Unemployment hysteresis in Middle East and North Africa countries: panel SUR-based unit root test with a Fourier function

Abstract

ABSTRACT Unemployment hysteresis of Middle East and North African (MENA) countries is investigated under a battery of unit root testing frameworks in the extant literature, including a recently proposed Panel SUR Dickey-Fuller-like unit root test with Fourier and Exponential Smooth Transition Regression (ESTR) nonlinearities. The Fourier function allows for smooth nonlinear breaks, while the ESTR nonlinearity allows for instantaneous breaks. The two nonlinearity types make the recent approach quite appealing. It has, however, been scarcely applied to empirically test the unemployment hysteresis hypothesis. Although we find conflicting stances from ADF, FADF and ADF-SB testing frameworks, evidence of unemployment hysteresis effect in Lebanon is consistent across all three tests. The ADF and FADF tests confirmed the hysteresis hypothesis in Kuwait and Lebanon, while FADF-SB rejected the unemployment hysteresis hypothesis across all the 19 MENA countries. The results from the KSS and FKSS unit root testing frameworks consistently affirmed the hysteresis effect in Oman and Turkey, while there are mixed stances for Kuwait and Lebanon. The results from SURADF and SURKSS only supported the hysteresis hypothesis in Turkey, while the same was confirmed only for Bahrain under the SURFADF and SURFKSS testing frameworks. The unemployment hysteresis hypothesis is confirmed for 12 (about 63.15% of the total number considered) MENA economies.