is a new open-source probabilistic programming language well suited for econometrics.
Compared to standard statistical programming languages, Stan is flexible (easy to switch between MLE and MCMC) and
fast (utilizes automatic differentiation to compute gradients). Bayesian methods using Stan are particularly useful when large datasets or high
dimensional unobservables make standard methods infeasible.
Basic examples that can be easily modified are below:
Stan is well suited for estimating demand models where some consumers lack full information about product characteristics (see research
for more information). A simple version of the model is below:
Code requires Matlab, MatlabStan
, and CmdStan
For information about using Stan for time-series models in economics see here