Isabel Casas, Jiti Gao, Bin Peng, and Shangyu Xie, "Time-Varying Income 
Elasticities of Healthcare Expenditure for the OECD and Eurozone", 
Journal of Applied Econometrics, Vol. 36, No. 3, 2021, pp. 328-345.

All files are ASCII files in DOS format. They are zipped in the file
cgpx-files.zip. Unix/Linux users should use "unzip -a".

Data description
----------------

The data used in this paper are publicly available and were downloaded
from the World Bank database (https://data.worldbank.org/). They can be
found in directory Code_and_Data/Estimation_RCode.

The period of this study starts in 1995 and runs until 2014 for 19
Eurozone countries and 34 OECD countries. Therefore, there are two
datasets: Eurozone.csv and OECD.csv. They are balanced panel
datasets in long formats with the same variables (no missing values),
but with different sets of countries.

Variables:

- Country
- Year
- lhe: log of total healthcare expenditure per capita for the country in
       $ and ppp-adjusted
- LogGDP: log of the GDP per capita in $ and ppp-adjusted
- Pop65: proportion of the population over 65 years
- Pop14: proportion of the population under 15 years
- Gov: public finance share of the total healthcare expenditure
- LifeExpBirth: life expectancy at birth

Instructions to replicate results
---------------------------------

The following instructions show how to use the R functions to plot all
figures and tables of our paper. The code needs the MASS, plm, xtable
and phtt packages installed in your computer to work. These files also
call the data files "Eurozone.csv", "OECD.csv", which can be found
also in directory Code_and_Data/Estimation_RCode. 

1. Plot Figures 1 and 6

> source("Figure1_JAE.R")
> source("Figure6_JAE.R") 

2. Construct Table 1

>source("Table1_JAE.R")
>Table1(zone = "OECD") 
>Table1(zone = "Eurozone")

3. Construct Table 2

>source("Table2_JAE.R")

4. Figures 2-5 and Table 4. The following command plots the
coefficient of TVFE and TVIFE estimates for the Eurozone with their
95% confidence intervals and for different control variables. If you
enter nvar = 1, then the model contains only lgdp as a regressor. If
nvar = 3 then it contains lgdp, pop65, pop14 as regressors. If nvar =
4, then it contains lgdp, pop65, pop14, public as regressors. The
files with the plot are automatically named
"Eurozone_LogGDP_nvar_*.pdf", "Eurozone_Pop65_nvar_*.pdf",
"Eurozone_Pop14_nvar_*.pdf" and "Eurozone_Gov_nvar_*.pdf" ( Figures 2,
3 and 4). Only 5 bootstrap samples are used for the confidence
intervals by default, to make it fast for testing. However, the
paper's figures are done with nboot = 100 and the user can change
parameter nboot in the function call to adapt it. In addition, it
plots the factor estimates in file "Eurozone_Factors_nvar_*.pdf"
(Figure 5). The same for the OECD is done for argument zone = "OECD".
The RMSE of each estimation is printed as well.

>source("estimation_JAE.R")
>estimation_JAE(zone ="Eurozone", nboot = 5, level = 0.05, nvar = 4)

5. Construct Table 3
>source("Table3_JAE.R")


Instructions to replicate simulation results in Appendix

This is done in Matlab. Run function ShellCode1.m in directory
Code_and_Data/Simulations_MatlabCode


Isabel Casas
icasas [AT] deusto.es