1. Some researches have been unable to find evidence of a causal relationship between patent reform and the rise in patents (Kortum and Lerner, 1998). Others acknowledge a link, but contend that the increase in patenting may be socially wasteful (Hall and Ziedonis, 2001), reflect inventions that are undeserving of protection (Bessen and Maskin, 2000) or impede future progress (Heller and Eisenberg, 1998). They caution that the recent U.S. patent reforms may have gone too far in protecting intellectual property rights.
2. Aim of the paper: After providing some background on the U.S. patent reforms, I evaluate the theory and evidence on the extent to which stronger patents stimulate innovation, encourage firms to disclose their inventions and facilitate efficient technology transfer. Administrative, legal and other transaction costs that arise from extending patent protection are identified, along with proposals for alleviating them.
3. Own aside: constant comments of the form: 'the conventional model', 'the original model'. Where this suggests some weighty, well derived and robust system derived after much research. In fact refers to Arrow (1962) / Nordhaus (1969) model that is very simple and rather trivial. It has straightforward monotonic relationship (not derived but assumed) between protection and innovation and it derives its result from trading this off against the monopoly costs.
4. Also constant elision (despite own warnings) between patent applications and innovation. E.g. p. 138: 'In an international analysis of the relationship between patent strength and innovation, Lerner (2001) examines 177 policy shifts in 60 countries over 160 years.' Yet all the policy shifts are shifts in the patent system (coverage, cost etc). The dependent variable (for innovation) is then the number of innovations issued. BUT this is ALL WRONG. change in patent system alters a) the propensity to patent b) the incentive to innovate (and thereby level of innovation). You are always in great danger of confusing a) and b). e.g. might get increase after change but how do we know this isn't just that more innovations are patented rather than that there are more innovations.
5. The argument least favorable toward patents is that they are ineffective instruments for protecting inventions. Supporting this explanation is survey evidence that, except in a small number of industries (pharmaceuticals, chemicals, medical equipment and some machinery industries), patents are considered less effective relative to alternative mechanisms for protecting intellectual assests, such as secrecy and lead time (Cohen, Nelson and Walsh, 2000). see also notes on dosi_1988.
6. Value of disclosure: How valuable is the information disclosed in patent applications? According to a survey of U.S. and Japanese .rms, the information spillovers from U.S. patent applications are relatively low (Cohen et al., 1998a). One reason is that for new, rapidly changing technologies, patent information is largely outdated by the time the patent is granted.Moreover, researchersmay intentionally refrain from reading patent applications, since prior knowledge of inventions may expose them to the risk of treble damages in patent infringement suits.
7. Own conjecture: patents become self-fulfilling. consider simple example: payoff matrix for innovator X

   payoff for X in each cell	X: no patent	X: patent
   other firms: no patent	5	6
   other firms: patent	0	6

   cost of obtaining patent is 2 say. So net payoff is:

   payoff for X in each cell	X: no patent	X: patent
   other firms: no patent	5	4
   other firms: patent	0	4

   Thus it the firm may well patent to protect itself but if this misses out on the best solution of 5 when no patents are possible