Teaching
Quantitative Finance
A course in portfolio construction (graduate)
The graduate edition of the portfolio-construction course: the same arc from return and risk measurement to full strategy backtesting, taken at master's depth. The graduate textbook carries the MSc main text plus PhD appendices, adding rigor on estimation, factor structure, and pricing tests.
Topics
- Mean-variance optimization, rebalancing, and estimation error
- Covariance and factor models
- Momentum, reversals, and the investment horizon
- CAPM, multifactor models, APT, and pricing tests
- Characteristics versus covariances; finding alpha
- The Black-Litterman model
- Event studies and strategy examples (betting against beta, quality minus junk, macro momentum)
- Backtesting and evaluation
Audience
Master's students comfortable with statistics and linear algebra; PhD appendices for research-bound students.
Materials
A complete graduate textbook (~375 pp., 23 chapters: MSc main text plus PhD appendices), lecture decks, team projects plus the Fund Project.
The graduate textbook extends the undergraduate course chapter by chapter, with starred advanced sections and PhD appendices for research-bound students. Worked Python examples run throughout, and graded work is built around team projects and the semester-long Fund Project with its own rules and deadline calendar.
Book Preface and Table of Contents Show
Quantitative Finance: A Course in Portfolio Construction — the graduate textbook.
From the preface. This book is an attempt to teach quantitative portfolio construction as a single coherent argument rather than as a parade of disconnected models. The argument begins with the simplest possible problem an investor can pose, choose portfolio weights to trade off expected return against variance, and ends with a fully realistic implementation that respects estimation error, transaction costs, and the limits of statistical inference.
The thread that ties the chapters together is the recognition that the Markowitz optimization problem, beautiful as it is, requires two inputs that are essentially impossible to measure well from data: the vector of expected returns μ and the covariance matrix Σ. The first part of the book is about how to do less badly at each of these. We attack Σ first because progress there is more straightforward: factor models (including a detailed treatment of the commercial Barra-style architecture), Michaud resampling, shrinkage estimators, and random matrix theory all reduce the estimation problem to manageable dimensions. We then attack μ, which is harder, by decomposing each asset's expected return into a sum of components, a risk-free rate, a market premium, exposures to other priced factors, characteristic-based mispricings, and finally an idiosyncratic alpha, and devoting a chapter to each.
The preface continues with the structure of the accompanying fund project and notes on using the book: pitched at master's level, mathematically careful but not measure-theoretic, with a Python worked example, exercises with full solutions, and a PhD-level appendix closing each chapter.
Contents
Part I. The Markowitz Problem and Why It Fails in Practice
- Mean-Variance Foundations
- Why Markowitz Doesn't Work Out of the Box
Part II. Improving the Risk Model
- Sample Covariance and Its Pathologies
- Factor Models for Covariance
- Michaud Resampling
- Shrinkage Estimators
Part III. First Strategies and First Views
- Momentum
- Reversal
Part IV. Combining Information
- The Black-Litterman Framework
Part V. Constructing the Expected-Return View from First Principles
- CAPM as the Benchmark Model
- The Risk-Free Rate
- The Market Risk Premium
- Other Factor Premia
- Characteristic-Based Mispricing
- Individual Alpha
- Individual Alpha and Valuation
Part VI. Event Studies
- Event Studies
Part VII. Synthesis
- Putting It All Together
Part VIII. Standard Strategies
- Quality Minus Junk
- Macro Momentum
- Betting Against Beta
- Other Strategy Ideas
Part IX. Evaluation
- Performance Evaluation
Appendix: Solutions to Exercises