Kaiyuan (Kai) Chen, FRM

Why ALMFES Exists

The industry built better models. It could not build better decisions. ALM remained largely a compliance tool — and in many institutions, still does. The SVB failure in 2023 sounded the alarm.

After seven years examining ALM practices across community and regional institutions, I had seen this pattern up close. Treasury teams had detailed EVE/NII sensitivity reports. What they lacked was a framework to connect those numbers to concrete decisions — on capital allocation, hedge selection, deposit pricing, funding mix.

I built that framework, initially to explain the risk better — a multi-step attribution sequence that traces how behavioral assumptions, hedge positioning, and funding structure compound through the balance sheet into capital outcomes. I used Silicon Valley Bank's public data to illustrate it.

What I had not anticipated: the same framework that exposes and explains risk also supports decisions. It tells Treasury not just what the risk is — but what to do about it. That realization is what gave rise to ALMFES.

How the Research Has Evolved

Capital Efficiency Case Studies

After SVB, I also turned toward institutions that got it right — to understand what capital-efficient balance sheet management actually looks like in practice. Case studies are drawn from publicly available disclosures across institutions, rate cycles, and regulatory environments — examining what disciplined balance sheet engineering actually looks like in practice.

ALMFES is building a growing library of these studies. Have a case worth studying? We'd like to hear it.

AI Governance for Treasury — "Glass Box" Framework

As the capital efficiency research developed, a parallel problem became impossible to ignore: AI tools for ALM processes were advancing faster than the governance frameworks needed to deploy them safely.

This led to the "Glass Box" framework — a phased, explainable approach to integrating AI into ALM processes, aligned with SR 11-7 and NIST AI RMF from inception. The core principle: AI augments validated baselines, it does not replace them.

Weekly Research Series

In parallel, I publish a weekly "AI in ALM" series on LinkedIn — tracking what is actually happening at the intersection of AI adoption and balance sheet management, and sharing it with a global audience of Treasury and risk professionals.

What We Do

ALMFES is an open research laboratory at the intersection of financial engineering, regulatory reality, and strategic decision-making. All research is based exclusively on publicly available data and disclosures. Our work is shaped by contributions from fellow practitioners — including veteran ALM professionals who bring insights and hard-won institutional experience to the research.

The Mission

Our mission is to engineer ALM into the core of how a bank makes decisions about risk, capital, and growth.

About Kai

Learn more about Kai's background and research on LinkedIn.