Humboldt-Universität zu Berlin - Wirtschaftswissenschaftliche Fakultät

Publications and work in progress

Most of my research is in empirical labor and development economics. A particular focus is on impact evaluation, statistical methods for causal inference, and meta analysis.

Recent work

The effectiveness of interventions to reduce informality in low and middle income countries

With J. Jessen, most recent version here.

Youth employment interventions in low and middle income countries

With colleagues from the World Bank, ILO, Oxford, published in World Development

Here is a World Bank blog with some key insights, and here the full systematic review report related to this research.

Labor market policies in Latin America and the Caribbean

A meta analysis of policies in LAC published in the Journal of Development Studies

Galpao Aplauso: RCT of technical and soft skills training

Experimental evidence on an innovative training program for favela youths in Rio de Janeiro. The training combines vocational skills and life skills components. The study is joint work with C. Calero, V. Gonzalez, Y. Soares (all IDB) and C. Corseuil (IPEA, Rio de Janeiro). Published in Labour Economics, special issue on field experiments

ALMP meta analysis

New meta analysis of Active Labor Market Program evaluations jointly with David Card and Andrea Weber, based on a new and expanded data set, published in the Journal of the European Economic Association. This is the data appendix. Here is a Brookings blog with some key results.

Experimental long term evaluation of youth training in the Dominican Republic

Long term impact evaluation of the youth training program "Juventud y Empleo" in the Dominican Republic. It is joint work with Pablo Ibarrarán, Laura Ripani, and David Rosas (all IDB). We follow-up individuals at six years after random assignment, published in Industrial and Labor Relations Review

Statistical employer discrimination

Evidence for Germany -- relating to women of childbearing age -- from a natural experiment, joint with J. Jessen and R. Jessen and published in Labour Economics

Parental benefits / RD

In a paper with Sebastian Schmitz (FU Berlin) we show that the German parental benefit reform of 2007 causes sizeable positive female labor supply effects in the medium run. Our evidence supports the idea that the reform generated an "anchor", i.e. a societally accepted point in time (=one year after childbirth, when benefit eligibility ends) at which mothers return to the labor market. Employers reward this early return by improving job quality for returning mothers. Published in the Industrial and Labor Relations Review

Better with Bologna? Tertiary education reform and student outcomes

In a paper with Sabrina Hahm (HU Berlin) we use a natural experiment design and IV to assess the impacts of the Bologna reform on students' educational outcomes. This version is from January 2019, forthcoming in Education Economics

 

Selected previous publications

Statistical Matching IV: Generalized Propensity Score (GPS)

In this paper we analyze the impact of training using continuous variation in the duration of the treatment. Since selection into different treatment durations is not random, we adjust for observable covariates using the generalized propensity score.

 

Natural experiment

This paper analyzes the labor market effects of a new parental benefit in Germany (Elterngeld). The quick legislative process generated a rather nice natural experiment around a cut-off date (1 January 2007, when the law came into effect): At the time of conception parents could not know that once their child is born they would be eligible for the benefit. Comparing parental behavior within three months before and after the cut-off therefore gives unbiased estimates of the reform effects.

 

Meta analysis: active labor market programs

This paper provides a statistical (meta) analysis of a new sample of impact evaluation studies, comprising 199 impact estimates from programs worldwide.

 

In this article, originating in a research project for the European Commission, meta-analytical methods are used to investigate the effectiveness of European active labor market programs. The results are surprisingly clear-cut: The program type matters, but contextual factors (unemployment rate, state of the economy, labor market regulations) do not. Only high employment protection legislation has a negative correlation with program effectiveness. As regards programs: training is good, wage subsidies may be even better (disregarding general equilibrium effects), public employment is bad, and job search services and sanction are also very effective in the short run.

 

Statistical Matching III / Cross-sectional DiD: Rural electrification

The study assesses the expected impacts of electrification interventions in rural Rwanda. Based on a sample of households in villages that have access to the grid (and part of which choose to actually connect, while others do not), we use propensity score matching methods to identify expected impacts in villages where connection is planned in the future. Cross-sectional DiD estimates strengthen the finding that rural electrification contributes to poverty reduction.

 

Statistical Matching II: Dynamic treatment - training and wage subsidies

This paper uses a dynamic matching estimator employing a “moving window” technique to analyze program effectiveness in a context of rapid change in the economic environment.

 

Statistical Matching I: Optimal Full Matching

This paper investigates matching methods with respect to two crucial implementation aspects: the choice of distance measure and the type of algorithm. We implement optimal full matching—a fully efficient algorithm—and present a framework for statistical inference.

 

Effects of the Hartz Reforms

The evaluation of the comprehensive set of labor market reforms (Hartz reforms) was likely the biggest evaluation project ever to be implemented in Germany. This article summarizes the key results.

 

Minimum wages: forecasting labor market and fiscal effects

A highly controversial topic in many countries, and especially in Germany. This paper estimates labor market and fiscal effects of a (at the time hypothetical) minimum wage in Germany.

 

More quantitative evidence on European labor market policy

This is a policy paper on active labor market programs, using a quantitative approach to assess key aspects in program effectiveness.