Macro-financial vulnerabilities and economic downturns : a comparison analysis between cost-sensitive learning and markov-switching approaches

Abstract EN

This paper examines the effects of macro-financial vulnerabilities on economic downturns in North Africa and GCC countries based on a Financial Stress Index (FSI).

The paper identifies episodes of financial turmoil according to FSI values, and proposes an analytical framework to assess the impact of financial stress on economic downturns by using two approaches : a Cost-sensitive learning neural network and a Markov-switching time-varying model.

It concludes that episodes of financial stress can be identified clearly by the two methods in a cooperative but not competitive way.