Abstract:
The purpose of this article is to demonstrate that the usefulness of econometric models is quite limited in the field of
social sciences, particularly in economic science. From Walras to our days, the economists imagine that macroeconomic
models can be built to serve as a decisional basis for economic policy. The social phenomena are characterized by
complexity and continuous change. The contemporary human societies are composed by a lot of people. The direct or
indirect contacts between individuals, possible by means of rapid communication, generate a complex and fluid social
order. The complexity and the fluidity of social order prevents the economic system representation in the form of
macroeconomic models. The main problem of such models is that they contain an excessive number of variables. The
huge amount of knowledge generated by the relations between individuals cannot be structured in a systematic manner.
Therefore, it can not be taken into account, modeled and introduced into macroeconomic models. They would be
helpful if there were the possibility of obtaining, processing and use real-time knowledge of the entire volume generated
by the social order. But this is impossible for two reasons. First, the human knowledge is manifested in time and space,
it is dissipated among members of the social order. The costs of concentrate it are impossible to cover. Second, the
human knowledge is essentially subjective, unspeakable and non-transferable. For these reasons, the macroeconomic
models can only be an oversimplification of reality. Basically, they cannot be used to make predictions. In the best case,
can help to realize qualitative predictions, but not a quantitative ones.
