Introducing the IZA Evaluation Dataset

International Journal of Manpower

ISSN: 0143-7720

Article publication date: 18 October 2011


Giulietti, C. (2011), "Introducing the IZA Evaluation Dataset", International Journal of Manpower, Vol. 32 No. 7.



Emerald Group Publishing Limited

Copyright © 2011, Emerald Group Publishing Limited

Introducing the IZA Evaluation Dataset

Article Type: Guest editorial From: International Journal of Manpower, Volume 32, Issue 7

About the Guest Editor

Corrado Giulietti Research Associate and Deputy Program Director of Migration at Institute for the Study of Labor (IZA), Bonn. He is also visiting associate at the ESRC Research Centre for Population Change (UK). His research interests include labor economics and applied econometrics, in particular the determinants and impacts of migration, the assimilation of immigrants, the role of social networks on immigrants’ outcomes and the estimation of migration flows. His publications have appeared in refereed journals and book chapters. He has also worked on several immigration research projects.

In recent years, the evaluation of labor market policies has become a crucial objective for central and local governments, and it has received increasing attention from economic research as well. Indeed, the quest for an efficient assessment of policies was amplified by the recent economic turmoil, during which the implementation and success of labor market interventions were hindered by both the rising unemployment rates and the contemporaneous contraction of government budgets.

A major challenge in the evaluation of labor market policy lies in the scarce availability of high-quality data. On the one hand, as more complex econometric methodologies are developed, more accurate information is needed to provide an efficient assessment of a given program; on the other hand, since both the policy and the evaluator typically target specific subsets of the population, an enormous amount of information is required in order to achieve meaningful statistical inference, and this often clashes with data confidentiality regulations.

The IZA Evaluation Dataset, created within a large research project, possesses the required virtues to provide an efficient evaluation of labor market policies. The dataset was developed by enriching administrative data from the Federal Employment Agency with longitudinal survey data in Germany. It consists of a large sample of unemployment entrants followed over time, for which detailed information on program participation, employment history and job search behavior are available. Furthermore, the dataset is enriched by numerous survey questions about personality traits, economic preferences, social networks and life satisfaction. To prevent the disclosure of the confidential content of the IZA Evaluation Dataset, access is restricted to project members authorized by the Federal Employment Agency and data must be processed in safe settings.

The aim of the special issue contained in this volume of the International Journal of Manpower is to present first evidence on the search behavior of unemployed individuals and its interrelation to labor market policies. Although the external validity of these studies is limited to the case of Germany, they provide important insights about unexplored aspects of job search and the impact of labor market policy.

The issue begins with the article “The IZA Evaluation Dataset: towards evidence-based labor policy-making”, by Caliendo, Falk, Kaiser, Schneider, Uhlendorff, van den Berg and Zimmermann. This is a useful companion paper for all the studies that will be based on the dataset. It contains a comprehensive description of how administrative data from the German Federal Employment Agency have been combined with complementary panel survey data to form the IZA Evaluation Dataset. Furthermore, the article contains an exhaustive summary of the sampling framework and of the available variables, along with an outline of some studies based on the dataset.

One of the potential explanations behind the observed gaps in the labor market performance of natives and immigrants in Germany is that these two groups differ in their job search behavior. This hypothesis is explored in the second article of the special issue entitled “Ethnicity, job search and labor market reintegration of the unemployed” by Constant, Kahanec, Rinne and Zimmermann. Here, the role of ethnic identity and ethnic self-identification in the process of job search is analyzed using the sample of recently unemployed natives and immigrants from the IZA Evaluation Dataset. One of the key findings is that separated foreign-born, i.e. those who have strong attachment to their home country and are scarcely assimilated in Germany, tend to experience longer unemployment spells. The relatively low search effort used by the immigrants is discussed as potential explanation behind this result. On the one hand, the findings of this article highlight important differences in the search behavior of immigrants, and on the other hand, they emphasize the need for labor market policies to take these differences into account.

The next article is entitled “The threat effect of participation in active labor market programs on job search behavior of migrants in Germany” by Bergemann, Caliendo, van den Berg and Zimmermann. The main feature of this article is that it exploits the comprehensive information on attitudes and expectations of the IZA Evaluation Dataset in order to elicit the search behavior of unemployed individuals in relation to their subjective probability of participating in active labor market policies. In particular, the authors explore whether differences exist between natives and immigrants from different countries of origin. An interesting finding is that immigrants from Central and Eastern Europe increase their job search effort if they perceive that they will be treated in the near future. Similarly to the article by Constant, Kahanec, Rinne and Zimmermann in this issue, these findings confirm the existence of substantial differences in the search behavior of natives and immigrants.

The fourth article “Social networks, job search methods and reservation wages: evidence for Germany” by Caliendo, Schmidl and Uhlendorff explores the impact of social network size on the choice of job search methods. Using the detailed information on job search available in the IZA Evaluation Dataset, the article provides evidence that unemployed individuals with larger networks are more likely to seek help from friends, family and colleagues, rather than relying on more formal methods, such as job centers or internet applications. A larger network is also associated with relatively higher reservation wages. The findings of this article shed light on the complex mechanisms linking social networks and labor market outcomes.

This special issue is concluded by the article “Economic preferences and attitudes of the unemployed: are natives and second generation migrants alike?” by Constant, Krause, Rinne and Zimmermann. The article investigates whether natives and second generation immigrants differ in terms of their risk attitudes, time preferences, trust and reciprocity. One of the interesting findings is that the omission of risk aversion covariates in the employment equation would lead to an overestimation of the employment gap between natives and second generation immigrants. While this bias is relatively moderate and does not compensate for the substantial gaps between natives and second generation immigrants, it suggest the importance of taking into account personal traits and economic preferences in the analysis of labor market outcomes of these two groups.

The articles in this issue shed light on many aspects of unemployed individuals and their job search behavior. At the same time, these works offer insight for posing new questions. The IZA Evaluation Dataset will allow not only assessing the impact of labor market policies on outcomes such as employment and wages, but also on subjective well-being. Moreover, the interaction between different types of labor market policies can be analyzed to investigate whether they are complement or substitute to each other. A common finding of the papers in this issue is that natives and immigrants are quite heterogeneous in terms of their search behavior, and that immigrants themselves differ depending on their origin and ethnicity. Hence, interesting questions are: How do natives’ and immigrants’ search behaviors differ in presence of prolonged or multiple unemployment spells? How do ethnic social networks influence program participation? How do the effects of active labor market programs differ between migrants and natives? The unique comprehensive information available in the IZA Evaluation Dataset, as well as the availability of its future waves, will allow investigating these and many more research questions in depth.

Corrado Giulietti