Economics is a social science. That might surprise you if you are an economist reading those lines. After all, economists use formalization (mathematics and statistics) like in physics, so it should be qualified as a natural science. To put it simply, this claim is at best fallacious. In Economics, we are not trying to explain the behavior of a gas or of a particle submitted to different conditions. In Economics, we are trying to explain the behavior of a consumer when she has to buy an ice cream during a hot summer or the reaction of a central banker facing a geopolitical risk shock due to a new war in the Middle East. You see. Economics may borrow tools and methodologies coming from the natural sciences, but ultimately, its object of study is human behavior for individuals and for societies. Consequently, Economics is a social science that uses rigorous and sophisticated tools, but it is not a natural science.
Now, we have seen an empirical turn in Economics with the credibility revolution:
On theory, my humble and personal view:
Theory is based on empirical observations, stylized facts. Then, theory (mathematically formalized or not) alone is never sufficient in the social sciences. The right way is to produce testable assumptions from your theory. These testable assumptions are put to the empirical test in order to evaluate the relevance of the channels underlined by the theory. So it starts with empirical observations and ends with empirical tests. Many theories that have been tested are not very relevant in economics. These theories should be discarded by the community. No methodological superiority of theory alone. The right way to do science is the hypothetico-deductive model presented and explained by Karl Popper.
So, the share of paper with theory alone in the economics papers has declined from 60 to 20 percent since the beginning of the 1980s. That’s a good sign that Economics is becoming increasingly scientific in the limits that we have in the realm of social sciences.