Estimating the Value of Creative Invention

29 June 2007 at 3:43 am hsiehc Leave a comment

| Chihmao Hsieh |

One final bit of shameless self-promotion, as I soon head off into the figurative sunset from all this blogging: an empirically-oriented working paper of mine entitled “The Identification of Opportunities and the Value of Invention” can now be found here at SSRN. Any comments or suggestions are absolutely welcomed! Description of the research lies below.

ABSTRACT: This paper highlights entrepreneurial opportunities as fundamentally composed of inventions, and following recent work it begins to investigate how the value of inventions might be estimated. Inventions are composed of sets of components, each marked by both a degree of interdependence and a degree of relatedness. Specifically, I distinguish between the degree to which connections between pairs of components are interdependent upon one another as regards value, and the degree to which components are related (i.e. ‘connected’). Those creative inventions made up of a set of moderately related components are most likely to be highly valuable. Sets of barely-related components are not likely to offer valuable configurations, and sets of highly-related components amenable to creative thinking are likely to correspond to an overwhelming variety of configurations, the most valuable of which cannot be found given boundedly rational actors. I also hypothesize that the expected value of a creative invention increases as more components are recombined, because more configurations are available. However, the rate is decreasing, again due to bounded rationality. Finally, like Fleming and Sorenson (2001) I also hypothesize that repeated search using familiar components leads to discovery of increasingly valuable inventions.

The empirical analysis supports two of the three hypotheses. Upon measuring relatedness via the identification of ‘connections’ among components, I find that creative inventions (as indicated by successful patent application) with moderately related components are most valuable, on average. I also find that the expected value of a creative invention increases as more components are recombined, but at a decreasing rate. However, the last hypothesis is not supported. According to patent data, repeated search using familiar components appears to lead to the discovery of decreasingly valuable inventions: for any given set of components, the most valuable inventions are found first.