Analysis on risk, global economics, geopolitics, and banking, written from a defensive-realist perspective and always carried through to what it means for a decision.
Tanker traffic through the Strait of Hormuz collapsed in March and has not recovered. Brent went to $117 and then gave most of it back. Both of those things are still true, which is the problem.
Bank credit to nonbank financial institutions has grown at 22.7% a year and passed a trillion dollars. Since December 2024 the Call Report has broken it into five categories, and almost nobody is reading it institution by institution.
The OCC and the FDIC have now narrowed what an examiner may write up. It is the second rule this month the Federal Reserve declined to join, and 698 banks holding $4.7 trillion sit outside both.
An arbitral tribunal ordered Niger not to sell the uranium. Roughly a thousand tonnes went out through Burkina Faso to the port of Lomé anyway. The enforcement point was never the tribunal.
Consolidation is usually written as a story about scale. The measurable consequence is narrower and more awkward: one bank in ten can no longer assemble a cohort of its own size in its own state.
Tioga-Franklin filed a Call Report seven weeks before the FDIC closed it. Read on that filing alone, its liquidity was the strongest thing about it, which is the part worth sitting with.
The FDIC would hand the industry about $4 billion a year. The part still worth reading in five years is a threshold, which moves 76 banks onto a different pricing method and is now indexed to inflation.
A repo facility exists so that Japan can raise dollars without selling US Treasuries. Japan let $123bn of them run off anyway, and borrowed almost nothing. Both of those numbers are published.
Structural realism is the most useful default frame for reading state behaviour, and it is wrong often enough that treating it as a theory of everything will cost you money.
Country risk that stays qualitative can't enter a model, and country risk that pretends to be precise can't be trusted. The useful path runs between them.
Not because the funds are competing for the same borrowers, though they are, but because of which borrowers they take, and because the channel everyone warns about turns out to belong to a different tier of the industry entirely.
Analysed as a commodity it makes little sense, and as an inflation hedge it disappoints. Read as a claim about the reserve system, its behaviour is much less mysterious.
Tariffs on a distant competitor tax an import. Tariffs on a partner whose supply chains cross the border repeatedly tax your own producers, several times, and the trade statistics will not agree about by how much.
The bond vigilante story has the mechanism backwards. Size does not let a manager punish a government; it forces the manager to keep buying, and the discipline comes from somewhere else entirely.
A hard peg to the euro removes most of the monetary tools a central bank is assumed to have. Understanding what is left explains where West African credit risk actually sits.
A defensive-realist read of the continent's position: the structural incentives that will outlast whatever is in the headlines, and the one variable that decides the rest.
Most Western coverage of the post-Soviet space is a summary of a summary. The primary documents are public, in the original language, and say different things.
Screening a counterparty against a list answers a narrow question. The exposure that causes trouble usually sits one or two ownership layers behind a name that screens clean.
It is a price, set by people with positions, constraints and mandates. Reading it as the market's collective prediction attributes to it an intelligence it does not have.
Every ratio conversation with a board or an examiner turns into a peer conversation. Whoever defines the peer group has already decided how it ends.
A liquidity ratio tells you how much. Institutions fail on when. The gap that matters is on a specific day, against obligations that do not reschedule themselves.
Counting hardware is the easy half and the less important one. Deterrence fails when the other side stops believing you will use what you plainly have.
A forward stress test can only ever tell you that the scenario you picked is survivable. Start from the loss that breaks you and work backward, and you find out what you are actually exposed to.
The yield tells you what the market charges. The holder base tells you who is willing to be exposed, and that is a considerably more interesting question.
The supervisory stress test projects nine quarters because of the question it was built to answer. For a community bank's own book, that window closes before the losses arrive.
Institutions rarely fail because everything deteriorated at once. They fail because of one thing they had too much of, correlated in a way nobody had measured.
Most community banks file the Call Report as a compliance chore and never read it back. It is the most complete description of the institution that exists, and you have already written it.
States lie about their alignments constantly and buy weapons honestly. Procurement is a decade-long dependency, chosen deliberately, and it is published.