Patent families as macro level patent value indicators: Applying weights to account for market differences

This paper introduces a methodology for the construction of a country level patent value indicator based on the family size of a country's patent profile at the level of technology fields. Because individual family members target different markets and technologies have a different propensity to internationalization, family size has been shown to have a restricted power to assess the quality of patent profiles of countries. We address this gap by weighting the members of patent families filed at different patent offices before calculating the family size indicators, to account for the market potential in which the patents of these families were filed. We apply different weighting schemes and test which scheme is best able to explain the export performance of countries. In order to conduct our analyses, a panel dataset, consisting of annual data (1990-2002) on international trade from the UN-COMTRADE database and patent data from the "EPO Worldwide Patent Statistical Database" (PATSTAT), was compiled. Several bivariate analyses reveal that weighted and unweighted family counts are highly correlated, meaning that statistics based on absolute (weighted or unweighted) family counts are barely affected by the chosen weighting factor. This, however, is different when using the average family size, where weighting the family members by imports, as well as GDP, can be shown to have a robust positive effect to explaining export performance. The imports and the GDP weighted average family size are thus able to act as a consistent indicator of patent value at the country and technology field level.