[Patents] Bessen & Maskin reply on software patent paper

[Jim Bessen jbessen@researchoninnovation.org]

Greg

(forgive me if I am repeating arguments that have been made before...I am
new to the list)

>
> Come on Jim, neither side has any clear and convincing evidence of the
> impact patents economically (outside the pharmaceutical industry). The
> impact of patents is in general a minor financial concern for companies
> (as compared to other business affairs), so measuring such a subtle
> influence is difficult, either to prove patents are good or bad.
>
If patents have no significant economic effect, then why have them at all?
Given that the software industry has been innovative without patents, the
burden of proof falls on those who seek to extend government involvement.

The argument for patent extension in Europe has largely been based on the a
priori "stronger is always better" argument, not on any empirical evaluation
of policy. But you have a somewhat different argument...

[BTW IBM does claim about $1 billion/year in licensing revenues from all
fields...this is not minor]

> The pro-software patent side has an easy case to make - all they have
> to show is that the impact of software patents in the software industry
> is comparable to the impact of patents in other areas of technology.
....
>
> The anti-software patent side has a harder case to make - to somehow
> show (with real data, not ancedotal stories and poorly researched
> press clippings) that software patents are disrupting both the
> software industry and consumers greater than in other industries.

Not true. The problem is this. Software patents do not have a substantial
negative impact YET in the U.S. But this does not mean that the 100,000
patents out there will not have a substantial effect in the FUTURE.

Indeed, I find it a bit strange for you, of all people, to argue that the
present U.S. patent regime will have no detrimental social impact now or in
the future. If this is so, then why call for better quality patents in
software?

It is, of course, possible that firms will not exert their patents in the
future. But clearly the firms have some economic motivation for spending
millions to obtain these patents. They apparently see either the possibility
of obtaining licensing revenues or of the need to defend against possible
patent litigation. If these firms are right, then there is, in fact, a
serious risk to future innovation.

The pro-patent forces in Europe need to demonstrate that there is real
benefit to be gained to offset this risk. Software patents represent a huge
implicit subsidy; the pro-patent forces need to show what society gets for
this give-away.

> And this is a hard argument to make, because the problems of bad
> software patents (and I agree there are way too many) for the most
> part pale in comparison to the problems with biotech patents (both
> for biotech companies that increasingly are blocked out of promising
> research areas because of patents, or consumers [especially in third
> world countries
> ] who find it difficult to be able to afford to buy the new drugs).
>

But then did you support these lousy biotech patents 10 years ago, before
there was actual evidence of blocking?  The logic is exactly the same:
firms, anticipating a legal opportunity/threat, spent millions obtaining
patents; a large number of bad patents were granted; these have been used to
block and litigate and things have begun to go bad and may get far worse in
biotech.

What makes you so sure there is no risk of a similar thing happening in
software? The patent problems in biotech hardly seem a good argument for
patents in software.

> That's why your paper is so disconcerting.  It models a very small
> subset of the dynamics of the software world, and analyzes a very
> small subset of software industry financial data, to make some rather
> strong conclusions about the impact of software patents.  I question
> then the statistical reliability of your conclusions.
>
 But again you are missing the point of the analysis. We did not try to show
that the software industry has been damaged by software patents. We showed
that the firms obtaining most of the software patents (and this is a small
subset) did not increase their R&D as the standard pro-patent argument
alleges.

The economic tally is clear enough from our limited data: society has
granted a small number of firms subsidies that may be worth billions. In
return society has not gotten a dime's worth of additional R&D. And in
addition, there is a significant risk of decreased innovation in the future.

As you pointed out, our paper does not argue that software patents are
necessarily bad. But I think our paper does make a good case against the
"stronger patents are always better" school. This has been the main argument
of the pro-patent forces in Europe and these forces have largely avoided
empirical arguments. As you know, these people squashed Brian Kahin's study
of software patent quality. This was my point to Francois.

You have raised another argument, that software IP should not be treated
differently from other fields. You have a point, but given that patents are
also screwed up in other fields, I am not sure this alone provides a
convincing reason for Europe to change its statute, at least not until some
of the other patent problems are solved.

I don't know what the best long-run policy is for Europe or for the U.S.
(which might well be different). However, if Europe does something other
than follow the USPTO whole hog (or at least delays), this may be the
clearest message that all is not right at the Crystal Palace (or wherever
they're moving).

Jim