Entrepreneurial Innovation: Patent Rank and Marketing Science
Shaffer, Monte J.
MetadataShow full item record
This dissertation examines innovation from the Austrian-economic perspective of the entrepreneur utilizing patents as a proxy. For over 60 years, scholars in marketing, strategy, and economics have been looking for a way to link patents to radical innovation; I present Patent Rank as that link. Patent Rank defines a patent's intrinsic value by considering its legally-defined relationships within the entire patent network; a patent's value is a function of its ancestry (backward citations), its heritage (forward citations), and the dynamics of the patent network (it updates over time). Using over 5.6 million patents, 320 million annual Patent Rank scores, and 100 million nonlinear models, I examine patent innovations at the patent level, innovation level, firm level, industry level, and market level. In the first essay, patents are defined as innovations using legal first principles; Patent Rank is presented based on network theory and demonstrated to be superior to the forward-citation-count metric. In the second essay, the mathematics of Patent Rank is further explored; a generalized model is derived based on the sorting, augmenting, partitioning, and solving of a linear system. A marginal-structure (ms) model is utilized to demonstrate that the Austrian-defined market process of entrepreneurial innovation exists: both Schumpeterian shocks and Kirznerian competition. In the third essay, a marginal-combined (mc) model is employed and a nonlinear model is derived to predict a patent's expected lifetime value, and ultimately a firm's annual patent stock from its portfolio of patents. In the final essay, Patent Rank is asserted to be an objective measure of radical innovation. This claim is validated using face validity (top patent innovations are radical), concurrent validity (top patent innovations align with Wharton experts), and predictive validity (abnormal market portfolio returns correlate with different levels of patent stock).