Research

Working Papers & Research in Progress

Echo Chambers in the Free Market for News: Evidence from Telegram

Job Market Paper — available by request

This paper studies the formation of echo chambers on social media in the absence of algorithmic recommendation. Using high-frequency data on millions of posts from hundreds of major political Telegram channels in Russian during 2022–2024, I document how forwarding networks and subscriber flows endogenously generate clustered communities. I show that echo chambers can arise even without platform recommendation systems: reciprocal forwarding, selective attention on the demand side, and the tendency to forward from similar channels are sufficient to segment the news market. To complement the reduced-form evidence, I develop a simulation model demonstrating that demand-driven mechanisms alone can replicate the observed clustering patterns. We also show that under certain conditions algorithmic feed leads to less polarization.

Work in Progress:

Why Don’t Babies Drill Enough? Evidence of Spatial Correlation Effects on R&D in the Gulf of Mexico

We estimate a spatial model of oil and gas exploration in the Gulf of Mexico (R–INLA/SPDE) and show that firms’ decisions to acquire tracts and drill exploratory wells depend strongly on activity in neighboring areas.

Testing the IPV Hypothesis in Oregon Timber Auctions

We develop a small-sample parametric test leveraging exogenous rival mill-to-tract distances; in Oregon Department of Forestry data, the test rejects IPV in favor of common-value components.

Paper available by request

Are All Points Created Equal? Evidence from Professional Tennis

We show that points within a tennis service game are strategically heterogeneous: players adjust serve speed and accuracy in line with incentives, serving faster and with greater accuracy when the marginal effect on winning probability is greatest, and serving slower with less focus when points are less pivotal.

Paper available by request