Is the patent system worth the barriers to entry it provides?

The answer is it depends, and the fact that it depends is what makes patent reform both complicated and interesting.

For purposes of this post, I'm going to divide the world of new technology into two boxes: "Computers" and "Pharmaceuticals," although I'm abstracting so far away from the details of the technology that it would almost be more fair to call them "Category X" and "Category Y."  The distinction I want to draw is a distinction between technology that is very difficult to reverse-engineer and technology that is very simple to reverse-engineer.  It is easy to observe that there is some software technology, for example, that is relatively difficult to reverse-engineer starting only from the publicly available versions of the software.  By contrast, by the time a pharmaceutical receives FDA approval, it is usually very easy to produce the exact same pharmaceutical compound using a slightly (or even substantially different) process.

How is this important to patent reform?  The difference explains why many (or most) engineers and business people in computer-related technology see the patent system as a burden, drawing off resources that would otherwise be devoted to innovation whereas pharmaceutical companies see patent law as a boon.  For computer-related technology the barriers to comptetitive entry are usually already high without the exclusivity of patent protection.  Ronald Coase observed how WIlliam Knudsen's human capital created competitive advantages and eventually barriers to entry for Ford by streamlining automobile manufacturing and assembly.  Similarly, computer companies gain competitive advantages and eventually barriers to entry by programming or building superior technology at the back end.  It was not easy to reverse engineer and duplicate what was done at Ford because it was the human capital rather than regular capital that provided the competitive advantage.  Something like that is also true of computer companies today, and perhaps of emerging technology markets in general.  While the markets are still immature and the technology is rapidly evolving, human capital becomes the key to achieving a competitive advantage and market share.  Differences in legal rights can become a rounding error.  Even differences in (regular) capital structure may be irrelevant.

Notice, however, that the opposite may be true of more mature technology markets, or markets where the technology is harder to reverse-engineer.  Without patent rights, the economies of scale in pharmaceutical markets may result in only a handful of very large, relatively undifferentiated competitors, and hence slower and less innovation.  A newer or smaller pharmaceutical company may need a guarantee of exclusivity in order to attact the (regular) capital necessary for completing the R&D process for a new drug or therapy.  Suddenly a system of patent rights doesn't seem like such a bad deal for investors in such a startup.

This is why patent reform is difficult.  Both sides have good points to make.  Now how do we resolve these differences?  The answer is we can't right now, but we should be able to over time.  The problem for computer companies is not that exclusive rights exist, but rather that the transactions costs of the system are relatively large compared to the marginal benefits they provide (such as the increased liquidity in the human capital of the engineers that work at these computer companies relative to a world in which only bilateral contracts and trade secret torts control how their ideas get used).  But transactions costs are going to continue to decline as the Internet grows, and the costs of our system of testing the valid and infringing scope of patents continues to decline with Internet-enabled technology.  Twenty years ago, it cost on average a few hundred dollars just to do a patentability search.  Now it's essentially free.  Similarly, patent ownership information is now readily available.  The patenting process is also growing more efficient as the PTO brings Internet-technology to bear on the problem of prior art searching.

If anyone from Congress is reading this post the message is this: Please don't demolish the patent system.  It's working for some right now, and for others, it's bound to become more efficient over time.

UPDATE: A recent empirical study supports my premise for the distinction between "Computer" and "Pharmaceutical" patents -- i.e., the relative ease of reverse-engineering.  The Internet, of course, is a technology that makes reverse-engineering easier.