SippData constructor scripts

This repository contains

1. bash scripts to download SIPP raw data and stata do-file instructions to read the data into .dta format
2. modified versions of the do-files as supplied by NBER. The modification is small and only allows the scripts to be called inside a loop one after the other.

useage

• you need a unix/macOS machine to run the bash scripts. (I think).
• downloads core and topical modules form SIPP
• The typical stata data for a panel has about 7GB
• the entire raw data is 50GB
• The structure of the files online is not uniform; file names change from one panel to the next, so don't expect uniform filenames in the bash scripts either.

Proceed like this:

1. run each of the setup_SIPPyy.sh scripts with ./setup_SIPPyy.sh on your command line. This should download all the raw data into a folder structure I set up at ~/datasets/SIPP. You can change that location by changing the variable dest inside the scripts
2. run each of the makeyyyy.do scripts in stata. this should read the raw data into stata format and apply the labels.

The bash scripts will check if a file exists, to avoid downloading large files again. if you think anything went wrong during the process, just delete a folder in ~/datasets/SIPP to recreate it.

Each panel has about 7GB in stata format

Disclaimer

The usual disclaimer applies, i.e. I do not guarantee anything about this software. I hope it is useful and feel free to use it.

Copyright

• The original do-files are all copyright NBER. Please see the link above.