Schools under Mandatory Testing Can Mitigate the Spread of SARS-CoV-2
Joint with Ingo Isphording, Reyn van Ewijk and Nico Pestel
2022, PNAS
Winner of the best paper award of the Deutsche Gesellschaft für Gesundheitsökonomie
For a German summary see IZA Standpunkt 101; Media coverage in Welt and Focus.
We use event study models based on staggered summer vacations in Germany to estimate the effect of school re-openings after the summer of 2021 on the spread of SARS-CoV-2. Estimations are based on daily counts of confirmed coronavirus infections across all 401 German counties. A central anti-pandemic measure in German schools included mandatory rapid testing multiple times per week. Our results are consistent with mandatory testing contributing to the containment of the viral spread. We find a short-term increase in infection rates right after summer breaks, indicating the uncovering of otherwise undetected (asymptomatic) cases through the testing. After a period of about two weeks after school re-openings, the growth of case numbers is smaller in states which re-opened schools compared to the control group of states still in summer break. The results show a similar pattern for older age groups as well, arguably as a result of detected clusters through the school testing. This means that under certain conditions open schools can play a role in containing the spread of the virus. Our results suggest that closing schools as a means to reduce infections may have unintended consequences by giving up surveillance and should be considered only as a last resort.
Is large-scale rapid CoV-2 testing a substitute for lockdowns?
Joint with René Glawion, Peter G. Kremsner, Timo Mitze, Gernot J. Müller, Dominik Papies, Felix Schulz and Klaus Wälde
2022, PLOS ONE
Various forms of contact restrictions have been adopted in response to the Covid-19 pandemic. Around February 2021, rapid testing appeared as a new policy instrument. Some claim it may serve as a substitute for contact restrictions. We study the strength of this argument by evaluating the effects of a unique policy experiment: In March and April 2021, the city of Tübingen set up a testing scheme while relaxing contact restrictions. We compare case rates in Tübingen county to an appropriately identified control unit. We employ the synthetic control method. We base interpretations of our findings on an extended SEIR model. The experiment led to an increase in the reported case rate. This increase is robust across alternative statistical specifications. This is also due to more testing leading initially to more reported cases. An epidemiological model that corrects for ‘more cases due to more testing’ and ‘reduced testing and reporting during the Easter holiday’ confirms that the overall effect of the experiment led to more infections. The number of rapid tests were not sufficiently high in this experiment to compensate for more contacts and thereby infections caused by relaxing contact restrictions.
Joint with Joshua Angrist and Glenn Ellison
2026, NBER Working Paper 35244
Media coverage in VoxEU column: Research value-added in elite economics PhD programmes, and NotebookLM Podcast
The 2025 university endowment tax hike and other sources of financial pressure may lead the schools that train the most prolific economics researchers to reduce graduate enrollment. Will this affect long-run research output? We use a novel sample of MIT Economics PhD program applicants to estimate the research value-added of eight elite schools. Our estimates mitigate selection bias by controlling for MIT admissions committee rankings—a remarkably strong predictor of long-run research success—and for applicant aspirations as revealed by their application portfolios. While rank controls substantially reduce estimated gaps between elite and non-elite graduates, large differences in value-added remain. Graduates of high-tax and other top-eight schools produce 60-75% more impact-adjusted publications than do comparable graduates from non-top-eight US schools. The elite-school advantage is especially pronounced for top five journal publications. Differences in research success within the elite tier, however, are relatively modest. The out-performance of elite-school PhDs does not appear to be explained by editorial connections or peer effects in elite programs.
Dissertation Paths: Advisors and Students in the Economics Research Production Function
Joint with Joshua Angrist
2024, NBER Working Paper 33281
Media coverage in VoxEU column: Prolific PhD advisors are no guarantee of graduate student research success, and VoxTalks podcast episode
Elite economics PhD programs aim to train graduate students for a lifetime of academic research. This paper asks how advising affects graduate students’ post-PhD research productivity. Advising is highly concentrated: at the eight highly-selective schools in our study, a minority of advisors do most of the advising work. We quantify advisor characteristics such as an advisor’s own research output and aspects of the advising relationship like coauthoring and research field affinity that might contribute to student research success. Students advised by research-active, prolific advisors tend to publish more, while coauthoring has no effect. Student-advisor research affinity also predicts student success. But a school-level aggregate production function provides much weaker evidence of causal effects, suggesting that successful advisors attract students likely to succeed–without necessarily boosting their students’ chances of success. Evidence for causal effects is strongest for a measure of advisors’ own research output. Aggregate student research output appears to scale linearly with graduate student enrollment, with no evidence of negative class-size effects. An analysis of gender differences in research output shows male and female graduate students to be equally productive in the first few years post-PhD, but female productivity peaks sooner than male productivity.
Editors decide which manuscripts are published in scientific journals, and ties between editors and authors may distort these decisions. I follow the professional networks of the 594 editors at 28 leading economics journals between 1996 and 2025, formed by their former coauthors, their departmental colleagues, and their former PhD students. Using variation in the timing of editorial appointments, I find that connected authors publish 8.6% more in a journal while a connected editor is in office. This additional work is cited no differently than the same authors' other publications in the same journal, so it does not reflect lower editorial standards. The publication gains of affiliated colleagues and of coauthors who keep working with the editor extend to similar journals and to their complete research output, consistent with editorial knowledge spillovers, while coauthors whose collaboration with the editor has lapsed gain nothing. I find no evidence of quality-lowering favoritism, yet the benefits of editorial knowledge concentrate at the few institutions that hold the most editorships.
Two organizational literatures make opposing predictions about category-spanning knowledge work. The knowledge-recombination tradition argues that combining atypical inputs generates breakthrough impact; the categorical-evaluation tradition argues that producers who span audience-recognized categories face an illegitimacy discount. We adjudicate between these predictions in a setting that holds the evaluating audience and producer incentives constant: the discipline of economics. We further decompose category-spanning into within-discipline (multi-field) and cross-discipline (interdisciplinary) variants. Using a matched Web of Science-EconLit panel of 32,845 economics articles published in 2011-12, we construct four indicators of knowledge integration: three novel multi-field measures and one interdisciplinary measure. Articles scoring one standard deviation higher on any multi-field measure face 3-27 percent lower odds of placement in top outlets and receive 6-26 percent fewer within-economics citations. Cross-disciplinary work faces similar within-discipline penalties but, unlike multi-field work, attracts a compensating reception from audiences outside the discipline in the broader WoS. Contrary to the recombination tradition's high-risk, high-reward prediction, the variance of outcomes compresses rather than expands with category-spanning. Neither time, prior collaboration, nor producer status attenuates the multi-field penalty. Category-spanning inside economics is not a high-variance bet on breakthrough impact. It is a predictably lower-mean one.
Leaving the Club: How Brexit Changed Who Does Science in the UK
Joint with Carolin Häussler and Charlotte Höltl
Academics’ Forgone Wages: Evidence from German Register Data
Joint with Johannes König