Submitted by Rodrigo Naumann (naumann.rodrigo@gmail.com) on 2018-04-19T17:10:40Z No. of bitstreams: 1 Dissertation_RNaumann.pdf: 979702 bytes, checksum: d1fa4b5a30cfc47fe78652e87c5fd4cf (MD5) / Approved for entry into archive by GILSON ROCHA MIRANDA (gilson.miranda@fgv.br) on 2018-04-20T20:16:09Z (GMT) No. of bitstreams: 1 Dissertation_RNaumann.pdf: 979702 bytes, checksum: d1fa4b5a30cfc47fe78652e87c5fd4cf (MD5) / Made available in DSpace on 2018-05-08T14:39:53Z (GMT). No. of bitstreams: 1 Dissertation_RNaumann.pdf: 979702 bytes, checksum: d1fa4b5a30cfc47fe78652e87c5fd4cf (MD5) Previous issue date: 2018-03-28 / This paper studies incentives for information gathering in a monoposonist pricing setting. Our motivation stems from public procurement contracts where the government is the single buyer, and the true cost of providing the good is ex ante uncertain to potential suppliers. We develop a simple bilateral, monopsonistic trade model, based on Roesler and Szentes (2017), where the seller only observes a signal about actual production cost. The model is intended to highlight how information about seller's cost affects resource allocation, price, and buyer welfare. More specifically, for any given continuous cost distribution, we characterize the seller-optimal learning. Taking the uniform prior as our benchmark case, we illustrate that seller's equilibrium strategy induces less information acquisition than would be desired by the buyer. We also show that efficient trade may not happen with probability one under some seller-optimal signal structures.