Bayesian spatial PSM

Bayesian spatial propensity score matching. This is an update of the original code I made in 2014, now updated to run in MatLab 2020a

Bayesian Spatial Propensity Score Matching (BS-PSM)

Rolando Gonzales Martinez

Updated and tested to run on MatLab 2020a (January 2022)

In order to run the BS-PSM algorithm you will need:

   (1) A n x n spatial contiguity matrix (W)
   (2) A n x 1 binary treatment vector(y)
   (3) A n x p matrix of potential explanatory variables (X)
   (4) A n x 1 variable that measures the impact (I) of the treatment

There is a need also to define the parameters of the MCMC simulation:

   - ndraws: number of draws (simulations) of the MCMC
   - nomit: burn-in 

By default, the prior of rho is elicitated in the positive range (0,1]

BS-PSM uses some functions of James LeSage Spatial Econometrics Toolbox

To run an example file check BSPSM_poverty_example.m