This handbook-style article describes various longitudinal methods used commonly in program impact evaluation. We discuss the advantages of each method, the assumptions required for implementation, and provide programming examples in STATA using data from Indonesia and Malawi.
The following program can be used to determine which of 11 parametric distributions best fits the empirical (conditional or unconditional) distribution of a variable of interest. The distributions mimic those used in Jones, Lomas, and Rice (2014), which are chosen to fit a healthcare expenditure distribution that is characterized by a long right tail. I have attached an edited (publicly available) data file from the 1996 Medical Expenditure Panel Survey (MEPS) as an example of how the code works.
Create your own unique website with customizable templates.