Monday, July 25, 2011

US Debt Ceiling Crisis: Rating the Rating Agencies

By guest author Marc Joffe*

News reports and punditry reveal a shocking ignorance of the role played by rating agencies in the US deficit debate. Depending on the commentator’s bias, rating agency actions have been either lionized or demonized, often inappropriately. After dispelling some unfortunate myths about rating agencies, I will offer the reader a more informed assessment of how the three dominant rating agencies are handling the debt ceiling crisis.

Commentators often confuse a credit downgrade with a prediction of default. If a rating agency downgrades an issuer from AAA to AA+, it is not predicting that the issuer will default. The rating change simply expresses the view that a default or a loss of principal/interest is more likely. Since rating scales typically have about 18 different grades, a transition from the highest rating to the second highest rating does not reflect an enormous change in risk. A downgrade from AAA to AA+ simply reflects a rating agency’s view that the credit risk of the US has gone from de minimis to very low. Thus, if a rating agency were to downgrade the Treasury to AA+ and a US bond default did not immediately follow, it would not be appropriate to say that the rating agency made an error. On the other hand, if rating agencies leave the US at AAA and it does default, that would be indicative of an error – similar to assigning AAA ratings to structured instruments that later missed payments.

While it is true that the rating agencies commenting on US creditworthiness are the same firms that assigned inaccurate ratings to mortgage backed securities a few years ago, this fact is not especially meaningful. The analysts who assign ratings to structured finance instruments and those that assess sovereign bonds are different people, working in different groups, using different methodologies. More importantly, the commercial considerations that might bias sovereign ratings are totally different from those that impact assessments of structured assets.

Ratings for mortgage backed securities, collateralized debt obligations and other asset backed instruments are purchased by a relatively small number of issuers. If a rating agency provides an “unsatisfactory” rating for a deal structured by a given issuer, the agency risks losing a sizable fee for the deal in question as well as revenue from all future deals marketed by that issuer. Thus, in the structured finance area, the reactions of a few issuers can materially affect the rating agency’s revenue.

This is not the case with sovereign ratings. Advanced economy sovereigns, such as the United States, pay little if anything for their ratings. Thus, all the concerns about the so-called issuer pays model do not apply to sovereign ratings. At the same time, other commercial considerations might impact them. For example, since rating agencies are regulated by the United States, European Union and other sovereign authorities, they may have reason to fear retaliation from their regulators. While such fears appear to have a basis in Europe where official criticism of the agencies has been frequent, we have yet to see a similar problem in the United States.

Second, sovereign rating changes may impact other ratings in ways that create commercial challenges for rating agencies and investors. Given the dependence of numerous bond-issuing entities on the US government, a Treasury downgrade may trigger a large number of municipal, corporate and structured finance issuer downgrades as well. This cascade of downgrades would impose challenges on a rating agency’s internal systems, staff research skills and relationships with affected issuers.

To the extent that certain institutional investors are restricted to investing in AAA securities, a Treasury downgrade would result in the forced liquidation of many assets. Institutional investors – who often purchase research, data and analytics from ratings firms – may react negatively to such a scenario. Moreover, such portfolio changes could substantially impact interest rates. If these interest rate changes are blamed on the rating agencies, they may suffer reputational consequences.

Such concerns may unduly retard rating changes that appear justified by the issuer’s credit status. In this connection, I am reminded of the Enron situation in late 2001. Back then, the concern was that downgrading Enron to a speculative grade rating would effectively shut the firm out of the credit market and thereby force it into bankruptcy – which is precisely what happened when the belated downgrades were announced.

I characterize the Enron downgrades as “belated” because they occurred long after the firm was identified as a “junk” issuer by quantitative credit models, like the one marketed by KMV Corporation –now owned by Moody’s. Since computer models do not worry about commercial implications of their calculations, they promise to provide more instantaneous and less biased credit assessments than human rating analysts can. While quantitative models for corporate and structured instruments are quite common, relatively little progress has been made in modeling municipal and sovereign risk. (One notable exception is Kamakura Corporation’s sovereign model, released in 2008.)

As with Enron, major rating agency admonitions about US Treasury creditworthiness have been late. Official warnings about the long term sustainability of the federal budget due to population aging date back to the early 1990s. In 2006, Boston University economist Laurence J. Kotlikoff warned that the US was headed toward bankruptcy. In 2009, SR Rating, a Brazilian rating agency assigned the US a rating of AA. This was followed in 2010 by Dagong, a Chinese rating agency, which downgraded the US from AA to A+ last November. Earlier this month, Egan Jones, a small US-based rating agency that employs an investor pays model downgraded the US from AAA to AA+.

Within the last few years, US debt ratios have clearly diverged from comparable AAA sovereigns such as Canada and Australia. For example, the CIA World Fact book shows that in 2010 the ratio of publicly held debt to GDP was 59% in the US, compared to 34% in Canada and 22% in Australia. These other two countries are also less exposed to the consequences of population aging issues and they shoulder a smaller military burden than the United States, so it is difficult to see why all three countries merit the same rating in systems that have 18 distinct credit grades.

Finally, the parliamentary system employed by other Western democracies is better equipped to address fiscal stress than the divided government we have in the US. In a parliamentary system like that of the UK, a single party or a coalition can more readily obtain full control of all the levers of power and use them to implement unpopular changes at the beginning of its term (this is less true in parliamentary democracies that have large numbers of significant parties and thus weaker governing coalitions). In the US, with its more frequent elections, division of responsibility between two houses of Congress and an Executive branch often held by opposing parties and the risk of Senate filibusters, fiscal consolidation is far more difficult. This disadvantage has been exacerbated in recent decades as bipartisanship on many issues has given way to party polarization. These issues of governmental effectiveness should be a part of any comprehensive credit assessment of US Treasury securities.

One justification for maintaining the US AAA rating in spite of these concerns has been refuted by the current crisis. The argument is that since the US manages the world’s reserve currency it is somehow insulated from default. Unfortunately, the US dollar’s reserve status has been under attack for several years, and is not guaranteed to persist over the 30-year term of the longest dated Treasury instruments. Further, the reserve currency argument implicitly assumes that the Federal Reserve would monetize Treasury debts should default risk appear. This argument ignores Fed independence, as well as the fact that the CPI indexing of many Federal benefits would impede the government’s ability to liquidate debt by printing money. Finally, I have not heard any responsible commentator suggest that the government address a failure to raise the debt ceiling by paying creditors with newly created money, so clearly reserve currency status provides no refuge in the current scenario.

While all three major rating agencies have been late to the downgrade party, there are notable differences among them in their handling of recent events. In my view, S&P has performed the best. They were first to take any action, assigning a negative outlook to US Treasury securities on April 18, 2011. Further, S&P has correctly maintained that a debt ceiling increase without meaningful budget reforms would still merit a downgrade. As suggested earlier, the long term US fiscal imbalance has been known to policymakers for 20 years. If a political crisis like the one we are currently experiencing is unable to motivate elected officials to substantially reduce the gap between out-year revenues and expenditures, it is hard to see what will, thus leaving a future default as a significant risk.

Moody’s recent pronouncements have also been largely on target, but the agency was slower off the mark. On February 24, 2011, Moody’s predicted that the debt limit would be raised before the ceiling was reached - on May 16. On June 2, the agency observed that the risk of default due to a failure to raise the debt limit was a rising but still very small risk. Finally, on July 13, Moody’s placed the US credit rating on watch for possible downgrade and also noted that it would assign a negative outlook if substantive deficit reductions were not implemented together with a debt ceiling increase. Although welcome, Moody’s stance is not as strong as that taken by S&P. While Moody’s is threatening to maintain a negative outlook in the absence of substantive action, S&P has warned of an outright downgrade.

Another worry about Moody’s position is that the firm is frequently represented in the media by economists from Moody’s Analytics, such as Mark Zandi. Moody’s Analytics is a distinct subsidiary within Moody’s Corporation from the rating issuing entity, Moody’s Investors Service. Consequently, Zandi and others at Moody’s Analytics are not involved in the sovereign ratings process, a fact often lost on interviewers. Further, Moody’s Analytics economists have previously been on record as supporting fiscal stimulus measures including the recent temporary reduction in the Social Security tax rate. While Keynesian policies may be justified on other grounds, they are not consistent with maximizing a sovereign issuer’s creditworthiness at least in the short run. It is thus unfortunate when the public gets the impression that Moody’s Analytics economists are somehow representing the views of the rating agency.

Finally, it has been disappointing to see the third rating agency joining the discussion so late. In a report dated June 8, 2011, Fitch stated that it would place the US on negative watch on August 2nd if the debt ceiling was not raised and suggested that outright downgrades would occur only in the event of an actual failure to pay scheduled interest or principal. The idea that a default would be needed to trigger a downgrade negates the value of credit ratings. If credit ratings are not supposed to hold predictive content, it is hard to see why investors would need them.

In short, the leading credit rating agencies are belatedly awakening to the fact that a dysfunctional political system and long term fiscal imbalances have created significant risks for Treasury investors. Now these agencies, led by S&P, are beginning to provide investors with insight into the unfolding situation, largely free of the biases that affected them during the 2007-2008 credit crisis. That said, investors would ultimately be better served by measures of advanced economy sovereign risk that react more quickly and are less burdened by potential conflicts.

* Marc Joffe ( is a consultant in the credit assessment field. He previously worked as a Senior Director at Moody’s Analytics. This article reflects his personal opinion of sovereign rating practice. Although previously employed by Moody’s Analytics, the author no longer works at Moody’s and, when he did work there, his area of professional responsibility was software development and data collection. He had no professional experience as a ratings analyst, and no knowledge of Moody’s ratings practices beyond what is in the public record.


Charlie said...

"Within the last few years, US debt ratios have clearly diverged from comparable AAA sovereigns such as Canada and it is difficult to see why all three countries merit the same rating in systems that have 18 distinct credit grades."

Maybe instead of downgrading the US to AA+, rating agencies should just raise Canada and Australia's ratings to AAA+++. I've seen in done on eBay. Why haven't the major rating agencies caught on yet?

Josh said...

"On the other hand, if rating agencies leave the US at AAA and it does default, that would be indicative of an error."

Does AAA mean a default is impossible?

Marc Joffe said...

That's a fair question, Josh. Default probability mapping tables I have reviewed in the past normally associate a rating of AAA with a default probability of 0.01%. This level would seem to allow for a default only in very extreme circumstances such as a nuclear war. If rating agencies are confronted with evidence of even relatively small financial risk, I don't see a justification for maintain a AAA rating. So I would say that the failure to downgrade in this particular case is an error.


Marc - I am trying to think of any instance in which Keynesian policies may be justified as you imply. I can not think of any. Can you?

Marty Nemko said...

Is the concept of "default" risk too all-or-nothing? There's a big difference between the U.S. government deciding to short-term delay payment of a small percentage of its debt than a permanent inability to pay all its debts. No?

Marc Joffe said...

Regarding the question from the International Ventures Group blog: I believe the Keynesian view is that near term deficit spending generates growth that would in turn lead to more tax revenue and less automatic spending later on. When I learned Keynesian economics in school many years ago, the idea seemed to be that the government should run a deficit during a recession and an offsetting surplus during the recovery. From this perspective, it seems to me the goal posts have shifted: now if we go from a $1.5 trillion deficit to a $1.0 trillion deficit in a given year, the policy is characterized as contractionary. Further, since the US is not even in a recession at the moment, the case for further stimulus is hard for me to understand. I was taught that unemployment was a lagging indicator, yet now we hear calls for stimulus based on statistically insignificant movements in that number. With respect to disappointing GDP growth, Reinhart & Rogoff have taught us that this is common after a banking crisis. Further, with baby boomer retirements, workforce participation and consumer spending are both facing long-term demographic headwinds.

Marc Joffe said...

Hi, Marty - We credit folks do take a pretty strict view of default. When we get beyond consumer lending, pretty much any delay of a bond or bank loan payment is characterized as a default. Corporations and other large institutions employ professionals to manage their accounts and are not expected to experience glitches when discharging even small portions of their obligations. Due to the interconnectedness of the financial system, less than totally reliable payers are frowned upon.

Anonymous said...

Hi. re KMV "predicting" the Enron failure. How far in advance did it actually do that? As Enron was fraud I find it hard to believe that it picked it up in a timely manner. So was it predicting a default one year ahead, or one week?

Marc Joffe said...

Anonymous - KMV's model, like other public firm quantitative models, is primarily driven by market cap. Enron's stock price started its rapid decline more than a year before it was downgraded. You can see a chart of Enron's stock price at,31813,2013797,00.html. I don't know exactly when the KMV Expected Default Frequency dropped into non-investment grade space, but it was certainly well over one month before the rating agency downgrades to non-investment grade levels.

Anonymous said...

Marc....yes the share price was in decline through 2001...but it actually lept from $25 to $40 from Sept to which case a mrkt cap model like KMV would have marked it as an improving credit...and if u r a long term lender what good is a system that tells u one month beforehand of a problem? And from the start of 2000 to the decline of Enron an awful lot of share prices were going down....(not to the Enron degree of course), but that was the trend in the equity market. What did KMV rate the cmpany 1 yr and 6 months before it went bust?

Marc Joffe said...

Anonymous - I am no longer with Moody's KMV (I worked there from 2002-2005), so I don't have access to the EDF history. In 2001, I was at a bank that had a large exposure to Enron and saw that the EDF took off about a month prior to the downgrade (which, in turn, was a few days prior to the bankruptcy). The bank could have bought credit protection or sold a portion of its exposure in the secondary market, but didn't. I think the debate over whether quantitative models add value is pretty well settled when you consider that Moody's paid over $200 million for KMV, and both Bloomberg and Morningtar have recently released similar models.

Hendrik said...

Thank you Marc for sharing your insights. I was rather surprised that advanced economies do not pay for getting rated. Do you have a source for that? And what about Emerging Economies? Sovereign Ratings do play a crucial role in determining their access to international debt markets. Uganda for instance is rated B+ by Standard & Poor's - surely they paid for the rating.

Hendrik said...

It might be a stretch to put Uganda into the Emerging Market category, but you get my point: at the lower end of the rating scale, payment must have an influence. Or doesn't it?