Once you can define ARMA models, you must estimate them and pick one. These slides derive the sampling behaviour of the sample ACF (±1.96/√T bands), the Ljung–Box portmanteau test, estimation of AR by OLS and MA/ARMA by MLE, and model selection with AIC and BIC — all applied to UK GDP growth data.
What these materials cover
- Sample ACF and correlogram confidence bands
- Ljung–Box portmanteau testing
- AR estimation by OLS; MA/ARMA estimation by MLE
- Model selection: AIC vs BIC and how they trade fit against parsimony
- Application to UK GDP growth data
Download
Free to download and use for personal study. Written for my own university teaching; shared here as evidence of teaching style and depth.
Lecture 3: estimation, testing & model selection (PDF)
Who this is for
Undergraduates fitting their first ARMA models, and anyone choosing lag orders for coursework or dissertations.
Working on this topic?
Send the module topic list, the problem set and where you're stuck. A free consultation diagnoses whether the difficulty is definitions, derivations or software output — and proposes a plan.
Related free resources
- Time Series Econometrics — full study-note hub
- ARMA estimation and model selection — study note
- How to read a correlogram — study note
- All teaching materials — notes, exercises and solutions
One-to-one help
For help with this material — or the module it belongs to — see econometrics tuition or university economics tuition. The first consultation is free, with no obligation.
Free worked video lectures: @economaths on YouTube.