It is something of a given in modern western thinking that competition is a prerequisite for progress, whether in the marketplace or in academia. That there is a general relationship, of the sort illustrated by the left hand part of the curve in figure 1, is shown by experience in rapidly expanding systems such as the United States economy after the Second World War, where, for example, removing barriers to competition via deregulation or lending with poor underwriting standards produces a rise in productivity. That is not, however, the whole story. In the absence of major external input, both growth and competition have limits. High-competition environments are costly in terms of resources, including human resources, relative to low-competition environments with a high degree of cooperation. At some point (b) these costs counterbalance the increase in productivity, and net growth ceases, although gross productivity continues to rise. If competition continues to rise past that point, net productivity falls, and the trajectory resembles the right hand side of figure 1.
Figure 1
It can be seen in the figure that (on a theoretical level at least) there is a point (a) on the curve where increase in competition produces the largest corresponding increase in productivity. It is a mistake to extrapolate very far in either direction and expect predictions based on the relationship in a narrow time frame to hold true. Diminishing returns produce disappointing results.
The essence of competition is the expectation that added effort on the part of one party will result in that party getting a larger share of the payoff than the competitors. If he does not, it’s not competition. As long as the total payoff is increasing at a rapid rate, the losers in the competition can also benefit. Inequity in distribution rises slowly, and the well documented adverse effects of large and increasing social inequity are muted. As net growth slows, there is less surplus from which to compensate the losers, and the payoff distribution becomes increasingly skewed, especially when competition rises rapidly, which it is bound to do if one still subscribes to the competition-progress equation established earlier in the growth curve.
High levels of competition encourage cheating. While the costs, if detected, tend to remain constant, and the likelihood of being detected actually decreases for a successful competitor who is better able to hide his tracks, the relative payoff to the successful cheater increases when legitimate avenues of competition – activities that add value to the enterprise – become insufficient to ensure success. I saw an illustration of this phenomenon when I was a teaching assistant in biology courses in the 1970’s, when we were obliged to be more vigilant with pre-medical students than with students contemplating entering non-medical fields including research science.
How does this basically economic model apply to science at American Universities? Consider the aim of this enterprise as the accumulation and analysis of observations and the production of ideas in the realm of science, and the overarching aim of bettering the human condition. The scientists also need to earn a living, and increasing autonomy and prestige are important factors. Most laboratories are nonprofit, so generating wealth for third parties is not, at least on the face of it, a major factor.
People capable of being research scientists are not an inexhaustible resource, and educating them is costly. In a low-competition environment, nearly everyone who has completed an advanced degree in the sciences finds steady, well-paid employment in the field, including people whose productivity is low, through lack of ability or motivation. Some competition for jobs and tenure should therefore increase the productivity of the remaining work force, but this is counterbalanced by the contribution (0) of people who lose out in the competition. The cost of identifying people with uncommon skills, educating them and then not using them can rise rather steeply.
The quality of output is critical in the sciences. Cheating in any social endeavor weakens it, and this is particularly dangerous in an institution we look to as central and authoritative, whether it be science or religion. In the sciences, we trust that information gathered is accurate and unbiased. The scientist who, to secure his position, fakes research results or manipulates them in a deceptive manner to satisfy a granting agency undermines the very aims he purports to further. The person who gives the appearance of high productivity by publishing large numbers of papers based on the same piece of research secures a competitive edge through illusion. There is an increasing disjunction between apparent contributions to progress and actual contributions to progress, and, if the illusion becomes pervasive enough, it can become an engine of denial of actual deterioration.
Appropriation of the products of research done by junior colleagues and graduate students is particularly pernicious to healthy scholarship. The appropriator is already a successful competitor and is acting unethically. Long experience in Western science over more than two centuries has shown time and time again that the most innovative ideas and dedicated hard work come from people early in their careers. In a highly competitive environment, failure to credit them with their work likely results in their loss to the profession, either through failure to secure an academic position or through burnout in a limbo of poorly compensated positions with little autonomy. The unethical superior, meanwhile, acquires an undeserved reputation for wisdom and is elevated to positions of increasing authority, where he can secure and enhance his position by more unethical behavior – for, by example, hiring decisions that eliminate rising lights who might, through innovation, challenge the superior’s reputation. I have seen principal investigators quite systematically undermine the academic reputations of individuals whose scholarship they have appropriated.
Based on my own observation and experience, and that of colleagues, I think American academic science reached the point of maximum return in the mid nineteen-sixties and has been following a trajectory of diminishing returns and increasing net costs ever since. We may have already ventured onto the right hand side of the graph, the point at which increasing competition not only fails to produce new useful ideas, but has begun to generate harmful memes for the benefit of unethical manipulators. It is not a hopeful scenario for a society that has elevated science to a religion.
Image Credit
Graph created by Martha Sherwood. All rights reserved.
Having been acquainted with your perceptive remarks in other, usually philosophic, venues on the web, I was surprised to encounter a graph. After puzzling with it for awhile, I decided just to ignore it. That was the right decision: your calm words and valuable insights are what I appreciate.
I am probably not prepared to appreciate your article, for the chart confuses me. It seems to me that the dotted line represents growth (because why would both lines represent productivity?), and that perhaps your references to point a and point b in the first two paragraphs are reversed.
That said, your personal experience in the practice of science shows the need to better balance incentives and rewards with the checks and balances of accountability.
The dotted line represents change in net productivity, and thus becomes negative when net productivity declines. I realize that the diagram is confusing but I haven’t figured out a way to visually represent the tipping point where diminishing returns actually start having a negative effect.
Interesting post! Do you perhaps have any scholarly references, e.g., for the curve in Figure 1
Nothing directly relevant in the way of a scholarly reference. I was auditing a course in Game Theory in the University of Oregon Economics Department, a subject about which I know nothing , and this article represents my own playing around with some of John Nash’s work.my background is in organismal biology, specifically plant taxonomy and the last mathematics course I took was during the Johnson administration. There are probably scholars in the field who have said similar things, but very cursory search didn’t turn up any. Intuitively, the diagram seems right.