1. Okay, then, let's do a study about how much time does independent invention need. Then patent durations must be tuned to that. For instance, if most independent re-inventions takes 2 years, then we shouldn't grant patents for more than 3 of 4 years. I suspect that according to this criterion, current durations are much, much too high.
2. I just take for granted that all other things being equal, having a monopoly is much worse than not having it: the (diffuse) cost to society at large is much higher than the (concentrated, visible) benefit to the monopolist.
I take it for granted that it is not defensible to have the society pay a significant cost just to reward a lone entity, no matter how much he "deserves a reward". Global costs must be balanced by global benefits. The very existence of an invention is such a global benefit. But unless the patent helped get the invention out sooner, that benefit is void, and the patent, unjustified.
Also, apparently, we haven't read the same "tons of studies". The arguments againts patents I have read so far (Against Intellectual Monopoly was a good read), sounded very compelling —not to mention grounded in evidence. Could you give me some pointers to the studies you speak of?
The term of a patent has nothing to do with independent invention, it is meant to allow an inventor an opportunity to capture the rewards of their invention. Both licensing and product development (even for software) can take many years to happen, and 4 or 5 years is laughably inadequate. Not to mention, of course, that patent offices take a few years for a patent to grant in the first place. What advantage is a patent if it expires just in time for others to start ripping you off?
In any case, I am not aware of any quantitative study of independent invention. But using patents as a sample, considering only 0.1% being involved in interferences and that only 0.1% of issued patents ever being asserted, one rough guess could be that most patentable ideas are never independently re-invented. On the other hand, given how narrow patents usually are, this is not surprising. Then again, most patents are licensed without litigation, and these deals are confidential, so this is a really rough guess.
> … Against Intellectual Monopoly …
Heh, I guessed that would come up. Unfortunately, that book contains many fabrications and mischaracterizations of other work, which calls into question anything they say. I've already pointed it out once:
It's unfortunate that many people believe their work without critical thought because it confirms their biases. The second comment linked above links a few studies showing the beneficial effects of patents and ends with a reference to a more balanced and honest overview of the current state of patent economics. I'd encourage you to read those instead.
> The term of a patent […] is meant to allow an inventor an opportunity to capture the rewards of their invention.
This is not a valid justification. Granting the inventor a monopoly costs more to society than it benefits the inventor himself. As such , the inventor doesn't matter. His invention does.
What would be a good enough justification, is if granting the patent hastened the creation of the invention. To me, it is not clear this is generally the case. First, there is this "independent invention" thing. Second, there are other ways besides monopoly. Being first to market can be a pretty good incentive for innovation. Third, patents can slow down incremental inventions —improvements to patented ideas. Fourth, granting a monopoly to the inventor could encourage him to get lazy, and stop or slow down his output. Fifth, inventors could seek out monopoly-friendly innovations, instead of more useful ones —rent-seeking behaviour. Sixth, in practice, many patents are filled after the invention has been completed for other reasons.
Finally, there's the sheer bookkeeping costs of the whole patent system. 0.1% of patents only are being asserted? Does that include licensing deals as well as litigations? Anyway, there's the patent office, the effort of writing the patent in the first place, patent lawyers, the judges and juries that take time to rule litigations, and of course any effort directed at patent circumvention (compression algorithms have several good examples). Such overhead is not trivial.
Innovation fostering has to be significant enough to compensate for the various costs (monopolies and overheads). This sets the bar rather high in my opinion.
> Granting the inventor a monopoly costs more to society than it benefits the inventor himself.
You cannot make that assertion without empirical evidence. You bring up good points about the costs of patents, but note that your language ("can", "could") is all in terms of potential.
The studies I've read consider all those points and then try to quantify them, even going as far as seeing how they play out in different industries. Other studies raise points which show the benefits of patent systems (increased R&D spending, increased diversity in the areas of innovation, increased probability of securing VC funding) and quantify those too. Other studies actually find counter-intuitive results about monopolies in general (e.g. monopolies can be more innovative than startups by some metrics -- think Bell Labs, Xerox, MS Research -- which casts doubt on the "monopolist gets lazy" argument). There are tons of studies presenting interesting results, and of course, there are studies that point out flaws in these.
The problem is, it is impossible to compare these results quantitatively because the data, methods, assumptions, interpretations, shortcomings and metrics are wide, varied and often orthogonal. How can you meaningfully compare a lower rate of stitch / minute improvements (that was used as a proxy for the cost of sewing machine patents) with the increased diversity of industries participating in a technical convention (which was used as a proxy to measure the beneficial effects of introducing a patent system)? And to top it off, both studies were by the same author (Petra Moser)! This is just one example -- there are as many metrics as there are academics researching this area.
2. I just take for granted that all other things being equal, having a monopoly is much worse than not having it: the (diffuse) cost to society at large is much higher than the (concentrated, visible) benefit to the monopolist.
I take it for granted that it is not defensible to have the society pay a significant cost just to reward a lone entity, no matter how much he "deserves a reward". Global costs must be balanced by global benefits. The very existence of an invention is such a global benefit. But unless the patent helped get the invention out sooner, that benefit is void, and the patent, unjustified.
Also, apparently, we haven't read the same "tons of studies". The arguments againts patents I have read so far (Against Intellectual Monopoly was a good read), sounded very compelling —not to mention grounded in evidence. Could you give me some pointers to the studies you speak of?