Model of the United States economy with learning MUSEL

The model presented here is an estimated medium-scale model for the United States (US) economy developed to forecast and analyse policy issues for the US. The model is specified to track the deviation of the medium- run developments from the balanced-growth-path via an estimated CES production function for the private sector, where factor augmenting technical progress is not constrained to evolve at a constant rate. The short-run deviations from the medium run are estimated based on three optimising private sector decision making units: firms, trade unions and households. We assume agents optimise under limited-information model-consistent learning, where each agent knows the parameters related to his/her optimization problem. Under this learning approach the effect of a monetary policy shock on output and inflation is more muted but persistent than under rational expectations, but both specifications are broadly comparable to other US macro models. Using the learning version, we .find stronger expansionary effects of an increase in government expenditure during periods of downturns compared to booms.