Elisabetta Ottoz, Franco Cugno
University of Turin

Copyright © 2008 by American Law and Economics Association; Elisabetta Ottoz, Franco Cugno

Different protection mechanisms may be employed at the same time when an innovation is comprised of separately protectable components. If patents and trade secrets can be mixed in protecting single innovations, a strengthening in patent breadth may induce a lower level of patenting, as innovators are more prone to rely on secrecy. (JEL K00, L49, O34)

1. Introduction

According to Cohen, Nelson, and Walsh (2000), for certain purposes the key difference between a complex and a discrete technology is whether a new product or process is comprised of numerous separately patentable elementsversus relatively few.
It is then well known that electronic products tend to encompass a large number of patentable elements--often hundreds--and, hence, may be characterized as complex. But product complexity is not a sector-specific feature: on the basis of an estimation of the number of patent applications per innovation, conducted on a survey of one thousand one hundred and sixty-five US firms realized in the mid-1990s, most innovations appear to be protected by more than one patent (Cohen, Nelson, and Walsh, 2000; Arora and Ceccagnoli, 2005). While the number of US patents per innovation is relatively small in biotech and pharmaceuticals (around two), it can rise to more than seven in fields such as semiconductors, transportation or rubber products, being on average 5.6 (Lévêque and Ménière, 2007). This suggests that product complexity--in the sense specified above--is pervasive, although its incidence differs among sectors.
In the case of complex innovations, firms can rely upon more than one mechanism to protect one product. Under some circumstances there is no choice, in that certain components are not patentable, but often the type of protection constitutes an option for innovators, who can choose the extent patents and trade secrets combine with one another. So, trade secret protection may be important not only during the pendency of a patent application, but also during the term of, or after the expiration of, the patent. As trade secret protection is relinquished to the extent an invention is disclosed in a patent application, there is sometimes motivation to minimize the disclosure made in a patent application in order to obtain broad patent protection and yet retain significant trade secret protection. In software terms, for instance, this can mean a patent disclosure that does not reveal any code.
In this paper, we present a model where an innovator, who possesses all the complementary pieces of the new technology and uses them directly, can choose the patent--secret mix. Assuming the technology proprietor aims at maximizing the present value of profits flows deriving from the chosen mix, we determine the conditions pertaining to patent and secret strength and to patent length allowing for an internal solution, that is to say a solution where patents and secrets actually protect a single product. We will see that, perhaps counterintuitively, an increase in the level of patent protection may induce the innovator to rely more on secrecy. Our results are obtained in a simple setting where signaling, à la Horstmann, MacDonald, and Slivinski (1985) or Anton and Yao (2004), plays no role.
The core intuition for our conclusions is that an increase in the patented and disclosed knowledge decreases the likelihood that a rival will invent around the patented knowledge, but increases the probability that the trade secret leaks out (since there is less knowledge that must leak). To enter the market as a competitor, the rival must do two things: circumvent the patent and get access to the secret. Rather reasonably, we assume that the probability of these two events together decreases with the disclosed knowledge, while the probability of keeping the trade secret forever also decreases with the disclosed knowledge. Because of these opposing effects, the optimal disclosure is (for many parameters) somewhere between none and all of the knowledge.
A second argument is that an increase in patent breadth (interpreted as a reduction in the probability of circumventing the disclosed knowledge, holding the amount of disclosure fixed) has ambiguous effects on the amount of knowledge disclosed. As for the first argument, two opposing effects operate, this time related to the policymaker's choice of patent breadth. First, when patent breadth increases, innovators tend to substitute patent protection for trade secret protection. But, second, an increase in patent breadth implies that innovators can enjoy the same total patent protection as with a lower patent breadth by disclosing a lower fraction of knowledge. This induces the innovator to rely more on secrets.
Section 2 contains a discussion of some legal aspects about the coexistence of patents and secrets and some historical examples. Section 3 expounds the model. The last section presents some concluding remarks.