A Crosswalk from the Business R&D and Innovation Survey (BRDIS) and the Survey of Industrial Research and Development (SIRD) to the Longitudinal Business Database (LBD) and the Standard Statistical Establishment List (SSEL)

The Survey of Industrial Research and Development (SIRD) and the Business R&D and Innovation Survey (BRDIS) provide a rich description of R&D at the firm-level. Unfortunately, linking BRDIS and SIRD to other Census data is not straightforward. Standard Census identifiers are often missing, while the identifiers used in BRDIS-SIRD are different in format than those used in other data sets like Longitudinal Business Database (LBD) and the Standard Statistical Establishment List (SSEL). In this project we develop a new crosswalk to address the problem. The crosswalk assigns to each firm-year pairs in BRDIS-SIRD the identifiers of corresponding observations in LBD or SSEL. To generate the crosswalk, we: (i) Infer standard CES identifiers (FIRMID) from variables in SIRD. (ii) Map from BRDIS-SIRD to LBD, and from LBD to SSEL. (iii) Combine the results of multiple linkages, each using a different identifier. The crosswalk allows connecting BRDIS-SIRD with any Census collected data set that uses the identifiers applied in LBD and SSEL. Further, it allows creating links from BRDIS-SIRD to external data using names and addresses appearing in SSEL. In this context, it improves researchers’ ability to use tools that were developed by the Census to connect SSEL to patent data for assigning patents to firms in BRDIS-SIRD. That, in turn, facilitates further study of the relation between R&D activity, reported in BRDIS-SIRD, and innovation outputs, as they are reflected in patenting.