Skip to content

nilshg/LearningModels

Repository files navigation

LearningModels

Julia Code for Life Cycle models with Learning

The aim of the project is to compare the simluated wealth distributions arising from different life cycle models that all feature income processes including an idiosyncratic, deterministic trend component. There are four basic models of which we want to compare the output:

  1. No Habits, no learning
A simple life cycle model in which households face idiosyncratic income risk in the form of permanent and transitory shocks, but know about their idiosyncratic drift component
  1. No Habits, learning
This corresponds to the model in Guvenen (2007, AER) - Households have to learn about their drift component
  1. Habits, no learning
As model (1), but with habits in the utility function
  1. Habits, learning
As model (2), but with habits in the utility function

About

Julia Code for Life Cycle models with Learning

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages