*************************************************************************
Automation Classification for

Induced Automation Innovation: Evidence from
Firm-level Patent Data
*************************************************************************
Hémous, Olsen, Zanella, Dechezleprêtre, Journal of Political Economy 2024
*************************************************************************

Status: 23.10.2024

This folder contains:
	- Our classification for aggregated CIPC codes in the default version used in the paper (1978-2018) as .csv files
		- files called "classified" are codes classified as the noted category with the noted percentage threshold
		- files follow the following naming scheme: e.g. auto90_ipc6xx = automation, 90%-threshold applied on ipc6xx codes
		- "pairs" indicates the pairing of two ipc4 codes
		- ipc4 indicates the combination of an ipc4 code with G05/G06 (and NOT an ipc4 code in isolation)
		- for each keyword category, each of these files contain the count and share of EP patents that include the keyword at least once in their text.
	- The files ipc6XX_tf, ipc4_pairs_tf, and ipc4_tf contain the same information for all 6 digit codes, pairs of 4 digit codes, and combination of 4 digit codes with G05/G06
		- these files include non-machinery codes for which the classification is NOT designed.
		- it also contains the names of the technological field and technological sector of each code and an index for whether a code (or code combination) is in machinery or not. Note that the classification is designed for machinery codes only.
	- A crosswalk from ipc and cpc codes of used EP applications to our combined cipc codes based on the ipc/cpc concordance (cipc_concordance)
		- We have modified the code from the replication package to produce a more accessible cpc/ipc to cipc crosswalk.
		- The code for this is enclosed, though two of the files used have been extracted from Patstat and cannot be shared.
	- A crosswalk from cipc6 codes of used EP applications to aggregate cipc6xx codes that encompass at least 100 patents (ipc6XX_mapping).
	- Files with the threshold for the automation ipc6xx categories

The enclosed cpc/ipc concordance was downloaded from 
https://www.cooperativepatentclassification.org/cpc/concordances/cpc-ipc-concordance.txt 
on March 12, 2023 and reflects the 2023.02 version.

For more information and code on the classification and the cipc to cipc6xx crosswalk
please consult the official replication package on the JPE dataverse (https://doi.org/10.7910/DVN/IMCSW1) 
or the paper itself (forthcoming).
