In order to develop reasonable measures to counteract the destructive consequences of economic crises or to prevent them, it is important to build reasonable and adequate models of the emergence and course of crisis processes. The formal compliance of existing models with socio-economic processes during the crisis is often incomplete or cannot be adapted to other conditions, economies, and markets. Therefore, the article is devoted to the expediency and possibility of transformation of the existing paradigm of modeling economic crises.Modeling the interdependence of two basic indicators, which characterize the crisis in the economic development of European countries (GDP growth rates and employment growth rates), refuted the possibility of their mutual determination. The interdependence of the dynamics of GDP growth rates and employment growth rates for all studied objects without a shift in time, determined for the period 1996-2020 with a high level of reliability, is of the same type and made it possible to determine the range of their mutual changes. According to this range and the parameters of the codependencies of GDP growth rates and employment growth rates, eight clusters of countries were determined by the method of finding concentrations. The emergence of crisis processes coincides with specific parameters of the interdependent dynamics of GDP growth rates and employment growth rates but is not related to the internal division of Eurozone countries into clusters.The obtained results gave reason to put forward an alternative paradigm for modeling crisis processes, according to which economic crises arise as a result of the impact of random destructive events on systemic coincidences of potential periods of shifts in the contingent regularities of partial self-similar dynamics of economic processes. Within this paradigm, modeling of economic crises and forecasting their occurrence will be based on the study of self-similarity, dissipativeness and contingency of the dynamics of economic processes.