## Macroeconometric Forecasting and Analysis (MFA)

## Source details

**Canonical URL:** [Macroeconometric Forecasting and Analysis (MFA)](https://www.imf.org/-/media/files/icd/mfa.pdf)

## Other formats

- [Markdown version](/-/media/files/icd/mfa.pdf.md)
- [Structured JSON version](/-/media/files/icd/mfa.pdf.json)

---

### Course objectives and target audience
- Two-week course aiming to provide government and related officials a rigorous foundation in the estimation of macro-econometric models and their application for forecasting and policy analysis in central banks, ministries and public research institutions.
- By the end of the course participants should be able to:
  - Understand the underpinnings of a number of model specifications.
  - Use EViews software to apply modeling techniques to country data and replicate results from a variety of important published research papers.
  - Apply the techniques learned to country cases from their region to forecast and analyze a policy issue.
  - Bring applicable econometric tools, implemented in the EVIEWS econometric package, back home for immediate application to their work or research projects.
- Target participants: staff economists involved in developing macro-econometric models and forecasting for analysis, design and implementation of macroeconomic policy; expected to have an advanced degree in economics or equivalent experience and background in econometrics.
- Recommended prerequisite: completion of the online Macroeconomic Forecasting (MF.x) course; participants are expected to complete the EViews component of MF.x prior to the start of the course if not already experienced EViews users.

### Performance evaluation and course team
- Performance evaluation:
  - Two multiple-choice tests will be given, one at the beginning and one at the end of the course.
  - Performance in these tests will be recorded in participants’ evaluations.
- Team:
  - Charis Christofides (lead), Adina Popescu, Christian Johnson, and Mikhail Pranovich (JVI).
- Reviewers:
  - Internal: Sam Ouliaris, Sunil Sharma (RES).
  - External: Massimiliano Marcellino (Bocconi University).

### Timetable and offerings
- Lectures to be completed by end-April, 2016.
- First delivery at the JVI, May, 2016.
- China offering, August, 2016, to be based on Chinese data.
- Morocco offering, November, 2016.

### Course structure and pedagogical format
- Mixture of lectures and hands-on workshops; final project with group work and plenary presentations.
- Introductory lecture:
  - L-0: 0.5 hours — Structure of the Course; short introduction to design, main elements, and objectives.
- Final Project:
  - O-1: Workshop 10.5 hours — Participants provided (and encouraged to bring) datasets for selected countries to apply taught models to forecast inflation or another key macro variable using single equation, factor, Kalman Filter, combination.
  - O-2: Workshop 3 hours — Project presentations in plenary session.
- Quizzes and administrative presentations: O-0, 2.5 hours.

### Unit-level topics, learning activities, and references
- UNIT 1: Stationary VARs, structural VARs and their application I: short-run restrictions
  - Lecture L-1: 1.5 hours — Introduction in SVAR: identification problem; Choleski decomposition and short-run SVAR restrictions; Impulse responses.
  - Workshop W-1: 1.5 hours — Evaluating effect of monetary policy shocks in “Choleski-ordered” SVARs and SVARs with the “institutionally-implied” short-run restrictions.
  - References include Sims (1992); Bernanke and Mihov (1995); Blanchard and Perotti (2002); Canova (2007).

- UNIT 2: Modeling of non-stationary variables, forecasting with VECMs
  - Lecture L-2: 3 hours — Testing variables for integration; Testing for cointegration and estimating VECMs.
  - Workshop W-2: 3 hours — Estimating long-run macroeconomic equilibrium relationships; Forecasting with VECMs.
  - References include Johansen (1988); Hamilton (1994); Martin, Hurn and Harris (2013); Ghysels and Marcellino (2015).

- UNIT 3: Structural VARs and their application for policy analysis II: long-run and other restrictions
  - Lecture L-3: 1.5 hours — Identifying Structural VARs using long-run restrictions; Other restrictions.
  - Workshop W-3: 1.5 hours — SVAR for evaluating effects of fiscal policy; supply and demand shocks with long-run restrictions; Tri-variate SVAR with sign restrictions.
  - References include Blanchard and Quah (1989); Fry and Pagan (2011); Ouliaris et al (2015).

- UNIT 4: VAR extensions II: FAVARs
  - Lectures L-4, L-5: 3.0 hours total — Basics of factor models; Small and large scale; selection of number of factors; Estimation and forecasting with FAVAR; Extensions; Unbalanced datasets; I(1) variables; nonlinearities.
  - Workshops W-4, W-5: 3.0 hours total — Estimating FAVARs on several macro-financial datasets (monthly industrial production; quarterly GDP growth; monthly inflation). Examples from both industrial and emerging economies.
  - References include Bernanke, Boivin and Eliasz (2005); Stock and Watson (2005); Fernald, Spiegel and Swanson (2014).

- UNIT 5: Conditional forecasting with VARs in small open economies
  - Lecture L-6: 1.5 hours — Conditional forecasting using VARs; Incorporating external forecasts and scenario analysis.
  - Workshop W-6: 1.5 hours — Conditional forecasting and scenario analysis with a VAR model for a small open economy.
  - Reference: Waggoner and Zha (1999).

- UNIT 6: Estimation of and forecasting with Bayesian VARs
  - Lecture L-7: 3 hours — Introduction to Bayesian econometrics; exercise on Bayesian estimation of moments of normal distribution; Estimating BVARs with analytical Minnesota and DSGE-VAR priors; Review of empirical results on BVARs forecasting performance.
  - Workshop W-7: 3 hours — Estimating BVARs with Minnesota, Normal-Wishart priors and DSGE-VAR priors; Forecasting macroeconomic variables with BVARs.
  - References include Carriero, Clark and Marcellino (2011); Litterman (1986); Del Negro and Schorfheide (2004, 2007); Giannone, Lenza and Primiceri (2015).

- UNIT 7: State-Space Models and the Kalman Filter
  - Lecture L-8: 3 hours — State-space representation; The Kalman filter; Maximum likelihood estimation and Kalman smoothing.
  - Workshop W-8: 3 hours — Applications: estimating business condition index, forecasting the yield curve, estimating equilibrium interest rate; Output gap estimation (e.g., HP filter, multivariate filter).
  - References include Aruoba, Diebold and Scotti (2009); Diebold and Li (2006); Laubach and Williams (2003); Ghysels and Marcellino (2015).

- UNIT 8: Forecast Combination
  - Lecture L-9: 1.5 hours — Motivation for combining forecasts; Implementation issues; Methods to assign weights.
  - Workshop W-9: 1.5 hours — Application of combination techniques to forecasting of macroeconomic variables.
  - References include Clemen (1985); Stock and Watson (2004); Timmermann (2006).

- UNIT 9: Univariate and multivariate models of volatility and their application
  - Lecture L-10: 3 hours — Estimating univariate volatility models (ARCH, GARCH) and descendants (TARCH, EGARCH); Estimating multivariate volatility models; Background for the workshop: Value-at-Risk analysis.
  - Workshop W-10: 1.5 hours — Estimation of univariate and multivariate GARCH models; Forecasting with GARCH models; Application of MVGARCH to Value-at-Risk analysis; Volatility impact on first moment prediction.
  - References include Bollerslev (1986); Andersen, Bollerslev, Diebold and Labys (2003).

### Time allocation and course totals
- Lectures (L's): 21.5 hours
- Workshops (W's): 19.5 hours
- Other (O's): 16 hours
- TOTAL: 57

---


_Source: https://www.imf.org/-/media/files/icd/mfa.pdf_