Arrow revisited
I recently had a chance to reread Kenneth Arrow’s 1959 classic paper, “Economic Welfare and the Allocation of Resources for Invention.” Originally written for the RAND Corporation (and now available on RAND’s website), a version of this paper was included in the 1962 book edited by Richard Nelson, The Rate and Direction of Inventive Activity.
This paper is often cited as arguing that patents facilitate technology transactions and a market in technology, that patents “resolve the information paradox for inventions” (the paradox that a buyer needs knowledge of the invention in order to value it, but once that information is revealed, the buyer may have already acquired much of that value).
I was struck by how Arrow identified so many subtle details about innovation that I missed on my first reading and which now seem so important. And Arrow’s analysis is hardly a simplistic “resolution” of the information paradox.
This paper was part of Arrow’s larger general equilibrium project, so it looked at the economics of invention as a problem in the economics of information. Although this is an abstract theoretical treatment, Arrow nevertheless identified many practical details about the invention process, including the possible significance of diverse research approaches, the cumulative nature of much innovation, and the difficulty of drawing property boundaries corresponding to inventions.
While recognizing that patents may improve technology trade to some degree, Arrow also puts his finger on some of the reasons why this improvement is likely to be limited. Treating invention as a form of information, he writes,
With suitable legal measures, information may become an appropriable commodity. Then the monopoly power can indeed be exerted. However, no amount of legal protection can make a thoroughly appropriable commodity of something so intangible as information. The very use of the information in any productive way is bound to reveal it at least in part. Mobility of personnel among firms provides a way of spreading information. Legally imposed property rights can provide only a partial barrier, since there are obviously enormous difficulties in defining in any sharp way an item of information and differentiating it from other similar-sounding items.
And later,
…the inventor will have in any case considerable difficulty in appropriating the information produced. Patent laws would have to be of an unimaginable complexity and subtlety to permit such appropriation on a large scale.
Although this idea was not fully developed (nor were the other ideas mentioned above), it is nevertheless impressive how a short theoretical treatment managed to identify, if but briefly, key difficulties in technology markets and property rights.
