Patent data with founding year variable
-
1.
Columbia University
-
2.
Cornell University
Description
This is the 10 July 2024 release of Founding Patents (see foundingpatents.com). The data adds a founding year for a large fraction of US-based corporate assignees using information from OpenCorporates. There are three files:
ocpb_assigneeatfiling.csv
This is the primary data that contains the founding year and venture capital indicator. Note that the venture capital indicator is time-invariant. The variables are defined here and in Table 1 of the accompanying paper:
| Variable | Type | Description |
| patent_id | Int | The year the assignee was founded according to our search for the patenting firm in OpenCorporates as described in Appendix Section A of the paper. |
| founding_year | int | The year the assignee was founded according to our search for the patenting firm in OpenCorporates as described in Appendix Section A of the paper. |
| founding_score | int | The confidence in assignee/OpenCorporates match (out of 10, see Appendix Section A in the paper). |
| vc_backed_assignee | flag (boolean) | A 0-1 variable that is equal to one if the company was ever VCbacked using our merge of Pitchbook to the assignee data as described in Appendix Section B of the paper. |
| vc_score | int | Confidence in the assignee/PitchBook match (out of 10, see Section B.) |
| initassignee_id | string | an ID created by us to cover original as well as reassigned assignees. see initassignee_dxwalk20220630pv for the assignee ID from the PatentsView 30 June 2022 release, where it was possible to find a match. |
| initassignee_organization | string | Assignee name, either from the 6/30/2022 release of PatentsView or from the Patent Assignment database |
| same_as_20220630pv | flag (boolean) | indicates whether the assignee name is the same as in the 6/30/2022 edition of PatentsView (0) or whether it was replaced via the Patent Assignment file (1). |
| assignee_organization_20220630pv | string | Assignee name from the 6/30/2022 release of PatentsView. For assignees replaced via the Patent Assignment file, this will not match initassignee organization. |
| assignee_id_20220630pv | flag (boolean) | Assignee ID from the 6/30/2022 release of PatentsView. PatentsView changes the assignee IDs with each release. |
| initassignee_idxwalk20220630pv | string | Assignee name from the 6/30/2022 release of PatentsView |
| foundinoc | flag (boolean) | was the initiassignee organization found in OpenCorporates? |
| company_number | int | Company identifier from OpenCorporates (if found). |
| gvkey | string | linkage to Compustat, from DISCERN Arora, Belenzon, and Sheer (2021) |
| cusip | string | linkage to Compustat, from DISCERN Arora, Belenzon, and Sheer (2021) |
| first_year_publicly_listed | int | The year an assignee that was ever publicly-traded appeared on a U.S. exchange. Data was sourced from (Ritter, 2023), supplemented by manual searches. |
| uo_discern | flag (boolean) | One if the firm was ever publicly-traded on a U.S. exchange, and the patent is assigned to that company (Arora, Belenzon, and Sheer, 2021). |
| sub_discern | flag (boolean) | A 0-1 variable for whether the assignee was acquired by a publiclytraded firm on a U.S. exchange. |
| univ_hospital_gov | flag (boolean) | A 0-1 variable for whether the assignee is a university, a hospital, or a government entity (independent of PatentsView having classified it as a firm) |
| is_us_assignee | flag (boolean) | A 0-1 variable for whether the assignee is based in the U.S. (independent of PatentsView having classified it as such) |
initiassignee_crosswalk202206030pv.csv
This file provides PatentsView Assignee IDs (as of the March 30, 2022 release) for as many of the original/initial assignees as we could find. Original/initial assignees are from the USPTO Patent Assignment database, which does not contain Assignee IDs. We used the 2022 March 30 release of PatentsView to generate this database.
assignee_crosswalk_20220630pv20240331pv.csv
PatentsView occasionally changes the Assignee IDs. This file provides a crosswalk between the version we used to build the data -- June 30, 2022 -- and the March 31, 2024.