Teaching
Financial Econometrics
Measuring risk and return
The econometric methods that produce the two inputs every portfolio needs: expected returns and risk. Sessions open with the expected-return toolkit: predictive regressions, the cross-section, the factor zoo, machine learning. They close with the risk toolkit: GARCH, realized volatility, large covariance matrices, tail risk. The two converge in portfolio construction under estimation error, with a live Python lab in every session.
Topics
- Stylized facts of returns; linear time series and forecast evaluation
- Predictive regressions, Stambaugh bias, out-of-sample discipline
- The cross-section: portfolio sorts, Fama-MacBeth, GRS; factor zoo and multiple testing
- Machine learning for expected returns: regularization, trees, neural nets, IPCA
- GARCH, realized volatility and HAR, DCC correlations
- Large covariance matrices: shrinkage and factor structure
- Tail risk: VaR, Expected Shortfall, EVT, backtesting
- Portfolio construction under estimation error; honest backtesting statistics
Audience
Master's (10×3h) and advanced undergraduate (13×2h) editions from a single source; graduate econometrics assumed, no prior finance required.
Materials
Dual-edition textbook in progress, 16 session decks, Jupyter labs on free open data (Ken French, Open Source Asset Pricing, JKP, Goyal-Welch, FRED).
Every session pairs a lecture with a live Python lab on free open data, and the course closes with a group capstone: a complete expected-return and risk pipeline, backtested under an honest protocol: deflated Sharpe ratios, turnover, and costs included. A student-poll elective session covers high-frequency data, text and LLMs, event studies, or term-structure econometrics.