Course contents *
The development of simulation-based estimators, along with the evolution of computational power, now allows the researcher to flexible formulate the a structural model based on micro-economic theory and estimate the structural model without relying on reduced form econometric models. The aim of this course is to introduce simulation methods to estimate such flexile discrete choice econometric models.
In this course we will focus on econometrics with limited dependent variables. This include, but is not limited to, discrete choice econometrics. As such, we will cover econometric methods that are a crucial part of the scientific toolbox in transportation research, but also in many other fields such as psychology, environmental economics and labour economics.
Theory for random utility economics, maximum likelihood estimation, logit, probit, censored probit, spatial probit, simulated method of moments, simulated maximum likelihood, method of simulated scores, and introduction to Bayesian methods, Metropolis-Hastings and Gibbs sampling, dynamic discrete choice models including dynamic programming models
Intended learning outcomes *
After completing the course the student should be able to
- formulate an appropriate and flexible discrete choice econometrics designed for a given problem at hand
- estimate a standard discrete choice econometric model using standard econometric software
- formulate an appropriate discrete choice econometric model, choose an appropriate estimator and estimate such a model, using high level programming language and mathematical software libraries
- understand the trade-offs between statistical and computational efficiency and take informed decisions of choosing an appropriate estimator for a given problem
- choose a reduced form or structural econometric model for a given research question
- choose and apply basic bootstrap techniques
No information inserted