Do firms face a trade-off between the quantity and the quality of their inventions?

Gaétan de Rassenfosse

Research Policy (2013), 42(5): 1072-1079

Research Question

Do innovative firms face a trade-off between the quantity and the quality of their inventions?

Why Do We Care

Understanding the relationship between the quantity and quality of inventions is critical for assessing research productivity in firms. Prior studies have shown varying results regarding this relationship, and without clarity, firms may misallocate resources in their innovation strategies. Moreover, knowing whether a trade-off exists could inform policy and management decisions in innovation practices.

What Did We Do

Using survey data from 772 patent applicants at the European Patent Office, we identified a model to explore the relationship between invention quantity and quality. We accounted for firms' propensity to patent by employing an instrumental variable approach, which allowed us to address the identification problems due to inherent differences in patenting motives among firms. We estimated parameters in a two-step process involving negative binomial regression for invention count and two-stage least squares for quality assessment.

What Did We Find

We found evidence of a trade-off between the quantity and quality of inventions, as indicated by a statistically significant negative coefficient associated with the quantity parameter. Specifically, the estimation revealed that firms with a higher rate of inventions tended to have lower average invention quality, with an estimated correlation coefficient of -0.22 representing the degree of this trade-off. The results imply that firms should be cautious in setting targets for quantity, as increasing inventions can adversely impact their quality.

Primary Data

  • Survey data from EPO applicants, N = 772, covering firms filing patent applications in 2005.

Assumptions

  • The propensity to patent is assumed to influence observed patent quality.
  • Invention value is treated as a random variable following an exponential distribution.
  • Firms' decisions to patent are influenced by a specific value threshold.