Hysteresis in Youth Unemployment in Sub-Saharan Africa: Focus on Gender, Income, and Region

ABSTRACT: Despite innumerable government policies to reduce youth unemployment in Sub-Saharan Africa, the problem persists. This study examines the extent to which the hysteresis effect exists in the youth unemployment rate in Sub-Saharan African countries during the period 1991 to 2021, disaggregated according to territorial delineation, income, and gender. First, we employ the univariate and panel unit root tests including the Augmented-Dickey-Fuller, Phillips-Perron, the Dickey-Fuller-Generalized Least Square by Elliott, Rothenberg, and Stock (1996), the ADFmax by Leybourne (1995), and the Kwiatkowski, et al. (1992) test, to enable comparison with previous studies. The Pesaran's (2007), test is then applied to account for cross-sectional dependence. Finally, the Zivot–Andrews single structural break by Zivot and Andrews (1992), Clemente–Montane–Reyes two structural breaks unit root tests by Clemente, Montañés, and Reyes (1998), as well as the Karavias-Tzavalis panel unit root test with structural breaks by Karavias and Tzavalis (2014) are employed. The univariate unit root tests generally support the hysteresis hypothesis (within a band of 60% to 97%) for almost all countries, and sub-regions in SSA. However, accounting for structural breaks, the hysteresis effect in total youth unemployment is only visible in 19 countries (40%), with evidence of gender disparities pointing to the hysteresis effect in females only in 16 countries (34%), and, in males only in 10 countries (21.3%). The lower-middle-income countries reject the natural rate hypothesis in total and both male and female youth unemployment rates. While only male youths suffer from the hysteresis in low-income countries, the total youth in upper-middle-income countries shares this effect. Accounting for cross-section dependence alters the finding of the hysteresis effect only in males in Southern Africa and upper-middle-income countries. A one-size-fits-all policy may not work. Instead, we recommend an integral approach aimed at the creation of employment opportunities for youth and females, including but not limited to diversification of economic activities in rural areas to reduce urbanization, skills training, political and economic stability, cultural transformation to remove unnecessary gender-related employment discrimination, as well as improved economic integration policies. For countries found to exhibit the natural rate hypothesis, it may not be necessary to engage in costly policies to return the youth unemployment rate to equilibrium since the shock is only temporary.