Julia Cajal Grossi
MEGA Salle Carine Nourry
Maison de l'économie et de la gestion d'Aix
424 chemin du viaduc
Timothée Demont : timothee.demont[at]univ-amu.fr
Eva Raiber : eva.raiber[at]univ-amu.fr
Multinationals’ reputation in high-income countries is increasingly tied to the behavior of their foreign suppliers. How do buyers find suppliers in low-income countries? Using customs records from Bangladesh’s garment sector, this paper shows that when starting to source a product, buyers experiment with potential suppliers through small-scale, short-lived interactions before forming lasting relationships. While large buyers experiment more than smaller buyers on average, they experiment less when quality dispersion among potential suppliers is high—a counterintuitive finding from the perspective of search theory. I rationalize these empirical facts with a model of search with reputational risk and highlight an important trade-off for international buyers: when quality dispersion is high, they experiment more in hopes of finding a high-quality recurrent trade partner but may unintentionally experiment with a low-quality supplier who damages their reputation. The model characterizes the optimal amount of experimentation and the threshold supplier quality at which buyers should settle. It yields two difference-in-differences relationships that I test in the data, exploiting exogenous variation in buyers’ reputation concerns after the largest industrial accident in the history of the Bangladeshi garment sector. This shock to reputation concerns led to less experimentation and worse matches among large (plausibly more reputation-sensitive) buyers in product categories with high supplier quality dispersion.