ST 473
Financial Data Analysis
This course will serve as a hands-on, computationally intensive introduction to the analysis of financial market data. Students will integrate knowledge developed in prerequisite courses in finance, probability and statistics in order to thoroughly and rigorously analyze financial data. Topics might include (but will not necessarily be limited to) some of the following:
● Using maximum likelihood and/or method of moments to fit a variety of parametric models to stock return and interest data. Students will discover that such data tend to exhibit heavy tails, asymmetry and non-stationary parameters.
● Using regression techniques to assess the predictive power of so-called factor models for stock returns, such as the Capital Asset Pricing Model and the Fama-French Three-Factor Model. Students will determine the predictive power (or lack thereof) of various factors (such as “beta” or “momentum”) in explaining the performance of various stocks.
● Assess the empirical performance of various portfolio optimization methods, such as mean-variance optimization and Black-Litterman approach.
● The difference between parameter estimation (using historical data) and calibration (using current option price data).
● Dimension reduction for high-dimensional data, such as principal component analysis and techniques for “cleaning” the spectrum of empirical covariance matrices.
● Volatility forecasting using GARCH and/or regime-switching models.
Prerequisites: MA270, MA307 or MA371, ST362.
This course will serve as a hands-on, computationally intensive introduction to the analysis of financial market data. Students will integrate knowledge developed in prerequisite courses in finance, probability and statistics in order to thoroughly and rigorously analyze financial data. Topics might include (but will not necessarily be limited to) some of the following:
● Using maximum likelihood and/or method of moments to fit a variety of parametric models to stock return and interest data. Students will discover that such data tend to exhibit heavy tails, asymmetry and non-stationary parameters.
● Using regression techniques to assess the predictive power of so-called factor models for stock returns, such as the Capital Asset Pricing Model and the Fama-French Three-Factor Model. Students will determine the predictive power (or lack thereof) of various factors (such as “beta” or “momentum”) in explaining the performance of various stocks.
● Assess the empirical performance of various portfolio optimization methods, such as mean-variance optimization and Black-Litterman approach.
● The difference between parameter estimation (using historical data) and calibration (using current option price data).
● Dimension reduction for high-dimensional data, such as principal component analysis and techniques for “cleaning” the spectrum of empirical covariance matrices.
● Volatility forecasting using GARCH and/or regime-switching models.
Prerequisites: MA270, MA307 or MA371, ST362.
This course will serve as a hands-on, computationally intensive introduction to the analysis of financial market data. Students will integrate knowledge developed in prerequisite courses in finance, probability and statistics in order to thoroughly and rigorously analyze financial data. Topics might include (but will not necessarily be limited to) some of the following:
● Using maximum likelihood and/or method of moments to fit a variety of parametric models to stock return and interest data. Students will discover that such data tend to exhibit heavy tails, asymmetry and non-stationary parameters.
● Using regression techniques to assess the predictive power of so-called factor models for stock returns, such as the Capital Asset Pricing Model and the Fama-French Three-Factor Model. Students will determine the predictive power (or lack thereof) of various factors (such as “beta” or “momentum”) in explaining the performance of various stocks.
● Assess the empirical performance of various portfolio optimization methods, such as mean-variance optimization and Black-Litterman approach.
● The difference between parameter estimation (using historical data) and calibration (using current option price data).
● Dimension reduction for high-dimensional data, such as principal component analysis and techniques for “cleaning” the spectrum of empirical covariance matrices.
● Volatility forecasting using GARCH and/or regime-switching models.
Prerequisites: MA270, MA307 or MA371, ST362.
