Yesterday it was Telephone and IBM. Tomorrow smaller and more vulnerable companies face Government action. 70
The cold light of hindsight compels a more positive evaluation. In the decade following the AT&T and IBM patent licensing decrees, constant-dollar company-financed industrial R&D expenditures increased at an average rate of 6.5 percent per year, and private sector output per work hour grew at the record-setting rate of 3.1 percent per year. 71 Both indicators grew much more slowly between 1967 and 1977, when compulsory licensing was ordered much more sparingly: 3.1 percent for R&D and 1.6 percent for productivity. Nor need the analysis turn on macromagnitudes, whose movements were affected by many other influences of potentially greater importance than antitrust. Concerned about the IBM and AT&T decrees, eight fellow Harvard Business School students and I undertook in 1958 a study of how compulsory licensing had affected business firms' R&D decisions. Among other things, we interviewed executives at twenty-two patent-holding corporations and obtained mail questionnaires from sixty-nine more. 72 To our surprise, we found that actual or feared compulsory licensing had very little impact on large U.S. corporations' investments in new technology, although companies saddled with decrees did move significantly toward more reliance on secrecy and less patenting. A follow-up study in 1977 using newly disclosed data found no support for the hypothesis that forty-four companies operating under significant compulsory patent licensing decrees spent less on research and development per dollar of sales than 635 corporations not similarly affected. 73
Thus, a massive antitrust attack on business firms' use of patents to monopolize markets or enhance profit returns appears to have had negligible adverse consequences for the vigor of innovative activity in the United States. It seems unlikely, therefore, that much more narrowly targeted actions with respect to companies like Xerox and du Pont could have generated widespread disincentive effects. And more generally, in my many years of looking, I have been able to find no evidence that the broader array of U.S. antitrust policies has retarded innovation to any significant or extensive degree.
There are a few things we know about the links between antitrust and efficiency, and too much we do not know. We know that many discussions of antitrust and efficiency have violated the New Testament injunction against beholding the mote and ignoring the beam. 74 X-efficiency is much more important quantitatively than allocative efficiency, and dynamic efficiency is almost surely even more important. We know that the links