Patents, market structure and welfare: A theoretical investigation into new dimensions of the patent system.

Chapter I: The relationship between inventor and the Patent Office is modelled as a 'patent regulation game' and it is shown that the conventional wisdom that the P.O. always maximizes welfare by playing the Stackelberg leader is incorrect. Other solution concepts are explored and it is found that, because of the patent life constraint, a reversal of roles may be beneficial. The result that social welfare can be maximized by the P.O. being a Stackelberg follower survives (albeit for a narrower range of values of the key parameters) even if the P.O.-leader is endowed with the additional instrument of a compulsory royalty rate. Chapter II: A new twist is added to the debate on the Schumpeterian competition hypothesis, by considering the structure of the final-product market as a policy Instrument, set by the Patent Office by manipulating patentability standards. It is found that for a vast range of demand functions and under constant returns to scale, a patentability standard that allows for more than one patent to be granted within a given product/process class is welfare superior to the monopoly-generating first-past-the-post current system. If patent life is beyond the P.O.'s control and/or there are increasing returns, no patentability standard is unambiguously preferable. Chapter III: When Research and Development are modelled as two analytically distinct stages, the choice between patentability standards (whether to grant patents to research prototypes or to fully-developed products) is shown to affect the allocation of resources between Research and Development. It is shown that under a single-patent regime, granting patents to research prototypes is unambiguously welfare-improving, whereas under a multiple-patent regime a change to patents being granted to fully-developed products and the attending increase in market uncertainty may raise welfare. Chapter IV: The economics of the 'integer constraint' is analysed and it is found that proper treatment of the indivisibility of firms may reverse the qualitative conclusions of interger-unconstrained models. As an example, a product quality oligopoly model is examined and it is shown that not only the Chamberlinian excess entry result does not apply but also that a free-entry oligopoly and a socially managed industry may produce goods of identical quality, irrespective of the values of cross-derivatives deemed crucial in the literature. Moreover, the integer constraint is shown to provide an explanation for a positive correlation between profitability and concentration in a Cournot oligopoly model with free entry