National Bureau of Economic Research

We thank Jintaek Song for helpful feedback. We thank Jingzhou (Ron) Huang, Taryn O’Connor, and Amanda Zeitlin for research assistance. This work was funded by the Food Policy Research Program at the University of Illinois Chicago, which is supported by a grant (2025-153247) from Bloomberg Philanthropies’ Food Policy Program (www.bloomberg.org). The contents of this publication do not necessarily reflect the views or policies of Bloomberg Philanthropies.

The views expressed in this paper are those of the authors and do not represent the views or policies of the Federal Reserve Bank of Minneapolis or the National Bureau of Economic Research. We thank Jason Dolatabadi, Harry Elworthy, Libby Miskin, Mariya Shappo, and Jonathan Wang for their excellent comments and research assistance.

We gratefully acknowledge financial support from Boston University’s Global Development Policy Center, the Institute for Humane Studies, Fonds de recherche du Quebec, Social Sciences and Humanities Research Council of Canada, and the MIT Blueprint Labs. For useful comments and suggestions, we thank seminar participants at UCLA, Universidad de Chile, Universidad de Buenos Aires, University of Michigan, Michigan State University, Princeton, ITAM, and University of South Carolina.

We are thankful to Duke University for financial support. We also thank Elliot Ash, Sinzi Dorobantu, Joey Engelberg, Runjing Lu, and Will Mullins, for comments, and to participants at the Management, Analytics, Data Conference, Strategy Science Conference, and Academy of Management. All errors and omissions are our own. Chatterji is currently an employee of OpenAI, on leave from Duke and the NBER.

M.K. received funding from the Graduate Fellowship for STEM Diversity. The views and conclusions in this document are those of the authors and should not be interpreted as necessarily representing the official policies, either expressed or implied, of any funding sources. We thank Dr. Alex Peahl for helpful conversations about how a prediction of time to spontaneous labor would be used in practice, which informed our framing of the task and our choice of evaluation metrics.

We are grateful to the Lowe Institute of Political Economy at Claremont McKenna College for funding and to Frida Navarro Grau and Vera Kratz for research assistance. The views expressed herein are those of the authors and do not necessarily reflect the views of the National Bureau of Economic Research. Copy Citation Karla Cordova and Jessamyn Schaller, "Local Labor Demand and Achievement Gaps: Evidence from the Great Recession," NBER Working Paper 35639 (2026), https://doi.org/10.3386/w35639.

We thank Maimoonah Shafqat for outstanding research assistance, Michael Zhang for providing human verification, Arnaud Costinot for valuable discussions, and the Editor and four anonymous referees for excellent and constructive comments. Errors are our own. The views expressed herein are those of the authors and do not necessarily reflect the views of the National Bureau of Economic Research.

The authors are grateful to seminar organizers and attendees at the University of Pennsylvania, University of Southern California, University of Alabama at Birmingham, Harvard University, Penn State University, and the American Society of Health Economists 2026 Conference for excellent feedback on earlier versions of this work. The authors additionally thank the Florida Agency for Health Care Administration (AHCA) for providing data resources.

We are grateful to Lin William Cong, Sean Foley, John Griffin, Campbell Harvey, Shiyang Huang, Bjorn Jorgensen, Jonathan Karpoff, Wayne Landsman, Chen Lin, Evgeny Lyandres, Mei Luo, Michael Sockin, Eugene Soltes, Charles C. Y. Wang, Gloria Yu, and Che Zhang for their invaluable feedback on this study.

We would like to thank Leah Boustan, Katherine Hauck, David Lagakos, Omer Moav, Allen Peters, Steve Ross, and Wayne Sandholtz for their insightful feedback on this paper. We would also like to thank Desmond Ang, Price Fishback, Paul Rhode, and Ken Sylvester for kindly sharing data. The views expressed herein are those of the authors and do not necessarily reflect the views of the National Bureau of Economic Research.