Course content

Economic forecasting is a critical tool in decision-making processes across both the public and private sectors. This course provides an in-depth exploration of the theory and practical applications of forecasting, combining traditional statistical methods with cutting-edge machine learning techniques. PhD students will learn both fundamental and advanced forecasting methods using state-of-the-art tools, including R and Python programming, as well as economic and financial databases such as FRED. The course covers a wide range of statistical methods and machine learning methods in time series forecasting.

Key topics include:

• Time Series Forecasting: The course covers key statistical methods for time series prediction such as descriptive statistics, regression analysis, ARIMA models, VAR models, and Bayesian VAR models. Additionally, students will explore advanced models like VECM, TAR/STAR, regime-switching models, and State Space models with Kalman filters, and models for data with different frequencies.
• Big Data and Machine Learning: Emphasis is placed on machine learning techniques for time series data, including variable reduction methods (e.g., LASSO, Ridge, Elastic Net), tree-based methods (Random Forests, Gradient Boosting), and neural networks (RNN, CNN). The course also covers feature selection techniques, dimensionality reduction (PCA), and probabilistic machine learning, with a special focus on Bayesian Neural Networks (BNNs).
• Forecast Evaluation and Combination: Students will learn how to evaluate forecasting errors, compare point forecasts to distributional forecasts, and combine forecasts from multiple models to improve prediction accuracy.
• Risk Management and Financial Forecasting: The course addresses financial risk modeling, including methods like Riskmetrics, Filtered Historical Simulation, EVT, GARCH models, and Quantile Regression. Students will learn about Value at Risk (VaR), Expected Shortfall (ES), stress testing, and scenario analysis. Special attention will be given to modeling risks in financial markets, including the use of copulas for complex risk factor dependencies.
• Monte Carlo Simulation: The course also covers Monte Carlo simulation techniques, including univariate and multivariate methods, as well as applications in modeling stochastic processes, covariance, correlations, and factor models. Students will learn how to implement Monte Carlo simulations for forecasting and risk assessment, including Least Squares Monte Carlo (LSMC) methods.

The course integrates theory with hands-on experience in data analysis and forecasting, enabling students to apply the techniques learned to real-world economic and financial forecasting challenges. Students will gain proficiency in statistical software and databases, and will be equipped to build and evaluate complex forecasting models using both traditional and machine learning approaches. By the end of the course, students will have developed a deep understanding of the statistical and computational methods used in modern economic forecasting and risk modeling.

Learning outcome

In this course, the candidate will acquire key knowledge in modern forecasting methods in economics and finance. The candidate will receive training to make a lecture of selected topics and present these. The candidate will also receive training in writing and presenting selected data, methods, and implementation of methods from the course. This will be an important part of the general PhD training for the candidate. The candidate will also get acquaintance to usage of databases and statistical software such as R and Python.

Learning methods and activities

The course will consist of a mixture of traditional lectures and practical exercises over a 1-2 week seminar in the fall. Lectures will also be available digitally.

Compulsory assignments

• Active participation in lectures

Recommended previous knowledge

Basic courses within economics and finance. General knowledge within mathematics, statistics and computer science.

Course materials

• Alexander, Market Risk Analysis 4 Volume books, Wiley

• Barber D., Bayesian Reasoning and Machine Learning, Cambridge University Press

• Brockwell P. J. and Davis P.A., Introduction to Time Series and Forecasting, Springer

• Christofferson P., Elements of Financial Risk Management, Academic Press

• Danielson J., Financial Risk Forecasting, Wiley

• Ghysels, E., & Marcellino, M. , Applied Economic Forecasting using Time Series Methods. Oxford University Press

• Fotis et. al, Forecasting: theory and practice, International Journal of Forecasting

• Gelman et al., Bayesian Data Analysis, Chapman and Hall

• Hamilton J., Time Series Analysis, Princeton University Press

• Hamoudia et al., Forecasting with Artificial Intelligence, Palgrave McMillan

• Hansen B., Econometrics, Princeton University Press

• Hastie et al. Elements of Statistical Learning, Springer

• Hyndman R.J. and Athanasopoulos, Forecasting Principles and Practice, Otexts

• Huang C. and Petukhina A., Applied Time Series Analysis and Forecasting with Python, Springer

• Metcalfe A.V. and Cowperwait P.S.P, Introductory Time Series with R, Springer

• Murphy, K.P., Machine Learning: A Probabilistic Perspective, MIT Press.

• Murphy, K.P., Probabilistic Machine Learning: An Introduction, MIT Press.

• Murphy, K.P., Probabilistic Machine Learning: Advanced Topics, MIT Press.

• Pfaff B., Financial Risk Modelling and Portfolio Optimization with R, Wiley

• Tsay R., Analysis of Financial Time Series, Wiley

• West and Harrison, Bayesian Forecasting and Dynamic Models, Springer

Credit reductions