Research

Abstract:

Social influence drives consumption, yet its effect on firm competition and pricing is understudied. Using large-scale de-identified mobile call records from a Chinese city, I estimate social influence in smartphone choices. Among purchasers, a 10-percentage-point increase in the share of friends using a model raises choice probability by two-thirds of the mean. Larger effects for affluent friends and visible attributes suggest status seeking. I estimate a structural demand model and a two-period pricing model. Counterfactuals show it reallocates demand toward higher-quality products, lowers first-period prices by 1.5 percent, raises second-period prices by 0.14 percent, and increases profits and consumer surplus.

Abstract:

We study how different communication technologies shape the labor-market consequences of social networks. We develop a simple framework in which communication technologies differ in the effectiveness with which pre-existing social ties facilitate workplace mobility. Using cellphone geolocation and sociodemographic data on 3.3 million workers in Beijing, we construct a measure of the relative importance of face-to-face interaction and exploit workplace transitions within friendship networks. Conditional on friendship and pre-existing relationship strength, a higher face-to-face communication share is associated with a significantly greater likelihood of workplace following. This relationship emerges only after the friend has already joined the destination workplace, implying that differences in communication patterns become economically relevant only when the social tie is in a position to influence workplace mobility. Our findings show that worker mobility depends not only on who workers know, but also on how they communicate, implying that communication technologies are imperfect substitutes in shaping the labor-market value of social networks.

Abstract:

Targeted technology controls aim to constrain sanctioned firms, but may diffuse inventive capacity beyond them. We examine this unintended effect using additions of Chinese firms to the U.S. Entity List and a comprehensive inventor–firm panel drawn from Chinese patents between 2014 and 2022. Listing increases inventor departures by 8.1 percentage points without lowering movers’ patent output, especially at firms that depend heavily on U.S. inputs or have narrower knowledge bases. Recipient firms absorb knowledge from movers’ former employers, redirect their innovation, increase incumbent patenting, and intensify competition in innovation markets. These spillovers offset more than one quarter of the sanctioned firms’ innovation losses.

Abstract:

Online disputes are pervasive on social media, yet their consequences remain unexplored. We study disputes on Weibo, a leading social media platform in China, using 8.9 million posts from 30,516 users affiliated with public figures involved in disputes. Exploiting plausibly exogenous shocks in a stacked difference-in-differences design, we show that disputes sharply increase toxic expression and late-night posting, with larger effects among younger users, women, and opinion leaders. Disputes also increase out-group toxicity and consumption-related support for public figures. These effect spill over across topics and persist over time, extending beyond the original conflict to the broader online discourse environment.