Causal Multi-Equation Modeling, Scenario Forecasting & Error Analysis

Catalog description

This course turns its attention to causal and non-causal multi-equation modeling and forecasting frameworks. We will use computer programming to build such models and use them to generate forecasts. We study how causality and data structures provide guidance for the structure of causal models. Additional topics include but are not limited to: consensus forecasting, subjective forecasting, intervention modeling, scenario forecasting, the Lucas Critique, in-sample and out-of-sample methods, forecast-error measurement, and combination forecasting. The Capstone modeling project consists of a multi-frequency, multi-platform, and multiple-model database and forecast system that creates various forecasts, conducts forecast-error analysis, and builds combination forecasts. The Capstone includes PowerPoint presentations and students build a professional-style written report.