Time Series Analysis, Predicting and Forecasting

Time series variables (e.g., presidential approval, public mood liberalism, GDP, inflation, education level) are extremely common in the social sciences. However, due to certain properties, these series cannot always be handled using standard regression approaches. This course serves as an introduction to the world of time series analysis and also aims to strenghten research studies with dynamic data analysis. 

Course level
Master, PhD candidates and professionals from all disciplines
Session 1 and Session 2
11 January to 18 January 2020
Co-ordinating lecturersDr Francisco Blasques
Form(s) of instructionInteractive seminars and workgroups
Form(s) of assessmentTwo practical assignments
ECTS3 credits
Contact hours30 contact hours 
Tuition fee

€ 800 - International students and staff

€ 500 - VU students and staff

Students or professionals with an interest in analysing time-series data, dynamic policy analysis, prediction and forecasting. This course will appeal to professionals seeking to gain knowledge of time-series data analysis, as well as PhD and master’s students in economics, finance, business, marketing, sociology, and other social sciences interested in quantitative methods and seeking to strengthen their research/studies with dynamic data analysis.  If you have doubts about your eligibility for the course, please contact us: [email protected]
This course presumes that participants are familiar with introductory econometrics, and basic probability and statistics. Linear regression analysis and statistical tests, such as the t-test, should be well understood.

This course covers the following aspects of time-series analysis and econometrics:

Basic properties of time series 

What is a time series? Examples and definition. 
Trend and seasonal components 
Stationarity concepts 
Autocorrelation function (ACF)

Estimation, specification and analysis of ARMA models

Estimation of ARMA parameters
Probabilistic analysis of ARMA models
Forecasting with ARMA models
impulse response functions for ARMA models

Distributed lags and error correction

ADL  models and Granger causality
Dynamic multipliers
Forecasting ADL models in triangular systems
Impulse response functions in triangular systems
Estimation and model specification

Unit roots, integration and cointegration

Spurious regression            
Non-stationary time-series
Integration order of time-series
Unit roots and cointegration
Engle \& Granger error correction model
The 2-step estimation procedure
Modelling strategies for cointegrated data

Performance of participants will be assessed using two practical assignments. The assignments require participants to work with real data and solve real-world problems in time-series analysis and econometrics.

By the end of this course, participants will:

•    Know how to model, predict and forecast time-series data;
•    Understand how to implement optimal dynamic modelling strategies;
•    Have experience in analysing and extracting relevant information from dynamic time-series models;
•    Have solved practical real-world time-series problems faced by companies, governments, central banks and research institutes devoted to data analysis.

Lecture notes, lecture slides, exercise book and data sets will be provided by the teacher at the beginning of the course.

Dr Francisco Blasques

Francisco Blasques is an Associate Professor in Econometrics at the VU Amsterdam. He completed his PhD in Econometrics in 2011 at the University of Maastricht. Since then, he has been teaching time-series analysis and dynamic modelling to both Bachelor and Master econometrics students at the VU. Francisco has been nominated multiple times for the Best Lecturer Award at the VU University Amsterdam. His research focuses on advanced time-series econometric methods and covers both the theoretical and practical aspects of statistical and econometric analysis.