Skip to content

Week 2 — The OLS Engine upcoming

June 29 – July 2, 2026 (4-day week — Fri Jul 3 is the Independence Day observance) Chapters 2.1 – 2.5 · Lab 2 · Mentor Session 2 · Problem Set 2

Ordinary least squares in matrix form, its guarantees (Gauss–Markov, FWL), and the three ways the textbook story breaks (heteroskedasticity, clustering, misspecification).

By Thursday close you should be able to derive \(\hat{\boldsymbol{\beta}} = (\mathbf{X}'\mathbf{X})^{-1}\mathbf{X}'\mathbf{y}\) on a whiteboard, run Fama–MacBeth on real CRSP returns, and tell the story of Petersen (2009) on standard errors.

Daily schedule

Block What
9:00 – 12:00 Read Ch 2.1 — OLS in Matrix Form. Work nb2.1 (OLS from scratch, NumPy vs. statsmodels).
1:00 – 4:00 Live derivation on Zoom: normal equations, hat matrix, projection geometry.
4:00 – 5:00 Office hours
Block What
9:00 – 12:00 Read Ch 2.2 — Gauss–Markov and the Meaning of "Best". Work nb2.2 (Monte Carlo).
1:00 – 4:00 Read Ch 2.3 — Frisch–Waugh–Lovell. Work nb2.3 (residualization).
4:00 – 5:00 Office hours
Block What
9:00 – 12:00 Read Ch 2.4 — Heteroskedasticity, Clustering, HC/HAC. Work nb2.4 (Petersen 2009 reproduction; t collapses from 5.70 to 2.19 under firm clustering).
2:00 – 3:30 Mentor Session 2 — "The standard-error ballgame"
4:00 – 5:00 Office hours
Block What
9:00 – 12:00 Read Ch 2.5 — Misspecification: OVB, Measurement Error, Functional Form. Work nb2.5.
1:00 – 3:00 Lab 2 — Fama–MacBeth on CRSP (2 hours live, build the rolling cross-sectional regressions).
3:00 – 4:30 Wrap PS 2.1 – 2.5. Open the Week-2 PR.
4:30 – 5:00 Retro + Week-3 preview
5:00 pm Deliverable cutoff (note: Friday is holiday)

HOLIDAY — Independence Day observance. No meetings, no Slack expectations.

What to read

Notebooks (notebooks/week-02/): nb2.1 through nb2.5. Lab: Lab 2 — Fama–MacBeth on CRSP. Mentor reading: Mentor 2 — "Standard-error ballgame".

What to submit (due Thu Jul 2, 5:00 pm ET)

  1. Problem Set 2 (PS 2.1 · PS 2.2 · PS 2.3 · PS 2.4 · PS 2.5).
  2. Lab 2 notebook with the Fama–MacBeth coefficient series + Newey–West-adjusted t-stat reported.
  3. Journal entry — journal/week-02.md.
  4. Pull request — week-02 — <your name>.

WRDS access required

Lab 2 pulls CRSP daily returns. You must have your GMU WRDS account active before Wednesday — see Resources → Data access. If WRDS is not ready, the lab includes a deterministic synthetic-CRSP fallback so you can still complete it.

Data we touch this week

Next week preview

Week 3 — Causal Inference I →. The Rubin causal model, propensity-score matching, entropy balancing, 2SLS, and the weak-IV pathology.