Working Paper Revision

Impulse Response Functions for Self-Exciting Nonlinear Models

Abstract: We calculate impulse response functions from regime-switching models where the driving variable can respond to the shock. Our focus is on nonlinear vector autoregressions with a variety of specifications for the transition function used throughout the literature. Using Monte Carlo simulations with different misspecifications, we identify the conditions under which impulse response function estimates exhibit significant bias. Furthermore, we extend the concept of model-average impulse responses to this nonlinear context and demonstrate their robustness to model misspecification. Applying these methodologies to the empirical estimation of regime-dependent fiscal multipliers, we find that the multipliers are generally less than one, with small differences observed across varying states of economic slack.

Keywords: generalized impulse response functions; threshold models; regime switching models; model averaging;

JEL Classification: C22; C24; E62;

https://doi.org/10.20955/wp.2023.021

Bibliographic Information

Provider: Federal Reserve Bank of St. Louis

Part of Series: Working Papers

Publication Date: 2023-08-29

Number: 2023-021

Related Works

• Working Paper Revision (2023-08-29) : You are here.
• Working Paper Original (2023-08-29) : Impulse Response Functions for Self-Exciting Nonlinear Models