88th International Atlantic Economic Conference
October 17 - 20, 2019 | Miami, USA

Corporate bonds: Uniform auction under risk and budget constraints.

Friday, 18 October 2019: 10:20 AM
Lamprini Zarpala Sr., M.Sc. , Banking and Financial Management, University of Piraeus, Athens, Greece
Since the onset of the financial crisis, corporate bonds have become an increasingly viable source of long-term financing and an attractive asset class for investors. The increase in corporate bond markets has been supported by regulatory initiatives and is consistent with the objectives of monetary policy by central banks. Insofar, much attention has been given to the optimal pricing auction mechanisms for treasury bonds, while less has been done for the pricing market mechanism used during the issuance of corporate bonds. Furthermore, most research emphasizes empirical evidence for book building, a current market mechanism for corporate bonds which approximates an auction. This paper contributes to the existing literature of auction theory by providing a rigorous game-theoretic model on corporate bond auctions. Specifically, a fully divisible corporate bond is allocated to a set of bidders with mean-variance preferences for their portfolio choice investment problem. Bidders are characterized by constraints with respect to both, their budget and, most importantly, the risk entailed in their portfolio. On top of that, we assume that bidders possess asymmetric information with respect to the secondary market's yield. In this setup, we choose to use uniform pricing mechanisms as opposed to discriminatory ones as the former generates more revenue and reduces the winner's curse. We prove the existence of an equilibrium and then provide the necessary comparative statics. We confirm that when bidders are risk averse uniform price auctions may ensure higher expected revenue than any other auction mechanism. Still, we anticipate that risk limits make bidders more prudent to the winner's curse. Last, we briefly consider possible extensions of our model, including the cost of information for variance-seeking preferences.