Optimal Policy with Occasionally Binding Constraints: Piecewise Linear Solution Methods

This paper develops a piecewise linear toolkit for optimal policy analysis of rational expectations models, subject to occasionally binding constraints on (multiple) instruments and other variables. Optimal minimises quadratic loss function under either commitment or discretion. The accounts the presence ‘anticipated disturbances’ model equations, allowing around scenarios forecasts that are not produced using itself (for example, judgement-based such as those often by central banks). flexibility applicability very large models is demonstrated in variety applications, including experiments version Federal Reserve Board’s FRB/US model.