2026-08-28 Observatory briefing

A Great Rebalancing Is Coming

Ref: O1 Item-ID: article-a4f1e51f90 Source: Foreign Affairs

Pettis's historical ledger is damning: every major imbalance episode except one ended with the deficit countries absorbing the losses. The 1950s exception worked because capital controls and productive investment meant debts generated their own repayment capacity. The 1920s, the 1970s, the 1980s all resolved through crisis, with the weakest bearing the costs. His point about the current moment is that the usual mechanics are inverted. China and Germany hold the surpluses, but the United States holds the dollar. That asymmetry matters more than any accounting identity.

Washington can inflate away its obligations or weaponise tariffs, forcing adjustment onto exporters. Beijing's growth model depends on suppressing household income to feed exports and investment, so a genuine rebalancing would demand redistribution from capital to labour within China itself. The Communist Party's legitimacy rests on the current arrangement, which is why its diagnosis blames American fiscal profligacy. The American diagnosis blames foreign currency manipulation, which conveniently ignores domestic wage stagnation and financialised growth. Each side's remedy protects its own ruling class.

China is most exposed because its surplus is structural, not cyclical. Europe possesses latent power but no mechanism to use it collectively. The adjustment will come through crisis rather than coordination, and the costs will fall on those with least capacity to resist: workers in the surplus economies whose wages were suppressed to generate those surpluses, and debtors in the deficit economies. They will pay twice, first through stagnant incomes, then through the consequences of a crisis their rulers refused to prevent.

Hormuz transits still limited even as Iran, Oman near deal; carriers eye return to Suez route

Ref: O2 Item-ID: article-24a729c03f Source: Hellenic Shipping News

The IRGC’s warning to the Indian tanker Haana, which turned around after attempting the southern Oman Corridor, captures the precise texture of the standoff. Five vessels transited on 25 August, all under the Iranian Unilateral Scheme, four of them shadow vessels. Iran and Oman may be close to an agreement, but the Islamic Republic’s conditionality — reopening the Strait if the US stops obstructing and accepts its terms — means the negotiations are less a diplomatic breakthrough than a restatement of the underlying leverage. Iran controls the waterway; the question is what price it extracts for reopening it.

The container shipping story runs on a different clock. Less than 2% of global container capacity passes through Hormuz annually, so the closure barely touches box rates directly. The pressure comes through bunker fuel prices, which climb as tanker rerouting tightens supply. Meanwhile the Suez return looks increasingly real: MSC has announced a partial return, and the Gemini Cooperation and Ocean Alliance are testing partial operations. Lars Jensen’s end-2026 normalisation estimate carries the caveat that carriers will keep some services around Africa to absorb the excess capacity a full Suez reopening would release.

That excess capacity is the material contradiction. Rates are at their highest since 2024, partly because the Cape route absorbs tonnage. A full return to Suez would slash voyage times, flood the market with hulls, and send rates down. The carriers know this. Their partial returns are a hedge: enough presence to test the security environment, not so much that they surrender the scarcity premium. The Houthis’ continued targeting of Saudi tankers at Bab el-Mandeb keeps that premium alive. For the chemical industry, the stakes are concrete — polymers and TiO2 move in containers, so a rate collapse would ease input costs, but only if the security picture holds.

Port congestion is getting worse in container shipping: who is responsible?

Ref: O3 Item-ID: article-80e57faabb Source: Hellenic Shipping News

Maersk’s Vincent Clerc blames fifteen years of underinvestment for the congestion now spreading across Europe, South America, West Africa and the Middle East. Drewry’s counter-analysis is more useful, not because it exonerates terminal operators but because it specifies the mechanism. Between 2019 and 2026, nine major ports expanded capacity by 21% while volumes grew 28%. That gap is the rational behaviour of private capital seeking returns. A terminal at 90% berth utilisation takes a week to absorb a one-day shock; at 75% it takes two days. The difference between those operating points is the difference between a competitive and uncompetitive return on capital. Buffer capacity is, by definition, capital that sits idle until a disruption justifies it, and no operator will carry that cost alone when shareholders reward utilisation.

The carriers are not innocent parties demanding better infrastructure. Their own yield-maximising tactics, blank sailings, ad hoc calls, extra loaders, concentrate arrivals into peaks that overwhelm whatever berth capacity exists. The system is built to optimise cost, not resilience, and every actor in it behaves accordingly. Waiting times have nearly doubled since 2019, and the share of port time spent queuing rather than working has grown, so the productivity loss is real and measurable.

What Drewry stops short of saying is that this is a collective action problem with no internal solution. Each operator optimises its own balance sheet; the aggregate result is a supply chain that cannot absorb shocks. The state, which might socialise the cost of spare capacity, appears nowhere in this account except as a concession delay in Santos. The congestion is not a malfunction of the market but its logical output, and the only actors capable of correcting it have no incentive to do so.

Leave Europe to the Europeans

Ref: O4 Item-ID: article-faba7e1fd6 Source: Foreign Affairs

The 1987 newspaper advertisement is doing a lot of work for Kavanagh and Logan. Trump's thirty-eight-year-old complaint about freeloading allies is presented as the intellectual seed of a coherent strategic doctrine, when in fact the authors are asking the reader to accept that a real-estate mogul's grievance has matured into a sound basis for dismantling the post-war order. Their case for a U.S. drawdown rests on a fiscal argument that is difficult to dispute: a state $40 trillion in debt spending $100 billion annually to garrison a continent whose combined economies dwarf its own is a misallocation of resources. The numbers are stark, and the authors wield them effectively.

Yet the analysis collapses exactly where it needs to stand firm. The authors treat European dependence as a moral failing, a case of allies "shirking their obligations" and capitalising on a "peace dividend" while Washington dutifully paid the bill. This framing inverts the actual relationship. The U.S. presence was never charity; it was the institutional form of American hegemony over Western Europe, a mechanism for ensuring that the continent's reconstruction proceeded on Washington's terms and that its militaries remained subordinate to NATO command structures. European under-spending was a feature the Americans cultivated, guaranteeing permanent dependence and permanent U.S. primacy within the alliance.

The contradiction in the authors' position emerges when they argue that a withdrawal will "force" Europeans to build their own forces. This assumes European states are simply waiting for permission to rearm, that their military weakness is a choice rather than a structural condition of the post-war settlement. The authors never ask whether a Europe forced to fund its own defence might also decide to chart its own strategic course, one less aligned with Washington's priorities in Asia or the Middle East. The posture review is framed as a technical exercise in burden-shifting, but the material question is whether the U.S. can cede financial responsibility without losing political control. Kavanagh and Logan assume it can. The history of the alliance suggests otherwise.

Is China Really a Champion of the Global South?

Ref: O5 Item-ID: article-e710fc5b15 Source: Project Syndicate

Adekeye Adebajo writes from Pretoria with a question that answers itself before the paywall cuts in. China’s claim to lead the Global South, he suggests, is less about overturning the international order than about inheriting it. The framing is useful because it refuses the romance of a Beijing alternative. China’s courtship of the Group of 77, its insistence on its own developing-country credentials, is a bid for position within a system the United States built and is now vacating, not a project to replace that system with something else.

The material basis for this is straightforward. Overaccumulation at home has pushed Chinese capital outward for two decades, and the Belt and Road Initiative is the most coherent expression of that pressure. But the political form that expansion takes is conservative. China needs the existing institutions of trade, finance, and sovereignty to make its outward investment legible and secure. It wants a larger share of the World Bank and the IMF, not their abolition. It wants the dollar system to remain stable enough that its own dollar holdings retain value, even as it hedges with bilateral swap lines and alternative payment rails.

Adebajo’s point about consolidation rather than transformation lands because it names the actual trajectory. The Global South is being offered a different manager of the same shop floor. Whether that manager proves more attentive to the workers than the last one is an open question, but the architecture of the workplace itself is not on the table. For states in Africa and Latin America, the choice is between two hegemons, one declining and one ascending, both operating within the same basic grammar of sovereignty, debt, and export-led growth. The G77’s embrace of China is a rational hedging strategy, but it is not a programme of emancipation.

Air NZ says engine issues 'substantially behind us' as groundings dwindle

Ref: O6 Item-ID: article-8a5cd72807 Source: FlightGlobal

Greg Foran’s declaration that engine issues are “substantially behind us” sits awkwardly beside the admission that Air New Zealand is “renegotiating new compensation terms” with Rolls-Royce. If the problem were truly closed, there would be nothing left to negotiate. The groundings have dwindled from five 787-9s to one, but the financial settlement is still being fought over, which suggests the operational recovery is real while the commercial damage remains unresolved.

The compensation talks reveal how the cost of the Trent 1000’s failures is being pushed back up the supply chain. Air New Zealand has already received payments for earlier groundings, but each new round of inspections and repairs has turned its 787 fleet into a source of recurring expense rather than reliable revenue. The aircraft, a fixed asset on the balance sheet, keeps failing to perform as one. The renegotiation is an attempt to convert that dead capital into cash, shifting the burden of unpredictability onto Rolls-Royce, which is already absorbing the reputational and financial weight of its engine’s durability problems.

For Rolls-Royce, the payout is a direct hit to margins, a tax on production capacity that promised more reliability than it could deliver. For the airline, the groundings have meant a smaller effective fleet than its accounts suggest. The wider pattern is that high-bypass engines have created a maintenance burden no warranty can fully indemnify. The risk has been repackaged and sold back to the manufacturer, but it has not disappeared.

American plans seven new transatlantic routes

Ref: O7 Item-ID: article-3891a15c55 Source: FlightGlobal

American’s summer 2027 schedule leans on the A321XLR, the narrowbody whose range was sold to airlines as a way to open thin transatlantic city pairs without the risk of a widebody’s fixed costs. Seven new routes announced in the same week as United’s own international expansion suggests the US majors are not merely chasing demand but positioning against each other on the North Atlantic, the world’s most profitable long-haul corridor. The XLR lets them do this with less capital committed per seat, which is precisely the point: capacity can be added in increments small enough to test a market without betting the aircraft’s utilisation on it.

The timing matters. These schedules are being locked in roughly eighteen months ahead, which means American is committing to aircraft utilisation, crew bases and airport slots on the basis of demand forecasts made in the middle of a fragile recovery. The transatlantic has been the one bright spot for US carriers, with premium cabins full and corporate travel returning faster than domestic. But the XLR’s economics depend on high load factors in a single-class configuration that leans heavily on premium economy. If the expected summer surge does not materialise, the aircraft can be redeployed, but the slots and the marketing spend cannot be recovered so easily.

United’s announcement a week earlier frames this as a race. Both carriers are chasing the same finite pool of European secondary cities, and the first mover gets the frequency advantage. For passengers this looks like choice; for the airlines it is a defensive scramble to protect share on a route where yields are already under pressure from Gulf carriers and the joint ventures that tie European and US airlines together. The XLR is a tool for this competition, not a response to it.

Discover Airlines to enter UK market with three new routes

Ref: O8 Item-ID: article-92fd78ea55 Source: FlightGlobal

Lufthansa Group's leisure arm will fly A320s from Bristol, Glasgow and Inverness from 2027, three routes that look modest but mark the first sustained attempt by a continental carrier to feed off UK regional airports since the pandemic rearranged the market. The destination choices carry the real weight. Bristol and Glasgow have long since lost their low-cost monopolies, yet easyJet and Ryanair still set the short-haul terms. Inverness is the outlier, a thin route whose economics depend on inbound tourism channelled through Frankfurt and Munich rather than head-to-head price competition.

The expansion reflects a quiet reorganisation inside the group. The premium long-haul core has struggled with cost inflation while leisure subsidiaries have become the reliable earners. Discover itself was assembled from Eurowings' long-haul remnants, and its UK push suggests the group sees more headroom in European point-to-point traffic than in defending its hub-and-spoke territory against Gulf carriers. That is an admission that the intercontinental premium model is no longer the growth engine it once was.

For regional airports, the announcement offers leverage. They have watched capacity migrate to Heathrow and Manchester for years, and a new international operator with deep pockets strengthens their hand in negotiations with the low-cost carriers that currently dictate terms. Whether the routes survive their first winter is another matter. Leisure services into Scotland and the West Country are seasonal by nature, and Discover must sustain load factors through the shoulder months or the aircraft will be pulled back to sun destinations with surer margins. The group's patience with underperformers has been limited, and UK air passenger duty does nothing to tilt the arithmetic in favour of thin regional services.

Does generative AI save time at work?

Ref: O9 Item-ID: article-a20166acfa Source: FRED Blog

The Federal Reserve Bank of St. Louis's new survey data shows generative AI adoption creeping upward: 39.2% of employed adults used the tools for work in the past week by Q2 2026, up from 28.2% in Q3 2024, with assisted work hours rising from 4.1% to 6.3% and self-reported time savings from 1.6% to 2.2%. The numbers rest on workers' subjective estimates of hours "saved," and the Fed's framing treats the trend as a slow-burning productivity story that may eventually surface in aggregate statistics.

What the tracker cannot capture is where those saved hours actually go. The survey measures assistance, not appropriation. A worker who drafts a report in half the time has not necessarily gained half an hour of leisure or even half an hour of additional output; under managerial pressure, the time freed tends to be refilled with more tasks, more iterations, more polish. The productivity gain is real at the level of the individual task, but its conversion into either higher wages or shorter working days depends entirely on the balance of power at the point of production. Nothing in the questionnaire asks who pockets the difference.

The Fed's interest in this data is itself a policy signal. Central banks track productivity because it is the lever that reconciles wage growth with price stability without squeezing profits. A technology that promises to raise output per hour without raising hourly compensation is, from the monetary authority's vantage, nearly ideal. The survey's steady upward slope offers reassurance that the productivity dividend may arrive without the industrial conflict that historically accompanied its predecessors. Whether workers experience that dividend as time regained or merely as the same wage for denser work is a question the tracker is not designed to answer.

The Economist's Embarrassing Proxy War Over AI Policy

Ref: O10 Item-ID: article-bc0d130745 Source: Project Syndicate

The Economist’s decision to smear Daron Acemoglu through unnamed critics interviewed over drinks suggests the magazine’s editors could not answer his arguments on the merits. Acemoglu’s sin is not poor scholarship but a refusal to treat AI’s trajectory as a fait accompli. His insistence that technological change can be steered through public policy threatens the comfortable assumption that the current wave of automation is a natural, inevitable force to which societies must simply adapt.

The substance of the disagreement matters more than the magazine’s manners. Acemoglu’s work on institutions and prosperity rests on a claim that political choices shape economic outcomes. If that holds for long-run development, it holds for AI too. The Economist’s position, stripped of its sneering, is that AI’s benefits are so vast and imminent that any attempt to direct its development through public policy is a costly indulgence. This is a bet on the present distribution of power in the tech sector, dressed up as technical optimism. The magazine’s editors would never apply the same fatalism to trade policy or labour law, where they happily advocate for state intervention to shape market outcomes.

The resort to ad hominem exposes the anxiety beneath the editorial bravado. Acemoglu’s framework, taken seriously, would open questions about who owns the productivity gains from AI and who decides how they are distributed. Those questions are uncomfortable for a publication whose readership includes the asset managers and tech executives who stand to capture those gains. The Economist is defending a particular allocation of the surplus, not economic science.

Buried in Meta's $18B settlement is a legal pass on kids' data

Ref: O11 Item-ID: article-979a8de073 Source: TechCrunch

The state attorneys general who negotiated Meta’s $18 billion settlement have traded away their most potent future weapon against the company on the very issue the case was meant to address. In exchange for Meta building an age-assurance model, the AGs have agreed “fully, finally, and forever” not to bring COPPA claims related to the retention and use of children’s data for that purpose. The guardrails are real on paper: no ad targeting, no marketing, no algorithmic optimisation using under-13 data, and an independent auditor will watch. But the settlement’s logic inverts the usual privacy framework. COPPA exists to minimise data collection from children because the incentive to exploit that data is structural, not behavioural. Meta’s business model monetises behavioural insight at scale; the carve-out asks it to build a system that identifies children’s behaviour signals and then isolates that understanding from every other system in the company. That is an organisational challenge as much as a technical one, and the settlement’s own vagueness on what data will be retained, for how long, and how the model will evolve concedes as much.

The legal pass matters less for what it permits than for what it signals. Future state enforcement against Meta over children’s data will now have to litigate whether a given use fell within the settlement’s terms, a dispute Meta can tie up for years. Peter Jackson’s phrase, “disincentivise future enforcement actions,” is the polite way of saying the AGs have handed Meta a jurisdictional shield. The broader AI industry pattern is visible here: every age-assurance, safety, or compliance system requires deeper data access to function, and each such system becomes a new legal justification for retaining data that would otherwise have to be deleted. The auditor will verify compliance with the letter of the agreement, but the agreement itself is the product of a negotiation where the company’s need for data was treated as a given rather than a problem to be solved.