This Panel Econometrics Toolbox is an object-oriented MATLAB package (R2021a+) designed for the estimation and simulation of advanced panel data models. It provides a unified workflow for handling static, dynamic, and limited dependent variable panel data structures. Key Features:
- Unified Data Handling: A smart
PanelDatacontainer handles sorting, unbalanced panels, and missing values automatically. - Simulation Engine: Built-in Monte Carlo generator for testing estimator validity under various data generating processes (DGPs).
- Advanced Estimators: Includes Fixed Effects, Random Effects, Correlated Random Effects (Mundlak), IV-2SLS, Arellano-Bond (Diff-GMM), Blundell-Bond (Sys-GMM), Panel Logit/Probit as well as DPF for dynamic panel estimation with fractional (limited) dependent variables (like debt ratios etc.).
- Specification Testing: Automated diagnostic tests (Sargan, AR(m), Hausman, Weak Instruments) included by default.
- Worked Examples: The toolbox contains a number of worked examples, using real-world data for comparison with Stata estimation, simulation evidence on the Nickell bias of ignoring endogeneity problems with a lagged dependent variable. These can be found in the examples subfolder*.*
- Code Validation: Monte Carlo backtests of all estimators are conducted. The file verify_monte_carlo.m can be found in the tests subfolder.
Release 1.2:
- Help Files: To use the newly added Matlab help files, you may have to switch to local help setting and restart Matlab. The help files then show (after clicking on "Help") in the Documentation Webpage und er "Supplemental Software"!
- Fixed Effects estiamtor with weighting and absorption method for large scale (arbitrary) fixed effects.
- Poisson Pseudo Maximum Likelhood estiamtor added (PPML).
- Error fixings, in particular improvements for the Blundell/Bond estimator.
- All estimator with Monte Carlo simulation tests.