This paper reports on an empirical venture into the darkest terra incognita of industrial organization analysis: the links between market structure and dynamic stability. Concretely, it seeks to determine whether capital investment expenditures were more stable between 1954 and 1963 in concentrated or in atomistically structured manufacturing industries, other things being equal. One can find numerous a priori conjectures on this subject, some suggesting that concentration is relatively conducive to stability, others the opposite. Representative of the first group is Tibor Scitovsky's assertion that pure competitors overreact to demand shifts because they ignore the direct and indirect effects of their investment on price. Oligopolists and monopolistic competitors are said to be less prone to overadjustment because they suffer no illusions about facing an unlimited market.' In a similar vein is G. B. Richardson's sophisticated new variant on the conventional cartel wisdom, holding that contrived market imperfections-overt price-fixing arrangements, large-scale mergers, and conscious parallelism-create an information system which helps call forth the proper amounts of investment by the right firms in response to demand changes.2 Others have implied that investment is likely to be particularly unstable under oligopolistic conditions. According to James Duesenberry, market position among oligopolists refraining from price rivalry depends upon such nonprice variables as reliability of service and personal buyer-seller ties. The firm with reserve capacity to serve customers in times of peak demand stands a good chance of retaining their loyalty permanently, and this motivates each oligopolist to try increasing its market share through competition in production capacity.3 To carry the logic a step beyond Duesenberry's explicit statement, fear of losing market share and future profit-earning potential to more aggressive rivals can lead oligopolistic producers to invest heavily in additional capacity when demand is growing. They may then cut back sharply when serious excess capacity appears. The net result is a volatile investment profile over time. In Joe S. Bain's somewhat different view, investment will tend to be unstable when barriers to new entry are modest while industry structure is sufficiently concentrated to support prices temporarily above the competitive level. Then producers may strive to maximize short-run collective profits, encouraging a wave of new investment which ends abruptly when the price structure collapses.4 A third possibility is that concentrated industries, with generally higher profits and cash flow, are better able to finance investment programs when funds markets are tight, and therefore will respond to increases in demand with a sharper acceleration of in* Work on this paper was supported by the Institute of Public Administration, University of Michigan. The author is indebted to Lowell M. Seyburn for research assistance, and to WV. H. Locke Anderson for critical comments. l Tibor Scitovsky, Welfare and Competition (Irwin, 1951), pp. 365-66. 2 G. B. Richardson, Information and Investment (Oxford: Oxford Univ. Press, 1960), pp. 67-70 and 128-36. I James S. Duesenberry, Business Cycles and Economic Growtl/ (McGraw-Hill, 1958), pp. 113-33. See also Donald H. Wallace, Market Control in the Aluminum Industry (Harvard Univ. Press, 1937), pp. 336-43. 4Joe S. Bain, Barriers to New Competition (Harvard Univ. Press, 1956), pp. 33-41 and 189-90.

For a seventeen-year panel covering 308 U.S. manufacturing corporations, we analyze firms' R&D spending reactions to changes in high-technology imports. On average, companies reduced their R&D/sales ratios in the short run as imports rose. Individual company reactions were heterogeneous, especially for multinational firms. Short-run reactions were more aggressive (i.e., tending toward R&D/sales ratio increases), the more concentrated the markets were in which the companies operated, the larger the company was, and the more diversified the firm's sales mix was. Reactions were less aggressive when special trade barriers had been erected or patent protection was strong in the impacted industries. Companies with a top executive officer educated in science or engineering were more likely to increase R&D/sales ratios in response to an import shock, all else equal. Over the full 17-year sample period, reactions may have shifted toward greater average aggressiveness.

Through a survey, private economic value estimates were obtained on 964 inventions made in the United States and Germany and on which German patent renewal fees were paid to full-term expiration in, 1995. A search of subsequent U.S. and German patents yielded counts of citations to those patents. Patents renewed to full-term were significantly more highly cited than patents allowed to expire before their full term. The higher an invention's economic value estimate was, the more the patent was subsequently cited.