Sources: Federal Reserve, Financial Stability Report, November 2026; Jonathan Hall, Principals and Agents

Two new publications, different observation clocks

Publication and evidence dates

01Fed report: Nov 13
02Fed data: Oct 28
03Hall speech: Nov 11
Publication and source-cutoff dates. [1,2]

Sources: [1] Federal Reserve, Financial Stability Report, November 2026 · Released November 13, 2026; generally information through October 28; [2] Jonathan Hall, Principals and Agents · November 11, 2026 speech

WASHINGTON

The Federal Reserve’s Financial Stability Report, released November 13, says hedge-fund leverage reached its highest level since 2013 and remained concentrated in the largest funds. It describes the banking system as resilient, while noting substantial fair-value losses on fixed-rate assets. Corporate credit spreads remained near the lower end of their historical range. The assessment generally reflects information available through October 28, not market prices at this edition’s cutoff. It identifies vulnerabilities rather than predicting the date of a crisis. [1]

LONDON

In a November 11 speech, Bank of England Financial Policy Committee member Jonathan Hall examined artificial-intelligence agents through the problem of delegated authority. He discussed access controls, monitoring and constraints intended to keep agents accountable to their principals. His argument concerns prospective governance risks as capabilities expand; it is not an announcement that an AI system caused a current financial failure. The speech and the Fed report are distinct official contributions, not a joint regulatory decision. [2]

ACTA DIURNA’s interpretation

The connection is exposure that can be difficult to observe from a headline result. A leveraged position can look profitable until its financing terms change. An automated instruction can look efficient until it uses authority in an unexpected way. These are different risks and need different evidence, but both challenge the assumption that a quiet period proves that controls are adequate. A useful assessment asks what can create a binding obligation, who can see it, and who has the power to stop it.

The missing map

Neither publication supplies a complete map of every institution’s counterparties, collateral and automated permissions. The paper therefore does not rank individual firms by danger or identify a current bank run. The reporting question is narrower: what would turn an adverse movement into a forced decision? Potential links include collateral calls, expiring financing and operational instructions that continue after their assumptions have failed. Each link needs evidence before it becomes a claim about an institution. Naming a mechanism is not proof that it is currently under strain.

Two clocks for resilience

Financial and operational adjustments can run at different speeds. A portfolio review may be scheduled monthly while its collateral obligations change daily; an automated workflow may act before a person has inspected the underlying input. Those are illustrative mismatches, not measured conditions in a named firm. Their significance is that a risk policy must specify both an acceptable exposure and a response time. A control that eventually detects an error may still fail to prevent the obligation that makes the error costly.

Sources & methodology

Bracketed numbers refer to the sources below. Analysis is original editorial interpretation, not a personalized recommendation.

  1. Federal Reserve, Financial Stability Report, November 2026 · Released November 13, 2026; generally information through October 28
  2. Jonathan Hall, Principals and Agents · November 11, 2026 speech