Chenchen (Sabrina) Peng

Welcome!

I am a Ph.D. Candidate in Economics at Yale University.

My research lies at the intersection of international trade, technology, and firms. My work combines micro data and causal evidence with quantitative models.

I am on the 2026–2027 job market.

My CV is available here. Email: sabrina.peng@yale.edu.

Chenchen (Sabrina) Peng

Working Papers

Offshoring and Knowledge Spillovers (Job Market Paper) [PDF]

Abstract

When offshoring strengthens foreign rivals and increases competition at home through knowledge spillovers, do firms offshore too much or too little? I study whether U.S. offshoring to China diffuses production knowledge to other local producers in China and raises their exports to the United States. Linking firm-level production records to transaction-level imports data from the U.S. Census Bureau, I identify offshoring relationships between the universe of 1997 U.S. manufacturers and their Chinese affiliates or suppliers. Using a staggered difference-in-differences design, I find that once offshoring reaches a city–industry, exports to the U.S. by local producers that never receive an offshoring order rise 16 percent over the following two to five years, relative to city–industries not yet reached. In a trade model with knowledge diffusion, offshoring firms do not internalize that these spillovers lower U.S. aggregate price index but erode other U.S. firms' profits, so the private level of offshoring may lie above or below what is optimal for the United States. A per-firm tax or subsidy on offshoring, scaled to the size of the spillovers, corrects the externality, while protection of trade secrets removes the externality without changing the level of offshoring.

Workforce Age Structure and Frontier Technology Adoption: Evidence from Newly Released Software [PDF]
with Tra Nguyen

Abstract

This paper provides causal evidence that aging slows technology adoption and sheds light on why. Using Lightcast job postings, we track local adoption of 15 software technologies following their releases, instrumenting local age composition with a Bartik-style measure based on historical age structure and national survival rates. Employers in older labor markets post increasingly fewer vacancies requiring the newest software: at year 4, a one-standard-deviation increase in the older-worker share reduces new-technology postings by about 15% of pre-release legacy postings share. The gap is also larger for more difficult technologies, which we measure using new LLM-based difficulty indices. We then develop a model to separate three potential channels through which aging affects technology adoption: shorter work horizon, legacy technology experience loss, and reduced learning efficiency. Removing experience loss reduces the age adoption gap by about 88%, more than removing differences in horizons or learning efficiency. The model also shows that policies affecting workforce age structure have different implications for adoption. Relative to replacement-level fertility with retirement at 70, and assuming zero net migration, a total fertility rate of 1.6 lowers aggregate adoption by 7.5%, whereas raising the retirement age to 75 has almost no effect.

Publications

Selective Entry in Highway Procurement Auctions [PDF]
Atlantic Economic Journal, 48(4), 519–533, 2020.

Teaching

Decision Making with Data (MBA)
Teaching Assistant for Prof. Peter Schott · Yale School of Management · Spring 2026

Long-Term Fiscal Sustainability
Teaching Assistant for Prof. Teresa Delgado · Yale College · Fall 2025

Herb Scarf Summer Research Program in Economics
Co-Director · Yale University · Summer 2024