Regardless of the fact that economics distinguishes itself from other social sciences by a high level of formal deductive modelling, it is a social science due to the essence of the economic process where a human is a subject and an object at the same time. In recent years this issue has been more frequently emphasized by economists in ongoing discussions. In the discussions, a good deal of time is devoted to economic models and, mainly, their relations with the socioeconomic reality and coherence of empirical evidence. The article presents a thesis that some mainstream economic theories have not always constituted the background to their practical applications, which led - and still can lead- to the dogmatic and inflexible use of model solutions for economic phenomena which are difficult to forecast in a non-variant rigid model. The aim is to critically analyse the beliefs about the usefulness of universal economic models in the economic reality advocated by mainstream economists, and to prove that not all economic models have constituted the background to their practical applications.