2026-08-13 Observatory briefing¶

The Next Global Economic Crisis Could Be Made in China¶

The arithmetic of Chinese overcapacity has finally caught up with its politics. Froman's argument is straightforward: a state-directed industrial machine that must export at two to three times the rate of global trade growth has run out of customers, and the political tolerance of its trading partners has run out alongside them. The numbers do the work here — a $1.2 trillion surplus, 30 percent of global industrial production heading toward 45 percent, a third of Chinese industrial firms operating at a loss. This is not a crisis of competitiveness but of the internal logic of a system that cannot stop producing.

The genuinely interesting material is the description of neijuan — involution — as a structural feature rather than a policy error. Local governments propping up loss-making factories to preserve employment and tax revenue, state banks rolling over debt for insolvent borrowers, firms cutting prices below cost to chase market share: this is a machine that cannot slow down because the fiscal survival of local state actors depends on its continued motion. The contradiction is concrete: the same subsidies that sustain employment at home generate the export flood that destroys the political conditions for continued market access abroad.

Froman's proposed solution — preemptive rebalancing toward domestic consumption — is the standard liberal prescription, and it misses the material basis of the problem. The CCP's anti-involution campaign and rural consumption initiatives are gestures against a system whose entire structure of accumulation, local government finance, and employment depends on industrial expansion. A shift toward consumption would require dismantling the very mechanisms that keep the system running, which is why Beijing's "recognition" of the problem has produced so little change.

The global implications are worth noting: if the export machine stalls, the shock will transmit through supply chains to economies far beyond the Asia-Pacific, and the United States — the only actor with the institutional capacity to stabilise the system — will be called upon to manage a crisis it has spent years trying to provoke.

The AI Growth Paradox¶

Source: Foreign Affairs

The US federal debt crossing $40 trillion is the occasion, but Rogoff’s real subject is the ideological function of the AI boom. The promise that artificial intelligence will grow the economy fast enough to dissolve the debt is not an economic forecast; it is a political convenience. It allows both parties to avoid the only question that matters: who pays for the state’s accumulated obligations? The answer, historically, has been savers and wage-earners, through financial repression and inflation. Rogoff gestures at this without naming it as a class question, but the mechanics are there in his account of the 2010s — the era of “secular stagnation” when ultralow rates made borrowing feel free, until the bill arrived in the form of 2.4 percent real yields and interest payments that now exceed defence spending.

The contradiction is concrete. AI-driven productivity gains, if they materialise, would push up interest rates as capital demand surges; higher rates raise the cost of servicing a debt stock that cannot shrink quickly. The state is thus caught between two futures: stagnation, in which the debt becomes unpayable, and boom, in which the boom itself makes the debt more expensive. Rogoff’s warning that tax cuts and spending hikes will “get out in front of the boom” is the sharpest line in the piece — it names the real dynamic, which is that the fiscal bonanza will be spent before it exists, because the political class cannot afford to wait. The AI narrative is not a solution to the debt problem; it is the latest form of the magical thinking that postponed the reckoning in the 2010s. The underlying issue is unchanged: a state that has chosen to finance itself through borrowing rather than taxation, and a ruling class that will resist any adjustment to the tax and spending trajectories that would actually address it. The crash, when it comes, will not be caused by AI’s failure but by the political refusal to confront the distributional question that the debt embodies.

Regional conflicts and their implications for charterparties¶

Source: Hellenic Shipping News

The legal framework for refusing a voyage into a war zone has always been a study in managed risk, but the current geography of conflict is stretching it in ways the drafters of standard clauses likely never anticipated. The piece correctly identifies the central tension: owners must demonstrate "objective reasonableness" in their judgement of danger, yet the threats described are no longer discrete events but a permanent condition of operating in these waters. When the Houthis can announce an embargo on Saudi vessels and hit two tankers in quick succession, or when Ukrainian drones strike 196 shadow fleet vessels in a fortnight, the distinction between a "real likelihood" of danger and mere possibility becomes a legal fiction.

What is analytically interesting here is how the clauses themselves have evolved to accommodate this new permanence. The 2013 and 2025 CONWARTIME revisions explicitly state they apply whether the risk existed at the time of charterparty formation or arose later. This is a quiet admission that war risk is no longer an exceptional deviation from normal trading but a baseline condition to be priced and allocated. The shift from 48 to 72 hours' notice under the 2025 clause similarly reflects a reality where rerouting decisions require more deliberation because the alternatives are themselves compromised.

The tanker forms reveal a different logic. Shelltime 4's clause 35 and Beepeetime 2's clause 40 both hinge on the "reasonable opinion" or "discretion" of owners, yet The Product Star established that this discretion cannot be exercised arbitrarily. The legal test, in effect, asks whether a shipowner's refusal to sail into a live conflict zone is a proportionate response to the risk. That question, posed against the backdrop of insurance premiums that have become a speculative market in themselves, suggests the real function of these clauses is not to protect crews but to allocate liability when the inevitable dispute arrives. The law has become a secondary market in risk, layered on top of the physical one.

Shipping Number of the Week , China’s iron ore mining cools as imports rise 6% and steel production weakens¶

Source: Hellenic Shipping News

The displacement of domestic Chinese iron ore mining by seaborne imports is not simply a story of comparative advantage, but of the spatial fix being enacted at a particular moment in the overaccumulation cycle. Domestic output fell 7% y/y while net imports rose 6%, pushing imports to 57% of supply. The logic is straightforward: higher-grade foreign ore is cheaper per unit of iron, and with steel demand weak, mills are cutting costs rather than volumes. This is capital restructuring under pressure, not expansion.

The interesting wrinkle is the ton-mile effect. Imports from Australia and Brazil rose modestly, but the real growth came from Guinea, Liberia and Peru — longer hauls that inflated dry bulk demand by 9% y/y even as cargo volumes grew less. The Simandou project, ramping toward 120m tonnes by decade’s end, is the material expression of this: a massive fixed-capital investment in West Africa predicated on Chinese steel demand that is, at present, contracting. The 79% jump in the Capesize T4 Index reflects this artificial tightening — demand for shipping services rising precisely because the commodity itself is in surplus.

The contradiction sits in the second half outlook. Port inventories are elevated, steel production is falling, and the property sector remains sluggish. Yet the freight market is being buoyed by distance, not volume. This is a fragile equilibrium: if Chinese mills shift back to domestic ore — which becomes more competitive as seaborne prices rise — the ton-mile support evaporates. The capesize rally is a rent on dislocation, and dislocation is not a permanent state. For shipping capital, the question is whether Simandou’s ramp-up outpaces the decline in Chinese steel intensity. The current data suggests the latter is winning.

Air Canada sells $2.5 billion stake in loyalty programme to Blackstone, other investors¶

Source: FlightGlobal

The valuation is the story. Blackstone and co-investors are paying $2.5 billion for a quarter of Aeroplan, pricing the loyalty scheme at $10 billion — a figure that likely exceeds the market capitalisation of the airline that owns it. Air Canada has spent years burning cash on aircraft, fuel and labour, yet the most valuable asset on its books is a database of points liabilities. That inversion is not an accounting quirk; it is the logical endpoint of an industry where the actual business of flying has been financialised into a customer-acquisition cost for the points-selling operation.

The deal is a sale-and-leaseback of the balance sheet itself. Air Canada converts a future stream of member fees and partner payments into immediate liquidity, handing Blackstone a claim on revenues that require no capital expenditure, no union contracts and no jet fuel. For the private equity firm, Aeroplan is a toll road: the points are already issued, the redemption obligations are already priced in, and the airline remains responsible for honouring them. The carrier gets cash to service its debt; the investor gets a yield with none of the operational risk. The contradiction sits inside the structure — the more successful Aeroplan becomes at selling points, the more seats Air Canada must give away to redeem them, and the more of that future cost now belongs to someone else's profit margin.

That Air Canada is selling its crown jewel weeks after swinging to a quarterly loss suggests the liquidity is not optional. The airline is not monetising strength but collateralising its last unencumbered asset. For the wider industry, the template is ominous: when carriers discover their loyalty programmes are worth more than their route networks, the flying becomes a subsidy for the points economy, and the next downturn will find airlines even more hollowed out.

Marshall aims to find buyer for aerospace business by year-end¶

Source: FlightGlobal

The Marshall family’s decision to offload its last remaining division by year-end marks the quiet completion of a long unwinding. The Cambridge firm, once a byword for British aerospace engineering, has been progressively shedding its identity for years; the C-130 sustainment work that remains is lucrative but captive to defence budgets and the slow decay of ageing airframes. A family firm selling its core business is not, in itself, remarkable. What is notable is the timing: the divestment is being pursued into a market where military MRO demand is robust, yet the buyer pool is thinning as private equity retreats from capital-heavy engineering assets.

The contradiction here is between the operational health of the business and its financial logic. Marshall’s sustainment contracts generate steady, predictable revenue — the kind of cash flow that should attract infrastructure-style investors. But the aerospace MRO sector is consolidating around scale, and a mid-sized player without a proprietary technology edge is increasingly a candidate for absorption rather than growth. The family’s decision to exit now, rather than invest in the capability upgrades needed to compete for next-generation platforms, suggests they have read the trajectory: the C-130 fleet is finite, and the transition to newer airlifters will favour primes with deeper balance sheets.

For the workforce and the broader UK defence industrial base, this is another step in the hollowing-out of domestic capability. The buyer will almost certainly be a larger group seeking to bolt on Marshall’s customer relationships and certifications, not to preserve its independent engineering culture. The state, which depends on this sustainment capacity for its own operational needs, will watch from the sidelines — content to let market forces determine the fate of a strategic asset.

Comac’s C919 completes first international flight¶

Source: FlightGlobal

The C919's first international hop, operated by Air China, is less a milestone in aviation engineering than a state-directed exercise in market signalling. Beijing is telling both domestic carriers and the global leasing community that the airframe can be trusted beyond the controlled environment of Chinese airspace — a necessary precondition for the programme to escape its current dependence on state-backed orders and begin amortising its enormous development costs across a wider customer base.

The timing matters. With Boeing still recovering from its production and safety crises and Airbus's delivery backlog stretching into the 2030s, there is a genuine window for a third supplier. But the C919's international certification remains the binding constraint. A single flight to a friendly destination does not constitute approval from EASA or the FAA, and without those certifications the aircraft cannot access the lucrative transatlantic and transpacific markets where the real money lies. The symbolic value of this flight is therefore aimed as much at potential buyers in the Global South — countries seeking to diversify away from the duopoly without fully aligning with either Washington or Brussels — as at Western regulators.

The deeper dynamic is inter-imperialist rivalry expressed through industrial policy. China is not merely competing on price or performance; it is building a parallel aerospace ecosystem — engines, avionics, certification standards — that could eventually fragment the current unified global market. For airlines in the developing world, this offers leverage in negotiations with Airbus and Boeing. For the duopoly, it represents a slow-burning threat to their pricing power, which has been a reliable source of super-profits for decades. The C919 will not displace them soon, but its existence already alters the terms on which they sell.

China’s Biggest Weakness on AI¶

Source: Project Syndicate

The Chinese state is caught in a bind that its own developmental model created. The court ruling against AI-driven dismissal, the suspension of robotaxi licences after the Wuhan failure, and state media’s admonitions to companies all point to the same structural problem: the party’s legitimacy rests on a promise of full employment, yet the accumulation strategy that delivered growth now demands the destruction of precisely those jobs.

Frey is right to identify the missing welfare state as the crux, but the framing undersells the depth of the trap. A robust social safety net is not merely a compensation mechanism that would smooth the transition; its absence is constitutive of how Chinese capitalism has functioned. The system has relied on the household as the ultimate shock absorber — rural land as a fallback, family savings as unemployment insurance, and a vast reserve army of migrant labour that can be expelled from the cities when demand falters. To build a genuine welfare state now would require a fiscal reorientation that threatens the very profitability of the state-owned enterprises and local government financing vehicles that anchor the political economy.

The court ruling and the Wuhan suspension are therefore not signs of a state turning against capital, but of a state trying to manage the contradiction between the social relations that sustain it and the technological imperative it has embraced. Each intervention buys time, but time is precisely what the AI race does not permit. The regime must simultaneously push automation to compete with the US while restraining it to preserve social stability. That is not a policy dilemma; it is the collision of two necessities that cannot both be satisfied. The robotaxi pause in Wuhan is the clearest evidence yet that the state will sacrifice technological momentum when the alternative is visible, organised unrest — a calculus that US capital, with its individualised risk and atomised workforce, does not face in the same form.

Global study reveals asset managers are concerned about the risks that AI introduces¶

Source: Hellenic Shipping News

The survey data here is less interesting for what it says about AI than for what it reveals about the institutional position of asset managers themselves. Sixty-two percent worried about lacking skills, 64% about implementation costs, 67% about data governance — these are not the anxieties of a sector being disrupted. They are the anxieties of a sector that has already decided to adopt the technology and is now trying to price the risks of doing so into its own operations.

Notice what is absent: no concern about what AI does to the underlying markets these managers trade. The worries are all internal — model transparency, bias, hallucinations, regulatory exposure. The technology is treated as a production input with compliance costs, not as something that might alter the competitive dynamics of the financial system itself. That is a telling blind spot. If AI genuinely improves alpha generation for some firms, it intensifies the zero-sum competition for returns that already defines asset management. The firms that lag on adoption will be punished by the market regardless of their caution; the firms that rush in will absorb the operational risks. Either way, the surplus extracted from the real economy — the shipping, iron ore, and steel movements that populate the rest of this publication — continues to be redistributed among financial intermediaries with a new technological veneer.

The 16% who admit they are unprepared for operational risks are the honest ones. The rest are expressing a confidence that the survey itself undermines. This is not a story about technological transformation. It is a story about capital absorbing a new tool into its existing circuits of accumulation, with all the familiar anxieties about who bears the cost of failure.