Topic Survey Methodology and Data-Science

Grenzen und Fortschritte indikatorengestützter Politik am Beispiel der Corona-Pandemie: Heinz-Grohmann-Vorlesung 2020/21

Indikatoren sollen der Steuerung von (sozialen) Prozessen dienen. Sie beschreiben jedoch die Realität in der Regel nur deskriptiv und unvermeidlich mit mehr oder weniger großen und systematischen Messfehlern behaftet. Insofern ist es im Allgemeinen alles andere als einfach mit Hilfe von Indikatoren zu steuern; insbesondere dann, wenn für Problembereiche (fehlerbehaftete) Zielwerte vorgegeben werden, ...

No Taxation without Reallocation: The Distributional Effects of Tax Changes

This paper investigates the dynamic effects of tax changes on the cross-sectional distribution of disposable income in the United States using a narrative identification approach. I distinguish between changes in personal and corporate income taxes and quantify the distributional effects on families and business owners. I document that tax changes affect incomes along the distribution differently and ...

Observing Many Researchers Using the Same Data and Hypothesis Reveals a Hidden Universe of Uncertainty

This study explores how researchers’ analytical choices affect the reliability of scientific findings. Most discussions of reliability problems in science focus on systematic biases. We broaden the lens to emphasize the idiosyncrasy of conscious and unconscious decisions that researchers make during data analysis. We coordinated 161 researchers in 73 research teams and observed their research decisions ...

German Financial State Aid during COVID-19 Pandemic: Higher Impact among Digitalized Self-Employed

In response to strong revenue and income losses that a large share of the self-employed faced during the COVID-19 pandemic, the German federal government introduced a €50bn emergency aid program. Based on real-time online-survey data comprising more than 20,000 observations, we analyze the impact of this program on the subjective survival probability. In particular, we investigate how the digitalization ...

Competition, Formal Governance and Trust in Alliances: An Experimental Study

We study the role of alliance governance in the behavior of partners in alliances with different degrees of competition. Using data from a lab experiment on 1,009 alliances and 31,662 partners' choices, we explore whether and how alliances succeed in different competitive scenarios, contingent on the use of formal governance mechanisms (termination clauses) and the number of partners in the alliance. ...

SOEP Annual Report 2021

Active Learning Improves Financial Education: Experimental Evidence from Uganda

We conduct a randomized field experiment to study the effects of two financial education interventions offered to small-scale retailers in rural western Uganda. The treatments contrast “active learning” with traditional “lecturing” within standardized lesson-plans. After six months, active learning has a positive effect on savings and investment outcomes, in contrast to small or zero effects for lecturing. ...

Financial Education Matters: Testing the Effectiveness of Financial Education Across 76 Randomized Experiments

Do financial education programs affect financial knowledge and behaviors? We examine this question using a meta-analysis that incorporates studies from the past decade, which saw a rapid increase in financial education research. When examining data from 76 financial education randomized experiments across 33 countries covering over 160,000 individuals, we find that financial education improves both ...

Heteroskedastic Proxy Vector Autoregressions: Testing for Time-Varying Impulse Responses in the Presence of Multiple Proxies

We propose a test for time-varying impulse responses in heteroskedastic structural vector autoregressions that can be used when the shocks are identified by external proxy variables as a group. The test can be used even if the shocks are not identified individually. The asymptotic analysis is supported by small sample simulations which show good properties of the test. An investigation of the impact ...

How Communication Makes the Difference between a Cartel and Tacit Collusion: A Machine Learning Approach

This paper sheds new light on the role of communication for cartel formation. Using machine learning to evaluate free-form chat communication among firms in a laboratory experiment, we identify typical communication patterns for both explicit cartel formation and indirect attempts to collude tacitly. We document that firms are less likely to communicate explicitly about price fixing and more likely ...