Labor-Market Scars When Youth Unemployment is Extremely High: Evidence from Macedonia

Abstract

The objective of this article is to assess how the duration of unemployment spells for Macedonian youth affects their later employment (employment “scarring”) and wage outcomes (wage “scarring”). The discrete-time duration method is used to determine the unemployment spell and the standard employment and Mincer earnings functions. The School-to-Work Transition Survey 2012 is used. The results robustly suggest the presence of employment scarring: youth who remain unemployed over longer time periods have lower chances to find a job afterwards. However, the article provides no evidence for the existence of wage scarring in the Macedonian labor market.

Keywords:

• employment scarring
• extremely high unemployment
• Macedonia
• wage scarring

JEL Classification:

• E24
• J24
• J64

Notes

1. This article does not investigate the underlying causes of youth unemployment; a good summary of the main theories and determinants of the issue is presented in Brada, Marelli, and Signorelli (2014).

2. In fact, wage is observed only in 49% of the total number of employed youth. The reason is that 30% and 7% of the employed in the sample are unpaid family workers and self-employed, respectively, and data on their wages are missing or unobserved. The rest claimed to have received their salary in kind or simply refused to state an amount.

3. In the present case, three endogenous dummies are instrumented with one instrument, the prime line of justification being that the three endogenous dummies reflect one single phenomenon—the unemployment spell. Since one might still object, the results of another estimator utilizing internally generated instruments are presented further on. In addition, the article relies on the repeated imputation technique in estimating the wage equation. Hence, ivreg (or a similar command) in Stata is technically not feasible. Therefore, the IV correction is conducted manually, as suggested in Wooldridge (2002) or in: http://www.stata.com/support/faqs/statistics/instrumental-variables-regression. Providing more than the usual under- and weak-identification statistics is not possible due to the specific methodological construct used here. However, some robustness checks are conducted below.

4. More precisely, the discussion uses the predicted unemployment spell duration from the equation where it is regressed on individual characteristics only, plus the regional unemployment rate upon school-completion. The former are variables entering the wage regression (included instruments), while the latter is an exclusion restriction (excluded instrument).

Additional information

Notes on contributors

Marjan Petreski

Marjan Petreski is an Associate Professor in the School of Business Economics and Management at the University American College, Skopje, Macedonia. Nikica Mojsoska-Blazevski is a Full Professor in the School of Business Economics and Management at the University American College, Skopje, Macedonia. Marcelo Bergolo is an Assistant Professor in the Instituto de Economia (IECON), Universidad de La República, Montevideo, Uruguay.

Nikica Mojsoska-Blazevski

Marjan Petreski is an Associate Professor in the School of Business Economics and Management at the University American College, Skopje, Macedonia. Nikica Mojsoska-Blazevski is a Full Professor in the School of Business Economics and Management at the University American College, Skopje, Macedonia. Marcelo Bergolo is an Assistant Professor in the Instituto de Economia (IECON), Universidad de La República, Montevideo, Uruguay.

Marcelo Bergolo

Marjan Petreski is an Associate Professor in the School of Business Economics and Management at the University American College, Skopje, Macedonia. Nikica Mojsoska-Blazevski is a Full Professor in the School of Business Economics and Management at the University American College, Skopje, Macedonia. Marcelo Bergolo is an Assistant Professor in the Instituto de Economia (IECON), Universidad de La República, Montevideo, Uruguay.