Building a firm level dataset for the analysis of industrial dynamics and demography

This paper illustrates the building procedure of a firm-level panel dataset that merges several sources of information concerning the various activities of business firms. The aim of this work is to achieve a detailed dataset able to shed light on firm demographics, in terms of survival, entry and exit processes, distinguishing between voluntary" and involuntary" exits. Moreover, the derived dataset allows to monitor the innovation activities of the firms and also to capture complementarities between two instruments of intellectual property rights (IPRs), namely granted patents and registered trademarks. We assess the validity of the proposed procedures resorting to the virtual universe of Italian limited liability companies as provided by Bureau van Dijk (BvD). The dataset covers more than 1 million companies operating in both manufacturing and service sectors and contain financial and economic information, as well as, among the others, the ownership structure and administrative procedures undergone by the firms, which may lead to firm exit. The main purpose of the paper is to provide a unified set of procedures to help the researcher dealing with the vast amount of information available on corporate firms and of ever increasing size. This will also facilitate the replication of empirical analyses, across researchers working on dataset with similar characteristics, although from different countries or data providers.