Introducing Asset Pricing with Omitted Factors for after-work quants

Takeaway. If your listed factors are incomplete, naïve betas and premia can lie—but you can still recover risk premia by combining returns information with a rotation / PCA-style approach to the latent space.

Asset Pricing with Omitted Factors (Stefano Giglio, Dacheng Xiu, 2021 and related working-paper lineage) is a modern classic for anyone who has ever typed sm.OLS on a short factor list and felt uneasy.

Fit for learners: Recover risk premia even when some true factors are omitted—PCA-plus-rotation ideas for factor notebooks.

This Hub note is a plain-language intro. Not investment advice. Pair later with Harvey–Liu–Zhu factor zoo and IPCA.

The problem

Textbook CAPM/FF regressions assume you observed the right factors. In practice many true drivers are missing. Omitted factors bias inference about risk premia—and “just add more named factors” does not magically fix identification.

The idea / core intuition

• Observed factors may span only part of the true factor space.
• Cross-section of returns still carries information about latent factors.
• With the right statistical structure, you can estimate risk premia for the factors you care about even when others are omitted (read the paper for assumptions and estimators—do not invent shortcuts).
• Mentally separate “named factor storytelling” from “what the return covariance actually spans.”

This is research methodology, not a trading recipe.

Why after-work Python learners care

Almost every first factor notebook quietly assumes the three- or five-factor list is complete. Giglio–Xiu gives vocabulary for why that assumption hurts—before you over-trust a single coef_ printout.

Tiny evening exercise: simulate returns from two latent factors but regress on only one observed factor; compare the “true” premium vs the naïve OLS premium—learning only, toy data.

One-evening reading checklist

1. Explain omitted-factor bias in one plain sentence.
2. Note what role PCA / latent factors play in the paper’s story.
3. Write why recovering premia ≠ inventing a free alpha signal.
4. List one assumption you must not hand-wave when coding later.

Download PDF

Download PDF

Landing / DOI: Journal page

Important disclaimer

Not investment advice. This Hub note is educational only. Giglio–Xiu is academic research—not a portfolio construction mandate or signal.

What this is not

• Not CEF / subsidy framing
• Not Quant or Macro course CTA
• Not a “secret factor” sales pitch

Next step

Single course CTA: SoloQuant Academy’s Introduction to Python Programming—get fluent with arrays and plots; then factor-matrix notebooks stop looking like hieroglyphs.