Description

This course is a foundational time series course for doctoral students in Economics. We will begin by covering the foundations of univariate linear time series models and will then discuss their estimation and use for forecasting. We will subsequently move to multivariate time series models (Vector Autoregressions) and methods for dealing with cointegration. Along the way, we will also cover models for heteroskedasticity, state space models and the Kalman filter as well as Markov Chain Monte Carlo methods. Extensive use of applications and examples will be made using financial data and the R statistical computing environment. Homework assignments will be designed to apply the methods learned in lecture to actual data - i.e. computing will be an essential component of the assignments.

Programming

We will use the R programming language to work with data and to solve problems. R is a free software environment used by academic and professional researchers in statistics. Learning R will be a major benefit to you in your graduate studies and in your post-graduation career.

Course Materials

There is no required textbook for the course. I will make my lecture notes available on this website. However, my notes will closely follow the textbook Time Series Analysis by James D. Hamilton. Any serious econometrician will own a copy of this text. An additional excellent time series reference with financial applications is Analysis of Financial Time Series, by Ruey S. Tsay.

Assignments

There will be four assignments, due every second Thursday, starting in the third week of the quarter: 18 April, 2 May, 23 May and 6 June. Late assignments will not be accepted and extensions will not be granted.

Exams

There will be two exams:

• Miderm: Thursday, 9 May 2019, in class.
• Final: Monday, 10 June 2019, Noon - 3:00 p.m.

There will be no make-up exams - it is your responsibility to arrange your schedule to be present for the exams.

Grading

Assignments: 35%; Mid-Term Exam: 30%; Final Exam: 35%. I will curve grades only at the end of the semester if the distribution is low enough and/or spread out enough. It is important for me to emphasize that curving will never hurt your grade - it will only work to your advantage.

Topics