Unemployment Persistence in The Turkish Labor MarketSelçuk Gemicioğlu, Hasan Şahin
The aim of this study is to investigate the existence of unemployment persistence in the Turkish labor market. In this regard, the existence and degree of genuine state dependence was analyzed using the data obtained from the 2013-2016 Survey on Income and Living Conditions and Household Labor Force Survey. The dynamic random effects probit model estimation results showed that there is genuine state dependence in the Turkish labor market. In other words, after controlling for observed and unobserved heterogeneity and initial conditions, past unemployment experiences increased the probability of being unemployed in the future. In addition, the marginal effect of genuine state dependence was found to be higher for the unemployed individuals than for the individuals employed in the initial period. As one’s education level increases, the likelihood of being jobless declines, however, men are more likely than women to do so. Age, gender, education level, marital status, and the number of household employees all have an impact on the likelihood of being unemployed. Demand-side factors such as the unemployment rate and growth rate of employment do not affect the probability of remaining unemployed.
Türkiye İşgücü Piyasasında Kalıcı İşsizlikSelçuk Gemicioğlu, Hasan Şahin
Bu çalışmanın amacı Türkiye işgücü piyasasında kalıcı işsizliğin varlığını araştırmaktır. Bu doğrultuda 2013-2016 Gelir ve Yaşam Koşulları Araştırması ve Hanehalkı İşgücü Araştırmasından elde edilen veriler kullanılarak gerçek duruma bağımlılığın varlığı ve boyutu analiz edilmiştir. Dinamik rassal etkiler probit modeli tahmin sonuçları Türkiye işgücü piyasasında gerçek duruma bağımlılığın olduğunu göstermiştir. Diğer bir ifade ile gözlenen ve gözlenemeyen heterojenlik ve başlangıç koşulları kontrol edildikten sonra, geçmiş dönemde yaşanan işsizlik deneyimlerinin gelecek dönemde işsiz olma olasılığını artırdığı ortaya çıkmıştır. Bunun yanı sıra gerçek duruma bağımlılığın marjinal etkisi başlangıç döneminde istihdam edilen bireylere göre işsiz olan bireyler için daha yüksek bulunmuştur. Eğitim seviyesi azaldıkça işsiz kalma olasılığının azaldığı görülürken kadınlara kıyasla erkeklerin daha yüksek işsiz kalma olasılığına sahip olduğu görülmüştür. İşsiz olma olasılığının yaş, cinsiyet, eğitim seviyesi, medeni durum ve hanede çalışan birey sayısından etkilendiği sonucuna ulaşılmıştır. İşsizlik ve istihdamın büyüme oranı gibi talep yönlü faktörlerin ise işsiz olma olasılığını etkilemediği ortaya çıkmıştır.
The aim of this study is to examine the existence of unemployment persistence in the Turkish labor market. In other words, it aims to investigate the existence of genuine state dependence, which means that past experiences of unemployment cause unemployment again in the future. In this study, we sought answers to several questions: Does the degree of genuine state dependence vary according to the initial conditions and unobservable heterogeneity? What is the expected duration of the spell of unemployment and the probability of entry and exit to unemployment, unemployment persistence, and steady state, according to individual characteristics that be constant over time such as gender and education level? What are individual factors that increase the probability of being unemployed?
Unemployment persistence is based on two factors (Arulampalam, Booth & Taylor, 2000; Stewart, 2007). The first factor is the heterogeneity, which refers to individual differences that are either observe or not. Genuine state dependence is the other factor. Genuine state reliance is when a past event raises the probability of experiencing the same event in the future (Heckman, 1981a). Theoretically, genuine state dependence may occur due to a loss of human capital, unemployment insurance disincentives, stigmatization by employers, a decline in search intensity, habituation, or discouragement (Ayllon, 2013). Empirical findings have shown that there is genuine state dependence on unemployment across many countries (Gangji & Plasman, 2008; Biewen & Steffes, 2010; Plum & Ayllon, 2015; Fok, Scutella, & Wilkins, 2015; Pasaribu, 2016; Cai, Mavromaras, & Sloane, 2018). The research investigated genuine state dependence on unemployment and utilise dynamic models and estimators. The Dynamic Random Effects Probit Model is the most preferred of these models. Among the estimators, Heckman, Orme, and Wooldridge estimators are used depending on the information provided by the data. As in previous studies (Arulampalam et al., 2000; Stewart, 2007; Gangji & Plasman, 2008; Pasaribu, 2016), we used the Dynamic Random Effects Probit Model to examine the existence of genuine state dependence. We estimated parameters using another version of the Wooldridge maximum likelihood estimator, which was suggested in Rabe-Hesketh and Skrondal (2013). This estimator offers reliable parameter estimates and enables control over both observed and unobserved heterogeneity. It also addresses the issue of beginning conditions. To investigate genuine state dependence on unemployment in the Turkish labor market, we estimate the Dynamic Random Effects Probit Model using the Survey on Income and Living Conditions and Household Labor Force Survey. The evidence shows that there is genuine state dependence on the Turkish labor market. In addition, according to the components of the unobserved heterogeneity, the conditional unemployment probability and the marginal effects of genuine state dependence vary significantly. The probability of conditional unemployment and the marginal effect of genuine state dependence are both significantly increased by being unemployed in the first period and being in the last two percent of the sample distribution in the part of the unobserved heterogeneity linked to the explanatory variables. Demand-side factors in the labor market such as the unemployment rate and employment growth rate do not affect the probability of unemployment, while factors such as gender, education level, marital status, and the number of employees in the household affect the probability of unemployment. When it comes to the probability of remaining unemployed, disadvantageous groups include men and those with low levels of education. To fight unemployment persistence, policymakers should develop policies that take into account factors such as a loss of human capital, unemployment insurance disincentives, stigmatization by employers, a decline in search intensity, habituation, or discouragement, which cause genuine state dependence. Making pre-graduate internships, which are currently applied in some departments of universities, compulsory for all departments to meet the experience requirement for newly graduated individuals by employers can accelerate and facilitate the transition to employment.
Unemployed individuals suffer from a loss of human capital, skills, and qualifications as they remain unemployed, and thus employers may adopt negative attitudes towards the employment of individuals who are unemployed for a long time. In order to protect the unemployed from these negative situations, governments can implement programs that protect and increase the skills and qualifications of the unemployed. These programs can be very important, especially for skilled workers working in technology-intensive sectors where high-added value can be created. Modifying unemployment insurance benefits is another policy that may be put into place to help the unemployed find work more quickly and to make sure that people stay on the job market and keep looking for work.