Pre-recession pattern of six economic indicators in the USA

Abstract

This paper applies a tightly parameterized pattern recognition algorithm, previously applied to earthquake prediction, to the problem of predicting recessions. Monthly data from 1962 to 1996 on six leading and coincident economic indicators for the USA are used. In the full sample, the model performs better than benchmark linear and non-linear models with the same number of parameters. Subsample and recursive analysis indicates that the algorithm is stable and produces reasonably accurate forecasts even when estimated using a small number of recessions.

Predictability of extreme events in a branching diffusion model

“…Im-portantly, these patterns are universal, common for complex systems of distinctly different origin. Similar premonitory patterns have been observed in socio-economic systems [39,40], dynamic clustering in elastic billiards [42], hydrodynamics, and hierarchical models of extreme event development [43][44][45][46][47][48]. We propose here a general mechanism that reproduces these universal premonitory patterns.…”

At-Risk Transformation for U.S. Recession Prediction

“…The evidence presented in this paper suggests that a simple binarization of predictors-the "at-risk" transformation-is a powerful tool for recession forecasting. Building on the foundational insight of Keilis-Borok et al (2000), who first applied this idea to a small set of indicators in a simpler setup, we demonstrate the effectiveness of a similar approach in a modern, high-dimensional forecasting environment. Our recursive out-of-sample analysis shows that models using these binary features are not only highly competitive but often superior to benchmarks that use standard continuous data, including machine learning methods like XGBoost.…”

Reverse tracing of short-term earthquake precursors

“…Similarly sporadic are many observed precursors to other critical phenomena, e.g. economic recessions (Keilis-Borok et al, 2000).…”