Economists studying innovation and technology have used patent data for decades. Patent data are exceptionally rich and - when properly understood - can generate powerful empirical insights.They are also difficult to analyze jointly with economic data because their underlying classification systems do not easily map into each other. 

With funding from the NSF SciSIP Program, several collaborators and I formulated a text-mining approach we call Algorithmic Links with Probabilities (ALP) to construct probablistic concordances (a.k.a. crosswalks) between patent, trade and industrial classification systems. We applied similar methods to construct ALP crosswalks for trademarks as well.

Published and Working Papers

Getting Patents & Economic Data to Speak to Each Other: An ‘Algorithmic Links with Probabilities’ Approach for Joint Analyses of Patenting & Economic Activity

Travis J. Lybbert and Nikolas J. Zolas. 2014. Research Policy 43: 530-542. Article | Slides

An ‘Algorithmic Links with Probabilities’ Concordance for Trademarks with an Application Towards Bilateral IP Flows

Nikolas Zolas, Travis J. Lybbert, and Prantik Bhattacharyya. 2016. The World Economy.Article | Slides

Tracking the technological composition of industries with algorithmic patent concordances

Nathan Goldschlag, Travis J. Lybbert and Nikolas J. Zolas. Economics of Innovation and New Technology. Article.

Downloadable ALP 'Crosswalk' Files

Zipped files of all the concordance files needed to map three different Intellectual Property Classifications (IPC, USPC, and CPC) into various Economic Classifications (ISIC, NAICS, SITC, HS) are available for download HERE. For a similar crosswalk files for trademarks (NICE), click HERE. For details on their construction, see the papers linked above.