Codes and public data used in papers

NBB B2B Transactions Dataset

This dataset contains information on firm-to-firm transactions between all VAT-liable firms in Belgium, currently for the years 2002-2014. The data is used in several research projects, including “The Origins of Firm Heterogeneity: A Production Network Approach” (Journal of Political Economy, 2022). The data itself is confidential, but we can provide some pointers that might be of interest.

We have cleaned and prepared the data in a scrutinised way. The procedures and some descriptive statistics are described in Dhyne, E.; Magerman, G. and Rubinova, S. (2015).
The dataset is based on yearly VAT listings data for VAT liable firms in Belgium. See templates here (Dutch) and here (French), and the description at the Federal Service of Finance here.

Correspondences of EU Product Classifications

We provide clean correspondences of 8-digit products in the EU Combined Nomenclature (CN) and Prodcom (PC) classifications. These mappings trace (i) products within classifications over time, and (ii) products across CN and PC. Classifications tend to vary from year to year, for several non-economic reasons. Incorrectly accounting for changes in these classifications leads to spurious entry and exit of products, price biases, and incorrect price and quantity indices. We characterize all singular and non-singular mappings, and all quantity units of measurement. Combined with product-level datasets, these mappings allow to calculate plausible unit values in CN and PC, changes in unit values, indices and e.g. domestic prices for internationally traded goods. All data and codes are available on Github.