It is funny, really, how the general perception of economists is that they are math geeks who simply draw conclusions (that is, inductively) from statistical data. Whereas there may be some truth to that, what is it in the math and stats that make economists different from other social scientists? What is it that makes such economics economics?

To say that economists do it with models is not really to point to an obvious distinction from other social scientists. Anyone going through advanced studies in political science, business management or (yes, even) sociology have to take courses in advanced statistical analysis (though they’re called different things: econometrics, psychometrics, and so on). So “doing it with models” is hardly an economics deal – it is a scientistic deal, it is how “we” do research in this highly positivistic, empiristic day and age.

It is distinct for economics in one way only: economics was the first of the social sciences to seriously adopt the “mathods” of the natural sciences. And, sort of consequently, economics therefore has the most developed and advanced methods – which is why for example many doctoral students in business get to (must) study econometrics. But it is really just statistical analysis applied on social sciences, though with a somewhat unhistorical view of social science.

So in a sense the view of economists as the boring math freaks who find data so fascinating that they spend their free time trying out “cool” new regression functions (models?) rather than have beers with people (models?) is not incorrect. It also is a rather embarrassing heritage in the discipline, especially considering the roots and origins of economic studies – and what distinguishes the social sciences.

Economics was and is referred to as the “queen” of the social sciences, but it really has only very little to do with the advanced statistical analyses that economists rely on. Economics is the foremost of the social sciences because it is much less arbitrary than the other disciplines: it studies social phenomena from a methodologically individualistic view, and can thereby address issues of emergence and decomposability of these aggregate level phenomena. “Culture” cannot be understood unless we consider the individuals that act in accordance with it; and decomposing culture shows clearly that it can have no other building blocks than individuals and their perceptions.

One does oneself as well as the whole world a disservice by starting on the aggregate level, such as e.g. sociologists do (they often assume a “structure” under and into which individuals are molded). One is too limited if only some types of phenomena are studied as were they distinct, purely separable, and different from other social phenomena (such as “political” issues). While it is true that a culture is “more” than the individuals, this is a conclusion only if we start with individuals, study their interaction, and how individuals’ choices and actions are influenced and affected by others’. There is no alternative to methodological individualism.

Economists realized this more than a couple of centuries ago, and it is the study of the interaction between individuals (interaction, exchange) and between individual and society (influence, culture) from the bottom up that made economics successful. It wasn’t the adoption of the natural science methods that made economics successful, though revisionist discipline historians would like to make this a new truth.

There are quite a few unfortunate consequences of adopting the view that “data” is what moves the world. One is that the necessary theoretical intuition, which used to be core to economic analyses (whether quantitative or not), becomes a rare quality. In my graduate program, when studying advanced microeconomics, the professor made a point out of having almost exclusively test problems with “corner solutions.” The latter are mathematical solutions that do not fit squarely with the model so that the results are unambiguous, but need interpretation. The purpose was to not only test the mathematical skills of students, but to make sure they would rely on their economic intuition and, thus, understanding for the economics (not the math) of the problem.

But as time progresses, fewer students will have come across real economic analysis. They know models, statistical techniques, and they’re really good at working with “data.”

The latter, unfortunately, has become a highly valued skill in the social sciences. I have personally had discussions with journal editors (in management) about the use of data sets in publications. Their claim was as clear as it is troubling and confusing: having used one specific data set once, it cannot again be published. In other words, a published journal article more or less “sells” the collected data. To get another publication one has to collect more/different data.

This, of course, makes no sense whatsoever from a methodological point of view. If it is the case that, as those of us in the social sciences who are affected by “physics envy” assume, social phenomena can be studied quantitatively and inductively through blind or semi-blind data collection, then certainly “data” should be considered objective measurements. What is then the problem of testing more hypotheses on the same data set? But that is a point that obviously escapes certain journal editors, who are focused on publishing “previously unpublished” articles – and this, in present data-centric academic research, translates to previously unpublished data.

This is, then, where we are currently at: social scientists as simple traders in data. While this will eventually lead to a remerger of the social sciences, which is perhaps overdue, it will happen for the wrong reasons and on the wrong bases: inductive data collection and manipulation as common denominator. In this sense, the “imperialism” of economics is as real as it is problematic. But it is not because economists are better in any real sense, but that they (we) for much longer time have been doing it with models.