This webpage describes how to replicate the results presented in the paper Aggregate productivity growth in the presence of (persistently) heterogeneous firms, authors: Giovanni Dosi, Marco Grazzi, Le Li, Luigi Marengo, and Simona Settepanella.
Latest update: Le Li, August 26, 2019
Table of Contents
- Software and system requirements
- Files and directories
- Toy examples
- Empirical examples
- Proprietary data
1. Software and system
- The majority of the codes are written in R
- Stata package levpet
- zon(n)ohedron written by Federico Ponchio
The codes have been tested on macOS Sierra (version 10.12.6) with R version 3.3.2 (2016-10-31) and Stata/MP 13.0
2. Files and directories
- toy_example.R contains all details on toy examples in the paper
- empirical_example.R contains all details on empirical examples in the paper
- directory codes contains codes files
- directory data contains raw data and cleaned data sample for our analysis
- directory results saves all results
3. Toy examples
toy_example.R generates figure 2 and Table 2, 3, 4, and 5 for the paper. See the comments in the file for more details.
4. Empirical examples
empirical_example.R generates Table 7 and 8 for the paper. See the comments in the file for more details.
5. Proprietary data
To provide the coordinates for replication, note that we employed Bureau Van Dijk AMADEUS database and in particular the Bluray physical support released in October 2015 Bluray. The media was accessed at the Scuola Superiore Sant'Anna. Since the dataset is proprietary, we cannot provide the data sample for the readers to conduct the replication. Instead, we provide the details (saved in ./codes/CleanData.R) on how we generated the data sample for our analysis from the Raw data. MD5 values of related data files are listed below,
- MD5 (data/data_replication.Rdata) = 60f9a91d164be576dda7af185f3f4784
- MD5 (data/Selected_7Sectors_FRIT_withMaterialCost.dta) = 6d4ea5cca9ae714f57474255c1485d75
- MD5 (data/TFP_all.csv) = eeaa00a5e4ee907e654171098735a792