Unemployment and under-employment:
the case of Switzerland
Dragana Djurdjevic *
SIAW, Swiss Institute for International Economics and Applied Economic Research,
First Version: August 2003
This Version: November 2003
I analyze the effect of unemployment on subsequent employment history for
Switzerland. Using administrative panel data from the unemployment
insurance system and the social security databases, I estimate a discrete time
hazard model for the exit from the different labor market states:
unemployment, employment characterized by either earnings losses, gains or
relatively stable earnings, and out-of-the labor force. I find that having
previously experienced unemployment increases the risk of persistent
unemployment. A further analysis based on personal characteristics such as
gender, age, foreign citizenship and qualification permits to identify two
profiles of unemployed persons. The “higher-risk” unemployed, namely the
female, foreign and less skilled workers are prone to remain trapped in bad
situations or to experience employment instability. On the contrary, the male,
younger and skilled workers are more likely to exit from unemployment and if
they experience earnings losses, it is more for transitory periods.
Keywords: exclusion, wages, discrete time hazard model
JEL Code: J64
Acknowledgements: I am grateful to the State Secretariat of Economic Affairs
of the Swiss Government (Seco) and the Bundesamt für Sozialversicherung
(BFS) for providing data. Financial support from the Swiss National Science
Foundation (4045-059673) is gratefully acknowledged. I am indebted to
François Laisney and Phu Nguyen Van for their help and encouragement. I
received helpful comments from Michael Lechner, Patrick Puhani, Ruth
Miquel and Blaise Melly. I also thank Heidi Steiger for her assistance in
preparing the data. I am finally grateful to Jacques Silber, Reinhold Schnabel
and Said Hanchane. All remaining errors are mine.

• Address for correspondence : SIAW, Dufourstrasse 48, CH-9000 St. Gallen , Switzerland ; phone : ++41 (0)
71 224 27 67 ; fax : ++41 (0) 71 224 22 98 ; e-mail : dragana.djurdjevic@unisg.ch ;
• web page :www.siaw.unisg.ch/lechner.
1 Introduction
Since the beginning of the 1990s, unemployment has become a source of worry among the
OECD countries. High unemployment rates for the period 1989-1999, as well as substantial
differences between countries, in particular between the US and the European countries, were
observed. In addition, the number of persons that employment agencies find difficult to place,
especially those hit by long-term unemployment increased substantially. This last decade has
thus shown a growing interest in the problem of long-term unemployment. While the existing
literature mainly focused on this topic, little attention was paid to the issue of repeated
unemployment. However, the movements between unemployment and lower paid
employment can lead to the same exclusion problems as long-term unemployment.
This paper investigates the impact of unemployment on subsequent employment
history for the case of Switzerland. The aim of the analysis is twofold: I identify the
determinants of the risk of exiting unemployment as well as those for the risk of re-entering
First, the analysis of exit from unemployment permits to address the question whether
unemployment exerts a negative effect on the stability of subsequent earnings, and to what
extent it may facilitate the withdrawal from the labor market. This question is relevant in
terms of policy implications: as unemployment may affect the workers differently, some
policies may not suit some workers. For instance, training programs are more likely to target
the male and the less skilled unemployed, while measures aimed at encouraging participation
tend to be more appropriate to female and elderly workers than to prime age male workers
because it seems that this latter category of workers always participates in the labor market.
However, the existing economic applications usually study the consequences of
unemployment on subsequent earnings without considering its impact on non-participation.
Most of the existing applications concern the US and focus on the effect of a job loss on
displaced workers (see the studies by Ehrenberg and Oaxaca, 1976, Addison and Portugal,
1989, Ruhm, 1991, Jacobson et al, 1993, Houle and van Audenrode, 1995, for the US and the
recent studies by Arulampalam, 2001, and Greggory and Jukes, 2001 for the UK). These
studies focus on workers highly attached to the labor market, namely the high-tenure and male
workers. They report evidence for significant long-lasting earnings losses associated with job
displacement. These findings are not surprising because little emphasis is laid on workers
weakly attached to the labour force. These workers like the female and the elderly workers are
more likely to withdraw from the labor market after the end of an employment spell. A survey
by Layard, Nickell and Jackman (1991) shows indeed that half of the unemployment spells in
the US end in withdrawal from