Working papers

Influencer Dynamics (job market paper)

Consumers use social media for entertainment and to discover new products. To reach potential customers, brands pay influencers to feature products in their content. Payment depends on the size of the influencer’s audience, and the effectiveness of the endorsement relies on trust. Excessive product recommendations may erode the relationship between an influencer and their followers. I develop a dynamic model in which an influencer produces sponsored posts (which include product recommendations) and organic posts (without product recommendations). Both affect the influencer’s growth and require effort to produce. A sponsored post incurs additional costs: it requires searching for and negotiating with brands, and followers are less likely to engage with it. Using 2,780,011 Instagram posts and 136,453 TikTok posts from 1,369 influencers, I quantify the engagement penalty by comparing identical posts across platforms. Analyzing the influencers’ career histories, I show that organic and sponsored posts have similar effects on follower growth. I leverage variation in the number and types of posts to estimate the unknown cost parameters in the model. An influencer with 100,000 followers optimally produces about 0.25 sponsored posts and two organic posts per week, and influencers with more followers produce more content of both types. Viral posts provide a transient boost to growth, but career progression largely depends on consistent content production. Regulating sponsored content can theoretically backfire by decreasing incentives to produce organic posts, but counterfactual simulations assuage this concern. A cost increase that reduces sponsored content by 25% only causes a 2.7% drop in organic content.

Works in progress

Platform Competition with Network-based Ad Targeting