Applied Time Series Analysis for Volatility Modeling

Responsible person: Assoc Prof Cristina Pizarro-Irizar, PhD
Language: English
Workload: 16 h attendance time
44 h self-study time
Credits: 2
Schedule: 27^th - 30^th April 2026
9:00 AM - 1:00 PM

For further information, please have a look at the detailed schedule.
Type of Examination: Assessment of own project
PDF report + Datasets + R script file(s)
Location: Platz der Göttinger Sieben 5
MZG 8.136
on Tuesday MZG 8.163
Number of students: 20
Registration: Please send an email until 05.04.2026 to gfa@uni-goettingen.de.

You are using statistics at an advanced level, are familiar with R, and now have a dataset containing time series such as climate data, growth data, or stock market prices. If so, this course will help you to develop a more professional understanding of time series and the methods applicable to their analysis.
This course provides an introduction to time series analysis, focusing on Generalised Autoregressive Conditional Heteroskedasticity (GARCH) models. These models are used to model and forecast time-varying volatility in time series where uncertainty evolves over time. This course is not only relevant for general data analysis, but also for evaluating energy markets and climate policy because both of these fields exhibit strong volatility clustering and shock persistence. The course includes estimation and forecasting, combining theoretical sessions with practical applications using real data and academic papers. Students will examine volatility clustering, shock persistence, volatility spillovers and volatility forecasting in order to analyse the impact of geopolitical shocks, supply chain disruptions, changes in demand, regulatory changes and weather-related events.

Target Group: This course is suitable for PhD students who are working (or wish to work) with statistical methods, including those involving financial, energy and climate data. Your statistical skills whould be at an advanced level and you have at least basic knowledge on time series analysis and R. There are no admission requirements, but a basic knowledge of statistics is recommended.

Tools: Students will need to bring a laptop to all sessions. Please install the latest version of R (≥4.3.0) and RStudio beforehand. Before the first session, please install the following packages (see the code below): rugarch (for univariate GARCH modelling), rmgarch (for multivariate GARCH models such as DCC), tseries and FinTS (for ARCH tests and financial time series tools), and forecast, zoo, xts, urca, moments and PerformanceAnalytics (for time series analysis and evaluation). For data manipulation and visualisation, please also install the ggplot2, dplyr and tidyr packages. Prior to the first session, students should verify that all packages load correctly in their R environment.

packages <- c(
  "rugarch",
  "rmgarch",
  "tseries",
  "FinTS",
  "forecast",
  "zoo",
  "xts",
  "urca",
  "moments",
  "PerformanceAnalytics",
  "ggplot2",
  "dplyr",
  "tidyr"
)
install.packages(packages)

Contents:

Characteristics of Time Series
Statistical Foundations of Volatility Models
Univariate ARCH models
Univariate GARCH models
Other conditional heteroscedastic models

IGARCH models
GARCH-M models
EGARCH models
GJR-GARCH models

Multivariate GARCH models (VECH, BEKK, DCC)
Estimation and Inference
Model Evaluation
Volatility Forecasting
Empirical Applications

Evaluation: The course assessment will be based on a final project consisting of an empirical exercise using data relevant to the student’s doctoral research*. Submissions must include the R code and a written report (between 7 and 14 pages) presenting the methodology, results and interpretation. The report should reflect students’ own analysis, interpretation, and decisions — copying pre-written text from external sources is not acceptable.

Submission format: PDF report + Datasets + R script file(s).
Deadline (online submission): May 18, 2026

To ensure originality and independent work, the report should include the following sections: 1. Introduction and Research Context, 2. Data Description and Exploration, 3. Model Selection and Justification, 4. Model Estimation and Diagnostics, 5. Forecasting and Evaluation, 6. Results and Discussion, 7. Conclusions and Further Research.

Reminder: In order to facilitate the development of the final project, students must come to the first session prepared with a brief description of their PhD research topic (1–2 slides), information on the type of data they work with (or plan to use), and details of the frequency, variables and research objectives. This information will then be used to begin planning the final project.

*Please note that if any students do not have their own data due to the nature of their research, the professor will provide them with data to complete the final assignment.

Admission requirements: Students of GFA, other PhD students if free places are available

Cancellation policy:

Your registration for courses and workshops offered by the GFA is binding. If you want to cancel your registration later than three weeks prior to the workshop start date, you have to provide a physician's note. A late deregistration without a reason for illness is only possible with the consent of the first supervisor. If non of these two conditions is met, you will be barred from registering for GFA courses for the period of one year.
Please be aware that with a late deregistration you block seats for other PhD students, which otherwise could have taken part in the workshop.

Absence policy

To earn credits for a workshop/course, you must not miss more than 10% of total contact hours. In most cases, this will be less than a day. If it is inevitable that you miss more than a whole day, please notify us well in advance (at least one week).