ppm of the Federation period), and class discourse at its highest mean. When institutional conditions are genuinely severe, class discourse crowds out the foreign-threat register.

Robert Menzies understood this with the cold precision of a barrister. The secret of Cold War political management was not to generate genuine class crisis but to maintain a manageable level of working-class pressure — enough to produce anxiety without producing the full crowding-out that would suppress the communist threat narrative. His use of the Red Scare as "electoral gold," his 1950 legislation to ban the Communist Party, his 1951 referendum — these were not responses to actual communist subversion. They were instruments for keeping the class-threat tension at the productive ratio. Menzies' genius was calibration.

The contemporary period introduces a structural shift that Menzies could only have dreamed of. My CAMS analysis of Australian institutions in the post-2006 period identifies what I call the Shield Hypertrophy Index (SHI): Shield Node Value minus the mean of Lore (knowledge institutions) and Flow (trade and exchange) Node Values. The SHI is the quantitative signature of a society in which security logic is crowding out knowledge and exchange — in which the institutions that could generate critical scrutiny of threat narratives are being progressively weakened relative to the institutions that produce and sustain them.

From approximately 2010–2012, a distinctive pattern emerges in Australia's CAMS scores. The Shield Node Value holds steady. Lore declines. Flow declines. The SHI rises. By 2017, the institutional imbalance is at its widest point in the post-2006 record. And the lagged correlation between SHI[t] and sino_ppm (Hansard)[t+1] is r = +0.697 (p = 0.0009, n = 18). Australian institutional imbalance in year t predicts Australian parliamentary China-threat discourse in year t+1. This is my primary empirical finding from the Epiphenomenon@Trove study.

What this means in plain terms is that the mechanism driving China-threat language in Parliament is not China. It is the widening gap between what the security apparatus produces and what the knowledge and exchange institutions can push back against. When Lore shrinks — when universities are defunded, public broadcasting is gutted, independent think tanks are replaced by security-funded ones — the epistemic capacity to challenge threat narratives contracts. When Flow is stressed — when the trading relationship with China generates anxiety rather than confidence — the emotional condition that makes threat narratives credible is amplified.

The most striking single result in the entire 105-year dataset is what I term the bilateral symmetry finding.

Across the Trove press record from 1900 to 1990, the Pearson correlation between Chinese institutional health (CAMS5 ensemble mean Node Value) and Australian Sino-discourse is r ≈ 0.10 — statistical noise. The discourse is indifferent to whether China is institutionally strong or weak, stable or chaotic, threatening or not.

After 2006 the correlation does not merely appear from the noise. It flips sign. The correlation between Chinese mean Node Value and Australian parliamentary Sino-discourse across 2006–2025 is r = −0.562 (p = 0.010, n = 20). Negative. When China's institutions are functioning well — when its governance is coherent, its economy productive, its social systems stable — Australian parliamentary China-threat language is higher, not lower.

The 2020 data point is the critical test of whether the structural transition is complete. In historical conditions — the Depression, the post-Soviet détente — genuine domestic crisis triggered crowding-out: the class register flooded in and the foreign-threat register retreated. In 2020, Australia experienced a genuine simultaneous domestic and international crisis. The pandemic drove institutional stress across every node. Chinese institutional health fell sharply (mean Node Value from 13.88 in 2018 to 9.81 in 2020). Both societies were stressed. The historical model predicts crowding-out.

Instead, Sino-threat discourse in Hansard surged. The crowding-out mechanism had been disabled. The strategic infrastructure assembled between 2014 and 2020 — the Foreign Influence Transparency Scheme, the Huawei ban, the ASPI apparatus, the Foreign Interference laws, the Five Eyes China-specific frameworks — had made the ecology structural enough to override the historical mechanism. The Anxiety Machine no longer required favourable conditions to produce output. It had become self-sustaining.

The 105-year arc of this research resolves into three phases of threat ecology, each showing the class-threat tension in a different configuration.

The Reactive-Prosperity phase (1900–1913; 1955–1970) operates from a position of institutional coherence. Threat discourse here is not anxiety. It is boundary performance — a confident society organising its self-definition through racial or ideological exclusion. The political class produces threat discourse from strength, as identity rather than as displacement.

The Reactive-Crisis phase (1929–1939; 1991–2006) is the crowding-out condition. Hands stress is severe, Lore is present but insufficient to redirect the energy productively, and the displacement infrastructure either does not yet exist or has been dismantled. Class discourse dominates. Foreign-threat discourse contracts. The Depression, paradoxically, was the era of least Sinophobia in the 125-year record.

The Structural-Infrastructure phase (2006–present) is the Anxiety Machine at full operation. Hands stress is real — wage stagnation, housing unaffordability, energy costs, casualised labour — but not yet at Depression severity. Lore is declining. Shield is dominant. The infrastructure for displacement was assembled and is now self-sustaining. The crowding-out mechanism has been disabled. Parliamentary Sino-discourse tracks Shield Hypertrophy, not Chinese behaviour.

The grammar remains the same across all three phases. The society facing its own internal stress requires a vocabulary for externalising that stress. What changes is the institutional infrastructure that determines whether the externalisation succeeds. In Phase III, it has succeeded more completely than at any previous moment in Australian history — because for the first time, the infrastructure for producing foreign-threat discourse is durable, well-funded, institutionally embedded, and politically coordinated with allied Five Eyes machinery that operates at a scale no single society could construct alone.

The essay began with measurement. It should end with it.

The Shield Hypertrophy Index for the United States in 2025–2026, the primary patron within whose managed threat ecology Australia operates, stands at 7.75 — more than double the previous worst period in the 126-year CAMS series, the Vietnam War at 3.75. US Helm stands at 0.4 (all-time series minimum). US Lore stands at 0.6 (all-time series minimum). Shield holds at 10.7.

A system with Helm at 0.4 and Lore at 0.6 but Shield at 10.7 cannot course-correct even if the impulse existed. The Australian Anxiety Machine does not operate in isolation. It is a subordinate node within a larger ecology whose own epistemic substrate — the Lore that carries the capacity to recognise and correct a trajectory — has been consumed alongside its governance capacity.

When working people are squeezed in this larger ecology — and across the Anglosphere, the Hands-node evidence of the last decade is unambiguous — the Anxiety Machine does not generate one nation's worth of displacement. It generates a coordinated, infrastructure-sustained, epistemically impoverished threat environment in which the answer — across 125 years of counting — is no. The cause of the squeeze is structural. The enemy is the product of the institution that produces the answer. And the institution that might tell us otherwise has a Lore score of 0.6.