IZA dataset workshopThe potential of the dataset was highlighted at a recent workshop on “Policy Lessons from the IZA Evaluation Dataset” held at IZA in Bonn. The main goal of the workshop was to summarize the existing results based on the IZA Evaluation Dataset and to serve as a platform to discuss how data access and quality could be further facilitated in order to create the basis for evidence-based policy advice in the future. The clear message sent out by this workshop was that high quality research requires access to high quality data. Klaus F. Zimmermann, Director of IZA and co-organizer of the workshop, emphasized that this dataset is a successful attempt to provide adequate data to the research community, allowing new studies on a wide range of research questions.
During the first day of the workshop the focus was on scientific contributions based on the IZA Evaluation Dataset. Data users from different international institutions presented their frontier research. The second day of the workshop addressed a policy oriented audience by presenting empirical evidence and resulting policy implications based on the IZA Evaluation Dataset to policymakers and practitioners in order to illustrate the importance and value added of access to high quality data. The panel discussion with practitioners and policymakers at the end of the second day finally served as a forum to exchange experiences and ideas, and to express the expectations held by researchers and policymakers.
After an introduction by Klaus F. Zimmermann, Annette Bergemann (University of Mannheim and IZA) gave the first presentation on the evaluation of ex ante effects of ALMP programs. For instance, individuals might prefer to avoid program participation due to different reasons and therefore increase job search activities. Those anticipation effects are usually hard to determine as this often requires unobservable information on the program assignment and the job search behavior of the unemployed. The IZA Evaluation Dataset provides the required information. Bergemann and co-authors find that anticipation effects exist in the sense that individuals adjust their reservation wage and search effort if they perceive a high treatment propensity.
IZA Program Director Amelie F. Constant (George Washington University) presented the findings of two papers. The first paper on “Ethnicity, Job Search and Labor Market Reintegration of the Unemployed” adapts a recently developed concept of ethnic identity, the ethnosizer, to shed light on the native-migrant differences in economic outcomes. The IZA Evaluation Dataset includes information on a large number of migration and job search variables allowing Constant and her co-authors to investigate the labor market reintegration, patterns of job search, and reservation wages across unemployed migrants and natives. The findings indicate that “separated” migrants have a relatively slow reintegration into the labor market, which is mainly due to low search effort and moderate reservation wages compared to other groups of migrants. The second study presented by Constant analyzes if reservation wages increase from first to second generation migrants. The hypothesis derives from two extensions of the basic job search model, namely an unknown wage offer distribution and different reference standards. In both extensions, changing frames of reference are identified as a channel through which the phenomenon of increasing reservation wages may arise. In as far as language skills or self-evaluated returns to characteristics reflect a person’s frames of reference, they indeed find empirical support for this mechanism to be present. Both studies clearly illustrate that ethnic identity has an important influence on the job search behavior and thus on labor market integration. Constant concluded that ethnic identity of individuals has to be taken into account when assigning unemployed individuals to ALMP programs.
The next session was devoted to studies on the job search behavior of unemployed individuals. Ricarda Schmidl (IZA) presented empirical evidence on the relationship between social networks and the job search behavior of unemployed individuals. She emphasized the importance of the IZA Evaluation Dataset for this study as it contains extensive information on job search behavior and direct measures for the social network of individuals. The authors find that individuals with larger networks use informal search channels more often and shift from formal to informal search, which confirms the theoretical expectations.
Arne Uhlendorff (University of Mannheim and IZA) addressed the question to what extent personality traits such as locus of control impact job search behavior. Standard job search theory assumes that unemployed individuals have perfect information about the effect of their search effort on the job offer arrival rate. The authors suggest an alternative job search model which assumes that each individual has a subjective belief about the impact of his or her search effort on the rate at which job offers arrive. The study indeed finds evidence that individuals with an internal locus of control search more. This has clear implications with respect to the assignment of ALMP. It suggests, for instance, that in particular unemployed individuals with an external locus of control need more job search assistance as they search less compared to individuals with an internal locus of control.
The presentation by Ulf Rinne (IZA) compared natives and second generation migrants among the unemployed with respect to economic effects of risk attitudes, time preferences, trust and reciprocity. The findings suggest that both groups mainly differ in terms of risk attitudes and positive reciprocity. Second generation migrants have a significantly higher willingness to take risks and they are less likely to have a low amount of positive reciprocity when compared to natives. Rinne discovered a significantly lower employment probability for individuals with a high willingness to take risks. Some evidence suggests that this result is channeled through reservation wages and search intensity. These results, as well, are highly important for the assignment process of ALMP programs.
Deborah A. Cobb-Clark (University of Melbourne and IZA) gave the keynote speech on “Evidence-based Policy: Do We Have the Data Necessary to Recognize Good Policy if We See It.” Using the case of Australia as an example, she highlighted the improvement in data quality and availability during the last decades. Today researchers have different data sources available for scientific purposes, including surveys, administrative data and a combination of both. This improved data availability led to a remarkable increase in publications over the last decades. As a consequence, empirical evidence is available on various issues that are relevant to policymakers. However, the abundance of studies makes it difficult for policymakers to absorb all this information and distinguish between reliable and unreliable evidence. Cobb-Clark therefore presented an “evidence hierarchy for Australian policymakers” based on criteria such as publication in high-quality journals, use of Australian data, recent publication date, and relation to the policy being considered. Furthermore, empirical evidence based on randomized trials and natural experiments should be preferred over non-experimental evidence. In this regard, Cobb-Clark encourages more interaction between researchers and policymakers to conduct more experimental studies in the field of labor economics. Policymakers are interested in reliable evidence while researchers want better data quality and access. Improving the alignment of both streams of interests would have a positive impact on the future availability of data.
The policy oriented part of the workshop started with presentations by IZA researchers Ulf Rinne and Steffen Künn. Rinne gave an overview of studies that analyze the role of migration background and ethnicity in the job search process. He emphasized the importance of the IZA Evaluation Dataset to conduct these assessments as it includes the required information on migration, ethnicity and job search behavior for a large sample of unemployed individuals. He pointed out that that ethnicity and cultural diversity have measurable impacts on economic outcomes. Furthermore, migrants’ complementary qualifications such as their specific problem-solving abilities, their creativity and their language skills constitute an enormous economic potential. Rinne argued that an industrialized economy like Germany should recognize this potential of a diversified population and more thoroughly think about how to fully activate it. This includes a faster reintegration of unemployed individuals with an immigrant background. His empirical results revealed important differences between natives and migrants in the job search process that should be taken into consideration in the making of labor market policy. For example, intensified monitoring and job search assistance would accelerate the labor reintegration of specific groups.
Künn presented first empirical evidence on the effectiveness of ALMP for unemployed youth in Germany. Surprisingly, no empirical evidence exists so far although the number of youths participating in ALMP each year is substantial. Data restrictions are the main impediment to such an analysis as analyzing young people as a subgroup of the labor market requires a large dataset. The IZA Evaluation Dataset fills this gap. Künn and co-authors consider participation in a variety of different programs such as short-term training, job creation schemes or wage subsidies. The results indicate positive long-term employment effects for nearly all measures under scrutiny, with the exception of public sector job creation which is found to be ineffective.
Panel discussion with practitioners and policymakers
The practitioners and policymakers who contributed to the panel moderated by Hilmar Schneider, Director of Labor Policy at IZA, included Eckart Hohmann (Former President of the State Statistical Office of Hesse and Member of the German Data Forum), Johannes Vogel (Member of the German Bundestag and FDP labor market expert) and Manfred Zauner (Expert for labor statistics at the Austrian Ministry of Labor and Social Affairs). The panelists commented on questions of data protection, access to highly sensitive data, and the need for political support of initiatives like the IZA Evaluation Dataset.
Schneider looked back on progress made over the past years in terms of data provision for scientific research. In cooperation with the established Research Data Centers, it has become possible to create and access much larger anonymized datasets. However, the problem of a general prohibition of data retention in Germany remains unsolved, which continues to impede the scientific use of administrative data. Schneider called on policymakers to enable researchers to access the data necessary to uncover inefficiencies and undesired effects of labor market policies and to identify those programs that work. “This would imply, of course, that politicians may find their own policies proven wrong – but this is a risk they should take,” said Schneider.
According to Hohmann, data access could improve in the wake of European legal harmonization. Zauner explained that access to administrative data in Austria is handled in a more pragmatic fashion without jeopardizing data protection objectives. It therefore comes as no surprise that international research has produced a number of important studies based on Austrian datasets. Vogel underscored the significance of facilitated data access for independent research from the viewpoint of policymaking in order to foster evidence-based evaluation of labor market programs. The experts agreed that this goal will become even more important in the future when it comes to analyzing the impact of demographic change on the labor market, employment and social security.
View the PDF version to read what experts say about the dataset.
