Working Paper: NBER ID: w20115
Authors: Drew D. Creal; Jing Cynthia Wu
Abstract: We develop new procedures for maximum likelihood estimation of affine term structure models with spanned or unspanned stochastic volatility. Our approach uses linear regression to reduce the dimension of the numerical optimization problem yet it produces the same estimator as maximizing the likelihood. It improves the numerical behavior of estimation by eliminating parameters from the objective function that cause problems for conventional methods. We find that spanned models capture the cross-section of yields well but not volatility while unspanned models fit volatility at the expense of fitting the cross-section.
Keywords: Affine Term Structure Models; Stochastic Volatility; Maximum Likelihood Estimation
JEL Codes: C13; E43; G12
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| spanned models (C23) | yield curves (E43) |
| spanned models (C23) | volatility (E32) |
| unspanned models (C29) | volatility (E32) |
| unspanned models (C29) | yield curves (E43) |
| estimation method improvements (C51) | stable parameter estimates (C51) |
| concentrated likelihood function (C25) | estimation efficiency (C51) |
| concentrated likelihood function (C25) | stability (C62) |
| estimation procedures (C80) | economic interpretation of model outputs (C51) |