Economics Has a Math Problem
Last sunday Dutch political discussion program Buitenhof featured a cringe-inducing discussion between an economics professor and two students seeking reform of their education. The students’ point was that economics treats models as sacred and one model in particular. The professor, mistakenly, seemed to assume the students were against a math-first education. He seemed to miss their point that, in their opinion, the education didn’t deal enough with choosing a model and dealing with the gray areas between models.
My own impression certainly agrees with the students: economists and econometricians seem to treat number as sacrosanct. I’m sure that is partly my own biased perception, but the professor underlined this perception by mentioning that mathematics works so very well for physics and then made a cheap jab by insinuating physics students like math so much more than economics students. He seems to mistakenly equate the rigor of (modern) economics with physics, where newer models are required to be strict supersets of older models. Moreover, most physicists agree broadly on the validity of the common models (relativity, quantum mechanics), whereas this seems to be not true of economics, where political preference seems to influence a scientists’ favorite model. Economics are of course much closer to politics than physics, hence the issue.
Bloomberg posted an article highlighting the lack of experimental testing of economics-models, and poses data science as the next step in some sort of model-less economics, where data rules and models can be cast aside. I’m sure that it’s not going to be that easy, after all, ad-hoc internal models of specific machine learning software training sets are not going to give a similar insight as a solid and rigorous model as may be found in physics. But my intuition is that they’re right that for economics to graduate to an exact science, ‘math-first’ is the smallest of steps towards that. Being critical of models and the way they overlap is the great strength of the science of physics. The apparent ‘model-islands’ found in economics with each their own tribe are hardly what constitutes exact science or, as the students mentioned, useful in answering todays important economical questions.
More discussion on HN.