Analysis of Financial Time Series
Instructor: Ruey Tsay
This course focuses on the theory and applications of financial time series analysis, especially in volatility modeling and risk management. Students are expected to gain practical experience in analyzing financial and macroeconomic data. Real examples are used throughout the course. The topics discussed include the following: (1) Analysis of asset returns: autocorrelation, business cycles, stationarity, predictability and prediction. Simple linear models and regression models with serially correlated errors. (2) Volatility models: GARCH-type models, GARCH-M models, EGARCH model, GJR model, stochastic volatility model, long-range dependence. (3) Forecasting evaluation: out-of-sample prediction and backtesting. (4) High-frequency data analysis (market microstructure): transactions data, non-synchronous trading, bid-ask bounce, duration models, logistic and ordered probit models for price changes, and realized volatility. (5) Nonlinearities in financial data: simple nonlinear models, Markov switching and threshold models, and neural network. (6) Continuous-time models: simple continuous-time and diffusion models, Ito's lemma and Black-Scholes pricing formulas and jump diffusion models. (7) Value at Risk and expected shortfall: Riskmetrics, extreme value analysis, peaks over threshold, and quantile regression. (8) Multivariate series: cross correlation matrices, simple vector AR models, co-integration and threshold co-integration, pairs trading, factor models and multivariate volatility models.
Computer program R is used throughout the course. No prior knowledge of the software is needed. All the programs used will be discussed in class and in review session.