1. Simulation Methods in Econometrics: Theory and Applications in MATLAB is a hands-on companion for learning and applying simulation-based techniques in econometrics. Each exercise in this repository explains the theory behind a topic and provides a worked application in PDF format, paired with the corresponding .m file. The book's dual focus on theory and practical implementation in MATLAB is reflected throughout, with an emphasis on intuition, reproducibility, and accessibility.
2. Fundamental concepts in econometrics rely on the notion of repeated sampling from a population, but in practice repeated sampling is nearly impossible to implement. This makes these concepts hard to demonstrate and therefore abstract. Simulation can mimic repeated sampling and make abstract concepts concrete. Apart from being a demonstrative tool, simulation has key uses in estimation. It is used to evaluate whether an estimator has the desired statistical properties. It is a method to stress test the performance of an estimator against violations of the assumptions made when deriving the estimator. Sometimes an estimation problem involves an expression that does not have an analytical solution, but it can be approximated using simulation. Yet, if there is an analytical solution, simulation can be used to validate it. An econometric model can be complex and hard to estimate, but it can be approximated by an auxiliary model and estimated using simulation. This page was created to make simulation methods in econometrics more intuitive, reproducible, and accessible to a wide range of learners. While the core topics are covered in existing literature, this project offers distinct value for several reasons. First, existing material on simulation methods is scattered across textbooks, journal articles, and lecture notes. These sources often vary in notation, depth, and clarity, making it difficult for learners to form a coherent understanding. This repository brings those ideas together in a unified, structured format. Second, it prioritizes accessibility. Concepts are broken down into annotated code, guided exercises, and concise explanations that help learners build intuition rather than just follow formulas. Third, it’s pedagogically driven. Simulation methods are inherently exploratory, and this resource encourages active engagement through experimentation and iteration, not passive consumption. Fourth, it supports reproducibility and transparency. By publishing openly on GitHub, the project aligns with the principles of open science, allowing others to verify, adapt, and extend the work. Fifth, it expands global access. Not everyone has the privilege of formal coursework or expensive textbooks. This repository helps democratize learning by offering high-quality content freely and publicly. Sixth, it offers a distinct teaching style. Even if the concepts have been explained elsewhere, the structure, examples, and interpretation here may resonate more effectively with certain learners. In short, this repository is more than a collection of code. It is a curated learning experience designed to unify scattered knowledge, foster understanding, and encourge users to explore simulation methods.
3. This page is a work in progress. More exercises on Monte Carlo integration, maximum simulated likelihood, method of simulated moments, importance sampling, and Markov Chain Monte Carlo will appear in years 2025 and 2026. As the content is actively being updated, occasional errors or inconsistencies may occur. These will be corrected over time as the material evolves.
4. MATLAB stands out as the programming language of choice due to its clean and intuitive syntax, and its excellent UI, which are superior to all alternatives. R or Python, for example, do not match the simplicity and elegance of MATLAB. These qualities make MATLAB particularly well-suited for educational purposes, aligning seamlessly with the pedagogical approach underlying all the exercises presented on this page.
5. The MATLAB code of an exercise is explained in a PDF file. A firm understanding of the MATLAB code, however, requires knowledge of programming in MATLAB. To support this, a MATLAB workshop has been created and is available in this repository. The workshop is self-contained, assumes no prior knowledge of MATLAB, and progresses to advanced programming concepts.
6. Some of the exercises make use of the MATLAB functions in the "exercise function least squares statsitics" or "exercise function least squares statsitics robust" folders. The code in these .m files require the Statisics and Machine Learning Toolbox of MATLAB.
7. To download the .m and PDF files, first click on a file and then click on the download button at the upper-right panel of the page.
8. Some of the exercises presented on this page are a collaborative effort. Contributors of an exercise are listed in the PDF file of that exercise. The contributors include Elisabeth Beusch, Akash Boelens, Jesper Eizenga, Hartog Horsch, Renata-Maria Istrătescu, Tunga Kantarcı, Quinten Salomons, Simonas Stravinskas, Jelmer Weiringa and Axel Zoons. To enhance clarity and readability, we used Microsoft Copilot to assist with language refinement. All ideas, methods, and instructional design are entirely our own.
9. If you find this project useful, giving it a ⭐ can help others discover it.