Transcript The Economics of Solicited and Unsolicited Credit Ratings

The Economics of Solicited and
Unsolicited Credit Ratings
Paolo Fulghieri
Gunter Strobl
Han Xia
discussion by
Adlai Fisher
Sauder School of Business
University of British Columbia
UBC Winter Finance Conference
March 5, 2011
Outline
1) Some institutional detail on credit rating agencies
2) The model and comments
I. Credit Rating Agencies
History of Ratings Agencies
• John Moody published first publicly available bond
ratings in 1909 (mostly railroad bonds).
– Poor’s Publishing, 1916
– Standard Statistics Company, 1922
– Fitch Publishing, 1924.
• Manuals sold to investors (“investor pays model”)
– Initial role purely market driven
History of Ratings Agencies
• Major change in role of U.S. ratings agencies in 1936
– Bank regulation prohibits investments in “speculative
investment securities” as determined by “recognized ratings
manuals” (Moody’s, Poor’s, Standard, Fitch)
– Insurance regulators of all 50 states follow suit over years
– 1970’s federal pension regulators adopt similar requirement
– Creditworthiness judgments of CRA’s achieve force of law
• SEC formalizes role of CRA’s in 1973
– Defines category of “nationally recognized statistical rating
organization” (NRSRO), grandfathers Moody’s, S&P, Fitch
– Only ratings of NRSRO’s deemed valid for determining
broker-dealer capital requirements
– Other regulators soon adopt SEC conventions
Industry Structure
• Over 1975-2000, SEC designates four additional firms
as NRSRO’s, but mergers restore to original three
– SEC criticized for opaqueness in criteria to achieve NRSRO
designation, creates additional barrier to entry
– Credit Rating Agency Reform Act (2006), requires SEC to
set up transparent criteria for NRSRO’s
– Seven new NRSRO’s designated in 2007
• Current industry structure remains highly concentrated
– Moody’s plus S&P have over 80% market share in U.S.,
adding Fitch concentration exceeds 95%
• Worldwide over 100 CRA’s, but high concentration
within geographical segments, S&P and Moody’s have
global reach and affiliations
Importance of CRA’s
• New York Times (1996), on sovereigns:
– “There are two superpowers in the world today... the United
States and Moody’s Bond Rating Service. The United
States can destroy you by dropping bombs, and Moody’s
can destroy you by downgrading your bonds.”
• Clear evidence that stock and bond prices react to
ratings changes
– Disagreement about how much of this is due to information
versus the importance of CRA’s in regulatory oversight
Role of CRA’s
• Characterize themselves as part of “financial media”
– Issue opinions, but these opinions are protected by First
Amendment rights to free speech
– Difficult to establish legal liability for opinions
• Have access to inside information for solicited ratings
– Exempted from Reg FD requirements
Evolution in Business Model
• Late 1960’s/early 1970’s: business model largely
changes from “investor pays” to “issuer pays”
– Difficulty of protecting subscription content
– “Ratings agencies have evolved from information providers
to purveyors of `regulatory licenses.’” (Council of Institutional
Investors)
• Moody’s, only free-standing major CRA, has about
70% of revenue from ratings, remainder from related
services
Current Issues
• Ratings shopping
– Especially relevant for structured products, where issuers
repeat business very frequently
• Conflict of interest under “issuer pays”
– S&P: “clear separation between those who negotiate the
business terms and the analysts who conduct the credit
analysis”
– Counterexamples in literature (Hanover re)
Unsolicited Ratings
• Tend to be lower than solicited
– Disagreement whether this remains controlling for issuer
characteristics
– Can be used as a punishment for not soliciting a rating
• Ratings agencies argue they are necessary to provide
a complete portfolio of ratings for investors
– Also appear to be used prominently when entering a new
market (e.g., Japan)
• Unsolicited ratings use only public information;
recently are distinguished as “unsolicited” by CRA
• Empirical evidence: between 5 and 20% of ratings
depending on market
II. Model and Comments
Model
• Firm types θ={G,B,N} drawn iid for t=1,2
–
–
–
–
G (p=βα): investment project succeeds with prob q
B (p=β[1-α]): investment project fails, payoff zero
N (p=1-β): no project, payoff V*≥0
Objective: “Maximize current market value of shares”
• CRA: honest (p=μ) or opportunistic (p=1-μ)
–
–
–
–
Propose solicited rating rs={h,l,0} and fee to issuer
Issuer accepts or declines
If solicited declined, give unsolicited rating ru = {h,l,0}
Type N firm may only receive rating r=0
• Investors: Bayesian update on firm and CRA types
after observing ratings and payoffs each period
Solicited-only Equilibrium
• Firms invest if and only if H-rating received
• Period 2: opportunistic CRA offers H-rating to all
issuer types; fee depends on CRA reputation
• Period 1: opportunistic CRA charges fee V(H)-I-V(0)
for H-rating, fee independent of issuer type
– G firm offered H-rating with probability 1
– B firm offered H-rating with probability k(H)>0
• Ratings inflation k(H) increases in payoff R of
successful investments and is inverse U-shaped in
prior CRA reputation
Equilibrium with Unsolicited Ratings
• Period 2: opportunistic CRA offers H-rating to all
issuer types; fee depends on CRA reputation
• Period 1: opportunistic CRA charges fee V’(H)-I for an
H-rating, fee independent of issuer type
– G firm offered H-rating with probability 1
– B firm offered H-rating with probability k’(H)>0
– If no solicited rating, CRA issues unsolicited UL-rating
• Results
– Fees higher when unsolicited ratings permitted (CRA can
separate L-type from 0-type with UL-rating, and L-type has
weakly lower value by assumption)
– For low values of V*, ratings are more accurate and social
welfare is higher; for high values of V* the opposite can hold
Comments
• Issuer objective function depends on current price
only, deserves more discussion;
– L-type willing to pay to withhold information when pooled
with 0-type, even though this has no impact on ultimate cash
flow of L-type
• The objective function of the ethical type is not fully
specified
– does not allow determination of fees paid to ethical type
– what happens if issuer refuses fee proposed by ethical type?
• Assume that fees may depend on rating
– CRA’s claim this is not accurate; deserves more discussion
Comments
• Assume CRA infrormation quality independent of
whether rating is solicited or unsolicited
– Seems at odds with inside information available to CRAs in
solicited ratings and exemptions from reg FD; deserves
more discussion
• How realistic is assumption that 0-type cannot be
rated?
– Firms are collections of projects and a very large proportion
of major firms have some basis on which they could be rated
– Using this device seems to unnecessarily limit scope of
model; if a similar result could be obtained with a less
restrictive assumption, would be preferable
• Duopoly?
Conclusion
• The paper tackles an important and interesting
institutional feature of information intermediation in
financial markets
• After years of observation, which firm, Moody’s or
S&P, is the ethical one?
– Tongue-in-cheek, but perhaps give more guidance about how to
relate the model to current financial markets