md4stata

Importing publicly available datasets into statistical software is often tedious and time consuming: the original dataset needs to be converted from its original format; longitudinal datasets need to be reorganized in a form suitable for analyzing panel data; and datasets from different sources need to be carefully merged. This last step is particularly time consuming because country codes and names tend to vary across datasets. Macro Data 4 Stata addresses these issues by homogenizing several commonly used macroeconomic datasets and importing them into Stata.

The AAA codes dataset is particularly useful for selecting subgroups of countries and merging the datasets already coded in Macro Data 4 Stata with other datasets. AAA Codes contains different ways to write country names, country groups (regional and income classifications), and IFS codes. Users should be careful and avoid running regressions or computing statistics using all the observations. Some datasets include observations that represent a group of countries, these observations need to be dropped when one wants to conduct country-level analyses.

If you find these datasets useful, please quote us as:

Catini, Giulia, Ugo Panizza and Carol Saade (2010), “Macro Data 4 Stata”
http://graduateinstitute.ch/md4stata

Compatibility with older versions of Stata: All datasets were created using Stata 10 and Stata 11. Therefore, they cannot be loaded into older versions of Stata (it is necessary to first open them in Stata 10 or Stata 10 and then save them by using the saveold command).

Last update: 6 October 2011

Disclaimer

We do not take any responsibility for the quality of the original and transformed data. While we did our best in transferring the data into Stata, coding errors are always possible and we would welcome any suggestion towards improving the data and correcting possible errors. Please contact us at: md4stata@gmail.com

Permission from the authors of the original datasets

Even though most of the original datasets coded in this website are in the public domain, we contacted the compilers of the original datasets and asked for their permission prior to posting the data. In most cases, we were given permission and we posted the data; in one case, we were not given permission and we removed the dataset from Macro Data 4 Stata and put a note mentioning that we coded the dataset but did not receive permission to post; in a few cases, we did not receive a reply and, after two weeks without a response, we assumed that the authors did not have any objection and we posted the dataset. If you are the author of one of the datasets that we coded and would like us to remove it from the website, please let us know and we will remove it.

* The variable WB_gb in the AAA codes dataset identifies observations that report aggregate data for specific country groups (such as Latin America, South Asia, Low Income countries, etc). These observations (coded as Group) should be dropped from the sample when performing country-level analyses. To do this, just merge your dataset with AAA codes and then type: drop if WB_gb=="Group"