Remarks

Richard C. Levin

Economics Department Reunion
April 18, 1999

My interest in technology and the economy derives from an earlier interest in economic history.  I encountered Bill Parker in 1970 at the stage of his career when he had ceased to care about increases in crop yields and had begun to reflect upon the big picture.  Long before the computer made such things easy, Bill was constructing elaborate diagrams depicting the complex relationships and feedbacks among technology, economic activity, demography, social structure, religion, political and legal institutions, art, and culture.

It is true that Bill's "big picture" pictures suffered, aesthetically, by comparison with John Fei's inspired use of colored chalk.  Nonetheless, fresh from a year's work on Max Weber, I found them intriguing, and it is with reference to Bill's framework, eventually described in print more than a decade later, that the principal work of my own foreshortened career as an economist can be best assessed.

Encouraged by Dick Nelson to relocate my interest in technology from the 18th to the 20th century, I wrote a dissertation concerned with the interaction of technical change and industrial structure.  Historians had long focused on the influence of the first on the second, but economists of the 1960s, following Ken Arrow's lead, were entirely focused on the reverse direction of influence.  By emphasizing technical change that altered the optimal scale of production, my dissertation was probably the first, and certainly the clumsiest, effort to model how technical change affects market structure even as market structure feeds back on the incentive to innovate.  I also found, in a variety of chemical industries, empirical evidence suggesting that the phenomena my models described were not entirely imaginary.  I suppose I thereby made some small contribution to understanding 5 to 10% of the Parkerian universe -- namely the interplay between two of its elements: technology and economic activity.

After an interlude working on railroad regulation, the oil industry, and some antitrust issues, I returned to the study of technology in the early 1980s by attempting to bite off a bigger chunk of the Parker picture -- namely the effects of law, specifically the system of intellectual property rights, on the relationship between technical change and economic activity.  Much of this work -- undertaken in collaboration with Al Klevorick, Dick Nelson, and Sid Winter -- involved mining the data set that has come to be known in industrial organization as the Yale Survey.  The project was actually a full employment act for graduate students, back in the days when it was possible to support four or five graduate students at a time on NSF grants.

Among the things we learned from our survey of 650 R&D executives in 130 industries was that the effects of the legal regime on market structure and economic performance depended in important ways on the nature of technology itself.  It was also clear, and has become more clear these past ten or fifteen years, that changes in the legal regime have been driven, at least in part, by changes in technology and market structure.   I mention this merely to observe that all the possible interactions and reciprocal influences among an expanded set of elements -- technology, economic activity, and legal institutions -- are empirically discernible.  To sum up, before I left economics for administration my work cast a dim light on perhaps 10 to 15% of the Parkerian universe.   One might with generosity conclude that I abandoned research in the field too soon.   But let's not kid ourselves, I was headed toward an asymptote somewhere below 20%.

One of lessons of our R&D survey was that patent protection provided a strong incentive to innovate in only a small subset of manufacturing industries -- specifically those where an invention could be both very precisely described and, in the absence of property rights, easily imitated.  To be concrete about this, patents seemed to serve their theoretical purpose especially well in the various chemical industries, including pharmaceuticals, where inventions were typically discrete molecules -- easily described, easily imitated in the absence of property rights, easily detected when produced by infringers.

We also learned that numerous other industries had rapid rates of technical advance without strong intellectual property rights.  Indeed, on close examination of some sectors, notably microelectronics, it seemed that rapid technical advance depended on routine cross-infringement of the property rights of others.  Most microelectronic innovations were cumulative, incorporating the features of previous innovations in their design.  The same attribute is present in software.

Our R&D survey was conducted fifteen years ago, at a time when the intellectual property regime had just begun to undergo a massive transformation.  In 1980, the Bayh-Dole Act permitted universities, small businesses, and nonprofit contractors to patent the results of research supported by government grants.  In the same year, the Supreme Court established the patentability of genetically-engineered life forms.  In 1982, a special appellate court -- the Court of Appeals for the Federal Circuit -- was established to hear all patent cases, a development that has tended to strengthen the position of inventors in litigation with alleged infringers.

The trend has continued unabated.  In 1984, the Semiconductor Chip Protection Act established property rights in integrated circuit designs, and the Hatch-Waxman Act extended the patent term of regulated pharmaceuticals.  In 1988, legislation blocked the importation of foreign products produced by methods infringing U.S. process patents.   In 1994, an international agreement strengthen intellectual property rights in certain nations by moving toward harmonization of national laws.  In 1996, the Economic Espionage Act made the theft of trade secrets subject to Federal civil and criminal penalties for the first time, and, in 1998, the Court of Appeals for the Federal Circuit upheld the patentability of business application software.

The cumulative effect of all this legislative and judicial change has been to strengthen substantially the rights of those who possess intellectual property.  In theory, this should enhance the incentive to innovate, and one might expect, as a first order effect, an increase in the rate of technical change, even if greater dead weight losses from prices in excess of marginal cost attend.  But both the manner in which intellectual property rights are strengthened and the nature of technology itself may attenuate or eliminate the first order positive effect of stronger rights on innovation and economic performance.

Consider, for example, the strengthening of trade secrecy protection. This may enhance the appropriability of innovation, but it ignores the other principal objective of patent law, which is to place new knowledge in the public domain and make it available to others.   In theory, those with a more socially productive use of patentable knowledge will buy (that is, license) it from those who own the rights, and social welfare improves.   If knowledge is kept secret, however, it will, other things equal, be used suboptimally.

Consider also the possible effects of stronger intellectual property rights in technologies that are inherently cumulative, like microelectronics and software.   Here yesterday's knowledge is a building block for today's advance, and whether stronger rights enhance incentives may depend on how well markets for these rights function and on how efficiently courts resolve disputes concerning ownership.  In electronics, the old patterns of widespread cross-infringement (or royalty-free cross-licensing) have been replaced by new patterns of widespread cross-licensing in which substantial payments are made, while the cost and penalties involved in dispute resolution have increased dramatically.

I would submit that we don't really know if the benefits of stronger intellectual property rights, as they have emerged in the past twenty years, outweigh the costs.   This is an interesting and important question, with obvious implications for public policy.  In the hope of placing the issue on somebody's research agenda, it will be the subject of a conference in Washington next year sponsored by the Science, Technology, and Economic Policy Board of the National Research Council, which I joined recently and where I have the luxury of suggesting topics for research rather than actually rolling up my sleeves.