For two of my dissertation chapters I am particularly interested in how the incentive effects of IP change when different inventions are not independent, but complementary:

The economic literature on IP has yet to catch up on this, however: compared to the well-developed literature on the effects of IP protection of traditional "discrete" inventions, the analysis of IP and "complex products" is still in its infancy.

The first chapter of my PhD dissertation contributes to the literature on the decision to patent as well as the measurement of innovation by using patent data. It is generally acknowledged today by far that not every innovation is patented, but the decision to patent is considered to be a purely discrete decision: an innovation is either patented or not, there is no "extent" or scope of patenting. If a firm has made the decision to patent an innovation, it is therefore assumed that this results in the whole innovation being patented, and thereby disclosed to the public.

At the same time, innovation scholars have acknowledged for at least two decades now that the nature of innovation is changing. An innovative product today is often not the result of a single, solitary invention, but rather the combination of multiple inventions. The research behind these inventions is often independent, and therefore the results can be independently patented. In order to introduce a new product to the market, though, a firm needs to combine these separate inventions. Such multi-component products are often labelled a "complex" product, and the underlying innovation activities are complementary in that they obtain (most of) their value from being combined with other innovations. This kind of complementary inventive activity is particularly evident in consumer technology sectors like computers and smartphones, but more and more lower-tech industries are characterised by this as well.

This dissertation chapter combines these two ideas: innovations consist of more than a single patentable invention, and each invention can be patented, but does not have to be. It thereby provides evidence of an "internal margin" of the patenting decision.

To do this, I require the following two main ingredients: patenting firms need to be induced to change the extent to which they patent, and we need a measure of the inventive content of patents. For the latter, I resort (mainly) to the number of "independent claims". Inventions are described in a patent in two important sections: the "description", which is aimed to explain "from engineer to engineer" how the invention technically works, and the "claims", which are legally binding definitions of what precisely is protected by the patent. If someone files a court suit for patent infringement, the court determines whether the patent is indeed infringed by comparing the alleged infringer's product against the claims of the patent. Furthermore, claims are examined (by the patent office at time of application) and considered (by the court when determining possible patent infringement) one-by-one. Infringing on claim 1 says nothing about whether claim 2 is infringed as well; in this respect the claims are entirely independent. There are further distinctions to be made between independent and dependent claims, which I distinguish from each other, and go into more detail about in the paper.

For the change in patent incentive, I exploit changes in US trade secret law that lowered the incentive to patent as it became easier for companies to keep business information secret. In reaction to this lower patenting incentive, the amount of information and "inventive output" (measured by the number of independent claims per patent) included in any patent document decreased. It is possible to show that this decrease happens at different firms than the decrease in the number of patents.

First, this means that there was a considerable decrease ongoing in both the protection provided by the US patent system as well as the amount of inventive information disclosed to the public that has so far not yet been considered. Second, I demonstrate how this endogeneity of patent content can affect the measurement of innovation output using patent data.

The second chapter is based on work that I started while working at Bochum. Shifting of legal fees to the losing party of a court suit has been suggested in the US as a remedy against the problem of "patent trolls", non-producing firms that acquire patents with the sole goal of exacting infringement damages from producing firms. This chapter takes a look at the effect of shifting of legal fees in patent litigation on the patenting incentive of producing firms of different size. The theory model shows that firms unambiguously decrease their patenting activity, and small manufacturing firms are hit harder when competing against larger potential imitators.

A second important result is that if we allow for the possibility of settling litigation before court judgement, and the extent of fee-shifting is large enough, then there is an incentive for very small firms to patent while anticipating "being infringed". This puts doubt on the effectiveness of legal fee-shifting as a policy tool against patent trolls. I am currently looking into the possibility of calibrating the model in order to provide more precise predictions for different industries.

The third chapter explores the impact that delayed publication of patent applications has on competing firms' R&D decisions. A widespread assumption in the theoretical economic literature on innovation and patenting is that once one firm applies for a patent, this immediately becomes publicly known. In reality, however, patent offices publish filed patent applications with a delay of 18 months. As long as a patent applicant does not announce the existence of an application themselves, its existence is private knowledge to the applicant (and the patent office). I analyse effects of this delayed publication in stand-alone R&D competition, cumulative and complementary innovation, and identify situations in which a patent application has an incentive to disclose the existence of a pending patent before the patent office would do so.