Search

The share of long-term unemployed in Greece, even during periods of low unemployment rate, is quite high. Even in 2008, when the unemployment rate stood below 8% at its lowest level, the long-term unemployed accounted for 40% to 45% of all unemployed. During the great crisis, three out of four unemployed remained unemployed for 12 months or longer, while in 2020 this share is estimated around 66%. Motivated by the fact that the long-term unemployment rate is systematically high, this study contributes to understanding Greek unemployment dynamics by documenting on the extent of unemployment persistence, as well as by analysing the extent to which observed unemployment persistence may be the result of stigmatization of unemployed. In the literature there is ample evidence that unemployed individuals tend to stay unemployed. This may be due to observed heterogeneity, such as low qualifications or lack of experience, or unobserved heterogeneity, such as lack of ability or motivation. As long as observed and unobserved characteristics persist over time, they will tend to increase the likelihood of future unemployment, creating a relationship between current and future unemployment incidence. However, one cannot rule out the possibility that the actual experience of unemployment has a genuine causal effect on future unemployment, in the sense that past unemployment causally increases the unemployment probability in the future regardless of other observed and unobserved factors. In other words, there may exist genuine state dependence. This paper provides evidence on the magnitude of state dependence in individual unemployment incidence. Moreover, we attempt to explain whether this persistence is a result of stigma effects by exploring if unemployed individuals face systematically lower chances of being hired because employers interpret their unemployment as a negative signal. The main hypothesis underlying our empirical approach is that the stigma effect of being unemployed is low when aggregate unemployment is high, as in this case individual unemployment is not a strong signal for low individual productivity, while high in times of low unemployment. To check our hypothesis, we utilize data from the Greek Labour Force Survey and estimate the probability of being unemployed conditional on the employment state of the previous period (i.e., the state dependence effect), an interaction of past unemployment and measures of cyclical unemployment risk, as well as other covariates. We use a dynamic nonlinear correlated random effects probit model taking into account the initial conditions problem. Moreover, in nonlinear models, the interpretation of the coefficient on the interaction term of past unemployment and cyclical unemployment, which is crucial in our analysis, is not as straightforward as in linear models and requires the estimation of the cross-partial derivative****. Our initial but firm results show that positive deviations of the unemployment rate from its trend are indeed associated with a significant lower level of state dependence. In other words, the penalty for being unemployed is smaller in times of high unemployment and larger in times of low unemployment, suggesting that there is a stigma effect in the Greek labour market. This stigma effect seems to vary by gender and age. The existence of stigma on unemployment persistence is confirmed with various robustness checks and tests. However, its magnitude is not sufficiently large to explain by itself the systematically big share of long-term unemployed in Greece. Our findings have important policy implications on the design of active labour market policies to fight unemployment. The fact that those longer in the unemployment face more disadvantages when striving to find a job because of stigma, especially in "good times", suggests that labour market policies should aim at preventing long-term unemployment and obstructing it from becoming permanent unemployment. In other words, short term policies to prevent unemployment and minimize its duration will have further long-term effects. It also suggests, given the vast pool of long-term unemployed in Greece, that radical interventions are required to mobilise the longterm unemployed and help them overcome the stigma. [ABSTRACT FROM AUTHOR]