Assessing the probability of economic slowdowns: a pattern-based approach

Forecasting periods of economic slowdown, or even contraction, is a crucial issue for economic and financial players.


As part of its research, EthiFinance Analytics has focused on assessing the probability of economic events based on macro-financial data. As a demonstration, Elizabeth Zuniga, David Labeurthre and Édouard Pineau applied their approach to assess the probability of economic slowdowns.

To do this, the leading index is transformed into a series of eight states, defined at each time step by the intensity of previous monthly variations. The states are then interpretable, from a strong decrease to a strong increase. From these states, events are defined as patterns of consecutive states. We can then calculate a slowdown probability by counting the co-occurrence of an event and an economic slowdown within the next six months, and dividing by the number of events.

The signal obtained shows that the intensity patterns of consecutive variations contain information on the position in the economic cycle and, in particular, make it possible to predict economic slowdowns.