syllabus
TRANSCRIPT
ECONOMETRIC METHODS II: TIME SERIES
SYLLABUS 2014
KRISTOFFER P. NIMARK
Class time and place: Wednesdays and Fridays 9.00-11.00 (20.017)TA sessions: Fridays 12.00-13.00, (Room TBA)Office hours: Thursdays 12.00 -13.00 Office 23.408 (Wellington building)Email Address: [email protected] website: www.kris-nimark.net
OverviewThis course aims at equipping students with the tools needed to produce applied macro economicresearch. Most of the material can be found in Time Series Analysis by James D. Hamilton(Princeton University Press, 1994), Bayesian Econometrics by Gary Koop (Wiley-Interscience,2003) and Cochrane’s “text” but reading articles may also be required.
Administrative mattersGrades will be based on the final exam (60%) and 3 homework assignments (2×10% + 1×20%).
Lecture 1 Introduction.
• What is Time Series Statistics and what is it good for?• Philosophical issues related to probability and statistics
– Why do we need probabilistic models to describe the world?– Is there true randomness in nature? Does it matter?
• Course overview• Basics
– Matrix algebra– Stochastic processes
• Difference Equations• Lag operators
Readings: Appendix A4-A5 Ch. 1-2 in Hamilton.
Date: April 1, 2014.
1
2 KRISTOFFER P. NIMARK
Lecture 2-3 Stationary ARMA processes, Prediction and Forecasting.
• Basics of stochastic processes• MA, AR and White Noise Processes• Auto Covariance Function
Readings: Ch.3 in Hamilton and Ch.3-4 in Cochrane.
• Conditional expectations• Prediction and ARMA processes• Interval forecasts
Readings: Ch.4 in Hamilton, Ch.3 in Cochrane and lecture notes on the Projection Theorem.
Lecture 4 Estimation of Linear Models.
• Maximum likelihood• Least squares• Linear projections
Readings: Ch.5 and Ch.8 in Hamilton.
Lecture 5-7 Vector Auto Regressions (VARS).
• Estimation– Lag specification– Hypothesis testing
• Identification– Contemporaneous (Cholesky) restrictions– Long-run (Blanchard-Kahn) restrictions– Sign restrictions
• Variance decompositions• Impulse response functions
Readings: Ch.9-11 in Hamilton and Ch.7 in Cochrane + Articles.
Lecture 8 Principal Components and FAVARS.
• Dimension reduction• Dynamic Factor Models• Forecasting
Readings: Stock and Watson (2010) and Bernanke, Boivin and Eliasz (2005)
Lecture 9-10 Unit Roots and Cointegration.
• Unit roots• Cointegration
Readings: Ch.18-19 in Hamilton and Ch.11 in Cochrane.
Lecture 11-13 State Space Models and the Kalman Filter.
• Latent variables• State Space Models• The Kalman Filter• Numerical optimization• DSGE models• Latent Factor models• Smoothing
ECONOMETRIC METHODS II 3
Readings: Ch.5.7 and 13 in Hamilton, lecture notes on the Kalman Filter, Goffe et al (1994).
Lecture 14 Introduction to Bayesian Statistics.
• What is Bayesian statistics?• Objects of interest: Prior, posterior, marginal likelihood, posterior odds ratio• Bayesian Computation
Readings: Koop Ch. 1-2
Lecture 15 Linear Regression Model with Natural Priors.
• Combining natural priors with sample information• Model Comparison• Monte Carlo integration
Readings: Koop Ch. 2-3
Lecture 16 Simulating posterior distributions.
• Gibbs Sampling• Inequality constraints
Readings: Koop Ch. 4
Lecture 17 Non-linear models.
• Metropolis-Hastings Algorithm• Model fit
Readings: Koop Ch. 5
Lecture 18-19 Bayesian estimation of State Space Models.
• Local level/Unobserved component model• DSGE models• Prior Predictive Analysis• Bayesian Model Averaging
Readings: Koop Ch. 8 and 11 and An and Schorfheide (2007)
Lecture 20 Course review.
References
[1] Cochrane, John, 2005, Time Series for Macroeconomics and Finance,http://faculty.chicagobooth.edu/john.cochrane/research/Papers/Time Series Book.pdf
[2] Hamilton, James D., 1994, Time Series Analysis, Princeton University Press.[3] Koop, Gary, 2003, Bayesian Economtrics, Wiley-Interscience.