Content uploaded by Emily Ekins
Author content
All content in this area was uploaded by Emily Ekins on May 08, 2020
Content may be subject to copyright.
Executive Summary
During the financial crisis of 2008, the finan-
cial markets would have been better served if
the credit rating agency industry had been more
competitive. We present evidence that suggests
the Securities and Exchange Commission’s des-
ignation of Nationally Recognized Statistical
Rating Organizations (NRSROs) inadvertently
created a de facto oligopoly, which primar-
ily propped up three firms: Moody’s, S&P, and
Fitch. We also explain the rationale behind the
NRSRO designation given to credit rating agen-
cies (CRAs) and demonstrate that it was not in-
tended to be an oligopolistic mechanism or to
reduce investor due diligence, but rather was
intended to protect consumers. Although CRAs
were indirectly constrained by their reputation
among investors, the lack of competition al-
lowed for greater market complacency. Govern-
ment regulatory use of credit ratings inflated the
market demand for NRSRO ratings, despite the
decreasing informational value of credit ratings.
It is unlikely that this sort of regulatory frame-
work could result in anything except misaligned
incentives among economic actors and distort-
ed market information that provides inaccurate
signals to investors and other financial actors.
Given the importance of our capital infrastruc-
ture and the power of credit rating agencies in
our financial markets, and despite the good in-
tentions of the uses of the NRSRO designation,
it is not worth the cost and should be abolished.
Regulators should work to eliminate regulatory
reliance on credit ratings for financial safety
and soundness. These regulatory reforms will,
in turn, reduce CRA oligopolistic power and the
artificial demand for their ratings.
Regulation, Market Structure,
and Role of the Credit Rating Agencies
by Emily McClintock Ekins and Mark A. Calabria
No. 704 August 1, 2012
Mark A. Calabria is the director of financial regulation studies at the Cato Institute. Emily Ekins is a research fellow
at the Cato Institute, director of polling at Reason Foundation, and a PhD student at the University of California,
Los Angeles.
2
Miscalculations
by issuers,
investors, credit
rating agencies,
and regulators
contributed
to the crisis,
but such
miscalculations
should have
come as no
surprise given
the incentive
structure of
our financial
markets.
Introduction
Starting in late summer 2007, credit rat-
ing agencies (CRAs) began extensive down-
grading of mortgage-backed assets and con-
tinued to downgrade them through the fall
of 2008. On September 15, 2008, Moody’s,
Standard and Poor’s (S&P), and Fitch down-
graded American International Group (AIG),
the nation’s largest insurance company, af-
ter which AIG’s stock price fell 61 percent.
The same day, Lehman Brothers and Merrill
Lynch filed for bankruptcy, just months after
Bear Stearns declared its insolvency. Several
days afterward, U.S. Treasury Secretary Hen-
ry Paulson and Federal Reserve Chairman
Ben Bernanke asked Congress for the ability
to purchase $700 billion in bad mortgage-re-
lated debt. President George W. Bush warned
that without the bailout the “entire economy
was in danger.”
The credit rating agencies’ downgrading
of assets did not cause the financial crisis of
2008, but they did shock the world in how
they revealed what were apparently inflated
ratings for corporate debt. Miscalculations
by issuers, investors, credit rating agencies,
and regulators contributed to the crisis, but
such miscalculations should have come as
no surprise given the incentive structure of
our financial markets.
This policy analysis does not seek to ex-
plain all the causes of the financial crisis. In-
stead, we focus on the role that regulations,
implemented in 1936 by the U.S. Comptrol-
ler of the Currency and in 1973 by the Secu-
rities and Exchange Commission (SEC), had
in creating regulatory dependency on desig-
nated credit rating agencies’ ratings.1 These
regulations resulted in reduced competition
and inflated market demand in the CRA in-
dustry, which in turn likely led to sustained
complacency in ratings’ methodologies and
potentially allowed for massive risk to go ig-
nored in the marketplace.
The financial markets would have been
better served if the credit rating agency in-
dustry had been more competitive. We pres-
ent evidence that suggests the Securities and
Exchange Commission’s (SEC) designation
of Nationally Recognized Statistical Rat-
ing Organizations (NRSROs) inadvertently
created a de facto oligopoly, which primar-
ily propped up three firms: Moody’s, S&P,
and Fitch. The NRSRO designation given to
credit rating agencies was not intended to be
an oligopolistic mechanism or to reduce in-
vestor due diligence, but rather was intended
to protect consumers. Although CRAs were
indirectly constrained by their reputation
among investors, the lack of competition al-
lowed for greater market complacency. Gov-
ernment regulatory use of credit ratings in-
flated market demand for NRSRO ratings,
despite the decreasing informational value of
credit ratings.
Our analysis demonstrates how well-
intentioned regulation can have unantici-
pated effects, and, in some cases, these unin-
tended consequences become part of a perfect
storm of other problems contributing to a fi-
nancial crisis. We must consider not just the
intended effects of regulatory proposals, but
also the second- and third-order effects of the
regulations. Once those have been considered,
we should conduct a cost-benefit analysis. In
the case of the NRSRO designation and its
use in regulations, it is unlikely that this sort
of regulatory framework could result in any-
thing except misaligned incentives among
economic actors and distorted market infor-
mation that provides inaccurate signals to in-
vestors and other financial actors. Given the
importance of our capital infrastructure and
the power of credit rating agencies in our fi-
nancial markets, and despite the good inten-
tions of the uses of the NRSRO designation, it
is not worth the cost and should be abolished.
Regulators should work to eliminate regulato-
ry reliance on credit ratings for financial safety
and soundness. These regulatory reforms will,
in turn, reduce CRA oligopolistic power and
the artificial demand for their ratings.
We review the origins of the credit rating
agency industry, what credit ratings agencies
and their ratings do in the marketplace, and
how the industry became regulated. We then
examine the evidence to determine if the CRA
3
industry became a de facto oligopoly with
complacent rating methodologies, whether it
had captive and inflated demand, and what
impacts it had on the financial markets.
Credit Rating Agencies
What Credit Rating Agencies Do
Credit rating agencies offer predictions
as to the likelihood that a particular debt
instrument will be repaid, in part or whole.
Ratings can be offered on general corporate
debt, in which case it is the overall financial
health of the company that is being ana-
lyzed. Ratings can also be based on the credit
risk of assets within a pooled security, such
as a mortgage-backed security. Prior to the
Great Depression, rating agencies also of-
fered some limited analysis of corporate eq-
uity values. The major rating agencies have
generally asserted that they assess only credit
risk and do not offer judgments as to losses
that arise from an instrument’s liquidity or
lack thereof. Debate continues whether the
losses witnessed during the recent financial
panic were more the result of credit or liquid-
ity losses, but we will not attempt to settle
that debate here.
In the case of an asset-backed security the
security’s issuers purchase loans from banks,
such as mortgages or corporate debt, and re-
package the risk into a financial product that
is purchased by investors as an investment ve-
hicle. Investors pay issuers for the financial
instrument and in return they receive the
principal and interest paid on the debt (see
Figure 1). However, debt is not always repaid
The major rating
agencies have
generally asserted
that they assess
only credit risk
and do not offer
judgments as
to losses that
arise from an
instrument’s
liquidity or lack
thereof.
Figure 1
The Role of Credit Rating Agencies
Source: Figure constructed by authors.
Key: Solid lines represent asset transfer, dashed lines represent indirect payment transfer, and dotted line repre-
sents information transfer.
4
Generally, ratings
change only if
the prospects
of default
actually change
by a significant
amount, rather
than if there is a
change in
short-run
credit risk.
and thus investors have an interest in know-
ing the probability that they will be repaid.
Credit rating agencies provide a solution to
the information asymmetries between inves-
tors and financial firms that issue financial
products by providing risk assessments in
the form of creditworthiness gradations for
financial products, such as with an ordinal
scale, with the intent to make it easier for
investors to compare different potential in-
vestments (see Table 1). Essentially, investors
want to know the credit risk associated with
debt instruments, such as corporate debt or
mortgage-backed securities, in which they
are about to invest. As lenders innovate new
ways of repackaging risk in complicated se-
curitized financial products, investors are
even more interested in the amount of risk
associated with these new products. A com-
parison between ratings scales for Moody’s,
S&P, and Fitch can be found in Table 2.
Credit rating agencies do not frequently
incorporate new information that may im-
pact currently issued credit ratings, thus al-
lowing these ratings to remain fairly stable
over time. Generally, ratings change only if
the prospects of default actually change by a
significant amount, rather than if there is a
change in short-run credit risk. This is called
“rating through the cycle” rather than rating
at a “point in time.”2 For example, parties
who intend to hold assets over a long period
Table 1
Generic Moody’s Credit Ratings Scale
Rating Rating Definition
Aaa Obligations rated Aaa are judged to be of the highest quality, with minimal credit
risk
Aa Obligations rated Aa are judged to be of high quality and are subject to very low
credit risk
AObligations rated A are considered upper-medium grade and are subject to low
credit risk
Baa Obligations rated Baa are subject to moderate credit risk. They are considered
medium grade and, as such, may possess certain speculative characteristics
Ba Obligations rated Ba are judged to have speculative elements and are subject to
substantial credit risk
BObligations rated B are considered speculative and are subject to high credit risk
Caa Obligations rated Caa are judged to be of poor standing and are subject to very
high credit risk
Ca Obligations rated Ca are highly speculative and are likely in, or very near, default,
with some prospect of recovery of principal and interest
CObligations rated C are the lowest-rated class and are typically in default, with
little prospect for recovery of principal or interest
Source: Moody’s Investors Service, “Rating Symbols and Definitions: January 2011,” http://www.moodys.com/
researchdocumentcontent page.aspx?docid=PBC_79004.
Note: Moody’s appends numerical modifiers 1, 2, and 3 to each generic rating classification from Aa through
Caa. The modifier 1 indicates that the obligation ranks in the higher end of its generic rating category, the
modifier 2 indicates a mid-range ranking, and the modifier 3 indicates a ranking in the lower end of that generic
rating category.
5
Frequently
changing credit
ratings could
potentially
raise the cost
of financial
instruments.
may be less interested in small changes that
may be reversed with regular market move-
ments. This method imposes fewer direct
costs on issuers and investors since they
may devote fewer resources to ensuring they
are in compliance with regulatory capital
requirements. Moreover, contracts between
firms and investors often include “rating
triggers” that are impacted by the credit rat-
ings. A rating trigger is a contractual obliga-
tion to repay more quickly, provide more
collateral, or any other such requirement.
Frequently changing credit ratings could
potentially raise the cost of financial instru-
ments, as they could be subject to greater
volatility and thus trigger faster repayment
or additional collateral. For instance, when
the CRAs downgraded AIG, the company
was required to raise additional collateral,
which contributed to its liquidity crisis.
In general, the original point of these rat-
ings was to help investors make better deci-
sions about what financial products to invest
in. They also pressure issuers to respect their
obligations, as well as assist in the market
pricing of debt instruments. To prove that
they add value beyond that already reflected
in market prices, credit ratings need to dem-
onstrate accuracy over time. As the actual
performance of the rating will not be discov-
ered until long after its issuance, if even then,
reputation can serve as an important source
of market discipline, with changes to reputa-
tion having significant impact on credit rat-
ing agencies’ profitability and success.
The CRA business model generally works
in one of two ways. Credit rating agencies
can rate financial products and make those
ratings available though a subscription ser-
vice to investors. In this business model,
CRAs’ primary clients are investors, and the
CRAs have an incentive to keep ratings ac-
curate so that investors will continue to pay
their accurate ratings. The other CRA busi-
ness model is for CRAs to sell individual
ratings to financial-product issuers. The is-
suers then sell the financial products to in-
vestors, using the associated credit ratings
as a form of “certification.” In this model,
CRAs’ primary clients are issuers, and CRAs
have some incentive to rate higher, but with-
in what is credible, since their reputations
matter to investors who purchase the rated
financial products from the CRAs’ primary
clients. In a competitive market, market par-
ticipants would diminish the value of rat-
Table 2
Comparison between Moody’s, S&P, and Fitch Rating Scales
Long-Term Rating Scales Comparison
Standard & Poor’s AAA AA+ AA AA- A+ AA-
Moody’s Aaa Aa1 Aa2 Aa3 A1 A2 A3
Fitch IBCA AAA AA+ AA AA- A+ AA-
Standard & Poor’s BBB+ BBB BBB- BB+ BB BB- B+ BB-
Moody’s Baa1 Baa2 Baa3 Ba1 Ba2 Ba3 B1 B2 B3
Fitch IBCA BBB+ BBB BBB- BB+ BB BB- B+ BB-
Standard & Poor’s CCC+ CCC CCC- CC C D
Moody’s Caa1 Caa2 Caa3 Ca C
Fitch IBCA CCC+ CCC CCC- CC C D
Source: Bank for International Settlements, “Long-term Rating Scales Comparison,” http://www.bis.org/bcbs/
qis/qisrating.htm.
6
There is some
evidence that
investors do
distinguish
among the
different rating
agencies on the
basis of quality.
ings inflation by discounting ratings that
were inflated simply to please issuers. There
is some evidence that investors do distin-
guish among the different rating agencies
on the basis of quality, as found in a recent
empirical study by Livingston, Wei, and
Zhou, which added to previous studies find-
ing similar results.3 The authors help to rec-
oncile previous findings that while issuers
view S&P as being more accurate, investors
favor Moody’s.4 Their findings suggest that
it is the view of investors that ultimately has
a greater impact on bond prices.
Typically, issuers purchase two credit rat-
ings for each financial product, and some-
times three if the first two are split.5 Inves-
tors, however, typically rely on only one
rating. When the investor is a commercial
bank and when multiple ratings exist for a
security, capital rules generally require the
lower rating to be utilized for calculating
risk-weighted capital. Financial regulations
usually require only a single rating when de-
termining whether a particular asset can be
“held,” and, in the case of multiple ratings,
allow the higher rating to be used to deter-
mine whether an asset is eligible to be held.6
The History of Rating Agencies
The present-day credit rating industry
has a long history, beginning in the 19th
century with financial publishing. The Mer-
cantile Agency, one of the first credit report-
ing agencies, was formed in 1841 and used
a network of agents to gather information
on operating statistics, business standing,
and creditworthiness on businesses. (In fact,
Lewis Tappan established the Mercantile
Agency to ameliorate information asymme-
tries that likely led to the financial crisis of
1837.) It then disseminated this information
to subscribers. Eventually what became to-
day’s credit rating agencies first emerged at
the turn of the 20th century, as publishers
from the financial press, such as John Moody
and Henry Poor, began collecting finan-
cial and operating statistics on the railroad
bond market and selling the information to
subscribers. This information collection in-
cluded examining the quality of a business’s
portfolio of opportunities and the manage-
ment’s success in pursuing these opportu-
nities, its ability to respect debt obligations,
and its tendency to honor debts.7 Eventu-
ally this information was used to create an
estimate of risk associated with corporate
debt. What we would now recognize as credit
ratings were first issued by Moody’s Analy-
ses Publishing Company in 1909, H. V. and
H. W. Poor Company in 1916, Standard Sta-
tistics Company in 1922, and Fitch Publish-
ing Company in 1924.8 Langohr and Lan-
gohr argue that the expansion of the credit
rating industry formed an information in-
frastructure necessary for bond markets to
expand throughout the 20th century.9
During the Great Depression, John
Moody’s and Henry Poor’s credit rating firms
performed well, with their highly rated bonds
being significantly less likely to default dur-
ing a period of high bond default rates. This
bolstered these companies’ reputations for
accuracy and dependability. Between World
War II and the 1970s, the financial markets
experienced relative stability, but innovation
stagnated. For this reason, CRAs were only
modestly profitable.
Major structural shifts took place during
the later part of the 1960s and the 1970s,
as market-based corporate funding became
more common and the demand for ratings
increased rapidly. The conduct of credit
analysis is a function of banks and credit
rating agencies. These entities compete in
the marketplace for the provision of credit
analysis. Commercial bank credit analysis
usually takes place in the context of a loan,
often held on its balance sheet, whereas rat-
ing agency analysis is performed in concert
with the issuance of a marketable debt in-
strument. As inflation began to accelerate in
the late 1960s and 1970s, commercial banks
and thrifts found it increasingly difficult to
raise deposits because of the legal limit, un-
der Regulation Q, on what they could pay
depositors. The extension of Regulation Q
to savings and loans in 1966 further restrict-
ed the ability of traditional lenders to pro-
7
The
inflation-induced
disintermediation
of bank business
lending led to
the growth of
the nonfinancial
commercial paper
market, which
tripled in size
between 1975
and 1980.
vide credit. Others, such as mutual funds,
found that banks’ funding problems pre-
sented a business opportunity in the provi-
sion of credit. As these other market partici-
pants generally lacked an infrastructure to
conduct extensive bank-like credit analysis,
they turned to the major rating agencies.
The first segment to undergo this tran-
sition was business lending. While railroad
bonds had served as the catalyst for the
growth of the U.S. corporate bond market,
until the 1960s other nonfinancial indus-
tries still relied almost exclusively on either
bank loans or internal funding for their fi-
nancing needs. The inflation-induced disin-
termediation of bank business lending led to
the growth of the nonfinancial commercial
paper market, which tripled in size between
1975 and 1980.10 The longer-maturity cor-
porate bond market grew by 80 percent dur-
ing these same years. With this growth in
the structured finance market also came an
increasing demand for credit ratings.11 Then
the boom of the U.S. residential mortgage-
backed securities and home-equity loan
markets rapidly increased the structured fi-
nance market. Starting in the 1980s, another
contributing factor to increased demand
was globalization, which enticed credit rat-
ings agencies to expand beyond the United
States. Agencies opened offices in the United
Kingdom, Japan, France, Australia, Canada,
India, Sweden, Russia, Mexico, and Austra-
lia. Thus, globalization and the introduction
of new financial products to be rated drove
greater demand for CRA products and ac-
counted for a large share of CRAs’ profitabil-
ity.12
In response to changes in the market-
place, in 1974 Standard and Poor’s shifted
its business model and began charging is-
suers for ratings rather than charging a
subscription service to investors. Lawrence
White points out that this business model
transition also coincided with the spread of
low-cost photocopying that led to free rid-
ing among investors, who would share rat-
ings.13 The 1970 Penn Central default on
$82 million in commercial paper, followed
by liquidity crises, also refocused attention
on the importance of credit risk. As issuers
wanted to assure investors of quality rat-
ings, they began actively seeking out rat-
ings. As market demand for ratings shifted,
CRAs began charging issuers rather than in-
vestors.14 This business model change also
coincided with the SEC introducing the
NRSRO designation and incorporating it
into regulations.
The History of Credit Rating Agency
Regulation
While the financial markets were chang-
ing and structured finance grew, govern-
ment regulators also played a key role in
the credit rating agency industry.15 Regula-
tors did not seek to regulate CRAs directly,
but rather used credit ratings as a means
to oversee the financial markets. Securities
and Exchange Commission regulators wrote
many rules that specifically identified CRAs,
thereby indirectly creating a regulatory
framework reliant on CRAs.
The first set of regulations involving cred-
it rating agencies went into effect after the
onset of the banking crisis in March 1931.
Banks were in need of greater liquidity fol-
lowing the onset of the Great Depression,
and so they dumped their lower grade bonds
on the market, which contributed to the
overall decline in bond prices. This lower val-
uation of bonds reduced the market value of
bank’s bond portfolios overall and contrib-
uted to bank failures, demonstrating that
bond values, rather than simply defaults,
also mattered to bank survival.16 Conse-
quently, the Office of the Comptroller of the
Currency (OCC) set out to regulate banks’
capital reserves with the hopes of preventing
future bank failures. To do this, the OCC set
minimum capital reserve requirements to
ensure banks did not become overleveraged.
To ensure compliance with the new reg-
ulations, the OCC looked for an outside
group with expertise in evaluating bonds to
determine how much risk, and thereby, how
much value, was associated with banks’ as-
sets. Credit ratings helped the OCC conduct
8
Obtaining a
designation was
not necessary
for credit
rating agencies
to operate,
but doing so
put them at
a significant
advantage.
a valuation of national bank bond portfo-
lios, and the OCC provided incentives for
investments highly rated by a CRA, increas-
ing demand for ratings. The comptroller
stipulated that national banks would not be
required to charge off depreciation to mar-
ket value on bonds receiving one of the four
highest ratings.17 This meant that publically
traded bonds rated BBB or higher by at least
one CRA could be valued at book value.18
This increased the demand for credit rat-
ings because otherwise defaults would have
counted 25 percent against bank capital. In
1936, the OCC and the Federal Reserve di-
rected that banks not hold bonds rated be-
low BBB by at least two credit rating agen-
cies. These rules introduced CRAs into the
financial regulatory framework. Banks were
now required to obtain credit ratings for
their assets to ensure they met federal capital
reserve requirements, and they also were pro-
vided additional incentives for having bonds
rated highly by CRAs.
In the late 1960s, a considerable increase
in volume on the New York Stock Exchange
overwhelmed the mechanisms that brokers
used to transfer securities. This, combined
with a subsequent trading volume decline,
drove nearly 100 brokerage firms out of busi-
ness.19 The SEC later concluded that inade-
quate access to liquid capital exacerbated the
crisis, and thus sought to enforce more strin-
gent capital requirements. Consequently, the
SEC adopted another wave of banking regu-
lation, beginning in 1973, with a uniform
net capital rule.20 Part of this wave included
the Net Capital Rule for broker-dealers (Rule
15c3-1), which was intended to ensure “that
registered broker-dealers have adequate liq-
uid assets to meet their obligations to their
investors and creditors.”21 To ensure com-
pliance, regulators turned to select credit
rating agencies to measure leverage. This es-
tablished a new designation for select credit
rating agencies called the Nationally Recog-
nized Statistical Rating Organization.
The broker-dealer net capital rule required
broker-dealers to deduct percentages of mar-
ket value from their proprietary securities’
position from their net worth, depending
on the ratings of these securities.22 This was
intended to safeguard broker-dealer propri-
etary securities from price fluctuation risks.
Obtaining investment-grade ratings from at
least two NRSROs reduced the requirement
of deducting particular percentages of mar-
ket value from the net worth of instruments.
This incentivized broker-dealers to invest in
higher NRSRO rated instruments because it
translated into higher net capital.
The SEC instituted the NRSRO designa-
tion to ensure that bank issuers would not
simply find credit rating agencies whose only
purpose was to deliver high ratings on finan-
cial instruments. Not all credit rating agen-
cies were bestowed the NRSRO designation;
instead, the SEC grandfathered in Moody’s,
S&P, and Fitch rating agencies. These com-
panies were selected because of their previ-
ous record of accurate ratings. The NRSRO
designations did not ensure future accurate
ratings, but instead were an endorsement
of Moody’s, S&P’s, and Fitch’s past achieve-
ments. Obtaining a designation was not
necessary for CRAs to operate, but doing so
put them at a significant advantage, as par-
ticular investors (including both public and
private pension funds, as well as insurance
companies) were legally mandated to pur-
chase investments highly rated by NRSROs.
Moreover, other investors were also incentiv-
ized to purchase investments highly rated
by NRSRO CRAs to obtain regulatory ben-
efits. This boosted demand specifically for
NRSRO CRA ratings.
The SEC did not grant many NRSRO des-
ignations, and those companies for which
they did often merged, keeping the total
number of NRSRO CRAs to about three to
four. In 1982, the SEC designated Duff &
Phelps, and in 1983, McCarthy, Cristanti
and Maffei (MCM). In 1991, Duff & Phelps
purchased MCM and spun off its credit rat-
ing business. That same year, the SEC des-
ignated IBCA, and in 1992, it designated
Thomson Bankwatch for banks and finan-
cial institutions specifically. However, IBCA’s
frustration with the SEC’s preventing them
9
Over time,
regulators
became
increasingly
dependent on
Nationally
Recognized
Statistical Rating
Organization
credit rating
agencies.
from expanding their designation beyond
bank ratings drove them to purchase Fitch in
1992. Fitch later bought both Duff & Phelps
and Thomson BankWatch in 2000. Mergers
quickly brought the number of CRAs down
to three by the turn of the millennium.
Over time, regulators became increasingly
dependent on NRSRO credit rating agencies,
and as Langohr and Langohr argue, “the use
of ratings in regulations is most widespread
in . . . the U.S.”23 In fact, by June of 2005, there
were at least 8 federal statues, 47 federal rules,
and 100 state laws referencing credit ratings
issued by NRSRO CRAs.24 Table 3 high-
lights several examples of regulatory uses of
NRSRO CRA credit ratings.25
CRAs came under fire in the early 21st
century with the implosion of Enron,
WorldCom, and Parmalat. The dominant
CRAs with NRSRO designations gave most
of these companies’ bonds investment-
grade ratings within a few days or months
of them declaring bankruptcy. Beyond these
evident errors, critics blamed CRAs for insti-
gating crises in particular industries. France
Telecom SA’s CEO Michel Bon blamed a
Moody’s downgrade for initiating a debt
crisis. Interest groups with a stake in credit
ratings, regulators, and legislators joined
in greater scrutiny of credit rating agencies
and regulatory dependency on the ratings.
The U.S. Senate began conducting investi-
Table 3
Sample List of Rules Referring to Nationally Recognized Statistical Rating
Organization (NRSRO) Credit Rating Agencies (CRAs)
Investment Company Act of 1940
Rule 2a-7 enacted in 1991
Less than 5 percent of money market mutual fund assets may
be invested in commercial paper that NRSROs assign lower
than the first or second highest tier. (NRSRO mentioned 29
times in Rule 2a-7.)
Securities Exchange Act of 1934
“Exchange Act” Rule 15c3-1
Required broker-dealers to deduct percentages of their pro-
prietary securities’ market value when computing net capital.
However, reduced deductions are required for particular
securities rated investment grade by at least two NRSROs.a
Secondary Mortgage Market
Enhancement Act of 1984
Congress used the term NRSRO in the definition of “mort-
gage related security,” requiring it to be rated in one of the
two highest rating categories by at least one NRSRO.
Simplification of Registration
Procedures for Primary Securities
Offerings, Securities Act 1992
The SEC uses NRSRO ratings to help distinguish between
different types of securities that may be issued using simpli-
fied registration procedures.b
Investment Company Act of 1940
Rule 2a-7
This rule requires money market funds to limit investments
to “eligible securities” that are rated in either of the top two
short-term debt rating categories by the requisite number of
NRSROs.c
Investment Company Act of 1940
Rule 3a-7
Issuers of fixed-income securities rated in one of the top four
rating categories by at least one NRSRO are exempted from
registering and complying with the Investment Company Act.d
Securities Exchange Act of 1934
“Exchange Act” Rule 10b-6
Exempts particular transactions in nonconvertible debt and
nonconvertible preferred securities from Exchange Act provi-
sions if the securities are rated investment grade by at least
one NRSRO.e
Continued next page.
10
Credit ratings
are intended
to reduce
information
asymmetries
between firms
and investors.
gations, holding hearings, and making pro-
posals for reforms. This eventually led to the
Credit Rating Agency Reform Act of 2006.
CRAs’ Dual Role
The regulatory use of credit ratings reveals
their dual role in the marketplace. As Beaver,
Shakespeare, and Soliman explain, credit rat-
ings have valuation and contractual uses.26
First, credit ratings are intended to reduce
information asymmetries between firms
and investors and improve the functioning
of financial markets. They are an opinion
about the measure of investment quality
and default probability. These, in turn, help
marketplace investors form valuations of
investments’ value and risks. Second, credit
ratings are used for contractual purposes,
primarily for regulatory compliance and
private contracts. In a sense, NRSRO credit
ratings are a regulatory license indicating a
debt instrument is eligible by law for par-
ticular kinds of investments and regulatory
perks. This license also helps insulate fidu-
ciaries’ liability when making investments.27
Under a regulatory licensing regime, favor-
able NRSRO ratings can reduce the costs of
regulation to market participants. As Coffee
explains, “such sales of regulatory licenses
need not be based on trust or reliance on the
rating agencies … but only on the short-term
cost savings realizable”; or, in other words,
the regulatory benefits realized.28 NRSROs
received their designations because of past
good performance, not necessarily because
of future good performance; as such, a regu-
latory license is not necessarily indicative of
the quality or risk of the financial instru-
ment receiving it.
Hypotheses
Oligopoly
A cursory overview of regulatory reliance
on NRSRO-designated CRA ratings suggests
Federal Deposit Insurance Act
Section 1831e
Congress defines “investment grade” corporate debt for sav-
ings associations as only securities rated in one of the four
highest categories by at least one NRSRO.f
Employee Retirement Income
Security Act of 1974
Required both public and private pension funds to base part
of their investment criteria on bond ratings provided by
NRSRO designated CRAs.
Retirement Income Security Act
of 1974
Mandated that both public and private pension funds base
part of their investment criteria on bond ratings provided by
NRSRO designated CRAs.
Source: Jonathan G. Katz, SEC Concept Release: Nationally Recognized Rating Organizations, Release Nos: 33-7085,
34-34616, IC-20508, International Series Release No. 706 (Washington: SEC, 1994)( File No. S7-23-94), p. 12.
a Please see 17 CFR 240,15c3-1, Adoption of Amendments to Rule 15c3-1 and Adoption of Alternative Net
Capital Requirement for Certain Brokers and Dealers, Exchange Act Release No. 11497 (June 26, 1975); and 40
FR 29795 (July 16, 1975). Please also see Securities and Exchange Commission, Securities Exchange Act Rel. No.
34-10,525 (Washington: SEC News Digest, 1973).
b Please see Adoption of Integrated Disclosure System, Securities Act Release No. 6383 (March 16, 1982) and
Adoption of Simplification of Registration Procedures for Primary Securities Offerings, Securities Act Release
No 6964 (October 22, 1992).
c See 17 CFR 270.2a-7.
d See Exclusion from the Definition of Investment Company for Structured Financing, Investment Company
Act Release No 19105 (November 19, 1992), 52 SEC Dkt. 4014.
e See 17 CFR 240 10b-6(a)(4)(xiii).
f See footnote five of 12 U.S.C. 1831e(d)(4)(A), enacted in 1989, at http://www.sec.gov/rules/concept/34-34616.pdf.
Table 3 Continued
11
In a competitive
market new firms
would enter
the industry to
capture a share
of economic
profits, but in an
oligopoly they
are prevented
from doing so by
barriers to entry.
these firms likely operated in an oligopolis-
tic market. Before examining the evidence
we must first define what an oligopoly is,
how it is created, and how it works.
An oligopoly is an industry dominated
by few firms that is protected from addi-
tional competing firms by either artificial
or natural barriers to entry, making it diffi-
cult for new firms to enter the market and
credibly compete with the dominant firms.
Reduced marketplace competition provides
oligopolistic firms greater market power
than competitive firms because they can set
price profitability above competitive levels
and reduce output. This allows oligopolistic
firms to enjoy higher profits than competi-
tive firms because fewer firms fulfill market
demand. In a competitive market new firms
would enter the industry to capture a share
of economic profits, but in an oligopoly
they are prevented from doing so by barriers
to entry (see Figure 2).
Figure 2
Lessons from a Monopoly Model
Source: Chart constructed by authors.
Note: The figure demonstrates higher prices and lower quantity (or lower quality) of credit risk information
with oligopolistic conditions compared with a competitive market. The shaded box represents economic profits
gained via oligopolistic market power. MC is the marginal cost curve and AC is the average cost curve of conduct-
ing research and analysis to determine the quality of financial instruments. D is the demand curve, represent-
ing the demand for credit-risk information for financial instruments, and MR is the marginal revenue for each
additional rated financial instrument. These, in turn, determine the price for credit ratings and the quantity (or
quality) of credit risk information supplied. One can also view quantity of information supplied as informa-
tional quality because more information improves the quality of credit risk estimation.
12
Reputational
factors create
natural barriers
to entry in the
credit rating
agency market,
but most barriers
result from
the regulatory
designation
of Nationally
Recognized
Statistical Rating
Organization
credit rating
agencies.
Generally, oligopolies result from either
natural barriers to entry or government-
created barriers to entry. Both tend to have
higher prices, higher profits, and reduced
quantity supplied than would otherwise be
the case in a competitive market. Oligopo-
lies sustained by natural barriers to entry
remain susceptible to new technology and
new entrants overtaking their oligopoly, cre-
ating a more competitive market. Oligopo-
lies sustained by government regulation are
also subject to competitive markets, but in
some instances, their oligopolies may be
sustained by government mandates.29
Although reputational factors create
some natural barriers to entry in the CRA
market, most of the barriers to entry result
from the regulatory designation of NRSRO
CRAs. Thus, we would expect the CRA mar-
ket to be generally resistant to competitive
pressure, be dominated by few firms, have
high profits, and have restricted output sup-
plied. Restricted output supplied can be in
terms of informational output (quality as
a dimension of quantity), rather than the
sheer number of ratings produced. Without
the threat of competition, oligopolistic CRAs
are likely to become more complacent in
their methodologies. Also, with fewer firms
and reduced competition, markets may find
it less likely to discover new tools to better
measure credit risk.
Reduced Investor Due Diligence
Regulations incentivize investors to
purchase financial instruments with high
NRSRO credit ratings, rather than credit rat-
ings with high informational value. Since it
is hard to determine the accuracy of credit
ratings, it is understandable why regulators
use high NRSRO credit ratings as a proxy.
Nevertheless, if investors were primarily
concerned with accurate credit ratings, they
would be more likely to make investment
decisions in line with their own risk-return
preferences and they would reward CRAs
that are better at innovating methodologies
to derive accurate ratings. However, if inves-
tors were primarily concerned with obtain-
ing regulatory benefits, they would be more
likely to make investment decisions based on
investments receiving high NRSRO credit
ratings, and they would reward CRAs that
were better at innovating methodologies to
derive high ratings. Although the NRSRO
firms charged with identifying investment
vehicles eligible for purchase and or regula-
tory perks may also be the firms most likely
to deliver accurate ratings, this is not neces-
sarily the case. The NRSRO designation was
given to firms for past good performance,
but that does not guarantee those firms will
continue to offer the most accurate ratings.
Given investors’ competing incentives to ob-
tain accurate credit ratings (for information)
and high NRSRO credit ratings (for regula-
tory benefit), it is not clear whether investors
would necessarily punish NRSRO credit rat-
ing agencies that offered low-quality ratings.
Inflated Demand
Regulatory reliance on credit ratings,
specifically NRSRO CRA ratings, suggests
regulation may have artificially boosted de-
mand for these ratings. Table 3 shows a se-
lection of regulations mentioning NRSRO
CRAs. Regulations would likely increase the
demand for high NRSRO credit ratings (rat-
ings’ contractual role), rather than increase
the demand for credit ratings with high in-
formational value (ratings’ valuation role),
because preferential regulatory treatment
was bestowed on high NRSRO credit ratings
rather than ratings with high informational
value (see Figure 3). We would expect pric-
es to rise above competitive levels, and the
quantity of information supplied to remain
below the competitive level.
Regulatory Impact:
Evidence
Regulations referencing NRSRO CRAs
created a de facto oligopoly in the ratings
market and regulation requiring or incen-
tivizing investors to purchase highly rated
products inflated and captured demand for
13
One of the
great strengths
of markets is
to convey a
tremendous
amount of
information in a
relatively simple
and efficient
manner via the
price mechanism.
high NRSRO ratings (even if those ratings
did not provide substantial informational
value). They also failed to incentivize inves-
tors to seek out ratings on the basis of high
informational value so they would invest in
high quality products. Instead, the regula-
tions incentivized investors to seek out high
NRSRO ratings in efforts to obtain regula-
tory privilege. This, in turn, reduced compe-
tition, increased prices, resulted in method-
ological complacency in credit risk modeling
innovation, and reduced investor due dili-
gence. The following section details evidence
of barriers to entry, high profits, restricted
output, inflated market demand, and inves-
tor due diligence.
Overview
One of the great strengths of markets is
to convey a tremendous amount of infor-
mation in a relatively simple and efficient
manner via the price mechanism. In markets
with relatively uniform goods and common-
ly shared knowledge, the price mechanism
is indeed quite powerful. In markets with
more heterogeneous goods and imperfect or
Figure 3
Lessons from a Monopoly Model: Shifting Demand
Source: Figure constructed by authors.
Note: The figure demonstrates higher prices and lower quantity (or lower quality) of credit risk information
with oligopolistic conditions compared with a competitive market. A shift in demand increases the difference
between competitive and oligopolistic results. The shaded boxes represent economic profits gained via oligopo-
listic market power. The lighter shaded box represents economic profits before demand shifts; the darker shaded
box represents additional economic profits after demand shifts.
14
Some investors
are legally
required,
and others
incentivized,
to invest in
the Nationally
Recognized
Statistical Rating
Organization’s
credit rating
agency products.
asymmetric information, nonprice mecha-
nisms often evolve in order to avoid Aker-
lof’s familiar “lemons problem,” or adverse
selection, in which some markets risk being
dominated by the sellers of shoddy goods.30
In markets where quality is revealed post-
purchase, producers may be able to commit
to providing high-quality goods by estab-
lishing a reputation for doing so. In essence,
reputation posts a “bond” that will be forfeit
if producers disappoint or cheat. That bond
can take the form of a brand name or cor-
porate goodwill, and the value of many cor-
porations depends heavily upon intangibles
such as these.
Goodwill, while reducing the information
asymmetries, can also end up as a barrier to
entry. Without a track record, how does one
compete with incumbent firms? One possi-
bility is for those with a strong reputation in
one line of business to expand into another.
Given that the quality of credit analysis
by a rating agency will likely not be discov-
ered for months or years after the initial
analysis, this would appear a market likely
to be dominated by a handful of firms with
significant reputations. And indeed, even
before the advent of extensive federal finan-
cial regulation, the market for credit ratings
was so dominated.
Regulation is often offered as a method
for ensuring the provision of a minimum
quality. While making no claims as to the
actual quality provided, the SEC’s involve-
ment in the market for credit ratings has
been an attempt to institute minimum qual-
ity standards. However, minimum quality
standards can also restrict competition, in-
crease concentration, and reduce consumer
welfare, and do not necessarily guarantee
the sought-after minimum quality. Whether
they do so, or whether they improve compe-
tition and consumer welfare, is ultimately
an empirical question, although theory can
guide the interpretation of relevant data.
When imposing minimum quality stan-
dards on an industry already characterized
by high concentration and strong repu-
tations, the standards’ net impact can be
evaluated by two metrics. First, minimum
quality standards should reduce the value
of incumbent firms’ reputational capital
(goodwill). If all new entrants need to do is
meet the new standard, then they should
not have to build reputational capital in
order to be competitive. For this reason, ob-
served market concentration should also fall
with the decrease in incumbent goodwill. If
concentration and incumbent reputational
capital increase with the imposition of qual-
ity standards, then such standards are likely
welfare-decreasing. We will come back to
measures of concentration and goodwill as
we examine the market for credit ratings.
Inflated Market Demand for NRSRO
CRA Ratings
Regulatory wording alone demonstrates
how regulations arguably drove additional
business to the NRSRO CRAs than would
have otherwise been the case (see Table 3).
Some investors are legally required, and oth-
ers incentivized, to invest in NRSRO CRA-
rated products. If demand were not inflated,
we would expect reduced relative informa-
tional value of ratings to be associated with
lower demand. Yet if the greater value from
NRSRO CRA ratings came from gaining
access to the regulatory license or to prefer-
ential regulatory treatment, then we would
expect the relative informational value to
matter less.
In fact, many people do question the in-
formational value of these ratings.31 Some
have argued that credit spreads could sub-
stitute for credit ratings, suggesting that
NRSRO CRA ratings may not provide sub-
stantial additional information.32 Moreover,
advances in information technology and the
increasing interconnectedness of society may
indeed have reduced the relative informa-
tional value of ratings compared with their
value in the past.33
Some believe credit ratings are “lagging
indicators of credit quality.”34 As James Van
Horne concludes, “[w]hile the assignment of
a rating for a new issue is current, changes
in ratings of existing bond issues tend to lag
15
While a few
misses are to be
expected, the
rating agencies
uniformly failed
to forecast a
major decline
in the housing
market,
illustrating that
the system is
subject to more
than just a few
random errors.
behind the events that prompt the change.”35
Packer and Cantor also find evidence to sug-
gest that agencies lag behind the market when
agencies initiate a ratings change.36 More-
over, credit ratings have proved themselves
to be quite inaccurate at times, such as with
WorldCom, Enron, Parmalat, and the 2008
financial crisis.37 While a few misses are to
be expected under any market structure, the
fact that the rating agencies uniformly failed
to forecast a major decline in the housing
market illustrates that the system is subject
to more than just a few random errors. Bruce
Lehmann at Columbia Business School ar-
gued that he has “never known a portfolio
manager who goes by the ratings.”38 A 2002
survey by the Association of Financial Profes-
sionals found only 29 percent of profession-
als believe rating changes to be accurate.39
In sum, it is not clear how ratings contin-
ue to offer significant additional informa-
tion if they lag behind alternative indicators,
and are, in some cases, underestimating the
ultimate risk of default. One would expect
their value in the marketplace to decline
following the bankruptcies of WorldCom
and Enron. One would also expect the de-
mand for credit ratings to decline or at least
remain fairly constant, at least among the
prominent NRSRO CRAs (as well as the
NRSRO CRAs’ values).
Nevertheless, NRSRO CRAs’ fees and prof-
itability have soared, especially from 2002 to
2008.40 It is estimated that the average rate of
return for credit rating agencies was slightly
over 42 percent from 1995 to 2000, with oper-
ating margins as high as 54 percent in 2006.41
Figure 4 demonstrates Moody’s skyrocketing
Figure 4
Historical Stock Prices 1998–2011: Moody’s vs. Dow Jones and S&P 500 Indices
Source: Compiled by the authors from Google Finance, Historical Data, http://www.google.com/finance/
historical?q=NYSE:MCO; and Yahoo Finance, Historical Data, http://finance.yahoo.com/q/hp?s=MCO+Hist
orical+Prices.
16
Figure 5
Moody’s and S&P Rated Residential Mortgage-Backed Securities (RMBS) and
Collateralized Debt Obligations (CDOs) in 2002 and 2006
Source: Carl Levin and Tom Coburn, Wall Street and the Financial Crisis: Anatomy of a Financial Collapse, Permanent
Subcommittee on Investigations (Washington: United States Senate, 2011), http://www.ft.com/cms/fc7d55c8-
661a-11e0-9d40-00144feab49a.pdf.
Rated CDOs per year
Rated RMBS per year
17
Regulations that
were intended
to incentivize
investors to
make better
decisions instead
incentivized them
to prioritize high
credit ratings
over quality
investments.
stock price in contrast to the Dow Jones and
S&P 500 indices. Figure 5 shows the jump in
the number of rated Residential Mortgage-
Backed Securities and Collateralized Debt
Obligations (CDOs) between 2002 and 2006
for both Moody’s and S&P. Figure 6 demon-
strates several jumps in Moody’s goodwill
between 1999 and 2010. Together, Figures
4–6 demonstrate that at least several NRSRO
CRAs continued to be highly valued as com-
panies and they continued to rate more prod-
ucts. This is surprising, because if a product
is becoming less valuable in the marketplace,
investors would most likely turn to other
sources of risk assessment.
Even those who argue that credit ratings
continue to have significant additional in-
formational value should expect that reputa-
tional values would have at least been tainted
throughout the crises of WorldCom, Enron,
and the 2008 financial crisis. Yet if only in-
formational value and reputational factors
matter, we cannot explain the evidence in
Figures 4–6.
Instead, a more likely explanation is one
described by Langohr and Langohr, who ar-
gue that “the regulatory uses of ratings . . .
created a captive demand for credit ratings
per se by market participants.”42 If investors
demanded credit ratings in efforts to meet
regulatory requirements and to obtain pref-
erential regulatory treatment, rather than
to obtain informational value, then it is not
surprising that demand for NRSRO CRAs’
ratings remained high during this time pe-
riod, as preferential treatment for NRSRO–
rated investments continued.
Investor Due Diligence
Regulations that were intended to in-
centivize investors to make better decisions
instead incentivized them to prioritize high
credit ratings over quality investments. Since
regulators used credit ratings as a proxy
for credit quality, it is understandable that
rules were implemented to incentivize high
ratings. Nevertheless, regulators could not
ensure that high credit ratings did indeed
Figure 6
Moody’s Goodwill from Balance Sheet
in millions ($)
Source: Moody’s Annual Reports.
Note: Goodwill is found on the company’s balance sheet by stating the difference between its purchase price and
the sum of the fair value of net identifiable assets, and could possibly be driven by reputation.
18
Credit ratings
often lag behind
market indicators
by as much as six
months.
represent high investment quality. The rules
incentivized firms to hold riskier assets with-
in a single credit rating. Instead of due dili-
gence, or seeking out accurate information
on investment quality, investors were incen-
tivized (or required) to focus on purchasing
instruments with high NRSRO ratings. In-
vestors faced a choice: they could prioritize
pursuing regulatory privilege by purchasing
investments with high NRSRO credit rat-
ings, or they could seek high informational
quality. They chose the former.
This explains why demand for NRSRO
CRAs did not seem to decline after high, yet
inaccurate, ratings were issued for World-
Com and Enron months, if not days, before
their announced bankruptcies. Figures 4,
5, and 6 show how NRSRO CRAs seemed
to continue to be rewarded with increas-
ing stock prices, more business, and higher
goodwill.
Conflict of Interest
It is possible that the 1973 regulatory
changes, instituting the NRSRO designa-
tions, may have played a role in incentiviz-
ing NRSRO CRAs to switch their business
models from investor-pays to issuer-pays.
It is easy to understand how an issuer-pays
business model is prone to conflicts of in-
terest.43 This model raises concerns about
incentives for CRAs to issue inflated ratings
to promote their business with clients.44 In
theory, CRAs could gain short-term profit by
overstating issuer or investment quality, but
this would come at the expense of long-term
reputational loss to their respective firms.45
Although reputational concerns may con-
strain NRSRO CRAs from engaging in overt
catering to issuers, their increased market
power may have allowed them to at least
marginally construct higher ratings. Also,
NRSRO ratings still offer regulatory benefits
regardless of whether investors or issuers
purchase them.
CRAs have assured regulators they man-
age conflict-of-interest problems, for in-
stance, by separating compensation from
revenue. Also, the SEC has written regula-
tions to try and manage conflict-of-interest
problems by prohibiting coercive actions and
anti-competitive behavior.46 Nevertheless,
evidence continues to mount demonstrat-
ing that non-NRSROs tend to respond more
quickly to information changes in the mar-
ket, and that NRSRO CRAs do tend to lag
in reporting downgrades.47 Moreover, other
empirical research finds that credit rating ad-
justments tend to adjust in favor of slightly
higher ratings, and that ratings often lag
behind market indicators by as much as six
months.48 Han Xia finds evidence to suggest
that the issuer-pays rating model contributes
to CRAs’ incentives to issue inflated ratings.49
Alternative explanations do exist for the
business model change. Rather than captive
demand, photocopying may have reduced
the profitability of rating subscription ser-
vices, as investors would share informa-
tion.50 Langohr and Langohr argue that the
increase in market-based financing in the
1970s made it difficult for CRAs to meet
demand with subscription-based services.51
Another explanation is that the early 1970s
liquidity crisis, subsequent to the 1970 Penn
Central default in commercial paper, moti-
vated issuers to assure investors of quality
ratings.
Although these explanations may very
well be partially correct, it still remains un-
clear why the market would continue to tol-
erate a conflict of interest between NRSRO
CRAs and issuers, especially after Moody’s,
S&P, and Fitch had rated Enron investment
grade four days before the company declared
bankruptcy.52 It also remains unclear why
investors would continue to trust the sub-
jective opinions of CRAs, who contend they
manage conflict of interest (which can result
in lower quality ratings) or that they believe
the SEC’s regulations are capable of prevent-
ing conflict of interest. CRAs’ management
of conflict of interest was also challenged in
2011, when the Permanent Subcommittee
on Investigations found that Moody’s and
UBS executives met to discuss when to start
downgrading and the problems it would
cause.53
19
Regulations
stipulating
that regulatory
preference only
be afforded to
those credit
rating agencies
designated by
the Nationally
Recognized
Statistical Rating
Organizations
created barriers
to entry.
The evidence demonstrates continued
NRSRO CRA profitability, despite their fail-
ure to detect problems in WorldCom, Par-
malat, Enron, and mortgage-backed securi-
ties, as well as potential conflict-of-interest
problems. Thus, if reputation is not the
most important factor driving profitability,
it suggests that regulations creating captive
demand for NRSRO ratings may fill that
role instead. The incentives to obtain regula-
tory privilege may explain the shift in market
power from investors to the NRSRO CRAs.
Oligopoly
Barriers to entry: few firms. We have ar-
gued that regulations bolstered market de-
mand for NRSRO CRAs and potentially per-
petuated an investor-pays business model.
These problems may have been mitigated
had there been greater competitive threat
from additional firms or new entrants. New
firms may have acted as whistleblowers,
which may have resulted in greater meth-
odological innovation in the marketplace.
However, regulations stipulating that regu-
latory preference only be afforded to those
using NRSRO-designated CRA ratings creat-
ed barriers to entry for non-NRSROs CRAs,
because NRSRO CRAs’ ability to essentially
sell “regulatory licenses” gave them a signifi-
cant market advantage over potential new
entrants.
Interestingly, the SEC did not prevent
non-NRSRO CRAs from issuing ratings, but
instead excluded them from offering regula-
tory privilege. The NRSRO-designated CRAs’
ability to offer regulatory privilege was a bar-
rier to entry to a very important and profit-
able area of the marketplace. Thus, Table 3
provides several examples for how barriers to
entry were constructed in the CRA market-
place.
Besides a cursory overview of regulatory
uses of NRSRO CRAs, additional evidence
points to barriers to entry: namely, sustained
high profits and high market concentra-
tion with few firms. As explained earlier, the
NRSRO CRA market has experienced in-
creased profitability. Economic theory sug-
gests that new firms would enter the market
to capture some of the increased profits.
However, the SEC did not often grant new
NRSRO designations, so increasing profit-
ability for only a few firms continued.
The Herfindahl-Hirschman Index (HHI)
is a quantitative measure of industry concen-
tration that accounts for both the number of
firms in the industry and each firm’s market
share. Specifically, the measure takes into ac-
count NRSRO revenues, the number of enti-
ties issuing NRSRO-rated debt securities, and
the dollar amount of new U.S.–issued asset-
backed securities. An HHI index lower than
1,000 is considered a competitive market.
Most of the HHI indices in Table 4, 5, and 6
are instead over 3,000, demonstrating high
market concentration. By the middle of the
decade, 80 percent of rated issues were rated
by only Moody’s and S&P, and 14 percent
rated by Fitch.54 Although sustained profits
should have attracted more firms to the mar-
ket, it did not, and high market concentration
remained. This suggests there are either regu-
latory or natural barriers to entry.
It is possible that the CRA industry tends
toward natural concentration because of net-
work and reputation effects. Lawrence White
explains how a few CRAs benefit from a “net-
work effect” because many users desire con-
sistency across ratings categories and tend
to use the same rating agency to achieve this
goal.55 Reputation effects may result from
the time it takes for a CRA to develop investor
trust. This would suggest early CRA entrants
have an advantage that continues even as new
incumbents enter the market. For this reason,
Langohr and Langohr argue that “a small
number of CRAs with the highest reputa-
tion for quality and independence will always
dominate” independent of regulations.56
Nevertheless, the dominant NRSRO CRAs
did not prove their quality ratings during
the WorldCom and Enron collapses or the
2008 financial crisis. Nor do their conflicts
of interest instill confidence in their indepen-
dence and role as objective evaluator of credit
risk. Moreover, if early entrant CRAs fail to
produce consistently accurate ratings, it does
20
The process
to obtain the
Nationally
Recognized
Statistical Rating
Organization
designation
was opaque,
without clear and
consistent rules.
not follow that the few dominant CRAs will
be the same firms over time.
Since reputation alone need not explain
CRAs’ small number and continued success,
it suggests that the small number of exist-
ing firms is not solely the product of natural
barriers to entry, but instead also regulatory
barriers to entry: namely the NRSRO desig-
nation.
Non-NRSRO. Considering the perspec-
tive of non-NRSRO CRAs provides addi-
tional evidence that the small number of
firms in the industry is not the result of nat-
ural barriers to entry. Non-NRSRO CRAs
believed the SEC restricted their entry and
prevented them from capturing increasing
economic profits.57 Typically, these firms
waited 2–7 years to receive a status deter-
mination.58 The process to obtain NRSRO
designation was opaque, without clear and
consistent rules. Figure 7 shows NRSRO
CRAs as of 2010, the number of years they
produced credit ratings, and when they re-
ceived recognition as NRSROs.
Table 4
Herfindahl-Hirschman Index Based on Dollar Value of Newly Issued U.S.
Asset-based Securities, January 2004–June 2010
Asset Class 2004 2005 2006 2007 2008 2009 2010a
All U.S. asset-backed securities 3444 3375 3469 3398 3396 2973 2809
U.S. Commercial Mortgage-
Backed Securities 3224 3222 3359 3212 3751 2916 2804
U.S. Traditional Asset-Backed
Securities 3374 3338 3314 3280 3305 3262 3046
U.S. Prime Residential-
Mortgage Backed Securities 3677 3672 3542 3376 3148 3222 4145
U.S. Nonprime Residential-
Mortgage Backed Securities 3390 3177 3344 3515 3531 10000b6009
U.S. Collateralized Debt
Obligations 3772 3944 4173 4253 4846 3795 5561
Source: Securities and Exchange Commission, Action Needed to Improve Rating Agencies Registration Program and
Performance-Related Disclosures (Washington: United States Government Accountability Office, 2010).
a The HHIs for 2010 are based on data through June 30, 2010.
b Only one deal was issued in 2009, and it was rated by a single NRSRO.
Table 5
Herfindahl-Hirschman Index for Nationally Recognized Statistical Rating
Organizations Based on Total Revenues, 2006–2009
2006 2007 2008 2009
All asset classes 3617 3511 3333 3324
Annual Percentage Change (%) -2.93 -5.08 -0.27
Source: Securities and Exchange Commission, Action Needed to Improve Rating Agencies Registration Program and
Performance-Related Disclosures (Washington: United States Government Accountability Office, 2010).
21
The only
stable Nash
equilibrium
in a Bertrand
duopoly is zero
profits for all
firms.
Sean Egan, CEO of Egan-Jones,testifying
before the Senate Committee on Banking,
Housing, and Urban Affairs, stated that the
“ratings industry is suffering from a state
that is hard to characterize as anything oth-
er than dysfunctional.”59 He contended that
the industry suffered from a lack of competi-
tion under the old NRSRO system—a “part-
ner monopoly” as he called it. Egan-Jones
had been quicker to downgrade ratings for
WorldCom, Enron, and the Ford Motor
Company than the dominant NRSRO com-
panies, yet Egan-Jones had been excluded
from the NRSRO designations while exist-
ing NRSRO CRAs retained their designa-
tions.60 Egan-Jones had applied for NRSRO
status in 1998; however, it only received the
designation in 2007—nine years later. Egan-
Jones argued the application process was
exclusive and categorically unfair. It believed
it had proven to have more accurate rat-
ings, superior risk assessment models, and
reduced conflict-of-interest problems since
the company used a subscriber-based busi-
ness model.
Dominion Bond Rating Service was
founded in 1976, but did not receive the
NRSRO designation until February 2003,
despite its application three years earlier. Do-
minion also argued that recognition from
the SEC would help it expand in the U.S.
credit rating market and put it on a level
playing field with its competitors.61 Larry
Mayewski, Chief Rating Officer at AM Best,
a CRA specializing in the insurance industry,
pointed out “there are still companies that
would like to see us with the NRSRO desig-
nation before they’ll do business with us.”62
These accounts from non-NRSRO CRAs
demonstrate the importance of the NRSRO
designation for CRAs’ ability to compete in
the marketplace. Since the NRSRO desig-
nation is a regulatory product, it could be
argued that regulatory, rather than natural
barriers to entry, structure the CRA market.
Competition and market power. Popular
accounts of the market for credit ratings
generally overlook two important interre-
lated considerations in the study of mar-
ket power. The first is that in concentrated
markets (particularly those where products
are homogeneous), fewer firms can actu-
ally reduce prices and, hence, profits. Mar-
ket dynamics characterized by two firms
with nearly identical products could, under
plausible assumptions, result in a Bertrand
outcome, where profits converge to zero be-
cause any firm could capture the entire mar-
ket by simply lowering its price below that
of its rival(s). The only stable Nash equilib-
rium63 in a Bertrand64 duopoly is zero prof-
its for all firms. The power of the Bertrand
model is not in disproving the existence of
duopoly rents, but in proving that the exis-
Table 6
Herfindahl-Hirschman Index for Nationally Recognized Statistical Rating
Organizations Based on Number of Issuers Rated, 2006–2009
Asset Class 2006 2007 2008 2009
Corporate Issuers 3069 2625 2596 2483
Financial Institutions 2773 2555 2550 2452
Insurance Companies 3353 3066 2826 2749
Issuers of Government Securities 3822 3820 3846 3889
Issuers of Asset-backed Securities 3602 3561 3553 3493
Source: Securities and Exchange Commission, Action Needed to Improve Rating Agencies Registration Program and
Performance-Related Disclosures (Washington: United States Government Accountability Office, 2010).
22
tence of such rents is dependent upon more
than just the number of sellers.65
The second issue is a basic confusion
about market power. In the standard con-
flict-of-interest narrative, rating agencies are
“corrupted” by the fact that they are paid
for the ratings by the debt issuer. Obviously
any issuer would prefer a higher rating, so
competition forces the raters to lower their
standards and sell inflated ratings. For this
outcome to hold, market power would have
to be on the side of the issuer, not the rat-
ing agency. In addition, this narrative never
explains why raters would only compete on
quality and not price. Once the raters only
sell the highest rating, then the agency sell-
ing the highest at the cheaper price should
capture the market, hence there are very
strong incentives for price competition in
the conflict-of-interest model, yet in the real
world we witness substantial profits on the
part of the largest agencies, indicating that
we are not in a Bertrand equilibrium or any-
thing like it. It is more likely that the rat-
ing agencies possess market power because
of various regulatory requirements, rather
than the market power being possessed by
the issuers.
Restricted output: quality information and
complacency in methodology. Regulatory bar-
riers to entry reduce competition in the CRA
marketplace. Without competitive threat,
the CRAs may have had reduced incentives
to innovate more sophisticated rating meth-
odologies to keep up with the changes in
structured finance. If this were the case, we
would expect NRSRO CRAs to have restrict-
ed, or the organizational structure in the
marketplace to have limited, the quantity or
quality of information in the marketplace
about credit risk relative to an open market
industry. Or, as John Coffee explains, the
“lack of competition is important [because]
. . . it permits these nominal competitors to
shirk, engaging in less effort and research
than if there were true active competition.”66
Figure 7
Credit Rating History of Current Nationally Recognized Statistical Rating Organizations
Source: Securities and Exchange Commission, Action Needed to Improve Rating Agencies Registration Program and Performance-Related Disclosures
(Washington: United States Government Accountability Office, 2010).
23
Opening up
the Nationally
Recognized
Statistical Rating
Organization
market without
addressing
inflated
demand may
not sufficiently
incentivize those
credit rating
agencies to
innovate better
methodologies
to measure credit
risk.
Also, had there been more NRSRO CRAs,
it may have increased the likelihood that
at least one of them would have innovated
methodologies better able to measure and
detect credit risk.
To be clear, NRSRO CRAs did not restrict
the output of the number of credit ratings;
on the contrary, they had strong incentives
to rate as many as possible. According to
the oligopoly model shown in Figure 5, with
lower competitive threat, NRSRO CRAs re-
stricted the quantity of information produc-
tion supplied to the market—in contrast to
what would have been supplied under a mar-
ket regime. In addition, conflicts of interest
between NRSRO CRAs and issuers, likely
maintained by inflated/captive demand,
also disincentivized NRSRO CRAs from
information production. The CRAs often
worked with issuers to structure financial
products that would obtain higher ratings
rather than focusing on instrument quality.
This incentivized NRSRO CRA methodolo-
gies that would formulate more highly rated
instruments rather than detecting increased
credit risks in more sophisticated and com-
plicated financial products.
This perception of NRSRO CRAs’ meth-
odological complacency was found in a
2002 survey by the Association for Financial
Professionals. According to the survey, 40
percent of individuals working for compa-
nies with rated debt thought rating changes
were timely, 29 percent found them to be ac-
curate, and 22 percent believed the ratings
favored the interest of investors.67
Another important issue with rating meth-
odologies is that two bonds could have the
same rating but vastly different returns. This
overlooks the risk-return tradeoff and sug-
gests that an Aa investment with a 12 percent
return is just as safe as an Aa investment with
a 6 percent return. Rating firms argue that
their ratings are “relative measures of risk”
and that’s why “the assignment of ratings in
the same categories to entities and obligations
may not fully reflect small differences in the
degree of risk.”68 However, there seemed to be
a significant difference in risk when Moody’s
Baa-rated CDOs had a default probability of
20 percent while their Baa-rated corporate
bonds had a default probability of only 2 per-
cent.69 In other words, CDOs’ debts with the
same rating were ten times as risky as similarly
rated corporate debts.70
Bo Becker and Todd Milbourn argue that
increased competition within this regulatory
framework actually reduces methodologi-
cal efficacy. In a regulatory regime in which
the existing firms enjoy captive and inflated
demand—which in turn helps bolster a lu-
crative issuer-pays business model—adding
an additional NRSRO actually may not im-
prove the quality of ratings.71 As NRSRO
CRAs seek to meet the needs of their cli-
ents, which are primarily issuers, Becker and
Milbourn find that the ratings agencies do
marginally inflate their ratings and allow rat-
ings’ quality to decline.72 This demonstrates
the importance of not merely opening up the
NRSRO market to more competition, but
also addressing the issues of captive/inflat-
ed demand, which spur conflict-of-interest
problems. Opening up the NRSRO market
without addressing inflated demand may
not sufficiently incentivize NRSRO CRAs to
innovate better methodologies to measure
credit risk.
Regulatory Regime Models
So far, we have analyzed the problems
resulting from SEC regulations creating a
de facto oligopoly of NRSRO-designated
credit rating agencies and a captive market
demand for NRSRO CRAs’ ratings. Our
findings also suggest that markets may have
been better served with a different regulato-
ry framework. We will now examine alterna-
tive regulatory scenarios including an open
access regime, a licensing regime, and a licensing
regime with captive demand.
Open Access
Not all industry-specific regulatory frame-
works are created by government. To the con-
trary, there are copious examples of industry
24
Without
preferential
regulatory
treatment for
particular
investments,
investors would
seek out credit
risk analysis and
expertise based
on reputational
factors.
self-regulation, in which companies and in-
dustries learn and standardize best practices
for their industry based on the incentives
faced. An open access regime is an industry-
specific regulatory framework not stipulated
by the state. According to Langohr and Lan-
gohr: “[N]o government or regulatory body
distorts the market outcomes of natural sup-
ply and demand conditions. These condi-
tions would not be distorted through the use
of selected agencies’ ratings in regulations
nor through licensing mechanisms.”73
Under these conditions, there would not
be NRSRO designations, and regulations
would neither require nor incentivize inves-
tors to hold certain kinds of investments.
In this scenario, investors would still face
information asymmetries between their risk
preferences and financial instruments’ risk
offerings. Since investors care about the
probability of default, as well as the value of
investment return, they would conduct their
own due diligence about investments’ risk.
To do this, they would likely seek outside
opinion about financial products and their
associated credit risks. This advice might
take the form of risk categories calculated
by credit rating firms, or it might include a
conglomerate of market-based measures for
credit risk, possibly including credit spreads.
Since investors would be the primary de-
mander of credit risk information, finan-
cial innovation would likely search for new
methodologies and technologies to meet in-
vestor demand. Without preferential regula-
tory treatment for particular investments,
investors would seek out credit risk analysis
and expertise based on reputational factors,
rather than exogenous designations like the
NRSRO. Most likely, investors would trust
the credit risk analysis of those with incen-
tives independent of those selling the invest-
ments (investor-pays, not issuer-pays). This
would devolve the responsibility for due dili-
gence to individual and group investors.
If those providing credit risk analysis de-
cided they could increase profits by selling
ratings directly to issuers, they would need
to somehow prove to investors that con-
flicts of interest would not dominate their
ratings. If credit risk analysts cheated by
giving in to conflicts of interest, if investors
became uncomfortable with the clear con-
flict-of-interest problems, or if investors lost
confidence in the analyses’ informational
value, investor demand for issuer-paid credit
risk analysis would decrease. Consequently,
issuers would have less reason to purchase
analysis from credit risk analysts.
As issuer demand declined, credit risk
analysts would either have to change their
business model to issuer-pays or lose market
share to competitors. Moreover, competitor
analysts would have incentive to offer a busi-
ness model not entrenched with conflicted
interests to capture investor demand. But
competitor analysts might also have an in-
centive to develop new methodologies to de-
termine credit risk. Investors would have an
incentive to shop around for more valuable
sources of credit risk information.
It might be that reputational and network
effects create an environment such that only
a few credit analysts emerge. Or it might be
that investor-pays/subscriber-based models
are not extraordinarily profitable, and thus
only a few credit analysts are attracted to the
market. If this were the case, the competitive
threat would still remain to provide an incen-
tive for credit analysts to maintain quality
risk assessment models and innovate new
methodologies to keep up with financial in-
novations. Credit risk firms might even offer
greater transparency to persuade investors to
purchase their risk analysis.
The overall effect of an open regime on
the market would be little to no cheating in
analysis, few conflicts of interests, and in-
centives for analysts to innovate to increase
the informational value of their credit analy-
sis. The result would be more accurate credit
risk analyses. With increased accuracy in
credit risk analysis, issuers would have an
increased incentive to offer higher quality
investments. Investors’ primary incentive
would be to create a portfolio representative
of their risk preferences, rather than to ob-
tain preferential regulatory treatment.
25
It might be that
reputational
and network
effects create an
environment in
which only a few
credit analysts
emerge.
Licensing Regime
Under a licensing regime, the state would
stipulate that credit risk analysis be used to
either require or incentivize investors to pur-
chase high quality financial instruments.
This assumes that the state knows how to
define high quality credit risk analysis, and
also assumes that credit risk analysts pro-
duce accurate analyses. In order to comply
with the state or obtain benefits from the
state, investors would seek to purchase high-
ly rated investments.
Suppose, at first, that credit risk analysts
sell their analysis to investors. In contrast
to the open market regime, where investors
care most about risk analysis quality, inves-
tors under a licensing regime would care
about two things: risk analysis quality and
obtaining a high rating. Investors would
benefit from obtaining accurate ratings be-
cause they would be more likely to achieve a
portfolio they desire, and by obtaining high
ratings they would be eligible for state-be-
stowed benefits.
Now suppose credit risk analysts deter-
mined they could increase profits by sell-
ing ratings directly to issuers. They would
need to prove to investors that conflicts of
interest would not dominate their ratings.
If credit risk analysts cheated by giving in
to conflicts of interest, if investors became
uncomfortable with the clear conflict-of-
interest problems, or if investors lost confi-
dence in the analyses’ informational value,
demand for issuer-paid credit risk analysis
would not necessarily decrease, since inves-
tors value both high ratings and risk analy-
sis quality. If credit risk analysts cheated by
inflating ratings sold to issuers, it would
decrease their reputational value among in-
vestors, but the inflated ratings would still
make their investments eligible for state-
bestowed benefits. Thus, the net impact of
cheating on demand would depend on the
negative impact of cheating compared with
the value of state-bestowed benefits. If inves-
tors became uncomfortable with the clear
conflict-of-interest problems, the net im-
pact on demand would also depend on neg-
ative impact of concerns compared with the
value of state-bestowed benefits. If investors
lost confidence in the analyses’ informa-
tional value, then the net impact on demand
would also depend on the cost of unhelpful
ratings or the extent to which the ratings are
uninformative compared with the value of
state-bestowed benefits.
As long as credit risk analysts could keep
the costs of cheating, conflicts of interest,
and low informational value below the ben-
efit of investors obtaining state-bestowed
benefits, then this business model would
likely continue.
However, if the costs of cheating, con-
flicts of interest, or low informational value
exceeded the value of state-bestowed bene-
fits, then investors would not value analysts’
credit risk analyses. In turn, issuers would
have less incentive to purchase analysis on
behalf of investors. Then credit risk firms
would either have to reduce cheating and
conflicts of interests, improve their analyses’
informational value, or begin selling analy-
ses directly to investors. If not, they could
lose market share to a competitor.
It might be that reputational and network
effects create an environment in which only
a few credit analysts emerge. Or it might be
that investor-pays/subscriber-based models
are not extraordinarily profitable, and thus
only few credit analysts are attracted to the
market. If this were the case, then some com-
petitive threat would still exist and could
still provide an incentive to maintain qual-
ity risk assessment models, but it would also
depend on the relative costs and benefits of
obtaining a high quality rating to obtain
state-bestowed benefits.
Some competition for investor business
would remain. Since, in some cases, inves-
tors would demand higher quality products
and products would be rated more accurate-
ly, issuers would have some increased incen-
tive to offer higher quality products.
The overall effect on the marketplace
would be some cheating in ratings, with
conflicts of interests and less informational
value associated with credit analysis com-
26
Investors would
benefit from
obtaining
accurate ratings
because they
would be more
likely to achieve
a portfolio they
desire.
pared with an open market regime. However,
competitive threat would reduce incentives
to cheat and increase incentives to improve
informational value. Credit risk analysis
would not likely be as accurate as that under
an open market regime, but it would tend
towards accuracy. Issuers would have less
incentive to offer high quality ratings than
they would under an open market regime,
but they would tend towards offering a high-
er quality analysis.
Designated Licensing Regime
Under a designated licensing regime, the
state would stipulate that credit risk analy-
sis be used to either require or incentivize
investors to purchase high quality financial
instruments from issuers. In addition, the
state would also stipulate, or “designate,”
whose credit risk analyses would be eligible
to be used to meet requirements or incentives
when purchasing financial instruments. This
assumes that the state knows how to define
high quality risk analysis and also assumes
that credit risk analysts produce accurate
analyses. It further assumes that the state
knows whose credit risk analysis is the best
and thus should be eligible to meet state re-
quirements or incentives. In addition, it also
assumes that the firms’ whose credit analysis
is best today will continue to be the best in
the future.
In order to comply with the state or ob-
tain benefits from the state, investors would
seek to purchase highly rated investments
from designated firms. Under this regula-
tory framework, investors want both credit
rating analyses to be accurate so that they
obtain the portfolios they desire, but they
also want to obtain high ratings from desig-
nated credit analysts.
Suppose, at first, credit risk analysts sell
their analyses to investors. In contrast to
the open market regime where investors
care most about risk analysis quality, and in
contrast to a licensing regime where inves-
tors care most about risk analysis quality
and obtaining high ratings, investors under
a designated licensing regime care about
three things: risk analysis quality, obtaining
high ratings, and obtaining high ratings
only from designated firms. Investors would
benefit from obtaining accurate ratings be-
cause they would be more likely to achieve
a portfolio they desire, and by obtaining
high ratings only from designated credit risk
analysts, they would be eligible for state-be-
stowed benefits.
Now suppose credit risk analysts deter-
mined they could increase profits by selling
ratings directly to issuers. They would need
to prove to investors that conflicts of inter-
ests with issuers, especially among analysts
with reduced competition, would not impact
their ratings. If credit risk analysts cheated
by giving into conflicts of interest, if inves-
tors became uncomfortable with the clear
conflict-of-interest problems, or if investors
lost confidence in the analyses’ information-
al value, demand for issuer-paid credit risk
analysis would not necessarily decrease, since
investors value both high ratings from desig-
nated firms and risk analysis quality. Under
this regime, credit risk analysts have fewer
competitors because the state only designat-
ed a few of the analysts. If credit risk analysts
cheated by inflating ratings sold to issuers,
it would decrease their reputational value
among investors, but the designated inflated
ratings would still make their investments
eligible for state-bestowed benefits. Analysts
would also be able to get away with more
cheating because of less threat of competi-
tion in the marketplace. Thus, the net impact
of cheating on demand would depend on the
negative impact of cheating compared with
the value of state-bestowed benefits. If inves-
tors became uncomfortable with the clear
conflict-of-interest problems, the net impact
on demand would also depend on the nega-
tive impact of concerns compared with the
value of state-bestowed benefits. If investors
lost confidence in the analyses’ informa-
tional value, then the net impact on demand
would also depend on the cost of unhelpful
ratings or the extent to which the ratings are
uninformative compared with the value of
state-bestowed benefits.
27
Holding credit
rating agencies
liable for ratings
does not address
the problems of
inflated demand
or reduced
competition.
As long as credit risk analysts could keep
the costs of cheating, conflicts of interest,
and low informational value below the ben-
efits of investors’ obtaining state-bestowed
benefits, then this business model would
likely continue.
However, if the costs of cheating, con-
flicts of interest, or low informational val-
ue in any combination exceeded the value
of state-bestowed benefits, then investors
would not value credit risk analysts’ analy-
ses. In turn, issuers would have less incentive
to purchase analyses on behalf of investors.
Then credit risk firms would either have to
reduce cheating and conflicts of interest, im-
prove their analyses’ informational value, or
begin selling analyses directly to investors.
If not, they could lose their market share
to a competitor. However, since the market
is much less competitive, they retain some
market power over how much cheating and
conflicts of interest will be tolerated, as well
as over informational value of ratings. Credit
risk analysts would still have some incentive
to increase the quality of ratings through re-
duced cheating and reduced conflicts of in-
terest, or to improve their methodologies to
provide value-added to investors. However,
this incentive would be mitigated because
there would be less threat of competition in
the marketplace.
If the analysts still could not bolster de-
mand, then they might have to change their
business model to an investor-pays model
or lose their market share to competitors.
However, with reduced competition in the
marketplace, they may be able to continue
the issuer-pays business model.
Although reputational and network ef-
fects may contribute to a smaller number of
credit risk analysts, the special designations
given to particular credit risk analysts also
significantly contribute to the concentrat-
ed market and reduced competition. Some
competitive threat would remain in the mar-
ketplace and provide incentive for credit risk
analysts to maintain quality risk assessment
models, but it would also depend on the
degree of market power the designated ana-
lysts maintained, as well as the relative costs
and benefits of purchasing high ratings to
obtain state-bestowed benefits from desig-
nated firms.
Since credit risk analysts have greater mar-
ket power than in the previous models, issu-
ers will recognize that there is less competi-
tive threat, with less incentive for the credit
risk analysts to rate financial instruments as
accurately as possible. As such, there may be
less reason to only offer the highest quality
products.
The overall effect of designated credit
analysts on the marketplace would be more
cheating in ratings, greater conflicts of in-
terest, limited competition, and less incen-
tive to improve informational value than in
a licensing or open regime. Credit risk analy-
sis would not be as accurate as it would be
under a licensed or open market regime. In
turn, issuers would have less incentive to of-
fer high quality products.
Reforms Not To Pursue
Some have argued that problems relat-
ed to credit rating agencies result from the
SEC passing off its responsibility to private
companies. But requiring the SEC to con-
duct valuations of credit risk would not deal
with the problem of inflated and captive de-
mand. Moreover, it is unlikely that the SEC
is equipped or could ever become equipped,
without major market stifling, to handle
credit risk valuation on its own. For exam-
ple, a 2002 Senate study found that SEC of-
ficials had managed to review only 16 per-
cent of the 15,000 annual company reports
submitted in the previous fiscal year, and
they had not reviewed Enron in a decade.74
Some contend that holding CRAs liable
for their ratings would force them not to
cheat and to innovate better ratings method-
ologies. First, holding CRAs liable for ratings
does not address the problems of inflated
demand or reduced competition. Second,
CRAs argue they are a “part of the media”
since they are financial publishers.75 For this
28
Another risk
from subjecting
rating agencies to
liability is that, in
order to protect
themselves, the
agencies would
utilize “consensus
forecasts” of
key economic
variables. Yet the
consensus could
be dangerously
off.
reason, they argue that the First Amendment
applies to them and protects their speech.
Third, it remains true that they do not have
a federal mandate, as NRSRO CRA analysts
contended at a Senate panel in the wake of
Enron’s collapse: “S&P operates with no gov-
ernmental mandate, subpoena power, or any
other official authority. It simply has a right,
as part of the media, to express its opinion
in the form of letters and symbols.”76 More-
over, since actual credit risk is nearly impos-
sible to know and CRAs are limited by issu-
ers’ financial disclosure, it would be difficult
to prove in court that CRAs purposefully in-
flated credit ratings.
Another risk from subjecting rating agen-
cies to liability for either their statements or
processes is that, in order to protect them-
selves, the agencies would adopt a “reason-
able man” approach. For instance, if the
agencies used government forecasts of house
prices in their mortgage default models,
then it is likely that any court would deem
such assumptions “reasonable”; after all,
these are the assumptions that regulators
rely upon. If such assumptions are, however,
grossly in error, as were the housing price
forecasts used by various federal agencies,
then the value of information created by the
rating agencies would also be reduced, if not
compromised. A reasonable-man approach
would also encourage rating agencies to uti-
lize “consensus forecasts” of key economic
variables. Yet the consensus could be danger-
ously off. The economic forecasting profes-
sion does not exactly have a great record at
predicting turning points, and it also missed
the decline in house prices. A system of liabil-
ity would likely destroy whatever additional
information the rating agencies bring to the
market, as the agencies would face tremen-
dous pressure to simply mimic widely held
beliefs, which themselves would already be
priced into the market.
Others have advocated restricting the
practice of “notching.” CRAs engage in puni-
tive notching if they compel issuers to pur-
chase more products by threatening to notch
down other rated financial instruments.
When or if notching does occur, it is anti-
competitive; however, prohibiting notching
also ignores the problem of inflated demand
and reduced competition. Moreover, CRAs
would have less ability to compel issuers to
buy their products if they had less oligopo-
listic market power.77
Another proposal has been to ban finan-
cial instruments that are determined to be
too complex for adequate risk assessment.
This also neglects the core problems of in-
flated demand and the de facto NRSRO
CRA oligopoly. It further assumes that fi-
nancial regulators and lawmakers have the
knowledge necessary to determine which
instruments are “too complex.”78 Finally, it
overlooks the significant opportunity costs
associated with preventing innovation in fi-
nancial markets.
Recent Development in CRA Regulatory
Reform
In the wake of Enron, the 2002 Sarbanes-
Oxley Act required the SEC to reexamine the
“role and function of rating agencies in the
operation of the securities market” and to
specifically address potential barriers to en-
try.79 In 2003, the SEC sought comments on
whether NRSRO credit ratings should con-
tinue to be used for regulatory purposes. Ac-
cording to the SEC, most of the 46 comments
responding to the 2003 Concept Release sup-
ported continuing the NRSRO designation
and expressed concern that “eliminating the
NRSRO concept would be disruptive to capi-
tal markets.”80 However, it is worth noting
the SEC’s primary citation for those in favor
of keeping the NRSRO designation was a
“Letter from Leo C. O’Neill, President, Stan-
dard & Poor’s, to Jonathan G. Katz, Secretary,
Commission (July 28, 2003).”81 The SEC
mentioned that only four commenters sup-
ported elimination of the concept, and cited
professors Frank Partnoy of the University
of San Diego School of Law and Lawrence J.
White of NYU’s Stern School of Business.82
From these comments, the SEC attempted
to clarify the process of identifying NRSROs
with a proposed definition:83
29
Barriers to
entry alone
were not solely
responsible for
problems in
market structure.
i. issues publicly available credit ratings
that are current assessments of the
creditworthiness of obligors with re-
spect to specific securities or money
market instruments;
ii. is generally accepted in the financial
markets as an issuer of credible and
reliable ratings, including ratings for a
particular industry or geographic seg-
ment, by the predominant users of se-
curities ratings; and
iii. uses systematic procedures designed
to ensure credible and reliable ratings,
manage potential conflicts of interest,
and prevent the misuse of nonpublic
information, and has sufficient finan-
cial resources to ensure compliance
with those procedures.
The non-NRSRO CRA community ex-
pressed concern that these proposed reforms
would, in fact, strengthen incumbent power
in the market rather than reduce barriers
to entry. First, the proposed rules would re-
quire CRAs to provide public credit ratings,
although this would be essentially offering
their products free of charge for subscrib-
er-based CRAs. The rule clearly catered to
the firms that used an issuer-pays business
model rather than a subscriber-pays model.
Second, the requirement for CRAs to be
“generally accepted” created something like
a chicken-and-egg problem for new firms.
As Banking, Housing, and Urban Affairs
Committee Chairman Sen. Richard Shelby
put it, “to receive the license a firm must be
nationally recognized, but it cannot become
nationally recognized without first having
the license.”84 As Lawrence White argues, in
order to be “generally accepted” it would re-
quire ratings be linked “to the views of the
predominant users of securities ratings.”85
Ultimately, the SEC did not move forward
with these proposals.
In 2005, the House considered legislation
to reduce barriers to entry in the CRA mar-
ket, and on September 29, 2006, President
Bush signed the Credit Rating Agency Act
of 2006. A primary result of the legislation
was to reduce arbitrary SEC power to desig-
nate NRSROs and instead set timelines for
SEC response. Under the 2006 law, any credit
rating firm issuing ratings for at least three
years could apply to the SEC to receive the
NRSRO designation. The SEC would need
to render a decision or set a timeline for eval-
uation within 90 days, and make a final deci-
sion within 120 days. The rules also aimed to
avoid bolstering a particular business model,
whether subscriber-pays or issuer-pays. The
law ensured that neither the SEC nor the
state could regulate credit ratings’ content,
procedures, or methodologies, and prohib-
ited NRSROs from allowing conflicts of in-
terest to impact rating integrity or “condi-
tioning ratings . . . on an issuer’s purchasing
other services from the NRSRO.”86
Despite these attempts to reduce barriers
to entry in the CRA marketplace, Langohr
and Langohr argue that this merely “re-
placed one form of controlled access to [the
CRA market] with another” by replacing the
opaque NRSRO subjective recognition sys-
tem with an objective recognition system.87
They argue that these “objective” measures
still used criteria that systematically favor
incumbent firms while continuing to disad-
vantage competitors. Still, there is evidence
that some ratings markets did experience
reduced concentration, as demonstrated in
Table 4.
Barriers to entry alone were not solely re-
sponsible for problems in market structure.
Instead, regulatory dependence on NRSRO
ratings also led to distorted incentives and
outcomes. As a result, policymakers have
considered reducing the role of NRSRO rat-
ings in regulation.
The Dodd-Frank Act
While each financial crisis seems to have
a cycle of complaints about failures among
the rating agencies, previous legislative re-
sponses, such as the Sarbanes-Oxley Act,
have relied mostly on further study rather
than wholesale reform of the ratings pro-
cess. The Dodd-Frank Act attempts to ad-
dress the quality of ratings via a variety of
30
The primary
focus of Dodd-
Frank’s changes
to the regulation
of rating agencies
is in attempting
to “insulate” the
agencies from
various perceived
conflicts of
interest.
mechanisms. The most extensive of these
are found in Section 932.
The primary focus of Dodd-Frank’s
changes to the regulation of rating agencies
is in attempting to “insulate” the agencies
from various perceived conflicts of interest.
For instance, Dodd-Frank requires improved
“internal controls” for the ratings process,
separating the sales and marketing functions
of the agencies from the ratings process, in-
creasing the number of independent direc-
tors on the agencies’ boards of directors, and
increasing the responsibilities of the ratings
agencies’ boards. Many of these features
mirror the expanded corporate governance
requirements for auditors imposed by the
Sarbanes-Oxley Act. This should not be too
surprising, as it was the same congressional
staffers who drafted the similar sections in
both acts. What is surprising is the expecta-
tion that such provisions would work any
better in improving credit ratings than they
did, or failed to do, in improving the quality
of financial audits.
The quest for board independence is a
repeated theme in corporate governance re-
forms. As mentioned, the Sarbanes-Oxley
Act increased the number of independent
board members for auditors, with similar
provisions covering rating agencies in the
Dodd-Frank Act. These repeated attempts at
independence, however, find little support in
the academic literature. In the case of banks
during the financial crisis, some research-
ers find that greater board independence is
actually associated with worse outcomes.88
Dodd-Frank further muddies the waters by
allowing some of the “independent” board
members to be users of ratings. This ignores
the fact that investors in rated securities have
their own incentives to avoid downgrades.
Instead of reducing conflicts of interests,
Dodd-Frank may very well simply be substi-
tuting one conflict of interest for another.
One of the Dodd-Frank rating agency re-
forms has already had tremendous negative
impact on our capital markets—so much so
that the SEC has effectively voided the pro-
vision. This Section, 939G, repeals SEC rule
436(g), which had exempted NRSROs from
being deemed part of a security’s registra-
tion statement for the purposes of securities
fraud. Rule 436(g) had protected NRSROs
from liability under Section 11 of the 1933
Securities Act. This protection actually in-
creased the flow and quality of information
received by investors by encouraging the
use of ratings in offering statements. Dodd-
Frank’s repeal of Rule 436(g) effectively shut
down the new offerings market for asset-
backed securities and corporate debt. It was
only the issuance of a “no-action” letter from
the SEC to Ford Motor Credit Company that
allowed this market to function. However,
this no-action letter is temporarily in effect,
leaving considerable uncertainty as to how
our debt markets will function in the ab-
sence of Rule 436(g), at least until such time
the markets evolve beyond the regular use of
credit ratings.
The Dodd-Frank Act, like the Sarbanes-
Oxley Act before it, attempts to remedy
regulatory failures with the increased use of
private litigation. Section 933 expands the
potential legal liability of rating agencies in
three ways. First, it established a private right
of action under Section 18 of the 1934 Se-
curities Act for any material misstatements
contained in reports to the SEC. Second, it
established liability for errors in factual as-
sumptions used in a ratings methodology.
An example would be the range of forecast-
ed house prices over the life of a mortgage-
backed security. Third, and last, there is es-
tablished legal liability under Section 21E
of the 1934 Securities Act for misstatements
in any forward-looking statements made by
the rating agencies. Of course, one defense to
these charges would be to adopt a reasonable-
man approach to ratings methodology and
predictions. Basing ratings on consensus, or
even government forecasts of key economic
variables, would likely provide some shield
to liability. Providing a consensus viewpoint
could, however, greatly reduce the informa-
tional value provided by ratings. Increased
liability could easily make rating agencies
risk-averse and less likely to offer unconven-
31
Overall, the
Dodd-Frank Act
is a mixed bag
when it comes to
the credit rating
agencies.
tional points of view. Agencies could also be
subject to suit by investors “harmed” by the
downgrade of assets which they hold. Un-
til these provisions are tested in the courts,
their ultimate impact on ratings’ quality will
remain unknown. It is likely, however, that
the increased incentives for risk aversion will
greatly reduce the value of ratings to our
capital markets, with potential harm to both
price discovery and liquidity.
Other provisions of Dodd-Frank are also
likely to reduce the utility of rating agencies,
with detrimental impacts on our capital
markets. For instance, Section 939B elimi-
nates the rating agencies exemption from
Regulation FD, which covers the “fair dis-
closure” of information. Regulation FD pro-
hibits senior executives of public companies
who regularly communicate with the public
from making selective disclosure of non-
public informational material to select per-
sons. Prior to Dodd-Frank, the rating agen-
cies were exempted, with the understanding
that the ratings process would be better in-
formed if the rating agencies had occasional
access to nonpublic information. Section
939B has the potential to reduce the flow of
information between public companies and
the rating agencies, with the result that rat-
ings become less informed.
Not all of the credit rating provisions
of Dodd-Frank are harmful or misguided.
In fact, the law takes a serious step toward
reducing the regulatory reliance on the rat-
ing agencies. Section 939A of Dodd-Frank
requires all federal agencies to review their
existing regulations and to provide alter-
native standards of credit risk. Although
the federal bank regulators have, as of the
publication of this paper, requested public
comments as to possible alternatives, these
same regulators have moved slowly on Sec-
tion 939A and have shown a general resis-
tance to abandoning their reliance on the
rating agencies. While Section 939A has
the potential to address some of the central
flaws discussed in this paper, it also leaves
considerable discretion to the very same
regulators who instituted those flaws. For
Section 939A to have real impact, however,
it may well take the continued involvement
of Congress.
Overall, the Dodd-Frank Act is a mixed
bag when it comes to the credit rating agen-
cies. Some provisions have a real potential for
reform, but their success is also contingent
on the same regulatory process that created
the problems. Unfortunately other more
concrete provisions of Dodd-Frank have al-
ready had a significant negative impact on
our capital markets. A repeal of these latter
provisions, particularly Sections 932, 933,
939B, and 939G, would protect the positive
capital market functions of the rating agen-
cies. Section 939A, which attempts to reduce
regulatory reliance on the rating agencies,
should be retained, but may be in need of
strengthening.
Proposal for Reform
The open-market regime provides results
closest to the regulatory ideal and best serves
the public good. It calibrates incentives in
such a way that issuers, investors, and credit
rating agencies have an incentive to promote
the public good while seeking their own self-
interest. The benefits likely achieved in the
open-market regime include higher quality
investment instruments, higher quality rat-
ings, increased methodological innovation,
and investor focus on informational value
rather than regulatory privilege.
To achieve open-market regime benefits
would require ending regulatory reliance on
credit rating agencies. It would also require
repealing the Nationally Recognized Sta-
tistical Rating Organization (NRSRO) des-
ignation and allowing competition in the
marketplace of credit rating firms. Market
participants may still use credit ratings to
evaluate credit risk, but they should also be
free to innovate their credit-risk evaluations.
Ultimately, these reforms will help reduce
CRA oligopolistic power and reduce artifi-
cial demand for credit ratings.
To the extent that this is not politically
feasible, or that policymakers worry that cap-
ital controls are required to prevent system-
32
One contributor
to the financial
crisis was a
significant
reduction in due
diligence on the
part of investors,
regulators,
financial
institutions,
borrowers, and
politicians.
atic bank runs, regulators should strongly
consider eliminating references to NRSRO
credit ratings as often as possible. Moreover,
when ratings are required, they could be
used to calculate net capital at the time of
purchase, rather than over time. Also, regula-
tors may consider innovations in credit risk
analysis, or market-based measures such as
credit spreads or market implied ratings, in
evaluating credit risk. In addition, many of
these measures of credit risk can be used by
investors themselves, promoting investor
due diligence in investment decisions.
Conclusion
A variety of factors contributed to the
worldwide financial crisis of 2008. One of
those was the mistaken belief that risky as-
sets, such as mortgage-backed securities and
sovereign debt, were actually risk-free. This
perception facilitated both the massive lev-
els of leverage and excessive degrees of asset
concentration witnessed within our finan-
cial system. In all likelihood, such leverage
and asset concentration would not have oc-
curred had the assets in question not been
blessed by the CRAs, or had financial regu-
lators not embedded the use of ratings into
the fabric of prudential regulation.
A lack of competition, in part the result
of regulatory barriers, along with mandated
usage by many financial market participants
has resulted in a dysfunctional credit ratings
industry. Entrenched market power has led
to the predictable result that ratings agen-
cies would reduce the quality of their servic-
es. Unfortunately, their services were signifi-
cant components in our financial regulatory
system. This reduction in ratings quality
also resulted in a reduction in the efficacy of
financial regulation.
Increased competition alone, however,
will be insufficient to address the failings in
the ratings market. In order to move toward
a functioning, competitive ratings market,
the users of ratings, particularly investors,
must be free to reject the use of ratings. In
fact, increased competition, coupled with a
continued “gate-keeping” role for CRAs, is
just as likely to increase financial fragility as
reduce it. To the extent that rating agencies
are performing police power functions of
the state, those functions should be trans-
ferred back to the state.
One contributor to the financial crisis
was a significant reduction in due diligence
on the part of investors, regulators, financial
institutions, borrowers, and politicians. To
some extent this reduction was facilitated
by a belief that rating agencies could pro-
vide such due diligence on behalf of other
market participants. Reducing the central
role of rating agencies in financial regula-
tion would undoubtedly increase the due
diligence cost of other market participants.
It would, however, greatly increase the qual-
ity and quantity of monitoring of financial
risk by market participants.
Ultimately, both taxpayers and financial
market participants would be better served by
rating agencies that were subject to competi-
tive market pressures. Such pressures would
most effectively be brought to bear by a re-
duction in regulatory barriers to entry and
the removal of artificial demand due to vari-
ous compliance requirements placed upon
other market participants, such as banks, in-
surance companies, and mutual funds.
Notes
1. For background information, please see
Securities and Exchange Commission, “Com-
ments on Proposed Rule: Rating Agencies and
the Use of Credit Ratings under the Federal Secu-
rities Laws,” http://www.sec.gov/rules/concept/
s71203.shtml.
2. Carol Ann Frost, “Credit Rating Agencies in
Capital Markets: A Review of Research Evidence
on Selected Criticisms of the Agencies,” Journal of
Accounting, Auditing, and Finance 22, no. 2 (2007):
25.
3. Miles Livingston, Diana Wei, and Lei Zhou,
“Moody’s and S&P Ratings: Are They Equivalent?
Conservative Ratings and Split Bond Yields,”
Journal of Money, Credit and Banking 42 (2010):
1267–94.
33
4. Kent Baker and Sattar Mansi, “Assessing
Credit Rating Agencies by Bond Issuers and In-
stitutional Investors,” Journal of Business Finance &
Accounting 29 (2002): 1367–98.
5. Herwig Langohr and Patricia Langohr, The
Rating Agencies and their Credit Ratings: What They
Are, How They Work, And Why They Are Relevant
(West Sussex, UK: John Wiley and Sons, Ltd.,
2008).
6. Richard Cantor and Frank Packer, “Multiple
Ratings and Credit Standards: Differences of
Opinion in the Credit Rating Industry,” Federal
Reserve Bank of New York Staff Report (New York:
Federal Reserve Bank,1996); Patrick Van Roy,
“Credit Ratings and the Standardized Approach
to Credit Risk in Basel II,” European Central Bank
Working Paper no. 517 (July 25, 2005): 1–55.
7. Langohr and Langohr, The Rating Agencies
and their Credit Ratings.
8. Lawrence J. White, “The Credit Rating In-
dustry: An Industrial Organization Analysis,” in
Rating Agencies in the Global Financial System (New
York: NYU Stern School of Business, 2001).
9. Langohr and Langohr, The Rating Agencies
and their Credit Ratings.
10. Federal Reserve, Federal Reserve Board Flow of
Funds (Washington: Federal Reserve Statistical Re-
lease, 2011).
11. Ibid.
12. P. E. Apgar, D. Arthur, and L. Monaco,
Moody’s, Equity Research Report (New York: Mor-
gan Stanley, 2005), pp. 1–16; B. Crockett, J. Jowe,
and F. Searby, Moody’s: Attractive Business, but
Growth Hiatus in 2004 (New York: North Ameri-
can Equity Rearch: JPMorgan, 2003), pp. 1–44.
13. Lawrence J. White, “The Credit Rating In-
dustry: An Industrial Organization Analysis” and
“Good Intentions Gone Awry: A Policy Analysis
of the SEC’s Regulation of the Bond Rating In-
dustry,” in Public Law & Legal Theory Research Pa-
per Series (New York: New York University School
of Law, 2005): 13–16.
14. White, “The Credit Rating Industry: An In-
dustrial Organization Analysis.”
15. For a concise discussion, see Lawrence J.
White, “A New Law for the Bond Rating Indus-
try,” Regulation 30, no. 1 (2007): 48–52.
16. Milton Friedman and Anna J. Schwartz, A
Monetary History of the United States 1867–1960
(Princeton, NJ: Princeton University Press, 1963).
17. Ibid.
18. Langohr and Langohr, The Rating Agencies
and their Credit Ratings.
19. Wyatt Wells, “Certificates and Computers:
The Remaking of Wall Street, 1967 to 1971,” Busi-
ness History Review 74, no. 2 (2000): 193–235.
20. Securities and Exchange Commission, Secu-
rities Exchange Act Rel. No. 34-10,525 (Washington:
SEC News Digest, 1973).
21. Michael P. Jamroz, “The Net Capital Rule,”
Business Lawyer 47 (1991–1992): 863–912.
22. Langohr and Langohr, The Rating Agencies
and their Credit Ratings.
23. Ibid.
24. Ibid.
25. For more information, see George Stigler,
“The Theory of Economic Regulation,” Bell Jour-
nal of Economics and Management Science (1971):
3–21. Stigler argues that one of the four primary
reasons industries lobby the state is to control en-
try into the market so they can reduce the compe-
tition and increase their profits.
26. William H. Beaver, Catherine Shakespeare,
and Mark T. Soliman, “Differential Properties in
the Ratings of Certified vs. Non-Certified Bond
Rating Agencies,” Working Paper (2006), http://
ssrn.com/abstract=59662.
27. Frank Partnoy, “The Siskel and Ebert of
Financial Markets? Two Thumbs Down for the
Credit Rating Agencies,” Washington University Law
Quarterly 77 (1999): 681–90; and John C. Coffee
Jr., Gatekeepers: The Professions and Corporate Gover-
nance (London: Oxford University Press, 2006).
28. Coffee, Gatekeepers.
29. George Stigler, “The Theory of Economic
Regulation.”
30. George Akerlof, “The Market for “Lemons”:
Quality Uncertainty and the Market Mecha-
nism,” Quarterly Journal of Economics 84 (1970):
488–500.
31. Partnoy, “The Siskel and Ebert of Financial
Markets?”; Claire A. Hill, “Why Did Anyone Lis-
ten to the Rating Agencies after Enron?” Journal of
Business and Technology Law 4 (2009): 283–94.
32. Frank Partnoy, “How and Why Credit Rat-
ing Agencies Are Not Like Other Gatekeepers,”
Legal Studies Research Paper Series (2006): 89–94.
34
33. Frank Partnoy, “The Siskel and Ebert of Fi-
nancial Markets?”
34. David Zigas, “Why the Ratings Agencies Get
Low Marks on the Street,” Business Week, March
12, 1990, p. 104.
35. James Van Horne, Financial Market Rates
and Flows (Upper Saddle River, NJ: Prentice Hall,
2000).
36. Richard Cantor and Frank Packer, “Determi-
nants and Impacts of Sovereign Credit Ratings,”
Economic Policy Review 2, no. 2 (1996): 45–46.
37. Hill, “Why Did Anyone Listen to the Rating
Agencies after Enron?”
38. Zigas, “Why the Ratings Agencies Get Low
Marks on the Street.”
39. Association for Financial Professionals, Rat-
ings Agencies Survey: Accuracy, Timeliness, and Regu-
lation (Bethesda, MD: Assocation for Financial
Professionals, 2002).
40. P. Jenkins, “DBRS to Challenge Big Agen-
cies,” Financial Times (London), January 10, 2006.
41. Lawrence J. White, “The Credit Rating Indus-
try: An Industrial Organization Analysis”; Coffee,
Gatekeepers.
42. Some argue that reputational factors mat-
ter more than conflicts of interest. See Daniel M.
Covitz and Paul Harrison, “Testing Conflicts of
Interest at Bond Ratings Agencies with Market
Anticipation: Evidence That Reputation Incen-
tives Dominate,” FEDS Working Papers (Washing-
ton: Board of Governors of the Federal Reserve
System, 2003): 1–23. However, we find this argu-
ment, operationalization, and empirical assess-
ment less than compelling.
43. Ibid.
44. Han Xia, “The Issuer-Pay Rating Model and
Rating Inflation: Evidence from Corporate Cred-
it Ratings,” Working Paper (2010): 56.
45. Covitz and Harrison, “Testing Conflicts of
Interest at Bond Ratings Agencies with Market
Anticipation.”
46. Langohr and Langohr, The Rating Agencies
and their Credit Ratings.
47. Beaver et al., “Differential Properities in the
Ratings of Certified vs. Non-Certified Bond Rat-
ing Agencies.”
48. John M. Griffin and Dragon Yongjun Tang,
“Did Subjectivity Play a Role in CDO Credit Rat-
ings?” McCombs Research Paper Series (2010): 13–
15; Gailen Hite and Arthur Warga, “The Effect
of Bond-Rating Changes on Bond Price Perfor-
mance,” Financial Analysts Journal 53 (1997): 35–
51; Frank Partnoy, “The Paradox of Credit Rat-
ings,” in Law and Economics (2001): 1–23, http://
www.riskcenter.com.tr/kredirisk/kredifiles/YM/
the_paradox_of_credit_ratings.pdf.
49. Han Xia, “The Issuer-Pay Rating Model and
Rating Inflation: Evidence from Corporate Cred-
it Ratings,” Working Paper (2010): 56.
50. White, “The Credit Rating Industry: An In-
dustrial Organization Analysis” and “Good In-
tentions Gone Awry.”
51. Langohr and Langohr, The Rating Agencies
and their Credit Ratings.
52. Hill, “Why Did Anyone Listen to the Rating
Agencies after Enron?”
53. Permanent Subcommittee on Investiga-
tions, “Wall Street and the Financial Crisis: Anat-
omy of a Financial Collapse,” eds. Carl Levin and
Tom Coburn (Washington: United States Sen-
ate, 2011).
54. Langohr and Langohr, The Rating Agencies
and their Credit Ratings.
55. White, “The Credit Rating Industry: An In-
dustrial Organization Analysis.” White contends
that the CRA industry is also concentrated as a
result of regulatory restrictions by the SEC deter-
mining who can be an NRSRO.
56. Langohr and Langohr, The Rating Agencies
and their Credit Ratings.
57. Sean J. Egan, “Statement of Egan Jones on
Credit Ratings Agencies,” letter submitted to the
SEC (2002).
58. Langohr and Langohr, The Rating Agencies
and their Credit Ratings.
59. Sean Egan, Managing Director, Egan-Jones
Rating Company, “The Role of Credit Rating
Agencies in the Capital Markets,” Testimony be-
fore the Senate Committee on Banking, Housing
and Urban Affairs (Washington: 2005), 1–12.
60. Egan, “Statement of Egan Jones on Credit
Ratings Agencies.”
61. Langohr and Langohr, The Rating Agencies and
their Credit Ratings.
62. J. Wiggins, “A Chance to Step into the
35
Light,” Financial Times (London), December 9,
2002.
63. A “Nash equilibrium” is a position in a game
or interaction involving two or more actors, in
which no actor has anything to gain by changing
only his own strategy unilaterally. Since there is
no unilateral action by either actor that would
improve their respective positions, the implica-
tion is that such an outcome will produce a stable
equilibrium.
64. A Bertrand duopoly is one in which a duo-
poly is sufficient to drive prices down to marginal
cost and produces a result consistent with that
found under perfect competition. While the Ber-
trand model is conditioned upon a number of
simplifying assumptions, it does decribe a situa-
tion where even a few competitors (a highly con-
centrated market) can mirror outcomes found
under perfect competition.
65. First articulated in Joseph Bertrand, review
of “Theorie Mathematique de la Richesse Sociale
and of Recherches sur les Principles Mathema-
tiques de la Theorie des Richesses” by Léon Wal-
ras, Journal des savants 67 (1883): 499–508.
66. Coffee, Gatekeepers.
67. Association for Financial Professionals, “Rat-
ings Agencies Survey.”
68. Fitch Ratings, “Definitions of Ratings and
Other Forms of Opinion” (2012).
69. Charles W. Calomiris, The Debasement of Rat-
ings: What’s Wrong and How We Can Fix It (New
York: Columbia Business School, National Bu-
reau of Economic Research, 2009), pp. 1–14.
70. Ibid.
71. Bo Becker and Todd Milbourn, “How Did
Increased Competition Affect Credit Ratings?”
NBER Working Paper Series (2010): 1–31; Bo
Becker and Todd Milbourn, “Reputation and
Competition: Evidence from the Credit Rating
Industry” Harvard Business School Working Pa-
pers (2009): 1–41.
72. Ibid.
73. Langohr and Langohr, The Rating Agencies
and their Credit Ratings.
74. Kevin Dowd and Martin Hutchinson, Alche-
mists of Loss: How Modern Finance and Government
Intervention Crashed the Financial System (West Sus-
sex, UK: John Wiley and Sons, Ltd., 2010).
75. Partnoy, “The Siskel and Ebert of Financial
Markets?”
76. James Van Horne, Financial Market Rates
and Flows (Upper Saddle River, NJ: Prentice Hall,
2000); U.S. Securities and Exchange Commis-
sion, Report on the Role and Function of Credit Rat-
ings Agencies in the Operation of the Securities Markets
(Washington: U.S. Securities and Exchange Com-
mission, 2003).
77. See further discussion in Fabian Dittrich,
The Credit Rating Industry: Competition and Regula-
tion (Cologne, Germany: University of Cologne,
2007), pp. 110–14, 153–55.
78. For more discussion about how insufficient,
and often unobtainable, knowledge makes cen-
tralized planning suboptimal, see F. A. Hayek,
“The Use of Knowledge in Society,” American Eco-
nomic Review 35 (1945): 519–30.
79. U.S. Securities and Exchange Commission,
Report on the Role and Function of Credit Ratings Agen-
cies in the Operation of the Securities Markets.
80. U.S. Securities and Exchange Commission,
“Proposed Rule: Definition of Nationally Rec-
ognized Statistical Rating Organization,” Federal
Register 70, no. 78 (2005): 21306–21323.
81. Ibid. See also, Securities and Exchange Com-
mission, “Comments on Proposed Rule: Rating
Agencies and the Use of Credit Ratings under
the Federal Securities Laws,” http://www.sec.gov/
rules/concept/s71203.shtml; and Leo C. O’Neill,
letter to Jonathan G. Katz, July 28, 2003, http://
www.sec.gov/rules/concept/s71203/standard
072803.htm.
82. White, “The Credit Rating Industry: An In-
dustrial Organization Analysis” and “Good Inten-
tions Gone Awry.”
83. U.S. Securities and Exchange Commission,
“Proposed Rule: Definition of Nationally Recog-
nized Statistical Rating Organization.”
84. Langohr and Langohr, The Rating Agencies
and their Credit Ratings.
85. White, “A New Law for the Bond Rating In-
dustry.”
86. Ibid.
87. Langohr and Langohr, The Rating Agencies
and their Credit Ratings.
88. Sanjai Bhagat and Bernard Black, “The Un-
certain Relationship between Board Composition
and Firm Performance,” Business Lawyer (ABA) 54
(1999): 1–3, 38–39; and Andrea Beltratti and René
M. Stulz, “Why Did Some Banks Perform Better
During the Credit Crisis? A Cross-Country Study
of the Impact of Governance and Regulation,”
NBER Working Paper Series (2009): 1–33.