Experimental evidence on a range of interventions in developing countries is accumulating rapidly. Is it possible to extrapolate from an experimental evidence base to other locations of policy interest (from "reference" to "target" sites)? And which factors determine the accuracy of such an...

We evaluate the impact of oral polio vaccines on the incidence of disabilities in India, focusing on polio-related disability. Polio was hyperendemic in India even as recently as the early 1990s but the country was declared wild polio virus-free in 2014. Average treatment effects on the treated...

We evaluate the impact of oral polio vaccines on the incidence of disabilities in India, focusing on polio-related disability. Polio was hyperendemic in India even as recently as the early 1990s but the country was declared wild polio virus-free in 2014. Average treatment effects on the treated...

Public child care is expected to assist families in reconciling work with family life. Yet, empirical evidence for the relevance of public child care to maternal employment is inconclusive. We exploit the introduction of a legal claim to a place in kindergarten in Germany, which was contingent...

Many commonly used treatment effects estimators rely on the unconfoundedness assumption ("selection on observables") which is fundamentally non-testable. When evaluating the effects of labor market policies, researchers need to observe variables that affect both treatment participation and labor...

We examine the consequences of compressing secondary schooling on students' university enrollment. An unusual education reform in Germany reduced the length of academic high school while simultaneously increasing the instruction hours in the remaining years. Accordingly, students receive the...

This paper exploits a unique universal educational policy - implemented in most German states between 2001 and 2008 - that compressed the academic-track high school curriculum into a (one year) shorter time span, thereby increasing time of instruction and share of curriculum taught per grade....

Many commonly used treatment effects estimators rely on the unconfoundedness assumption ("selection on observables") which is fundamentally non-testable. When evaluating the effects of labor market policies, researchers need to observe variables that affect both treatment participation and labor...

This paper uses the control function to develop a framework for testing for selection bias. The idea behind our framework is if the usual assumptions hold for matching or IV estimators, the control function identifies the presence and magnitude of potential selection bias. Averaging this...

In many European countries, labor markets are characterized by high regional disparities in terms of unemployment rates on the one hand and low geographical mobility among the unemployed on the other hand. This is somewhat surprising and raises the question of why only minor shares of unemployed...