Contact:
Email: si.huang@queensu.ca
Department of Economics
Queen's University
Kingston, ON K7L 3N6
Job Market Paper
In markets for tied goods, consumers buy a durable base good and then repeatedly purchase compatible consumables. I study how compatibility between base goods and consumables affects demand, using the U.S. razor and blade market as an empirical setting. Allowing more of a firm's consumables to fit a given base good changes demand through two channels. First, it expands the market: a base good that fits more consumables is more valuable to own, attracting new buyers and making existing owners less likely to switch away. Second, it intensifies cannibalization: owners can choose among more of the firm's consumables, so a newly compatible line takes sales from the firm's existing lines. The balance between these channels matters for interoperability policy and for firms' product decisions. Reduced-form evidence shows that compatibility shapes adoption and substitution: households that own a compatible handle more readily adopt new lines, and households substitute strongly between compatible cartridges. To separate the two channels, I estimate a dynamic demand model in which forward-looking households choose each month whether to buy a razor pack, a compatible cartridge pack, or nothing. Counterfactuals at fixed prices show that market expansion dominates cannibalization when a firm makes its product lines compatible, both across families that differ in price and quality and among similarly priced lines within a family.
with Robert Clark and David Davis
Motivated by the FTC's ongoing investigation, this paper examines whether infant formula manufacturers colluded in bidding for WIC infant formula contracts. Using auction data from 2003 to 2021, we compare bidders' behavior during a competitive benchmark period (2003–2013) with that observed in the alleged collusive period (2016–2021). Descriptive patterns reveal that participation declined in the alleged period—consistent with reduced competition—yet when Abbott and MJ met head-to-head, winning bids fell even more sharply. Incumbents were more likely to retain contracts, but only by bidding more aggressively, not through rivals' restraint. To formally test for collusion, we implement state-of-the-art collusion screens that target two behaviors central to the FTC's allegation: bid rotation and cover bidding. These include tests from Kawai et al. (2023), Porter and Zona (1993), Huber et al. (2019) and Clark et al. (2025). Across these methods, we find little evidence of systematic bid rigging in WIC auctions during the alleged period.
Competition authorities around the world rely on statistical screens to detect potential bid rigging and identify cases that warrant further investigation. Although many screens have been proposed in the literature, little is known about their relative effectiveness. This paper addresses this gap by simulating competitive and collusive bidding data using a structural auction model, applying widely used screens, and systematically comparing their performance across different collusive scenarios.