Financial Econometrics and Geoinformatics in Social Sciences using R 2400-ZEWW952

The main aim of the course is to make students familiar with the broad variety of data science methods for financial and spatial data, using geoinformatics methods and financial econometrics.

Course is divided into two parts. The first one will be conducted by visiting scholar: prof. Josip Arnerić and the other will be conducted by another visitng scholar: prof. Dimitris Ballas.

The course will be coordinated by an onsite lecturer – mgr Maria Kubara, while the whole class material will be delivered by the visiting professors.

The course will be taught in an intensive workshop setting over the course of two weeks between 12 and 23 May 2025. The students are asked to bring their own laptops with R v.3.3.0+ and RStudio Desktop installed in order to take active part in the practical live code exercises discussed during the class
Day 1 - Financial time-series and asset-pricing
Loading financial time-series from public sources and financial platforms into R. Preparation and transformation of raw data. Properties of financial time-series. Volatility clustering. Heavy tails phenomena. Asymmetry of information. Heteroscedasticity. Long memory. Price jumps. Microstructural noise. Fitting a return distributions and examination of their characteristics. A single-factor and multi-factor asset pricing models are considered (CAPM, ICAPM, APT and their variations), as well as forecasting models ARIMA, TAR, SETAR and regime switching models. Demonstrated financial applications in R using actual data.
Day 2 - Univariate GARCH models
Estimating and measuring volatility (realized, implied and integrated variance approaches). Univariate GARCH models. ARCH effects. Leverage effect, cross-correlation and sign bias test. Symmetric GARCH models. Asymmetric GARCH models. Lag selection criteria. News impact curve. Long-run volatility. Volatility persistence. Estimation method and algorithms assuming different distributions of innovations. Stability conditions. Goodness-of-fit, diagnostic checking and forecasting. Selected non-linear volatility models. FIGARCH model, GARCH model with regime switching and neural network GARCH. Demonstrated financial applications in R using actual data.
Day 3 Multivariate GARCH models
Multivariate GARCH models. Time varying covariance matrix. Direct generalization of univariate GARCH models (VEC, DVEC, BEKK). Non-linear combination of univariate GARCH models (CCC, DCC). Quasi-maximum likelihood estimation method and algorithms. Diagnostic checking and forecasting. MGARCH vs panel GARCH vs realized covariance. Demonstrated financial applications in R using actual data (examining time-varying betas and safe-haven properties of various assets).
Day 4: Realized volatility and realized co-volatility
Nonparametric volatility estimators. HL estimators. OHLC estimators. Financial properties of high-frequency data. Daily patterns. Data cleaning. Sampling frequency selection. Realized variance and realized covariance. Realized range estimator. Signature plot. Estimators of integrated variance which are robust and non-robust to microstructure noise and/or price jumps (RCOV, BPCOV, ROWCOV, RTCOV, TTSCOV, and others). Different synchronizations schemes. HAR model. HAR-GARCH model. Heavy model. Demonstrated financial applications in R using actual data.
Day 5: Risk measures and extreme value theory
Different approaches to estimating Value-at-Risk and Expected Shortfall, the two most common risk measures in practice, are discussed (RiskMetrics, GARCH, and quantile-based methods, among others). Pros and cons of parametric and non-parametric estimation techniques are summarized. Theory of extreme values. Modelling extreme values above the threshold. Generalized distribution of extreme values (GEV and GPD). Additionally, methods for estimating parameters of heavy-tailed distributions are examined. Demonstrated financial applications in R using actual data.