The (Empirical) Science Isn't Settled
"Because Science!" has become a popular refrain. I'm not just talking about climate studies. Headline writers love to say this. Apparently "Science" recently proved that we should continue to put two spaces after a period, as we did in the old fixed-width typewriting days.* (No word yet on whether it is OK to put pineapple on pizza.) This kind of language is meant to be a conversation-stopper. If a scientific study — i.e., a laboratory experiment or a robust statistical analysis of secondary data — has proven something, it must be accepted as fact.
Hayek memorably described this mentality as scientism, the claim that only the methods of the natural sciences are legitimate for addressing questions of fact. As Hayek emphasized, scientism is itself profoundly anti-scientific, when "science" is understood correctly as true knowledge about the world.
More generally, the Austrian tradition in economics has placed strong emphasis on economic theory. Mises used the term praxeology to describe his approach to economic theory, in which economic laws are worked out, by logical deduction, from a few self-evident axioms. Of course, Austrians are profoundly interested in empirical phenomena, policy analysis, and other forms of applied work. Menger started his career as a financial journalist. Böhm-Bawerk was finance minister in the Austro-Hungarian empire, and Mises was a full-time policy analyst and, besides his theoretical writings, produced applied works on intervention, monetary policy, international relations, and more. Rothbard wrote classic studies on the Great Depression, monetary and financial history, and a host of current events. But Austrian economists maintain that empirical phenomena — the "data" of history — do not speak for themselves, but must be interpreted in light of economic theory.
While most economists, Austrian or not, acknowledge the importance of theory, the Austrian view has often been mischaracterized as being hostile to empirical work. The eminent historian of economics Mark Blaug described Mises's methodological views as "cranky and idiosyncratic" but did not offer any specific complaints (and was himself a critic of excessive formalism). More famously, or infamously, Milton Friedman gave a particularly uncharitable reading of Mises's methodological writings. Friedman thought that logical statements and deductive arguments cannot be debated and discussed, while quantitative empirical work can provide definitive answers to economic questions.
Suppose two people who share von Mises' praxeological view come to contradictory conclusions about anything. How can they reconcile their difference? The only way they can do so is by a purely logical argument. One has to say to the other, "You made a mistake in reasoning." And the other has to say, "No you made a mistake in reasoning." Suppose neither believes he has made a mistake in reasoning. There's only one thing left to do: fight. Karl Popper — another Austrian like Mises and Hayek — takes a different approach. If we disagree, we can say to one another, "You tell me what fact, if they were observed, you would regard as sufficient to contradict your view. And vice versa. Then we can go out and see which, if either, conclusion the evidence contradicts. The virtue of this modern scientific approach, as proposed by Popper, is that it provides way in which, at least in principle, we can resolve disagreements without a conflict.
I once wrote — though I cannot locate the source — that Friedman sounds here like someone who has not actually done any empirical research. As the coauthor of the influential Monetary History of the United States, Friedman should know better. The statement (which Friedman repeated many times in oral presentations) that theoretical disputes can only be resolved by "fighting" is just plain silly; entire academic disciplines (philosophy, mathematics) are dedicated to analysis, dispute, revision, etc. via logical argument. Moreover, I am quite certain that there is more "fighting" among empirical researchers than among praxeologists!
In practice, empirical work in economics (and, I presume, in most other disciplines) is mainly about judgment and interpretation. How should the research question be formulated? What kinds of data are legitimate? How are key constructs operationalized and measured? What analytical techniques are appropriate? How should the results be interpreted? Are they robust? What conclusions can reasonably be drawn? Look at any empirical journal in economics or another social science. Follow arguments on Andrew Gelman's blog or on social media (here is an interesting example — the one that inspired this post, actually). The "evidence" never speaks for itself the way Friedman implies.
So, empirical evidence is important to applied economics, policy analysis, economic history, and so on. But empirical work relies on theory and the subjective judgments of empirical researchers, as Mises explained. Empirical science is vital, but never "settled." Even if you don't have to fight about it.
* Of course, as all civilized people know, when using a proportionally spaced typeface one should put one space, not two, after the period. See discussion here.