1

I am trying to estimate the impact of a grant on firm survival using the first propensity score matching to create a matched sample of similar organisations along a wide range of observables. Then I am using a cox proportional hazards model to estimate the hazard ratio between those that received and did not receive the grant while controlling for other characteristics which may affect firm survival. The issue is that I have two cohorts: one receiving the grant in the financial year 2012 and the other in 2013. So would it make sense to append these matched samples together and run a cox proportional hazards model to estimate the hazard ratio of interest? Or should I run to models separately?

Many thanks

Michael hall
  • 11
  • 1
  • Are the grant receiving firms, across both years, a small fraction of the firms in your data? 2) Do you have panel data on the control firms? 3) Do you know the criteria for assigning the grants?
    • – BKay Mar 28 '17 at 14:41