arrives at its investment decisions would pinpoint those areas in which management has insufficient information. Another rewarding study certainly would be to ask management what, with the benefit of hindsight, it should have had in the way of economic data to avoid contributing to the severity of the current recession.

CONCLUSION

I could go on with this agenda, but time moves me along to my finish. By way of summary, I would repeat the concluding comments I offered on a similar occasion:

Business and economic research now accounts for only a small fraction of the Nation's total research expenditures. The great bulk of research outlay is now directed toward the physical sciences. The cost of an accelerated business research program which I look for in the future will inevitably be borne largely by business itself; indirectly, in its contributions to educational and research institutions and in its tax payments; directly, in its own research payrolls.

This is certainly as it should be, since business has an overwhelming stake in the research effort. And I venture to guess that it will, in the not-so-distant future, get a large return in the form of increased stability of total business through increased capacity of the individual firm to minimize its own fluctuations. This would act to make government intervention in the business system less necessary and less frequent. At the same time, research can assure that government intervention, should it become necessary, will be more intelligent, more effective, and more economical.

That was three years ago. My remarks today have, I hope, served to emphasize two points with which I close. First, the scope and the content of our impressive research and development program should be broadened to embrace more fully all branches of the social sciences.

Second, private industry particularly must intensify its resource allocation for economic research and related sciences to do a better job with its managerial functions.

Character of Research and Development in a Competitive Economy

CHARLES J. HITCH

Chief, Economics Division, RAND Corporation

Chairman Neal: Our next speaker is Dr. Charles Hitch, who hails from Arizona, by way of long service as a student and teacher at Queens College, Oxford. He served with the War Production Board, with the Office of Strategic Services, with the Office of War Mobilization and Reconversion, and in other capacities before going to the RAND Corporation in 1947. Since 1948, he has been Chief of the RAND Corporation's economic division. A trained economist who for a decade has been closely connected with research and development in many forms, Dr. Hitch is also eminently qualified to discuss his subject, which is the character of research and development in a competitive economy.

M

Y OWN interest in the character of research and development in the competitive economy stems from my major interest, during the past ten years, in military research and development. For one thing, as you know, most military research and development is conducted on contract by private firms: there is a continuing problem, which has never been very satisfactorily solved, of making the most effective use of research and development in private, competitive firms to advance military technology. Beyond that, it has seemed to me that the military services-and the government generally-might learn something from the way research and development is conducted and managed in the competitive economy that they could, to advantage, emulate. The competitive economy, because it is competitive, has some tendency to penalize and discard inefficient procedures and managements, and to encourage and promote efficient ones. All sorts of things interfere with the ideal functioning of Adam Smith's invisible hand but it might work just well enough in the American economy

to justify some attention to the practices of firms and industries that have survived and prospered in a competitive environment.

MILITARY AND CIVILIAN RESEARCH

Judging from the press, there is general agreement that all is not well with military research and development, and, again judging from the press, there is rather surprising unanimity in diagnosing what is wrong. Apart from a small but distinguished group of dissenters, mainly from science and industry, practically everyone charges that military research and development is uncoordinated and inadequately planned, and plagued by duplication, competition, secrecy, and waste. The remedies are alleged to be obvious: there must be strong central direction and coordination; more and better central planning; tough-minded decisions to eliminate duplication; suppression of interservice and other competition; probably some "Czars" to knock heads together.

This diagnosis and prescription, apparently self-evident to officials, commentators, editorial writers, and most Congressmen, are really rather curious. As a nation we are presumably committed to free competitive enterprise. Anyone who proposed that we eliminate competition and duplication in research in, for example, the American chemical industry by central planning, coordination and knocking heads together would be denounced as either a socialist or a promoter of cartels. But whenever we think we have reason to be dissatisfied with military research and development we criticize it, as I will attempt to demonstrate, for precisely those features which it has in common with research and development in the competitive economy. Instead of taking the best practices in our more progressive industries as a guide for the military services, we plump for a highly centralized bureaucracy.

Let us take a hard look at the character of research and development in the competitive economy. It is not coordinated. There is no central planning or direction. There are no "Czars." There is no "weeding out" of unpromising projects at any level higher than the individual competitive firm. There is intense competition accompanied by secrecy and poaching, an enormous amount of duplication, and, from the vantage point of hindsight, much apparent obvious waste. In fact, in addition to the lack of coordination among firms, a good many companies noted for their achieve

ments in the research and development area (e. g., Bell Laboratories, General Electric, parts of Radio Corporation of America, and DuPont) deliberately decentralize effective control of research to the laboratory or research center level, each center receiving a block budget within which it has complete freedom and responsibility. Although the initiation of new research projects involving relatively large sums is subject to higher level review, these projects usually grow out of small spontaneous efforts in the laboratories, so that idea initiation and initial exploration are subject to a very responsive and decentralized decision-making apparatus.1

The antithesis could scarcely be more striking. In the competitive economy, there is complete decentralization and no high level planning. But the critics of military research and development want to solve the problems there by centralizing and planning at the top. There are three possible extreme explanations of this antithesis which I would like to examine. The first is that the organization of research and development in the competitive economy is wrong; that it is inefficient and wasteful. The second is that the critics of military research and development are wrong, that they have made a bad diagnosis. The third is that military research and development is so different in character from the research and development conducted by competitive firms that utterly different techniques of managing it are appropriate. I will conclude that the truth lies somewhere between the second explanation—that the critics of military research and development are wrong—and the third-that the problems are different-but perhaps nearer the second.

There is a good deal of strong prima facie evidence against the first explanation that research and development in the competitive economy is just inefficient and wasteful. The lack of high level planning in the American competitive economy does not result in anarchy: the forces of the market provide a discipline and screening that seem to be reasonably effective in promoting efficiency. There are few American industries that are not technologically more advanced as well as more productive than their equivalents in Russia—or indeed in any other country, including Western European economies which place more stress on coordination and cooperation and less

1

Some firms follow very different organizational and management principles for that part of their research and development that is on contract with the military. They are forced by the way the government administers contracts to more or less ape the government, and go over to a highly centralized, rigid, slow moving structure with all decisions funneled up and little discretion at the working level.

484767-58- -10

on competition. The exceptions are instructive. They include some military areas (e. g., rockets and tanks) and some nonmilitary ones highly dependent on basic research (e. g., metallurgy and, at least until recently, a good many chemicals and drugs). As we will see later, there are some reasons for believing that the competitive American economy is more proficient in supplying research and development at the engineering development end of the research and development spectrum than at the basic research end.

UNCERTAINTY IN RESEARCH AND DEVELOPMENT

In my opinion the most important thing to understand about research and development is the dominant role played by uncertainty. Uncertainty in research and development is greater and more pervasive than many researchers and research administrators permit themselves to imagine. Predictions by "experts" of the results or usefulness of particular research and development projects are highly unreliable. Research seems to attract an optimistic breed. Despite the fact that developments almost always take longer and cost more than predicted (by factors of 2 to 50), and that most fail in whole or in part, the predicted schedules of new development projects are almost always taken seriously. Of course there are so-called developments that are highly predictable (like some marginal product improvements) but research and development resulting in significant scientific or technological advances is universally uncertain, with occasional happy and frequent unhappy surprises. In addition to technological uncertainty, research and development shares with other kinds of time-consuming investment what is called environmental uncertainty—uncertainty about the kind of new product that will be saleable or useful in the unknown environment of the future years in which it will be available.

What constitutes sensible behavior when we are confronted by gross uncertainties? No one has succeeded in defining it with precision to general satisfaction. But certain of its characteristics are well understood. It makes sense to hedge, to preserve flexibility. Where there are several possible paths to an objective, say several competing research and development ideas, each of which is uncertain, it frequently makes sense to try several. In fact it can be proved that under many circumstances the multiple-path approach produces results more quickly and more cheaply than attempting in advance to choose the optimal path and concentrating all one's efforts in