A rules-based, long/short process in U.S. bank stocks, built on regulatory financial data and operated within written risk limits.
U.S. banks file detailed financial reports with their regulators every quarter. Our process reads those reports, dated by when each one actually became public, and draws on loan quality and its trend, loan concentrations, funding sources, capital and earnings, together with market prices and selected public filings. The same data is read the same way for every bank.
The process ranks publicly traded banks from strongest to most fragile. It may hold positions in banks it ranks among the strongest and, when market conditions favor it, sell short banks it ranks among the most fragile. Positions are few, and they are sized so that trades stay small relative to each stock's daily trading.
Exposure, position-size, liquidity and permitted-instrument limits are set in writing and are to be checked automatically before any order reaches the broker. The Firm does not borrow to buy securities and does not trade options or leveraged products. Protective rules reduce exposure when bank-stock markets become unusually volatile.
We use artificial-intelligence tools to help read and cross-check bank filings. Investment decisions come from written rules, not from AI. We are testing, on a forward basis with each assessment recorded before outcomes are known, whether AI-assisted reading adds useful information. It has not yet been shown to, and we say so.
The Firm operates under written policies governing personal trading, artificial-intelligence governance, hypothetical performance presentation, marketing and advertising, cybersecurity, business continuity, anti-money laundering, and general compliance. Changes to any element of the investment process are governed by a documented change-control process with compliance sign-off, and the written policies are subject to an annual compliance review.