🏛 Topics in Structural Econometrics

Short course at the University of Oslo, January 2026

This course provides a graduate-level introduction to the structural estimation static and dynamic disrete choice models. It covers the foundations of random utility models, maximum likelihood estimation, dynamic programming, and various estimation methods, including Nested Fixed Point Algorithm (NFXP), two-step methods based on conditional choice probabilities (CCP), and Nested Pseudo Likelihood (NPL).

Course page at UiO: ECON9104C

🧑‍🏫 Instructor

• Fedor Iskhakov

• Professor of Economics, Australian National University

• Email: fediskhakov@gmail.com

• Web: Personal page

📆 Lecture schedule

Date	Theory	Practice
January 12	Static random utility models	Computational setup, Python refresher
January 13	Maximum likelihood estimation	MLE implementation for logit models
January 14	Dynamic programming	Implementing various solution methods
January 15	Rust-Zurcher engine replacement model	Implementing NFXP estimator
January 16	Consumption-savings model	Endogenous gridpoint method and method of simulated moments estimator

Download initial syllabus

🏡 When and Where

• 5 days: approximately 2 hours of theory + 3 hours of practical exercises

Date	First session	Second session	Location
Jan 12 Monday	11.00 – 12.00	14.15-17.45	Room 1249, Eilert Sundt’s building
Jan 13 Tuesday	10.15 – 12.00	14.15-17.00	Room 101, Harriet Holter’s building
Jan 14 Wednesday	10.15 – 12.00	14.15-17.00	Room 101, Harriet Holter’s building
Jan 15 Thursday	10.15 – 12.00	14.15-17.00	Room 1249, Eilert Sundt’s building
Jan 16 Friday	10.15 – 12.00	14.15-17.00	Room 101, Harriet Holter’s building

🎓 Exam

• Individually or in a group of two

• Set up and solve a problem using tools from the course

• Submit a short proposal of what you intend to write about by Friday, January 23rd

• Final assessment will be based on video-presentation of the project (10 min max)

• The video + project paper + code have to be submitted by Monday, February 28th, 2026

Ideas for projects:

• Implement and estimate a static or dynamic discrete choice model of your interest using real or simulated data. Ideally the following steps should be included:

  • Theoretical model description

  • Solver code

  • Simulator code

  • Estimator code

  • Application of the model to the data

  • Counterfactual simulations of some sort

• Project with steps as above using one of the models in the course and labs:

  • Static random utility model (mixed logit, nested logit)

  • Extension of Rust-Zurcher bus engine replacement model (additional state variable, specification search for cost functions, etc)

  • Inventory management model (add simulator and estimator to existing implementation, or also reformulate the solver to account for the symmetry in the policy function we observed in the lab)

  • Consumption-savings model (extend the model with additional features, i.e. additional state variables to account for agent heterogeneity, etc.)

  • Location choice model from lab 2 (possibly modify to the model, simulator code, counterfactuals, etc.)

Short proposal

• Should give a brief description of the project you intend to carry out

• Should clearly lay out the roadmap of steps you intend to take

Final submission

Submission should include:

• Project document based on the proposal (pdf), does not have to be long

• Code developed for the project, ideally well documented and structured (Jupyter notebook, Python files)

• Video presentation summarizing what you have done, and describing the main steps and findings

  • Use Zoom, Loom or any other screen recording software

  • Make sure your face appears in the video (as overlay or in the introduction)

  • Upload the video on shared files system or YouTube, and include the link in your submission

Submission should be done via a Pull Request to the course GitHub repository (similar to student survey you did in the first day).

Please, submit on time. Grades for late submissions without appropriate justification will be penalized.

Grading

• Pass / Fail

• The main criterion is to verify your understanding of the steps of a typical structural estimation project

I will provide grades and some feedback within two weeks after the submission deadline (above)

🫶 Feedback and student evaluation

Please, share your feedback anonymously via this Suggestion Box

📷 Class photo