Asymmetric Information and Sovereign Debt:Theory Meets Mexican Data

Abstract

We combine novel data and theory to show that asymmetric information amonginvestors is an important friction in primary sovereign debt markets. We exploit aunique dataset of Mexican auctions for Cetes bonds. Auctions are pay-your-bid, andour data includes all bids made by all individual bidders from 2001 to 2017. We doc-ument that the largest bidders tend to bid at higher prices (that is, their bids are morelikely to be accepted), but on average they do not pay more for the bonds they buy(that is, their accepted bids are executed at the average price). We construct a modelin which investors can differ in wealth, risk-aversion, market power and/or informa-tion. Only heterogeneous information can qualitatively account for our findings. Wecalibrate the model and find that heterogenous information about rare disasters canquantitatively match key aspects of Cetes yield dynamics and bidding behavior.