Extended Data (till 2020) following Kogan, L., Papanikolaou, D., Seru, A. and Stoffman, N., 2017

This data provides an updated data series for KPSS values and forward citations, a patent number to the CRSP "permno" match and a patent number to CPC class match following the paper Kogan, L., Papanikolaou, D., Seru, A. and Stoffman, N., 2017. Technological innovation, resource allocation, and growth. Quarterly Journal of Economics, 132(2), pp. 665-712. The paper is available at https://academic.oup.com/qje/article/132/2/665/3076284.

Data

Data Versions:

The version released on August 9, 2020 is the latest data that updates and adds data for the second half of 2019.

The version released on August 21, 2020 adds filing date information for each patent.

The version released on September 6, 2020 updates filing date information for each patent.

The version released on June 8, 2021 is the latest data that updates and adds data for 2020.

Data Description:

We provide three data sets constructed from the paper here:

KPSS_2020_public.csv: Patent level panel data from 1926 to 2020
patent_permco_permno_match_public_2020.csv: Patent-CRSP permco and permno match data
patent_number_cpc_2020.csv: Patent-CPC class match data

For the patent level panel data, the variable definitions are:

Variable name	Definition
patent_num	Patent number
permno	CRSP permno
issue_date	Issue date of patent (mm/dd/yyyy)
filing_date	Filing date of patent application(mm/dd/yyyy)
xi_nominal	Value of innovation(ξ) in millions of nominal dollars
xi_real	Value of innovation(ξ) deflated to 1982 (million) dollars using the CPI
cites	Forward citations

For the patent-permco-permno match data, the variable definitions are:

Variable name	Definition
patent_num	Patent number
permco	CRSP permco
permno	CRSP permno

For the patent-CPC class match data, the variable definitions are:

Variable name	Definition
patent_num	Patent number
cpc	Full CPC class of patent

Notes:

The value of innovation(xi_real) is equal to the nominal value of innovation(xi_nominal) deflated to 1982 (million) dollars using the CPI. In other word, we can simply calculate the xi_real through the following equation: xi_real = xi_nominal / CPI * 100.
We drop patents in the dataset that were assigned to multiple CRSP permnos.
The updated KPSS values differ slightly from the data in the QJE paper for any of the following reasons:

a) Changes in the CRSP dataset over time.

b) Improvements in the matching process to firms.

c) Re-estimation of the parameters needed to compute the KPSS value:
- the parameter gamma is re-estimated in the full sample to 2020: the new estimate is 0.0138, while the estimate of gamma in QJE paper is 0.0146.
- stock return volatility is now computed over the last 52 weeks (as opposed to the previous calendar year in the QJE paper)
The number of forward citations is updated to 2020.
Till 2020, 3319 patents do not have any CPC class information. These patents have missing CPC class value in the patent-CPC class match data.

Contact:

Please contact Dimitris Papanikolaou (d-papanikolaou@kellogg.northwestern.edu) or Amit Seru (aseru@stanford.edu) for any questions regarding the data.