Many major economic journals publish models that can neither generate
operational statements nor be challenged by evidence. Authors sometimes motivate
these enterprises by allusions to “stylized facts.” Often, it is only in concluding
remarks that authors provide vague directions about how “future research”
might allow their results to operate in the realm of evidence. Coelho and Mc-
Clure (2005, 562-564) present evidence that in the American Economic Review 1963
through 1996 “[m]athematically complex articles were less operational and were
less likely to be cited in articles containing operational statements.” Empirical
research suggests that the probability of subsequent articles appearing with refuting
data, or any data, is substantially lower than in less mathematically complex
articles.2 Another reason to doubt that mathematically complex models can generate
operational research is that their assumptions are often complicated, substantively
obscure, and unworldly.3
Economics uses evidence to assess theories. Theories that do not provide
evidentiary or testable propositions at reasonable costs are usually disregarded.
The ability to formalize refutable statements and find evidence for or against
them is operationalism. Statements that cannot be operationalized are what Wolfgang
Pauli called “not even wrong.” Refutations instruct us on what is wrong;
non-operational statements lack this virtue.