Efficient Solution and Computation of Models With Occasionally Binding Constraints

Structural macroeconometric analysis and new HANK-type models with extremely high dimensionality require fast and robust methods to efficiently deal with occasionally binding constraints (OBCs), especially since major developed economies have again hit the zero lower bound on nominal interest rates. This paper shows that a linear dynamic rational expectations system with OBCs, depending on the expected duration of the constraint, can be represented in closed form. Combined with a set of simple equilibrium conditions, this can be exploited to avoid matrix inversions and simulations at runtime for significant gains in computational speed. An efficient implementation is provided in Python programming language. Benchmarking results show that for medium-scale models with an OBC, more than 150,000 state vectors can be evaluated per second. This is an improvement of more than three orders of magnitude over existing alternatives. Even state evaluations of large HANK-type models with almost 1000 endogenous variables require only 0.1 ms.