On October 11, the Royal Swedish Academy of Sciences will award the 53rd Nobel Prize in Economic Sciences, or, as it is correctly called, the Sveriges Riksbank Prize in Economic Sciences in Memory of Alfred Nobel, since it is not one of the prizes originally endowed by Alfred Nobel but by the Swedish Central Bank. Some therefore never tire of emphasizing that it is not a real Nobel Prize, even if the selection criteria and the selection procedure are identical. Often, what they really mean to say is that economics is not a science.
Admittedly, the controversies that often accompany the awarding of the Economics Prize are more similar to those surrounding the Peace or the Literature Prize than to the controversies surrounding the Physics or Chemistry Prize. In the latter case, it is usually the omissions of key contributors that cause controversy, not the research or the researchers themselves. Alfred Nobel’s great-grandnephew Peter Nobel therefore accused the Riksbank of having laid an unwelcome cuckoo’s egg in the nest of the renowned Nobel Prize.
The status of economics as a science was hotly debated in the context of the 2013 prize, when it was given simultaneously to Eugene Fama and Robert Shiller, two economists who seem to hold opposing theories. Fama was awarded for the efficient markets hypothesis, Shiller for challenging it. However, the awarding of a Nobel Prize simultaneously to scientists holding opposing theories has also occurred in medicine, namely in 1906 for contradictory theories about the nervous system. And physics holds on to both the general theory of relativity and quantum mechanics, although the two theories are fundamentally incompatible. Similarly, the fierce criticism of DSGE modelling and its prominence in modern macroeconomics by Paul Romer (2018 Laureate) in a 2016 speech entitled “The Trouble with Macroeconomics” was inspired by physicist Lee Smolin’s 2006 critique of string theory and its prominence in modern theoretical physics entitled “The Trouble with Physics”. Both Romer and Smolin criticized the lack of empirical support of the respective theories and therefore questioned their status as science.
Unfortunately, the matter of empirical support is not quite so simple. The first to come up with the idea of using empirical support as a criterion for delimiting science were the logical positivists in the 1930s. They argued that a scientific theory must be empirically testable, and that theories are confirmed if their observed consequences are shown to be true. However, by means of such confirmation it cannot be proved that the theory is true, because that would be a logical fallacy (fallacy of affirming the consequent). There are always alternative hypotheses equally compatible with the observation. Karl Popper therefore proposed to replace the criterion of verification or confirmation with the criterion of falsification: a theory can never be confirmed with a finite number of observations, but it can be falsified with an observation that disagrees with it. In economics, this position was expressed very clearly by Milton Friedman in his 1976 Nobel Lecture, in which he also addresses the question of whether economics is a science. Friedman and Popper knew each other well; they were both members of the Mont Pèlerin Society.
Willard Van Orman Quine, however, dealt the deathblow to this idea too, even if it died out slowly, especially among economists. In his influential essay “Two Dogmas of Empiricism” (1951), Quine showed that a theory could never be refuted unambiguously by empirical evidence. Theory is always underdetermined by data. Moreover, the physicist Thomas Kuhn’s classic “The Structure of Scientific Revolutions” (1962) nourished fundamental doubts about the possibility of science to uncover truth. Scientific research is dominated by paradigms that cannot be objectively compared with each other. Science is thus not cumulative, progressive, or truth-tracking. If there is progress, it is only progress within a paradigm. While Lakatos tried to add some rationality to Kuhn’s approach by means of progressive and degenerating research programs, Feyerabend pushed Kuhn’s approach with his Anarchistic Theory of Knowledge (1975) even further. In any case, it was obvious that the project of demarcating science from non-science had failed.
Lee McIntyre has recently made a new attempt. In “The Scientific Attitude” (2019), he argues that what is special about science is the attitude of scientists rather than a specific method. Scientists are willing to change their theories in light of new empirical evidence. This criterion at least gives a necessary condition for what science is, even if it is not a sufficient one. In other words, it can only be used to determine what is not science, not what is. Crucially, however, this scientific attitude is maintained not only by individual scientists but also by the scientific community, which has developed and institutionalized tools for this purpose: publication of research, peer review, data sharing, pre-registration of research plans, and replication. As for economics, McIntyre argues, it is not quite there yet.
While publication and peer review have long been standard practice in economics, data sharing is more recent. However, as Edward Miguel shows in “Evidence on Research Transparency in Economics” in the current issue of the Journal of Economic Perspectives, data sharing has become widely practiced over the past few years. Data sharing not only enables replication, but also helps detect and thus prevent data fabrication. A high-profile case became known just this summer when a group of scientists published a meticulous data analysis of an influential 2012 publication on their blog datacolada.org, concluding that the data in said publication was fabricated. To make matters worse, the publication in question was an empirical study on honesty.
Pre-registration, that is specifying the research plan in advance and submitting it to a registry, is increasing but not yet common in economics. Pre-registration should help separate hypothesis generation from hypothesis testing and thus improve the credibility of research findings. Replication, on the other hand, is still uncommon in economics. That replication can be illuminating, however, even when research is done correctly, honestly, and with the best of intentions, is demonstrated by a recent research project published in Organizational Behavior and Human Decision Processes (Schweinsberg et al., “Same data, different conclusions”). A team of researchers demonstrated that scientists could reach contradictory conclusions even when they use the same data. The analysis showed that this was mainly due to different but perfectly reasonable operationalization of the variables involved. For example, it makes a difference whether “active participation in a discussion” is measured by the number of requests to speak or by the number of words uttered. Quine sends his regards.
According to Alfred Nobel’s will, the Nobel Prize should go to those whose contributions have brought the greatest benefit to mankind. Can economics live up to this noble goal? There is certainly no lack of trying. Unlike physicists, economists not only want to understand the world, they also want to improve it. That, of course, is often the cause for disagreement. Like medical science, economic science has no miracle cures, but it does have recipes against the worst economic maladies. Economists may be a quarrelsome bunch, but on this, they agree.
Robert Lucas, the 1995 laureate whose work contributed significantly to the demise of Keynesian economics, said in a 2003 speech the following about Keynes and his 1936 magnum opus, “The General Theory”: “Keynes wrote in a situation where people were ready to throw in the towel on capitalism and liberal democracy, … he showed a way to respond to the depression that was consistent with capitalist democracy. … This was a great political achievement. It gave us a lasting image of what we need economists for.”