BLADE for productivity research
Staff working paper
This paper by Henry McMillan and Colin Burns was released on 5 May 2021. It discusses the uses and limitations of the Business Longitudinal Analysis Data Environment (BLADE) for economic research. BLADE is a comprehensive administrative dataset covering almost all Australian businesses between 2001 and 2019 with much potential.
Download the paper
The accompanying code is intended to create longitudinal datasets out of BLADE components with proper R object formatting. The code parses required BLADE variables and imputes missing values with XGBoost. This version of the code was designed for the 2017-18 release of BLADE and may require tweaking for newer releases. All code is subject to the attached licence file (MIT).
The core components of the code are:
- BLADE parser functions (contained in the R folder)
- Initial BLADE parse and missing value imputation model matrix creation
- Missing value imputation model estimation
- Application of model to BLADE.
Further information is documented in the zip file.
Download the code
This publication is only available online.
We value your comments about this publication and encourage you to provide feedback.