Methodology

The Corporate Invention Board project builds on robust methodological choices in relation to:

The kind of patents to be analyzed;
The delineation of the perimeter of industrial groups;
The matching of financial and patent data.

Priority patents applications reveal technological potential at actor’s or territory’s level

To explore technological strategies of multinational corporations, the Corporate Invention Board chose to focus on the analysis of priority patents applications. Priority patents applications are markers of a technical newness. A priority patent corresponds to the first application of the patent in a national office. Patents application happens at the beginning of patents’ examination while granting of patents is only at the end of this process. Furthermore, priority patents applications are made before any other request for extension whether in the same office (for technical reasons) or in another office (for geographical coverage reasons). In addition, these documents have a filing date closer to that of the invention - contrary to the extensions which are generally applied for at a later date than the time of the innovation. Priority patents applications are thus better adapted to stay close to the initial time of innovation. For more details see: Methodology for patents’ analysis.

Groups perimeter definition criteria

The Corporate Invention Board project uses the Orbis financial database to identify the perimeter of world industrial independent groups. To aggregate related entities (subsidiaries, companies acquired after a fusion and acquisition process) we use the “Global Ultimate Owner” (GUO) information. When a GUO has a share higher than 50,01% in a firm, it is considered to be one of its subsidiaries. Following this procedure, we have identified 2400 groups and 170,000 subsidiary companies. They constitute the population of companies analyzed in this project. For more details see: Methodology for identification of the industrial world independent groups.

Company name processing

As information comes from two distinct databases (financial and patents), additional processing to match this information was required. This was done by developing a specific in-house program. It first cleans company names of any spelling irregularity (for example, double spaces). It then removes the name of the companies’ legal designations (Ltd, Inc., Spa…) and uses names present in the Orbis database to fish those that are related in the Patstat database. For more details see: Methodology for matching the financial and patent databases