Insurance asset managers backing Funding Agreement-Backed Notes
(FABNs) have traditionally relied on static, buy-and-hold allocations that cannot respond
as market conditions shift, leaving spread income unrealized and regulatory capital
suboptimally deployed. This paper develops a dynamic reoptimization framework that
formulates the FABN-backing portfolio as a linear program, maximizing statutory net
investment income net of National Association of Insurance Commissioners (NAIC)
Risk-Based Capital cost, transaction cost, and cash-flow shortfall, subject to asset-
liability duration matching, tail-risk, and issuer-concentration constraints, and paired with
an at-the-money pay-fixed interest-rate swap overlay that hedges the book’s rate
exposure so that it is compensated for credit risk rather than for interest-rate
movements.
Applied to a real institutional FABN liability and an investment-grade corporate bond
universe over a two-year window, a walk-forward backtest shows that disciplined,
regulation-aware reoptimization delivers a meaningful and recurring improvement in
cumulative net statutory income over the static buy-and-hold baseline, achieved without
breaching any duration, capital, or concentration constraint. A perfect-foresight analysis
bounds the maximum attainable value and reveals substantial additional head-room,
while a decomposition establishes that the advantage originates in rate-rotation timing
recognized through the Interest Maintenance Reserve rather than in bond selection. The
framework is delivered through a dashboard and a conversational agent that serves
strictly as an interpretability layer over the optimizer, never participating in the financial
computation.
Watch the team present this project at 02:03:10 in the session recording here.
Keywords: Portfolio Optimization, Asset-Liability Management, Dynamic Reoptimization,
Risk-Based Capital, Fixed Income, Agentic A
Faculty Advisor
Francisco Azeredo is a financial economist and decision scientist with over 15 years of experience working for leading financial and consulting firms in the Chicago area.
He is presently an Associate Clinical Professor at the University of Chicago. Previously, he was the senior vice-president at Northern Trust, working on research and development of quantitative strategies for fixed income and equities at Northern Trust Asset Management.
Prior to joining asset management, Francisco led strategic and quantitative driven projects across Northern Trust’s three major business units. His experience included leading analytics projects on foreign exchange global revenue forecast, wealth management client deposit modeling, wealth management digital client behavior modeling, treasury deposit modeling, derivatives and hedging solutions and foreign exchange high-frequency algorithmic execution cost. He also conducted research on foreign exchange algorithmic execution methods and securities lending price modeling.
His prior work experience includes credit risk, operational risk and macroeconomic modeling for stress testing and capital planning, analysis of complex financial instruments, interest rate modeling, and econometric and predictive analysis for litigation matters involving financial institutions and multinational corporations.
His research in financial economics was recognized by leading economists in the Handbook of the Equity Risk Premium, Intermediate Financial Theory, and an encyclopedia of behavior economics. He currently teaches reinforcement learning, statistical analysis, non-linear statistical modeling, and optimization methods in the MS in Applied Data Science program at the University of Chicago and has five years of prior experience teaching at Northwestern University, the University of California, Santa Barbara and Riverside. He holds a PhD in Economics from the University of California, Santa Barbara.