An event study can find an abnormal return after a takeover announcement, yet still be answering the wrong question if its benchmark omits a sector shock, forces a curved relationship to be linear, or pools a pre-crisis regime with a post-crisis regime. Unlike the covariance failures in heteroscedasticity and serial correlation, specification failure can corrupt the conditional mean itself. Robust standard errors then give a more honest uncertainty statement about a coefficient that may have no useful economic interpretation.
This note treats specification as a sequence of testable modelling claims. It first uses Ramsey's RESET as a broad alarm, derives omitted-variable bias, then shows how intercept and interaction dummies encode observable regimes. It closes with the Chow test for a known break date and explains why an unknown break is a different problem.
Learning objectives. By the end of this note you should be able to:
- run and interpret Ramsey's RESET test as a general specification diagnostic;
- derive the direction of simple omitted-variable bias from the omitted effect and regressor correlation;
- specify and interpret intercept and slope dummies without falling into the dummy-variable trap;
- construct a Chow test from restricted and unrestricted residual sums of squares;
- distinguish a known-date structural-break test from a search for an unknown break date;
- choose a repair that changes the mean model rather than applying an inappropriate covariance correction.
In this lesson
- A diagnostic is an alarm, not a repair
- Omitted variables: when the coefficient, not its standard error, fails
- Dummies turn observable regimes into model restrictions
- Chow tests a known structural break
- A specification decision map
Before you start
- The Classical Linear Model and Where Finance Breaks Gauss–Markov
- Heteroscedasticity: White's Test, Robust Standard Errors, and Weighted Least Squares
- Serial Correlation: Durbin–Watson, Breusch–Godfrey, and Newey–West
The full lesson, with worked examples and exercise solutions, is for members.
Read the full lessonNew to SelfMadeQuant.com? See membership options or browse the Econometrics and Financial Time Series syllabus.