Milton Friedman once said that “a theory is the way we perceive ‘facts,’ and we cannot perceive ‘facts’ without a theory.” This is rarely a new topic in the philosophy of science, however the debate around use of theory in economics is still an ongoing one. In the book, Economics Rules, Dani Rodrik scrutinizes such debate along with the question of to what extent are economists allowed to simplify the world around them; in other words, the questions of whether the assumptions in economics are too unrealistic to mimic the world.
The book is a great analysis of the uses and misuses of models in economics. The author, who himself has published numerous quantitative studies, recognizes the limitations of the use of mathematical modeling in the discipline. He points out while simplification of a socio-economic phenomenon is a necessary step in understanding the problem, overuse of assumptions can lead us to results that may suffer both from external and internal validity.
Models in economics, like in any other discipline, follow a “if-this-then-that” procedure —”this” referring to the assumptions and “that” referring to the results. As much as we care about “that” we should also care about “this,” or simply the assumptions that led us to the results. We should not take assumptions as granted. While economists such as Milton Friedman, favor use of even unrealistic assumptions as long as the underlying model has some predictive power, Rodrick warns us that using unrealistic “critical” assumptions should not be a norm in economics.
On the other hand, Rodrick points out that “the antidote of a bad model is a good model” and not no model, and models can be helpful if are based on legitimate and transparent assumptions, and are accompanied by analysis of the results under modified assumptions. While natural sciences benefit from a vertical progress in development of theory, i.e. models are built upon each other and do not serve as alternatives, most of the progress in economic theory is horizontal, meaning that models typically serve as alternatives for each other. Their use depends on the context under which they are employed. For instance, we have multiple growth models such as the neoclassical growth model, the endogenous growth model, and models that emphasize the role of institutional quality, each leading to different development strategies. Each of these models capture a contextual truth and not an absolute truth. Therefore, any policy drawing based on such models will depend as much on our judgments about the context as on the validity of the model itself.
Lastly, we should remind ourselves that no economic model perfectly captures the underlying mechanism it is trying to describe and we will never find “the model” but rather “a model.” We can never rest our cases in understanding problems concerning human relations, unlike what John Stuart Mill’s famous statement that “happily, there is nothing in the laws of value which remains for the present or any future writer to clear up; the theory of the subject is complete.” The theory of the subject in economics will never be complete.