B8121 Statistics for Investments
Course Description and Syllabus – Fall 2013 B-Term
PRELIMINARY
Professor Paul Glasserman
403 Uris Hall
pg20@columbia.edu
Overview
This course discusses statistical concepts for investment analysis. We will use specific tools (especially
regression), but the emphasis will be on statistical thinking about financial data rather than on
technique. In particular, the course will focus on broadly applicable features of market data and the
statistical concepts that help us understand them. The course is organized around four applications:
equity factor models, volatility trading, interest rate dynamics, and credit ratings. Each of these
applications will be paired with one or more statistical tools. All of the concepts and tools developed in
the course will be firmly aligned with industry practice. No familiarity with statistics beyond the MBA
core course is required.
What this course is not: This is not a course in how to pick stocks, nor is it a course on “technical
analysis” or “charting,” in the sense of trying to find predictive patterns in price data. The tools we
discuss are however useful in the disciplined analysis and evaluation of these and other investment
ideas.
Prerequisites
The only formal prerequisite is the core class in Managerial Statistics. If you exempted out of the core
statistics class, be sure you are comfortable interpreting regression output. I will assume familiarity with
financial markets and terminology at the level of Capital Markets and Investments, so I strongly
recommend taking that class in parallel with this one if you have not taken it before. This course will
also have points of contact with the core courses in Corporate Finance and Decision Models/Business
Analytics.
Course Work and Grading
There is no textbook for the course. The course will be taught from lecture notes and background
readings. We will have frequent homework assignments, most of which will involve some data analysis,
building on what we do in class. Homework assignments may be done individually or in pairs, and you
may work with different partners on different assignments. We will have a final exam, which will be
based closely on the homework assignments. My intention is that if you have kept up in class and
understood the homework assignments, you will find the final exam straightforward; and if you have not
kept up in class or not been conscientious about the homework, you will find the final exam difficult.
Style and Level of Difficulty of the Course
A good way to get a feel for what this course will be like is to look at Chapters 1-3 of B.R. Fischer and R.
Wermers (2013), Performance Evaluation and Attribution of Security Portfolios, Academic Press, which
you can read on-line through the Columbia library. These chapters overlap with the first topic we’ll be
covering, equity factor models.
Class Schedule – Preliminary
1. Introduction




Course overview
Quantitative investment management
Review of the CAPM and regression – what works and what doesn’t work in the CAPM and why
it’s relevant to investment analysis
Using the market factor to evaluate performance and risk
2. Equity factor models




Why factor models?
Fama-French factors – size and value
Momentum
Alternative beta
3. Equity factor models



Fundamental factors
Macroeconomic factors
Evaluating fund performance
4. Equity investment performance analysis



Sharpe ratio, information ratio, and related measures
Performance attribution
Skill versus luck
5. Time series analysis




Properties of macroeconomic data
Trend, seasonality, stationarity
Autocorrelation
Autoregressive models
6. Volatility



Measuring volatility: realized, implied, VIX
Persistence; GARCH
Leverage effect
7. Volatility and Trading


Volatility risk premium; covered calls; tail risk
Low-risk anomaly; risk parity
8. Dynamics of interest rates




Properties of interest rate data
Yield curve risk
Nelson-Siegel model; level, slope, and curvature
Forecasting the coefficients
9. Dynamics of interest rates


Macro factors in term structure models
Taylor rule
10. Modeling multiple curves


Credit spreads and risk-free rates
Bond portfolios
11. Credit risk and credit scoring

Altman Z-score; discriminant analysis
12. Credit risk and credit scoring


Logistic regression
Credit ratings
The final exam will be held during exam period.