Pattern of Macroeconomic Indicators Preceding the End of an American Economic Recession
The Journal of Pattern Recognition Research (JPRR) provides an international forum for the electronic publication of high-quality research and industrial experience articles in all areas of pattern recognition, machine learning, and artificial intelligence. JPRR is committed to rigorous yet rapid reviewing. Final versions are published electronically
(ISSN 1558-884X) immediately upon acceptance.
Pattern of Macroeconomic Indicators Preceding the End of an American Economic Recession
V. I. Keilis-Borok, A. A. Soloviev, M. D. Intriligator, F. E. Winberg
JPRR Vol 3, No 1 (2008); doi:10.13176/11.106 
Download
V. I. Keilis-Borok, A. A. Soloviev, M. D. Intriligator, F. E. Winberg
Abstract
This study concerns the fields of economics and the dynamics of complex systems, specifically the process of recovery of American economy from recession. We identify a robust pattern of six macroeconomic indicators that appears within 6 months before the end of each American recession since 1960 and at no other time during these recessions. Its definition is formal and reproducible; as a precursor to the incipient recovery it is corroborated by sensitivity analysis (i.e. variation of its adjustable parameters) and application to out-of-sample data; noteworthy, it emerged before the end of the 2001 recession, which was not being considered while its definition developed. That pattern identified here appears through the whole time considered despite extraordinary changes in the economy. This reflects a well-known general feature of complex systems: they exhibit regular collective behavior patterns, transcending the complexity. Like many other complexity studies, identification of such patterns requires a robust holistic analysis; accordingly we used here a methodology combining pattern recognition of infrequent events and techniques developed in non-linear dynamics. This methodology inherits some of the features of the diffusion indicators of classical business cycle analysis, reformulating these features in the robust pattern recognition language. That includes formulation of the prediction problem per se, given the time series up to a moment t, to recognize whether this moment belongs or not to the last Δ months of recession. In terms of time series analysis our targets of prediction are extreme point events, and prediction is a discrete sequence of alarms; this is different from more traditional (Kolmogoff-Wiener) formulation, where prediction targets and predictors are continuous functions. That methodology is complementary to and compatible with other approaches to predictive understanding of recessions. The present study is a natural continuation of our previous one, aimed at predicting the start of a recession. We find that precursory trends of financial indicators are opposite during transition to a recession and recovery from it. To the contrary, precursory trends of economic indicators happen to have the same direction (upward or downward) but are steeper during recovery.
JPRR Vol 3, No 1 (2008); doi:10.13176/11.106 | Full Text  | Share this paper: