opensourceecon / ogstoch-inc Goto Github PK

Let me know what you all think, but I suggest we have different lifecycle profiles and layer stochastic income of top of this. So there are both permanent and transitory differences in the income of households.

The idea is we have an average life-cycle profile for each "life-time income group" and then there is a stochastic process for some residual earnings (that follows a first order markov process).

This would affect how we compute the distribution, but I think it's pretty feasible.

cc @jorgebarro @rickecon

We should discuss naming convention of variables and how we want to structure the modularity of our functions as well as modules. @jorgebarro uses a convention of long descriptive names, whereas @jdebacker and @rickecon often use a shorter naming convention that is tied to the theoretical representation of the model.

Example: beta vs. discountFactor.

Also, we should discuss how to optimally divide up functions into Python modules as well as how to divide up functionality by functions.