Testing the Hysteresis Effect in Age- and Sex-Disaggregated Unemployment Rates in Türkiye: Evidence from the Next Generation Autoregressive Neural Network Unit Root Test

Muhammet Daştan

Article Details

Testing the Hysteresis Effect in Age- and Sex-Disaggregated Unemployment Rates in Türkiye: Evidence from the Next Generation Autoregressive Neural Network Unit Root Test

Authors:

Muhammet Daştan

Publisher:

Akdeniz University

Main Topics:

İstihdam

Journal:

Akdeniz İİBF Dergisi Vol.24 Issue 1, pp.27-38

Document Type:

Article

Publication Date:

5/2024

Document Language:

Turkish

Document ID (Org No.):

2001311665

pISSN:

1302-9975

Keywords:

Abstract:

 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.

 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.

 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.

 Bu çalışmada, Türkiyede yaş ve cinsiyete göre ayrıştırılmış işsizlik oranlarındaki olası histeri etkisi Ocak 2005  Ekim 2023 dönemi için aylık veriler kullanılarak araştırılmaktadır. Çalışmada serilerin durağanlık özellikleri test edilirken, öncelikle Augmented Dickey-Fuller ve Phillips-Perron testleri gibi geleneksel birim kök testleri ile Zivot-Andrews, Lee-Strazicich ve Narayan-Popp testleri gibi serilerde bir veya iki yapısal kırılmaya izin veren birim kök testlerinden faydalanılmıştır. Ardından, işsizlik serilerinin doğrusal olmayan bir süreç izlediği göz önünde bulundurularak Yaya vd. 2021 tarafından geliştirilen yeni nesil ADF tipi Otoregresif Sinir Ağları ARNN-ADF birim kök testi kullanılmıştır. Çalışmada model tabanlı yöntemlerin aksine veri odaklı ve sistematik olarak doğrusal, kuadratik ve kübik şeklinde üç ayrı bileşenin tek bir doğrusal olmayan teste dönüştürüldüğü ARNN-ADF birim kök testinin kullanılması, güvenilir sonuçlara ulaşılabilmesi açısından önem arz etmektedir. Ampirik bulgular, tüm bu testlerin işsizlik oranlarında birim kök sıfır hipotezini reddetmede başarısız olduğunu göstermekte ve böylece Türkiyedeki işsizlik oranlarındaki histeri etkisinin yaş ve cinsiyet farklılıklarından bağımsız olarak geçerli olduğunu ortaya koymaktadır. Bu sonuçlar, Türkiye ekonomisindeki ani şoklardan sonra işsizlik oranlarının trend patikasına geri dönmediğine yönelik kanıt sunarken aynı zamanda işgücü piyasasının olumsuz şoklara karşı dayanıklılığını güçlendirmek amacıyla tasarlanan istikrar politikalarının başarılı olabileceğine işaret etmektedir.

Article Details

Testing the Hysteresis Effect in Age- and Sex-Disaggregated Unemployment Rates in Türkiye: Evidence from the Next Generation Autoregressive Neural Network Unit Root Test

Authors:

Muhammet Daştan

Publisher:

Akdeniz University

Main Topics:

İstihdam

Journal:

Akdeniz İİBF Dergisi Vol.24 Issue 1, pp.27-38

Document Type:

Article

Publication Date:

5/2024

Document Language:

Turkish

Document ID (Org No.):

2001311665

pISSN:

1302-9975

Unemployment Hysteresis

ARNN-ADF Unit Root Test

Turkish Economyİşsizlik Histerisi

ARNN-ADF Birim Kök Testi

Türkiye Ekonomisi

Issues

Article Details

Testing the Hysteresis Effect in Age- and Sex-Disaggregated Unemployment Rates in Türkiye: Evidence from the Next Generation Autoregressive Neural Network Unit Root Test

Authors: Muhammet Daştan

Publisher: Akdeniz University

Main Topics: İstihdam

Journal: Akdeniz İİBF Dergisi Vol.24 Issue 1, pp.27-38

Document Type: Article

Publication Date: 5/2024

Document Language: Turkish

Document ID (Org No.): 2001311665

pISSN: 1302-9975

Unemployment Hysteresis

ARNN-ADF Unit Root Test

Turkish Economyİşsizlik Histerisi

ARNN-ADF Birim Kök Testi

Türkiye Ekonomisi

Issues

Authors:

Muhammet Daştan

Publisher:

Akdeniz University

Main Topics:

İstihdam

Journal:

Akdeniz İİBF Dergisi Vol.24 Issue 1, pp.27-38

Document Type:

Article

Publication Date:

5/2024

Document Language:

Turkish

Document ID (Org No.):

2001311665

pISSN:

1302-9975