Financial Econometrics

• demonstrate knowledge of the econometric methods used in finance
• compare and contrast the merits as well as the limits of econometric approaches to financial data
• locate, retrieve, pre-process and visualize relevant financial data using R
• implement an analytical approach to learning from financial data
• structure and solve financial problems by combining relevant knowledge from the areas of finance, statistics and mathematics
• estimate financial models, test hypotheses and generate responsible forecasts using R
• communicate the results of the analysis to specialists as well as to non-specialist audiences
• track innovations in finance and follow new developments in the programming language R

The course starts with a primer on formulating statistical models in R. By means of examples, the R packages relevant for the course are introduced. 

A discussion of the key concepts of statistical inference provides the conceptual basis of the course. Estimation methods (least squares, maximum likelihood, and generalized method of moments), principles underlying hypothesis testing (likelihood ratio, Wald, Lagrange-multiplier) and principles of Monte-Carlo simulation will be motivated and put to work in the context of the multiple linear regression model. In this context, the main stages of a model building process, including model diagnoses, and specification testing will be demonstrated. A module on forecasting completes the treatment of this first approach to time series regression.  

Next, by adopting a dynamic process perspective, univariate linear time series models will be introduced. After exploring key concepts as trends, cycles, stationarity, and autocorrelation in financial data, the focus will shift to estimation, testing and forecasting in the context of autoregressive moving average (ARMA) models. In the sequel, methods for studying the relationships between several financial variables across time, as vector autoregressive models (VARs) and techniques for modelling the long-run (error correction models and cointegration techniques), will be discussed and applied.  A module on modelling volatility in financial data (ARCH, GARCH) will provide a first look at non-linear time series analysis. Finally, a first glimpse at statistical techniques and models for financial panel data finalizes the course. 

The course consists of lectures and group work sessions. Expected total workload: 200 hours.

We use a dual approach to introduce the student to modelling techniques suitable for financial data. Theoretical arguments presented in the lectures are supplemented by student projects. A project typically involves the application of statistical procedures (open source statistical program R) to real-life financial data sets. Each project is designed to give the student an alternative access to a theoretical issue and to train specific data analytical skills. Through this applied focus, we increase the students’ awareness of the potential as well as the limitations of state-of-the art econometric tools for finance.