Designing Gaps as Early Warning Indicators: Accuracy, Predictive Power, and Robustness

23 Pages Posted: 6 Feb 2019

See all articles by Daniel O. Beltran

Daniel O. Beltran

Federal Reserve Board

Mohammad R. Jahan-Parvar

Board of Governors of the Federal Reserve System

Fiona A. Paine

Board of Governors of the Federal Reserve System

Date Written: January 2019

Abstract

Credit gaps, or deviations of credit from its long-run trend, are commonly used in early warning indicators of financial crises. When the gap breaches a certain critical threshold, it triggers a signal. Previous studies set the threshold to balance the trade-off between accuracy and predictive power. We show how this trade-off improves significantly by varying the smoothing parameter inherent in different statistical approaches for estimating gaps. In addition, we compare the performance of alternative gap estimates for advanced and emerging market economies separately, while taking into account the robustness of these gaps in real-time to the arrival of new information.

Keywords: Credit, Credit Gap, Optimization, Predictive Power, Robustness, Trend-Cycle Decomposition

JEL Classification: C22, E39, G28

Suggested Citation

Beltran, Daniel O. and Jahan-Parvar, Mohammad R. and Paine, Fiona, Designing Gaps as Early Warning Indicators: Accuracy, Predictive Power, and Robustness (January 2019). Available at SSRN: https://ssrn.com/abstract=3326228 or http://dx.doi.org/10.2139/ssrn.3326228

Daniel O. Beltran

Federal Reserve Board ( email )

20th Street and Constitution Avenue NW
Washington, DC 20551
United States

Mohammad R. Jahan-Parvar (Contact Author)

Board of Governors of the Federal Reserve System ( email )

20th Street and Constitution Avenue NW
Washington, DC 20551
United States

HOME PAGE: http://sites.google.com/site/mrjahan/

Fiona Paine

Board of Governors of the Federal Reserve System ( email )

20th Street and Constitution Avenue NW
Washington, DC 20551
United States

Register to save articles to
your library

Register

Paper statistics

Downloads
15
Abstract Views
140
PlumX Metrics