Presenter: Junior Maih (Senior Adviser at Norges Bank)
Co-authors: Christian Matthes (Indiana University)
Title: Behavioral Expectations in Nonlinear DSGE Models
Abstract: This paper extends the application of Cognitive Discounting (CD) beyond linear frameworks to encompass nonlinear, nonstationary, and regime-switching Dynamic Stochastic General Equilibrium (DSGE) models. We provide a robust framework for analyzing how bounded rationality influences macroeconomic dynamics and policy implications in realistic settings. Our approach leverages the RISE toolbox and introduces a behavioral expectation operator applicable to the entire DSGE system, enabling a nuanced analysis of how expectations evolve and impact economic outcomes in complex environments. Our findings indicate that results from linear models do not necessarily generalize to nonlinear contexts. Notably, forward guidance is less powerful, and the Zero Lower Bound (ZLB) can be more costly depending on the specification of cognitive discounting. Optimal policy strategies can mitigate differences between rational expectations and cognitive discounting, while higher-order approximations reveal significant asymmetries and variable- and regime-specific effects of cognitive discounting.