Paula E. Stephan †
The Economics of Science
Journal of Economic Literature
Volume 34, Issue 3
Sept. 1996, 1199-1235.
SCIENCE COMMANDS the attention of economists for at least three reasons. First and most important, science is a source of growth. The lags between basic research and its economic consequences may be long, but the economic impact of science is indisputable. Second, scientific labor markets - and the human capital embodied in scientists - offer fertile ground for study. Third, a reward structure has evolved in science that goes a long way toward solving the appropriability problem associated with the production of a public good. Another reason to study science relates to the large amount of resources employed in the enterprise. In 1991, for example, more than 85,000 Ph.D. scientists were engaged in research in the physical, environmental, and life sciences in the United States (National Science Foundation 1994, table 10, p. 18). An undetermined but substantial number of physicians were also engaged in research. Basic research budgets in these fields were approximately 13 billion dollars, applied research budgets about 17 billion. 
Early work in the economics of science focused almost exclusively on the relationship between science and technology and the ways technology affects growth and responds to economic forces. This work led to the realization not only that science makes technological innovations possible, but that science itself is affected by technology. For example, technology provides apparatus to understand physical phenomena better. This work also led to an appreciation that to a considerable extent the scientific enterprise evolves in disciplines that from their beginnings have been closely tied to fields of technology.
The enhanced respect with which science emerged from World War II underscored the need to understand better the workings of scientific labor markets. The advent of human capital models in the early 1960s created a framework for
† Paula E. Stephan
Department of Economics and Policy Research Center
Georgia State University
The author would like to thank William Amis, David Audretsch, Dave Boykin, Eileen Collins, Paul David, Ronald Ehrenberg, Alan Fechter, Julie Hotchkiss, Mary Frank Fox, Vincent Mangematin, Edwin Mansfield, Rubin Saposnik, F. M. Scherer, Frank Stafford, Mary Beth Walker, Harriet Zuckerman, and two anonymous referees for helpful comments. Some of the ideas expressed in this essay have evolved from extensive conversations and collaboration with Sharon G. Levin. The author, however, bears sole responsibility for the opinions and conclusions expressed here. Stephen Everhart and Janet Keene provided research assistance. This essay was begun when the author was a visiting scholar at the Wissenschaftszentrum Berlin fur Sozialforschung. Financial support was received from the Andrew W. Mellon Foundation and the College of Business Administration, Georgia State University.
1. Research expenditures for these broad fields are estimated from data found in National Science Board (1993).
their study and a second line of inquiry concerning science-related issues was firmly launched. A third line of inquiry had its genesis in the work of sociologists, who, at a slightly earlier time, had begun to study the reward structure in science and the behavior that it engenders. This work has provided economists with a basis for understanding how a reward structure has evolved in science that encourages the production of the public good “knowledge.” Other useful concepts and ideas have also been imported to economics from the sociology of science, such as the observation that processes of cumulative advantage operate in science.
This essay attempts to bring together these (and other) lines of inquiry concerning science and to incorporate into the discussion salient facts about science and scientists that have been observed by colleagues working in other disciplines. We begin by discussing the public nature of knowledge and characteristics of the reward structure. Special attention is given to the recognition that priority of discovery is a form of property right. We then explore the winner-take-all nature of scientific contests and the inequality that characterizes such contests. Efficiency considerations follow. This leads to a discussion of how the incentives to disclose information in a timely fashion relate to the type of property right sought. We demonstrate that, contrary to popular belief, it is not uncommon for scientists in industry to publish, nor is it unknown for scientists working in the nonprofit sector to “privatize” information.
The second half of the essay begins with a discussion of scientific labor markets. This includes an examination of life-cycle models of the labor supply of scientists and empirical tests of life-cycle models. A portion of the essay is devoted to a discussion of the complexities underlying the production of scientific knowledge. The importance that resources play in this process leads to a consideration of attributes of different funding regimes. The essay ends with a discussion of empirical studies relating scientific research to economic growth. We also argue that a case can be made that science, by having endogenous aspects, figures prominently in the new growth economics. We conclude by suggesting topics for further study.
2. The Public Nature of Knowledge and
the Reward Structure of Science
In his 1962 article concerning the economics of information, Kenneth Arrow discussed properties of knowledge that make it a public good. Others (for example, Partha Dasgupta and David 1987, 1994; Harry Johnson 1972; Richard Nelson 1959) have also commented on the public nature of knowledge: it is not depleted when shared, and once it is made public others cannot easily be excluded from its use.  Moreover, the incremental cost of an additional user is virtually zero  and, unlike the case with other public goods, not only is the stock of knowledge not diminished by extensive use, it is often enlarged.
Economists were not the first to note the public nature of knowledge. More than 180 years ago Thomas Jefferson
2. Research findings only become a public good when they are codified in a manner that others can understand. The distinction, therefore, is often drawn between knowledge, which is the product of research, and information, which is the codification of knowledge (Dasgupta and David 1994, p. 493).
3. In reality, the marginal cost of use is greater than zero because users must incur the opportunity cost of time as well as the direct cost of access to journals or attendance at meetings. Information, of course, is only of use to those who possess the requisite intellectual framework. Michel Gallon (1994) argues that the public nature of science is greatly overstated. Tacit knowledge (discussion to follow) can be more costly to learn than knowledge that is codified.
(1967 edition, p. 433, section 4045) wrote:
If nature has made any one thing less susceptible than all others of exclusive property, it is the action of the thinking power called an idea, which an individual may exclusively possess as long as he keeps it to himself; but the moment it is divulged, it forces itself into the possession of every one, and the receiver cannot dispossess himself of it. Its peculiar character, too, is that no one possesses the less, because every other possesses the whole of it. He who receives an idea from me, receives instruction himself without lessening mine; as he who lights his taper at mine, receives light without darkening mine.
A cornerstone of economic theory is that competitive markets provide poor incentives for the production of a public good, because providers cannot appropriate the benefits derived from use. This observation, however, relates to rewards that are market-based. An important contribution of the sociologists of science and the economists who have extended their work is the demonstration that a non-market reward system has evolved in science that provides incentives for scientists to behave in socially responsible ways. In the sections that follow, we analyze the components of that reward system as well as the behavior it encourages.
A. The Reward Structure of Science: The Importance of Priority 
As economists we owe a substantial debt to Robert Merton for establishing the importance of priority in scientific discovery. In a series of articles and essays begun in the late 1950s, Merton (1957, 1961, 1968, 1969) argues convincingly that the goal of scientists is to establish priority of discovery by being first to communicate an advance in knowledge and that the rewards to priority are the recognition awarded by the scientific community for being first. Merton further argues that the interest in priority and the intellectual property rights awarded to the scientist who is first are not a new phenomenon but have been an overriding characteristic of science for at least three centuries.
The recognition awarded priority has varied forms, depending upon the importance the scientific community attaches to the discovery. Heading the list is eponymy, the practice of attaching the name of the scientist to the discovery. Haley’s comet, Planck’s constant, Hodgkin’s disease, the Copernican system are all examples. Recognition also comes in the form of prizes. Of these, the Nobel is the best known, carrying the most prestige and the largest purse (approximately $1 million in the early 1990s), but hundreds of others exist, a handful of which have purses in excess of $300,000.  Many countries also have societies to which the luminaries are elected: the National Academies of Science, Engineering, and Medicine in the United States, the Royal Society in England, the Académie des Sciences in France.
Publication is a lesser form of recognition, but a necessary step in establishing priority. A common way to measure the importance of a scientist’s contribution is to count the number of citations to an article or the number of citations to the entire body of work of an investigator. And while eponymy or a prestigious prize are perceived by most to be beyond their reach, the reward of publication is within the reach of most.
It is important to stress that recogni-
4. Parts of Sections A and B draw on joint work with Levin (Stephan and Levin 1992).
5. Zuckerman (1992) estimates that approximately 3,000 prizes in the sciences were available in North America alone in the early 1990s. This is five times the number awarded 20 years earlier.
tion in science depends on being first.  There are no awards for being second or third. The behavior such an incentive structure elicits is one of the themes of this essay. One consequence is the perceived need to rush work to a journal. It is not unknown for scientists to write and submit an article in the same day. Neither is it unknown to negotiate with the editor of a prestigious journal the timing of publication or the addition of a “note added” so that work completed between the time of submission and publication can be reported, thus making the claim to priority all the more convincing (Stephan and Levin 1992). Another consequence of a priority-based reward system is the energy scientists devote to establishing priority over rival claims. Moreover, such practices are not new. Merton (1969, p. 8) describes the extreme measures Newton took to establish that he, not Leibniz, was the inventor of the calculus.
The importance accorded priority and the response priority elicits bear a striking similarity to the practice of offering an award to the first firm to complete successfully a well-defined project (Brian Wright 1983). More generally, the race for priority can be compared to patent races, the essence of which is described in work by Morton Kamien and Nancy Schwartz (1975). Both are extreme forms of winner-take-all contests (Robert Frank and Philip Cook 1992) in which the winner is determined solely on the grounds of being first. 
Two characteristics of science account for the winner-take-all nature of scientific contests. The first is the difficulty that arises in monitoring scientific effort (Dasgupta and David 1987; Dasgupta 1989). This class of problem is not unique to science. Edward Lazear and Sherwin Rosen (1981) have investigated incentive-compatible compensation schemes where monitoring is costly. A second characteristic of science that fosters a winner-take-all reward structure is the low social value of the contribution made by the runners-up:
there is no value added when the same discovery is made a second, third, or fourth time. To put it sharply (and thus somewhat inaccurately), the winning research unit is the sole contributor to social surplus. (Dasgupta and Eric Maskin 1987, p. 583)
B. The Reward Structure of Science: Financial Remuneration and the Satisfaction Derived from Puzzle Solving
Financial remuneration is another component of the reward structure of science. Because the winner-take-all nature of the race places much of the risk on the shoulders of the scientist, it is not surprising that compensation in science is generally composed of two parts: one portion is paid regardless of the individual’s success in races, the other is priority-based and reflects the value of the winner’s contribution to science. While this clearly oversimplifies the compensation structure, the role played by counts of publications and citations in determining raises and promotions at universities is evident from the work of Arthur Diamond (1986a) and Howard Tuckman and
6. There are, of course, different levels of contests and repeated contests to enter. Many scientists choose to play in the minor leagues, working in the backwaters of science, or, as some would say, functioning as ditchdiggers. See discussion page 1204.
7. Substantial differences also exist between patent races and priority races. For example, there is no reward for reverse engineering in science and consequently no incentive to play the type of waiting game discussed by William Baldwin and Gerald Childs (1969).
8. The inaccuracy of the quote relates to the fact that replication and verification have social value and are common in science.
Jack Leahy (1975). The Diamond estimates, for example, suggest that the present value of publishing another article for a 35-year-old mathematician is (in 1994 dollars) about $6,750; the present value of an additional citation to a 35-year-old physicist’s work is about $2,225.  Unfortunately, we know little about the reward structure for scientists in industry or in government labs, particularly as that reward structure relates to priority.
The flat profile of earnings in science (at least for those employed in academe) is frequently noted. Ehrenberg (1992), for example, calculates that the average full professor in the physical and life sciences earns only about 70 percent more than the average new assistant professor. This arguably relates to monitoring problems and the need to compensate scientists for the risky nature of their work. On the other hand, if earnings are expanded to include compensation outside the institution, the profiles are in all likelihood not nearly as flat as is often assumed. A variety of extra-institutional rewards awaits the successful scientist in the form of prize money and speaking and consulting fees. Successful patents can also generate a significant income stream for their scientific inventor, and in recent years it has become standard practice for eminent scientists, particularly in the life sciences, to serve as scientific advisors and directors of new companies. Stephan and Stephen Everhart (forthcoming) demonstrate that a handful of scientists realize extraordinary returns from the stock they hold in such companies and that a substantial number have the potential of realizing nontrivial sums of money by exercising stock options. A fruitful area for further research would be to investigate what happens to the earnings profile when the definition of income is broadened to include these extra-institutional forms of compensation.
The other reward often attributed to science is the satisfaction derived from solving the puzzle. To quote Warren Hagstrom (1965, p. 16), “Research is in many ways a kind of game, a puzzle-solving operation in which the solution of the puzzle is its own reward.” The philosopher of science David Hull (1988, p. 305) describes scientists as being innately curious and suggests that science is “play behavior carried to adulthood.” This suggests that time spent in discovery is an argument in the utility function of scientists. Robert Pollak and Michael Wachter (1975) demonstrate that maximization problems of this type are generally intractable, because implicit prices depend upon the preferences of the producer. While this provides a rationale for excluding the process of discovery from models of scientific behavior, the failure of economists to acknowledge the puzzle as a motivating force makes economic models of scientific behavior lack credibility.
3. Inequality in Science
A defining characteristic of winner-take-all contests is extreme inequality in the allocation of rewards. Science, too, has extreme inequality with regard to scientific productivity and the awarding of priority. One measure of this is the highly skewed nature of publications, first observed by Alfred Lotka (1926) in a study of nineteenth century physics journals. The distribution that Lotka found showed that approximately six percent of publishing scientists produce half of all papers. Lotka’s “law” has since been found to fit data from sev-
9. These calculations assume that the rewards are incorporated into the base salary and that the real interest rate is three percent.
eral different disciplines and varying periods of time (Derek de Solla Price 1986). 
Inequality in scientific productivity could be explained by differences among scientists in their ability and motivation to do creative research. But scientific productivity is not only characterized by extreme inequality at a point in time; it is also characterized by increasing inequality over the careers of a cohort of scientists, suggesting that at least some of the processes at work are state dependent. Yoram Weiss and Lee Lillard (1982), for example, find that not only the mean but also the variance of publication counts increased during the first ten to 12 years of the career of a group of Israeli scientists.
Merton christened his explanation for inequality in science the Matthew Effect, defining it to be
the accruing of greater increments of recognition for particular scientific contributions to scientists of considerable repute and the withholding of such recognition from scientists who have not yet made their mark. (1968, p. 58)
He argues that the effect results from the vast volume of scientific material published each year, which encourages scientists to screen their reading material on the basis of the author’s reputation. Other sociologists (Paul Allison and John Stewart 1974; and Jonathan Cole and Stephen Cole 1973, for example) have argued that additional processes of “cumulative advantage” are at work in science, such as the ability to leverage past success into research funding as well as the “taste” for recognition that success engenders. While we have yet to understand these processes completely, a strong case can be made that a variety of factors are at work in helping able and motivated scientists leverage their early successes and that some form of feedback mechanism is at work (David 1994). This observation is consistent with other work in winner-take-all contests. Frank and Cook (1992, p. 31) observe that “in all their manifestations, winner-take-all effects translate small differences in the underlying distribution of human capital into much larger differences in the distribution of economic reward.”
4. The Choice of Scientific Contests
The winner-take-all character of scientific contests dictates that scientists choose the contests they enter with care. The probability of being scooped is a constant threat. This is particularly true in the case of “normal” science where the accumulated knowledge and focus necessary for the next scientific breakthrough is “in the air.”  Young scientists, in particular, must choose their contests with care if they are to successfully signal their ability or “resource worthiness” and set in motion the processes of cumulative advantage described above (Alan Garner 1979).
Scientists can minimize the threat of being scooped by seeking ways to monopolize a line of research. During the seventeenth and eighteenth centuries, discoveries in process were sometimes reported in the form of anagrams for the
10. Lotka’s law states that if k is the number of scientists who publish one paper, then the number publishing n papers is k/n2. In many disciplines this works out to some five or six percent of the scientists who publish at all producing about half of all papers in their discipline. Although Lotka’s Law has held up well over time and across disciplines, David (1994) shows that other statistical distributions also provide good fits to observed publications counts.
11 Note the distinction between social and individual risk. Because accumulated knowledge is an important input in the process of discovery, normal science is not especially risky from the social point of view (Dasgupta and David 1987, p. 526; Arrow 1962). From the individual investigator’s point of view, however, the risks can be substantial: being in the air is entirely different from being in scientist X’s air.
“double purpose of establishing priority of conception and of yet not putting rivals on to one’s original ideas, until they had been further worked out” (Merton 1957, p. 654). It was also not uncommon to deposit a sealed and dated manuscript with a learned society to protect both priority and idea. More recently, the ownership of apparatus or strains has proved to be a convenient way to monopolize a line of research. Scientists can also minimize the threat of being scooped by choosing to work on problems that fall outside the mainstream of “normal science” or by working in “the backwaters” of research (Stephan and Levin 1992). The downside of such a strategy is that, while the low number of competitors increases the probability of being first, the contest that is won may be of little interest to the larger scientific community and hence receive minimal recognition.
Researchers must choose not only a line of research. They must also choose a research strategy, because more than one method can be used to address the same question (Dasgupta and David 1994). Here, too, uncertainty enters the equation. The use of a novel method, for example, can prove rewarding, but the risk of coming up empty-handed can be quite large when an unorthodox approach is employed.  The uncertainty associated with the process of discovery also can be substantial. The outcome may not have been envisioned, neither may the outcome relate to the original objective of the researcher. In the process of trying to solve some very practical problems concerning fermentation and putrefaction in the French wine industry, Pasteur established the modern science of bacteriology (Nathan Rosenberg 1990). 
Basic research often provides answers to unposed questions.  Consequently, the risk associated with such research can be lessened by shifting goals during the course of research. Nelson (1959) argues that this strategy is more appropriate for scientists working in a nonprofit-based environment than for scientists working in the profit sector because the former can more easily capture the rewards regardless of where the research leads. On the other hand, companies having a broad technological base can benefit from research that is not directed to a specific goal. At the time General Electric developed synthetic diamonds, for example, it was the most diversified company in the United States.
A number of institutional arrangements have evolved in science to help minimize risk or provide some insurance against risk. Some of these, such as the ability to monopolize a line of research, have already been noted. Others include the adoption of a research portfolio that contains projects with varying degrees of uncertainty, the formation of research teams and networks and the practice of “gift giving” whereby scientists, by acknowledging intellectual debts to their colleagues (via citations), pay “protection money” to insure that those colleagues “won’t deny their grants, spread slander,
12. A consequence is that rival teams often select highly correlated research strategies. From a social point of view, highly correlated research strategies produce inefficiencies by failing to provide the kind of portfolio diversification that society would choose if it were allocating resources in a way to maximize the probability of success (Dasgupta and David 1994). The gains to society from sponsoring multiple lines of independent research are examined by Scherer (1966).
13. Serendipity plays a role in discovery when in the course of research an unintended outcome is observed. The following up, of course, is not accidental. Chance, according to Pasteur, favors only the prepared mind. Bernard Barber and Renée Fox (1962) discuss the role played by serendipity in science.
14. The unpredictable nature of scientific discovery is explored by Michael Polanyi (1962).
or – worst of all - ignore their work altogether” (Steve Fuller 1994, p. 13).
5. Efficiency Considerations
A. The Functional Nature of the Reward System
The socially desirable properties attached to a reward system that is priority-based are substantial. Shirking is rarely an issue in science. The knowledge that multiple discoveries are commonplace makes scientists exert considerable effort.  A reward structure based on priority requires that scientists share information in a timely fashion if they are to establish priority. Such a process in turn permits peer evaluation, which discourages plagiarism and fraud and builds consensus in science (John Ziman 1968; Dasgupta and David 1987). The process also provides scientists the reassurance that they have the capacity for original thought (Merton 1957) and encourages scientists to acknowledge the roots of their own ideas, thereby reinforcing the social process. Reputation also serves as a signal of “trustworthiness” to scientists wishing to use the results of another in their own research without incurring the cost of reproducing and checking the results. It also serves as a signal of trustworthiness to foundations. As such, reputation provides an answer to the agency problem (Stephan Turner 1994) posed by Ronald Coase. 
From an economist’s point of view, the most appealing attribute of a reward system that is rooted in priority is that it offers non-market-based incentives for producing the public good “knowledge.”Dasgupta and David (1987, p. 531), the first to make the observation, say it well: “Priority creates a privately-owned asset - a form of intellectual property - from the very act of relinquishing exclusive possession of the new knowledge.” Arrow (1987, p. 687), commenting on their work, articulates the cleverness of such a system:
The incentive compatibility literature needs to learn the lesson of the priority system; rewards to overcome shirking and free-rider problems need not be monetary in nature; society is more ingenious than the market. 
A reward system based on reputation also provides a mechanism for capturing the externalities associated with discovery. The more a scientist’s work is used, the larger is the scientist’s reputation and the larger are the financial rewards. It is not only that the reward structure of science provides a means for capturing externalities. The public nature of knowledge encourages use by others, which in turn enhances the reputation of the researcher (Stephan and Levin 1996).
B. Are There Too Many Contestants in Certain Contests?
The conventional wisdom holds that because of problems related to appropriability a public good such as knowledge will be underproduced if left to the private sector.  A common rationale for government laboratories and government grants for research rests squarely on this
15. The prevalence of multiples in science is discussed below. Mary Frank Fox (1983) and Hull (1988) discuss the effort and work patterns of successful scientists.
16. This is not to say that the reward structure is without problems. Fraud and misconduct occur with some frequency in science (Alexander Kohn 1986). Susan Feigenbaum and David Levy (1993) discuss the market for (ir)reproducible results; Mary Frank Fox and John Braxton (1994) discuss other issues related to fraud. There is also the considerable issue that the reward structure in science appears to have favored white men over women and members of minority groups.
17. Merton (1988, p. 620) also makes the connection when he speaks of reputation, saying that in science “one’s private property is established by giving its substance away.”
18. Uncertainty and indivisibilities provide two other reasons why knowledge will be underproduced (Arrow 1962).
premise. The production of knowledge can also be stimulated through the granting of property rights to the discoverer. With rare exception, patents have been the primary form of intellectual property rights that economists have examined, arguing that patents provide for appropniability while placing knowledge (eventually) in the public domain.  Moreover, it has been shown (Dasgupta and Joseph Stiglitz 1980) that under a wide array of circumstances social inefficiency results from patent races among rival groups. This inefficiency manifests itself in “excessive duplication of research effort (or)… too fast a pace of advance of the frontiers of knowledge” (Dasgupta and David 1987, p. 532).
The recognition that priority is a form of property rights leads to the question of whether there are “too many” contestants in certain scientific contests. Would the social good be served by having fewer? In a classic speech delivered at a conference commemorating the 400th anniversary of the birth of Francis Bacon, Merton detailed the prevalence of what he called “multiples” in scientific discovery. And Merton was not the first to note their presence. In what Merton calls a “play within a play,” he gives 20 “lists” of multiples that were compiled between 1828 and 1922. Moreover, Merton is quick to point out that the absence of a multiple does not mean that a multiple was not in the making at the time the discovery was made public. This is a classic case of censored data where scooped scientists abandon their research after a winner is recognized. Indeed, Merton argues that “far from being odd or curious or remarkable, the pattern of independent multiple discoveries in science is in principle the dominant pattern rather than a subsidiary one” (1961, p. 356).
The presence of multiple discoveries is due in part to the free access scientists have to knowledge and in part to the fact that uncertainty associated with who will make a discovery leads scientists to choose research portfolios that are correlated (Dasgupta and Maskin 1987).  The knowledge that multiples exist keeps scientists from shirking and moves the enterprise of science at a rapid pace. Such observations invite the question of whether science moves at too rapid a pace and whether certain contests attract too many entrants. Dasgupta and David (1987, p. 540) argue that the priority system can create excesses, just as the patent system does, provided the “reward to the discoverer... is tempting enough.”  They make no effort to define the boundary of temptation, but one wonders if the general knowledge that certain contests deserve the Nobel Prize does not attract an excessive number of scientists. 
19. While neither goal is perfectly achieved by the patent process, the goal of disclosure arguably suffers the most. “The imperfections we have examined in the patent as a device for rewarding disclosures of knowledge are not at all surprising; a stone flung at two birds really ought not be expected to make a clean strike on either” (Dasgupta and David 1987, p. 534).
20. Despite the popularity of patent race models, multiples are arguably more common in science than technology. The reason is that science is concerned with laws and facts, while technology is looking for practical ways to solve problems. Hence, while there is often only one answer to a scientific question, there usually are a variety of distinct ways of solving the practical problem.
21. Another efficiency concern relates to whether scientists direct excessive amounts of time to research as opposed to teaching. The fact that only a handful of scientists contribute the lion’s share of output suggests that substantial inefficiencies arise when yeomen scientists devote long hours to research. Other efficiency concerns exist. One is discussed in footnote 12. Another concerns whether the process of cumulative advantage excludes talented individuals from making contributions. Dasgupta and David (1994, pp. 506-07) discuss additional efficiency issues.
22. On the other hand, the common lament of interest groups that there are not enough entrants in certain races of apparent Nobel proportions (e.g., a cure for breast cancer) leads one to be cautious in making broad generalizations. It is, of [course, possible that such groups are expressing the concern that victory is undervalued by the community. It is also possible that a cure is not “in the air” and applying more resources to the contest would be inefficient.]
HHC: [bracketed] displayed on page 1208 of original.
C. The Incentive to Share Knowledge in a Timely Fashion
Despite the similarities between priority rights and proprietary rights such as patents, they differ markedly in the incentives they provide to disclose research findings in a timely fashion. On the one hand, the quest for priority requires scientists to share discoveries quickly, because it is only by sharing that priority rights can be established. The quest for proprietary rights, on the other hand, discourages the rapid sharing of information, because the very purpose of proprietary rights is to provide a means for capturing the economic rents attached to a new product or technology. And, while some forms of proprietary rights require the sharing of knowledge in recognition of its public nature (e.g., the patent process), incentives to divulge the knowledge quickly are not present. 
The distinction is so crucial that Dasgupta and David (1987, p. 528) argue that the two types of property rights, and the implications they hold for appropriability and disclosure, differentiate science from technology.
If one joins the science club, one’s discoveries and inventions must be completely disclosed, whereas in the technology club such findings must not be fully revealed to the rest of the membership.
This distinction between science and technology often leads to the (erroneous) conclusion that science is done by scientists at universities and public labs and results in published knowledge, while the focus of scientists working in industry is the development of proprietary technology (Nelson 1982). While location does correlate with the incentive to share knowledge in a timely fashion, the relationship is far from perfect. Some firms make the results of their research public; some academics engage in practices that lead to the “privatization” of knowledge. In many instances agents can eat their cake and have it too, selectively publishing research findings while monopolizing other elements with the hope of realizing future returns. Rebecca Eisenbeng (1987) argues that such behavior is more common among academics than might initially be presumed because they can publish results and at the same time keep certain aspects of their research private by withholding data or failing to make strains available upon request. 
The ability to eat one’s cake and have it too is not only facilitated by the fact that publication is not synonymous with replicability. It is also facilitated by the fact that techniques can often be transferred only at considerable cost, in part because their tacit nature makes it difficult, if not impossible, to communicate in a written form (or codify).  This pri-
23. Many nations require publication a year or so after the application for a patent. In most nations patents cannot be obtained if publication has occurred prior to application.
24. Eisenberg (1987) suggests that the patent process may be more congruent with the scientific norms of disclosure and replication than the publishing process in certain areas of the life sciences. This is because patents in the biological sciences require that the material in question be placed on deposit. This is not a requirement for publication; neither are the materials themselves part of the published text.
25. Some aspects of technical knowledge have a strong tacit component, meaning that they cannot be completed codified and made explicit in the form of blueprints or instructions, but instead must be learned through practice. Nelson and Sidney Winter (1982) discuss tacit knowledge, particularly as it relates to skill. Dasgupta and David (1944) use the term tacit somewhat differently to connote knowledge that, for whatever reason, is not codified and argue that the boundary between what is codified and what is tacit is not simply a question of epistemology. Rather, as suggested above, the boundary is “a matter, also, of economics, for it is determined endogenously by the costs and benefits of secrecy in relation to those of codification” (p. 502).
ate aspect of technology is a major reason patents are not a necessary condition for successful research and development and underlies the willingness of industry to share knowledge through publication.
There are other reasons why firms engage in disclosure. Foremost among these is recruitment of talent. Scientists and engineers often see the ability to publish as a condition of employment in industry, knowing that if they are not permitted to do so their career path will be severely restricted and they may fail to achieve prestige among their peers. The reputation of the lab, which is directly related to publication activity, also affects the ability of the company to hire scientists and engineers (Scherer 1967); it may also affect its ability to attract government contracts (Frank Lichtenberg 1986). Stephan’s work on biotechnology (1994) suggests that a firm’s publications can also play a role in signaling capital markets. Diana Hicks (1994) explores a number of other factors leading companies to opt for disclosure through publication. She points out that a critical element in this process is the company’s ability to screen the material that is published, thereby insuring that its proprietary interests are maintained. In the process, however, the firm must be mindful that delays can lower morale among research scientists. David Hounshell and John Smith (1988, p. 369) describe the loss of morale that occurred at Du Pont when research managers implemented what turned out to be a de facto moratorium on publishing.
6. Scientists in Industry
Firms engage in basic research for a variety of reasons.  In some instances, basic research is a by-product of the development of a new product or process (Rosenberg 1990). In other instances the production of generic knowledge is, itself, the goal and is motivated by the belief that a particular new product or process innovation will result from that knowledge. In still other instances basic research is needed if the company is to stay abreast of developments in relevant scientific fields and more readily absorb the findings of other scientists (Wesley Cohen and Daniel Levinthal 1989). Sometimes firms are motivated by the expectation that fundamental research will provide a scientific foundation for the company’s technology. Firms have even been known to engage in basic research because of a concern that the fundamental knowledge required for the industry to advance is lacking and unlikely to be forthcoming from the academic sector. When Charles Stine made his presentation to the Executive Committee of Du Pont in 1926, for example, he argued that fundamental research was necessary because “applied research is facing a shortage of its principal raw materials” (Hounshell and Smith 1988, p. 366). 
This means that the research of some scientists and engineers in companies like IBM, AT&T, and Du Pont is virtually indistinguishable from that of their academic counterparts. Not surprisingly, a number have received the top honors that their field can bestow. Bell Labs, Du Pont, IBM, Smith Kline and French,
26. The demand for scientists in industry relates to the demand for research and development. Here we focus on the narrower issue of the demand for basic research.
27. The payoff to a firm’s performance of scientific research often takes the form of first-mover advantages (Rosenberg 1990). Thus, even if the research findings eventually spill over to competitors or cannot be protected through proprietary rights, the firm performing the research has the opportunity of being the first to use the information for the basis of decisions, new products, etc. Despite the evidence concerning the effects of basic research on productivity (Mansfield 1980), recent years have seen a notable reduction in the amount of basic research supported by industry.
Sony, and General Electric have each been the research home to scientists who have subsequently won the Nobel Prize. In 1994, 3.8 percent of the 2,088 members of the National Academy of Sciences came from industry. Twenty-four of the members were at AT&T Bell Laboratories.
Table 1 [HHC: not included] gives the institutional origin of authors of U.S. scientific and technical articles published in 1991 for eight fields of science and engineering. While the vast majority of articles are authored by scientists working in the academic sector, industry produces a sixth of the literature in chemistry and physics and a fourth of it in “engineering and technology.” The blurred boundary between academics and industry is further indicated by the fact that 35 percent of articles with an industry address have a coauthor from the academic sector (see last column). Moreover, this proportion grew by more than 50 percent between 1981 and 1991. This trend undoubtedly relates to the increasing number of research alliances that have been formed between industry and academe since Monsanto in 1977 gave Harvard $23 million in research funds. Such alliances are particularly prevalent in biotechnology (David Blumenthal et al. 1986). 
28. 1 also points out the important research role that the nonprofit sector plays in clinical medicine and that the federal sector plays in biology and earth and space sciences. The Na-[tional Institutes of Health and NASA are important government research sites for these fields, respectively. The importance of Federally Funded Research and Development Centers (FFRDGs) in physics is also clearly demonstrated. These include Fermi National Accelerator Laboratory in Chicago, Brookhaven National Laboratory on Long Island, and the Stanford Linear Accelerator in Palo Alto.]
HHC: [bracketed] displayed on page 1208 of original.
The reasons for industry to publish research findings, as well as the economic incentives for adopting a basic research agenda, have been noted above. This should not, however, be taken as an indication that economists (or others, for that matter) have adequately studied scientists in industry doing “science.” Many questions remain unanswered and, perhaps even more fundamental, unposed.  For example, why do companies adopt compensation strategies that impair the productivity of scientists by tying salary increases to the assumption of managerial responsibilities? Does the strategy adopted by IBM and DuPont of creating well paid research fellow positions help alleviate the problem? What role do publications play in facilitating movement between the industrial and the nonprofit sector? There is also the question concerning how basic research in industry is monitored. The unpredictable nature of research, as well as the belief that creativity requires freedom of choice, suggests that success is hampered if managed too closely. Yet firms can ill afford to fund research that has little promise of (eventually) relating to the company’s objectives. Scherer (interview) reports that Bell Labs solved this problem by giving “the glassy-eyed stare” to scientists who were seen as straying too far from the Labs’ purpose. Recipients knew that they had the choice of either modifying their research or being ostracized. Finally, given the collaborative nature of science, there is a need to study the laboratory as a unit of analysis, instead of focusing exclusively on individual scientists.