The IJC method and Bayesian MCMC

I solve the Loyalty Program model using the IJC method in Julia. The documentation briefly describes the model, method, code, and estimation results. This project features:

  1. Bayesian Markov Chain Monte Carlo (MCMC), which is an efficient method to estimate structural models with many parameters such as those with heterogeneous individuals.
  2. The “IJC method” described in Imai, Jain and Ching 2009 which is a computationally efficient way to approximate the value function in dynamic models which are being estimated with Bayesian MCMC.
  3. The “Loyalty Program” which is a basic discrete dynamic choice model where consumers decide between products in the current period, and can earn “Rewards” after having chose a product a certain number of times (for example, a “buy 5 get one free”).
  4. Clean and optimized Julia code using best practices of the language, to the best of my knowledge.
Raymond Chiasson
Raymond Chiasson
PhD student in Economics

PhD student in economics at the University of Toronto.