This paper explores the potential hysteresis effect in age and gender-disaggregated unemployment rates in Türkiye by using monthly data for the period January 2005 – October 2023. When examining the stationarity characteristics of the series, the study employs both conventional unit root tests, such as the Augmented Dickey-Fuller and Phillips-Perron tests, and unit root tests that permit one or two structural breaks in the series, such as Zivot-Andrews, Lee-Strazicich, and Narayan-Popp tests. Then, the study employs a new generation ADF-type Autoregressive Neural Network (ARNN-ADF) unit root test introduced by Yaya et al. (2021), considering that unemployment series potentially follow a non-linear trajectory. Contrary to model-based procedures, the ARNN-ADF unit root test, which is data-driven and systematically converts three discrete linear, quadratic, and cubic components into a single nonlinear test, is critical for producing credible outcomes. The empirical findings indicate that all these tests fail to reject the null hypothesis of a unit root in unemployment rates, suggesting that the hysteresis effect in unemployment rates in Türkiye is valid, regardless of age and gender differences. These outcomes provide evidence that unemployment rates are not reverting to their trending path after sudden shocks in the Turkish economy while at the same time pointing out that stabilization policies designed to strengthen the resilience of the labor market against adverse shocks can be successful.

Keywords