This paper describes the methodology used to construct a new sample of patents matched to Compustat firms for the period 1980–2015. We compare our data to existing NBER data sets and underscore several methodological improvements, including dynamic matching, company name changes, and ownership structures. We examine how our improved match changes results obtained from the ’01 and ’06 NBER patent files using comparable samples. Overall, we find that improved patent assignment leads to slightly higher estimates of patent value in market value regressions, as well as to higher estimates of the R&D elasticity in patenting regressions.

中文翻译：

---

与Compustat公司匹配的专利，1980-2015年：动态重新分配，名称更改和所有权结构

本文介绍了用于构建与Compustat公司在1980年至2015年期间匹配的专利的新样本的方法。我们将数据与现有的NBER数据集进行比较，并强调了一些方法上的改进，包括动态匹配，公司名称更改和所有权结构。我们使用可比较的样本检查了改进的匹配如何改变从'01和'06 NBER专利文件中获得的结果。总体而言，我们发现改进的专利分配会导致在市场价值回归中对专利价值的更高估计，以及在专利回归中对R＆D弹性的更高估计。