64 K. Li and N.R. Prabhala
- The pricing of public debt offerings
Companies making a debt issue must make several decisions related to the offering
such as the terms and structure of the offering, the type of the underwriter for the issue.
Private information held by the issuer or the intermediaries participating in the offering
could affect the choices made by firms. If such information has value, it affects the
prices at which issues can be sold. A fairly wide range of self-selection models have
been used to address the existence of private information and its effect on the pricing of
debt issues. We review some of the applications and the key findings that emerge.
8.1. Bank underwritings and the Glass–Steagall Act:Puri (1996)
The choice of an underwriter is an area that has been extensively analyzed using self-
selection models. An early application isPuri (1996), who investigates the information
in a firm’s choice between commercial banks and investment banks as underwriters of
public debt offerings. Commercial banks are often thought to possess private informa-
tion about their borrowers. If they use the private information positively, commercial
bank underwritten offerings should be priced higher (the “certification” hypothesis).
Alternatively, banks could use their information negatively to palm off their lemons to
the market, in which case the markets should discount commercial bank underwritten
offerings (the “conflicts of interest” hypothesis). Selection models are natural avenues
to examine the nature of these private information effects.
Puri models the private information in the underwriter choice using a probit model,
viz.,
C=CB≡Wi=Ziγ+ηi> 0 , (44)
C=IB≡Wi=Ziγ+ηi 0 , (45)whereCBdenotes a commercial bank,IBdenotes an investment bank, andηiis the
private information in offeringi. Markets price issueiat yieldyiwhere
yi=xiβ+i, (46)
E(yi|C)=Xiβ+πλC(Ziγ). (47)Equation(47)follows from equation(46)and the assumption thatandηare bivariate
normal. The above system is, of course, the standard Heckman model of Section2,so
the sign of the covariance coefficientπdetermines the impact of private information on
offer yields. Ifπ>0, markets demand higher yield for CB offerings, consistent with a
conflicts of interest hypothesis, whileπ<0 supports the certification hypothesis.
The data inPuri (1996)are debt and preferred stock issues prior to the passage of the
1933 Glass–Steagall Act. She includes issue size, credit rating, syndicate size, whether
the security is exchange listed, whether it is collateralized, and the age of the issuer as
determinants of the offer yield. She finds thatπ<0, consistent with the certification
hypothesis. Additionally,πis more negative for information sensitive securities, where