Yikang Shen
Ph.D. Candidate in Economics, Carnegie Mellon University
2026–27 Academic Job Market
Embedded Files
Digital platforms · Industrial Organization · Political Economy
I study digital platforms and algorithmic markets, focusing on how algorithms and digital intermediaries shape information provision, firm behavior, and market outcomes. I combine structural modeling, causal inference, economic theory, machine learning, and large-scale data to study these questions.
Platforms as Editors: How Algorithmic Curation Shapes Online News Slant
Recommendation algorithms change not only what users see, but what news outlets choose to produce.
Using matched print and online headlines from eleven major newspapers, Facebook’s 2025 algorithm change, and a structural model of users, platforms, and news outlets, I find:
After Facebook restored political content, the online–print slant gap for political stories increased by about 67%.
Outlet responses account for about 80% of the additional slant in user exposure.
An alternative ranking design eliminates roughly 90% of the excess online slant with modest welfare losses.
After Facebook restored political content to recommendations in 2025, the online–print slant gap increased sharply for political relative to non-political stories.
Selected presentations:
NBER Political Economy Program Meeting · ES NA Summer Meeting · ZEW Economics of ICT · ESIF Economics and AI/ML · CES North America
Forthcoming: Conference on Information Systems and Technology (CIST), 2026
Working Papers
Getting the Agent to Wait
with Maryam Saeedi and Ali Shourideh
We study how an information provider controls the timing of disclosure when it values user engagement and users prefer to learn quickly. Optimal disclosure depends on relative patience and differences in beliefs; when beliefs are private, non-personalized communication leads to faster information revelation, while personalized communication delivers higher-quality information. The framework applies to digital platforms, personalized content, and other settings in which intermediaries strategically manage information provision.
Automated Exchange Economies
with Bryan R. Routledge and Ariel Zetlin-Jones
We study price discovery in decentralized exchanges where automated market makers replace traditional order books with algorithmic pricing rules. Using 19.2 million transactions from 31 Uniswap v2 pools and a dynamic model of liquidity provision, we show that liquidity providers are not purely passive: they strategically adjust prices in response to trades, with important implications for price formation in algorithmic markets.
Page updated
Google Sites
Report abuse