Learning to be Edison? How individual inventive experience affects the likelihood of breakthrough inventions

raffaele conti
bocconi university, management department

alfonso gambardella
bocconi university

myriam mariani
bocconi university

     Full text: PDF
     Last modified: January 19, 2010

Abstract
Learning to be Edison? How individual inventive experience affects the likelihood of breakthrough inventions

Raffaele Conti, Bocconi University
raffaele.conti@unibocconi.it
Year of enrolment: 2006. Expected final date: September 2010

By using data on 6943 European inventors, we study how inventive experience affects their ability to generate breakthrough inventions. For an inventor, the likelihood of producing a technological breakthrough depends on the number of inventions produced, on one hand, and on the probability that any of these inventions will be path-breaking, on the other hand. Our theory posits that experienced inventors produce a larger number of inventions, but each one of them is less likely to be a breakthrough. The former effect is a natural consequence of the impact of experience on inventive productivity, while the latter effect stems from myopia due to the exploitation of a well established technological path. The net effect of experience on breakthroughs is positive, since path-breaking inventions are largely unpredictable and thus producing many patents is the most effective strategy to achieve them. We first show, through a conditional logistic regression, that experience has a negative effect on whether any given invention is a breakthrough. A negative binomial regression with fixed effect then shows that the rate at which new inventions are produced increases with experience; finally, by using a fixed effect panel logit model, we find that expert inventors are more likely to achieve a breakthrough. We conclude by discussing managerial implications about how to organize the production of ideas in order to enhance the generation of breakthroughs.