Marcos Escobar-Anel
Title: Portfolio Optimization in Affine GARCH models
Date: Friday, May 22nd 2026
Time: 1:30PM (PDT)
Location: ASB 10900
Abstract: Affine GARCH models, pioneered by Heston and Nandi (2000), combine the estimation advantages of discrete-time GARCH with the closed-form tractability usually reserved for continuous-time stochastic-volatility models. This talk surveys a half-decade-long research program showing that, once the self-financing condition is approximated in a way that preserves the affine structure, the HJB recursion closes in exponential-affine form and analytical portfolio solutions become available across the entire family. I present the core method, anchor it with two results: a univariate CRRA allocation and its multi-asset extension under a parsimonious CAPM-style specification. I also tour four directions in which the framework has been pushed: richer innovation dynamics, alternative preferences (HARA, mean鈥搗ariance, CPT), additional instruments (derivatives, bonds, commodities, VIX), and Bayesian estimation.