This paper argues that the sophisticated techniques presently used by economists to forecast macroeconomic variables behavior (non-structural forecasting methods, in general, and regime-switching models, in particular) do not seem much effective for anticipating radical regime shifts as recently happened in the world economy. Thus, in order to improve their accuracy, it seems that they should be complemented by more holistic approaches. The general purpose of the paper is to show that the system dynamics methodology, which allows identifying the critical feedback loops that drive complex systems' dynamics, seems to be especially fitted to be one of those complementary approaches. To reach that goal, we present a systemic algorithm which allows identifying regime shift processes as the ones that take place when an economy is hit by the effects of a financial bubble burst.
Regime-switching models; System dynamics; Loop dominance analysis; Macroeconomics; Economic growth models
