The Role of the Credit Rating Agencies in the U.S. Subprime Crisis : Too Much Competition ?

A known fact is that the Big Three, namely, Moody’s, Fitch Ratings, Standard & Poor’s, have provided inaccurate ratings that were obviously too generous. These CRAs (not to be confounded with Community Reinvestment Act), contamined by conflicts of interest, had played a role in the subprime crisis. But not only their role in the crisis is exaggerated, the causes as for their deficiencies seem to be attributed to poor regulation rather than by regulation itself. The common, but erroneous view is the suggestion that financial markets need more, better regulation, while in fact it may be the regulation itself that had caused these deficiencies.

Obviously, if money stops overflowing thanks to central banks’ policies or other types of regulations that lead to credit over-expansion, this subprime crisis wouldn’t happen in the first place. Higher demands for bank credits, all else equal, lower the savings readily available to serve as loans. In reaction to the drop in money available to the banks, the interest rates will increase and the “overheating economy” process will cease as fast as it had even begun. But considering the regulations presently, problems accumulate and become more and more complex. Obviously, the false confidence that housing prices would continue to go up would not have been possible without money over-expansion.

In a sense, for the investors, CRAs are useful to the extent they give reliable information about the risk associated with the interest-earning assets, debt securities we have purchased. Included are government bonds and asset-backed security or ABS (e.g., collateralized debt obligations, CDO, or mortgage-backed securities, MBS) and almost any debts. The goal of the ABS is to diversify the risk. Individually, the interest-earning assets to be pooled are unable to be sold. But when pooled, the risk is lessened because each security represents a small portion of the portfolio (i.e., collection of investments). The riskier parts (tranches) being generally sold to the hedge funds. In general, the buyers are, insurance companies, pension funds, investment managers, banks, investment banks and hedge funds. The reason why banks sell the CDOs to investors is that the money they receive allow them to make more loans while (by loosening their reserve requirements) at the same time it shifts the risk of defaulting from banks to investors. Fundamentally, a CDO does no harm. But under credit over-expansion, it becomes a time bomb.

But for this purpose, the investment bank creates a new institution, called special purpose entity (SPE), in order to sell the securities to investors. At this point, we should remember that there are two types of securities. Bonds are debt, whereas stocks are equity. By purchasing equity (stock) an investor becomes an owner in a corporation. Ownership comes with voting rights and the right to share in any future profits. By purchasing debt (bonds) an investor becomes a creditor to the corporation or government. It’s debt that is the focus. By selling the debt securities, the banks remove the assets from the banks’ balance sheet to the CDOs. Since the regulation imposes a legal minimum of reserve requirements, the CDOs lower those requirements, allowing them to expand their credits even more. These SPEs serve to hide debt and help to obscure relationships between entities. It renders more difficult for investors to decipher a company’s actual debt exposure. Another purpose is realized through arbitrage CDO. As the name indicates, the purpose is to seek profit from the spread between the yield (e.g., 10%) on the assets underlying the CDO and the yields (e.g., 5%) paid out to investors in the tranches of CDO. If the difference is positive for the issuer of CDO, it yields profits. Still another kind is the synthetic CDO, which aims to generate money through contracts and options such as credit default swaps (CDS). In this situation, the SPE does not hold any assets at all. It merely sells CDS protection to investors, again, in tranches. They are some sorts of insurance or compensation the buyers take to protect themself against financial risks given the plausible situation of credit default (i.e., the borrower is unable to repay its debt) they might expect. The buyer of a CDS pays regularly some fees and if an accident happens on a given loan, the buyer is already insured. In doing so, the CDOs appear virtually risk free, as they are backed by a pool of these contracts. At least, when all is not expected to went bankrupt, in which case, every parts will be pulverized. Synthetic CDOs consist thus in backing CDO by securities that were already backed by mortgages; this creates leverage on leverage.

Imagine (although this is what really happened) that investment banks built bad products by packaging the good and bad debts together into a very unstable portfolio to make them somewhat indistinguishable from one another, specifically through CDOs. The CDO is divided in several categories called debt tranches with each having its own risk level, e.g., equity (highest risk), mezzanine, senior (safest). Because interest rates reflect the level of risk, equity bears the highest interest rates and senior the lowest interest rates. Under the situation where the cash collected through CDOs is insufficient to repay all of the CDO investors, the “equity” tranches suffer the first % of losses on the CDO porfolio, then the “mezzanine” being the intermediate and lastly the “senior” tranches that are the last to lose payment from default. Specifically, the senior are supposed to be repaid the first while equity is repaid the last. If no bonds default, each tranches receive all the interests they bear. For example, we have a bond/asset pool containing a total of $9000 in assets but divided in three tranches with the senior yielding 5% interests or 0.05*4000=$200, mezzanine 10% or 0.10*3000=$300, and finally the equity yielding 15% interests or 0.15*2000=$300. But when several loans default, imagine the interest income now totals only $400. What happens then is that senior is remunerated first, with $200 in interests, mezzanine with the remaining $200 and equity earn $0. In this scenario, mezzanine earns only a portion of its “rights” while equity earns nothing at all.

The complicated part of the CDO becomes even more obvious when it comes to determine the price of each category of CDO. How could we evaluate the risk of these products ? In reality, some of these debt defaults are more or less correlated; for example, debts on different enterprises but within state, or within region, or within a given sector. When a firm defaults on its obligations to another firm, this other firm will also likely be unable to pay its own obligations. In this case, there is positive correlation. But when a firm X goes out of business, its competitors Y and Z can get and bring back more customers at the expense of the defaulted firm, thus lowering the probability of default of the competitors. In this case, there is negative correlation. Overall, the ratings must take into account the pattern, number and strength of these correlations. To complicate matters, some companies can default regardless of the business cycles, i.e., when the performance of the economy on the whole is high, because they didn’t keep pace with changes in technology or unsound strategies. Obviously, the performance of companies and that of the economy as a whole must be strongly correlated. Most of the companies won’t close when the performance of the economy is good. So, at first glance, this won’t be a problem, except for the fact that (investment) banks take for granted that the economy will do well continuously, that is, they do not expect the forthcoming recession while in fact this result cannot be avoided. Austrian economics explain convincingly that credit over-expansion causes investment and consumption to be greater than what available savings can normally allow; sooner or later, the future products and goods can’t be sold because people don’t have enough savings to absorb this additional amount of goods in the future.

Globally, the following pictures illustrate the process in two phases :

The Financial Crisis Inquiry Report (2011) Figure 5.3

The Financial Crisis Inquiry Report (2011) Figure 8.1

The logic is summarized by the Financial Crisis Inquiry Report (2011) :

The securities firms argued — and the rating agencies agreed — that if they pooled many BBB-rated mortgage-backed securities, they would create additional diversification benefits. The rating agencies believed that those diversification benefits were significant — that if one security went bad, the second had only a very small chance of going bad at the same time. And as long as losses were limited, only those investors at the bottom would lose money. They would absorb the blow, and the other investors would continue to get paid.

This is how they get their triple-A from the CRAs. Lowenstein (2008) uses the following analogy to make things easier to get : “Imagine a seaside condo beset by flooding: just as the penthouse will not get wet until the lower floors are thoroughly soaked, so the triple-A bonds would not lose a dime unless the lower credits were wiped out”. This sort of segregation of payments enables CRAs to classify them as AAA, for example.

Without being rated, nobody would buy these products. But the banks can be willing to pay a great amount of money in order to receive an AAA. However, because of the near-impossibility to know precisely what these financial products contain, even the CRAs couldn’t be able to provide accurate ratings. In case like this, the best, other alternative is to give generous ratings.

But this needs not to be a necessity. We will explain. There exists two types of business models. The issuer-pays and subscriber-pays. Under the issuer-pays, the issuer will pay the agency to receive in exchange a creditworthiness assessment. This is the model mostly prevailed today (99% apparently). Its disadvantage is the obvious conflict of interest. Under the subscriber-pays, or investor-pays, the ratings is not freely availabe to the market because the access is restricted to a subscription fee. This is the model prevailed until the early 1970s. Its disadvantage is that the information is private and the investor must pay to access it. The actual issuer-pays model does, however, not necessarily induce conflicts of interest.

It is usually held that in the case of CRAs, competition doesn’t work as in other sectors of the economy. This reason has a name : conflict of interest. The more competition we have and the more likely the CRAs will inflate their ratings above what is reasonable. If one of the institutions don’t play the game, the other will. Because if one tries to give accurate ratings, this is not exactly what the clients want. What they want is the best rating possible. If the institution persists in not giving a good enough rating, he will lose clients, business and profits. Stated otherwise, bargaining power of the CRAs decreases as competition increases. So in the end, as soon as there are more than one agency, this dilemma remains insolvent. This was probably the conclusion Becker & Milbourn (2010) tended to make. One would be even tempted to believe that, in order to eliminate any possibility of conflict of interest, the government must ban competition and establish a monopolistic system. This implies that the opposed views that government had promoted conflict of interest by restricting competition in this sector, since there were only a few institutions operating, was in the wrong.

Now, remember. The major institutions, known as the Big Three, Moody’s, Fitch Ratings, Standard & Poor’s, have been given a special status, namely, Nationally Recognized Statistical Rating Organization (NRSRO), by the Securities and Exchange Commission (SEC) known as the authority of stock regulation, if those agencies are considered to be credible enough by the SEC himself. But the problem is : the market is not given. To obtain the confidence of the public these CRAs must first of all build their own reputation and keep it as high as possible. Needless to say, reputation can be lost very easily whereas it needs time to build it. In a sense, bad conduct is costly. Under government regulation however, things are a little bit different.

In all likelihood, economic agents, e.g., investors, would think that the ratings given by the three CRAs must necessarily be very reliable, just because they are entitled with the envious status NRSRO, a label that signals to everyone how safe their ratings are, indirectly certified by the authority of stock regulation, i.e., the SEC. In the case where the SEC gives to the investors the illusion that these NRSROs agencies are highly reliable in their activities, the regulator creates an additional demand for those ratings. Furthermore, because of the illusory trust created by the SEC, the investors will be less vigilant (but more careless) about CRAs. On the other hand, because of the carelessness of the investors, these agencies have greater incentive to inflate their ratings. Necessarily, these factors that could have helped to promote honest ratings have been pushed back. This situation is similar to what had led to the Enron Scandal in 2001. The impact of NRSRO status is to be taken seriously. Investors generally don’t want ratings from non-NRSROs. Such measures of the government to prompt people to seek credit ratings discourage CRAs to sell honest opinions (i.e., ratings) because it guaranteed them a persistent demand. This has been made clear by Lowenstein (2008) :

Then as now, Moody’s graded bonds on a scale with 21 steps, from Aaa to C. (There are small differences in the agencies’ nomenclatures, just as a grande latte at Starbucks becomes a “medium” at Peet’s. At Moody’s, ratings that start with the letter “A” carry minimal to low credit risk; those starting with “B” carry moderate to high risk; and “C” ratings denote bonds in poor standing or actual default.) The ratings are meant to be an estimate of probabilities, not a buy or sell recommendation. For instance, Ba bonds default far more often than triple-As. But Moody’s, as it is wont to remind people, is not in the business of advising investors whether to buy Ba’s; it merely publishes a rating.

Until the 1970s, its business grew slowly. But several trends coalesced to speed it up. The first was the collapse of Penn Central in 1970 — a shattering event that the credit agencies failed to foresee. It so unnerved investors that they began to pay more attention to credit risk.

Government responded. The Securities and Exchange Commission, faced with the question of how to measure the capital of broker-dealers, decided to penalize brokers for holding bonds that were less than investment-grade (the term applies to Moody’s 10 top grades). This prompted a question: investment grade according to whom? The S.E.C. opted to create a new category of officially designated rating agencies, and grandfathered the big three — S.&P., Moody’s and Fitch. In effect, the government outsourced its regulatory function to three for-profit companies.

Bank regulators issued similar rules for banks. Pension funds, mutual funds, insurance regulators followed. Over the ’80s and ’90s, a latticework of such rules redefined credit markets. Many classes of investors were now forbidden to buy noninvestment-grade bonds at all.

Issuers thus were forced to seek credit ratings (or else their bonds would not be marketable). The agencies — realizing they had a hot product and, what’s more, a captive market — started charging the very organizations whose bonds they were rating. This was an efficient way to do business, but it put the agencies in a conflicted position. As Partnoy says, rather than selling opinions to investors, the rating agencies were now selling “licenses” to borrowers. Indeed, whether their opinions were accurate no longer mattered so much. Just as a police officer stopping a motorist will want to see his license but not inquire how well he did on his road test, it was the rating — not its accuracy — that mattered to Wall Street.

Another insidious effect of SEC regulation has been advanced by Carden & Murphy (2008). Specifically, the price of an asset diminishes when short selling outnumbers long buying (i.e., short versus long position, or the selling/buying of a security such as a stock with the expectation that the asset will fall/rise in value, although it should be noted this trading is done through borrowed money). In principle, short selling consists in borrowing shares from someone (e.g., broker) and write a contract to sell x shares of stock at x dollars each. The investor will need to go into the spot market at some point in order to purchase the assets and cover his position. If this price is lower than stipulated in the contract, the investor earns a profit. When the SEC has banned short selling, it has prevented the investors to hedge themselves against risky financial firms. Otherwise, investors could have bought CDS to protect themselves against likely bankruptcies. As they write, if a big bank wasn’t getting hit by a wave of “attacks” from short sellers, e.g., speculators, this meant it was probably a relatively sound institution. Thus, a crucial information has been removed by the SEC, rendering the financial market more vulnerable to Enron-like episodes. Obviously, economic agents usually adapt to the new regulation, circumventing it, and causing the regulator to add newer regulations, which cat & mouse game continues if the market continually adapt to the new rules. But in the end, due to hampered market information, the winners of this kind of regulations are the poorly managed firms.

Now, when a crisis hits the economy, there is tendency to say “we need better regulation”. No one thinks about the alternative of no regulation. In this case, deficiencies of the CRAs are thought to be attributed to the SEC, not directly to the CRAs. When the SEC operates, there is no competition for credibility among the CRAs or market checks for risky behavior.

There is no even reason why a monopolist established by law would perform best. This is because a monopolist established through competition (see Rothbard, among others) must be the result of efficiency. A given firm that outcompeted all his competitors because his prices are the lowest or the products of better quality because, among the most obvious reasons, his production cost was the lowest.

Given this, if competition doesn’t work with CRA, this is no reason to conclude that regulation is necessary. But if competition works, that means the CRAs are able to provide honest ratings. And no natural monopoly will emerge. In the end, the widely criticized issuer-pays model might work. The investor-pays model is also another alternative. For example, the Egan-Jones Ratings Company, recently (in 2007) recognized as NRSRO, was the first to downgrade both Enron and WorldCom. Some time ago, already, Johnson (2003) found that EJR performs better than Standard & Poor’s. Using regression analyses, he “compares the grades to which Standard and Poor’s regraded firms from BBB- with Egan-Jones’ ratings of the same firms in the 3-10 weeks before S&P’s regrades”.

An Examination of Rating Agencies' Actions Around the Investment-Grade Boundary (Johnson 2003) Figure 2

A comparison between S&P’s and EJR’s ratings shows that, conditional on S&P’s upgrading or downgrading a firm from BBB-, its new grade was correlated with the grade EJR had awarded at least ten weeks earlier. This suggests that S&P defines its grade BBB- more widely in terms of default probabilities than EJR. It also suggests that S&P’s large downgrades from BBB- did not occur immediately after negative surprises to firms, but rather after a steady accumulation of bad news which EJR’s ratings reflected.


Becker Bo, & Milbourn Todd (2010). How did increased competition affect credit ratings?.
Carden Art, & Murphy Robert P. (2008). The SEC Short Sells Us Down the River.
Financial Crisis Inquiry Report (2011). Final Report of the National Commission on the Causes of the Financial and Economic Crisis in the United States.
Johnson Richard (2003). An Examination of Rating Agencies’ Actions Around the Investment-Grade Boundary.
Lowenstein Roger (2008). Triple-A Failure.

This entry was posted in Economics and tagged . Bookmark the permalink.