Felix Salmon

Is it time to abolish the triple-A rating?

By Felix Salmon
July 25, 2011

It’s looking increasingly likely that the US is going to lose its triple-A credit rating at some point in the next 12 months or so, whatever happens to the debt ceiling. And right now, nobody knows what the consequences of that would be: it’s never happened before. Paul Krugman is trying to put a brave face on things and point out that hey, sovereign borrowing costs didn’t rise in Japan. (Not, of course, that turning out like Japan is exactly desirable.)

But the fact is that Japan never got much benefit from having a triple-A rating in the first place. Japan borrows overwhelmingly from its own citizens, who have lots of savings — they don’t give two hoots about what Moody’s thinks of their country’s creditworthiness.

The US, by contrast, is very different. Treasury bonds are the ultimate global asset class: they’re the epitome of risk-free safety for investors all over the world. Look at what happened when the subprime bubble burst and the US economy was plunged into the greatest recession in living memory, causing a plunge in tax revenues, a spike in unemployment, and a surge in government debt: Treasury prices went up. That’s what happens when you’re the risk-free asset of choice: you’re the beneficiary of the flight-to-quality trade whenever markets get spooked.

Which brings me to the question of what exactly a triple-A credit rating means and does. To fully understand this, it’s important to realize just how lazy and/or overworked institutional fixed-income investors are. A lot of their time is spent drumming up new investments, holding existing clients’ hands, and the like — and when they do get time to do real analysis, the first thing they worry about is always interest-rate risk. If you’re managing a bond portfolio, you’re obsessed with the duration of your portfolio, where your maturities are, and the like: you’re constantly putting on steepeners or flatteners or otherwise trying to manage what’s happening with interest rates in whatever currency you’re investing in.

Only once they’ve got the rates situation worked out do fixed-income investors start thinking and worrying about credit risk. Credit is an overlay — a way of boosting returns over and above what you’d get from rates alone. Sometimes you want more credit risk, sometimes you want less; when you want more, you buy lower-rated bonds with higher yields, and when you want less you buy higher-rated bonds with lower yields.

But the triple-A credit rating is not a credit rating like any of the others. It’s basically a sign saying “no credit risk here” — a way of investing in fixed-income assets without taking credit risk. Triple-A bonds are in this sense a pure interest-rate play — while they can rise or fall in value, and they can get illiquid at times, the idea is that they’ll never default.

In their heart of hearts, fixed-income investors know that there’s no such thing as no credit risk — especially in a world where more than half of the bonds being issued are rated triple-A. But that’s a worry for academics and people with time on their hands. If there’s credit risk in triple-A bonds, it’s too small to price or to worry about, and so, by convention, it’s ignored.

That’s how the financial crisis happened: enormous quantities of triple-A-rated nuclear waste was issued and placed with fixed-income investors who were looking for a modest yield pickup over Treasuries with no real risk of default. They outsourced their due diligence to the ratings agencies: after all, these structured products were highly complex, and precious few bond investors had the ability to even try to work out the credit risk in them on their own.

Nowadays, people are a bit warier when it comes to structured products. But the triple-A mystique remains, and amazingly continues to adhere to Treasuries even as the current debt ceiling debate brings us closer and closer to default. More importantly, it adheres to municipal bonds: individual investors — people who certainly have no ability to gauge credit risk — demand triple-A ratings precisely because they don’t want to have to worry about credit risk. Similarly, people are happy keeping their money in the bank because it’s insured by the FDIC, which in turn is backed by the full faith and credit of the US government. Triple-A.

So what happens if the US loses its triple-A? The simple answer is that nobody knows. But it will certainly cause second thoughts among people who up until now have been very good at ignoring the question of credit risk in triple-A assets.

The problem is that the world of finance is built on mass suspensions of disbelief. Money itself has value only because we as a society all opt to believe that it does. When we deposit money in the bank, we happily leave it there and believe that it’s available to us on demand, despite the fact that the bank has taken that money and lent it out to someone else. Similarly, we believe that our investments are worth whatever they’re changing hands for today, even though we’re not selling them today, and any state of the world where lots of people wanted to sell those investments is a state of the world where they’d likely be worth much less than they are now.

The single biggest beneficiary of this suspension of disbelief is the USA. Because everybody believes Treasury bonds to be risk-free, they’re a risk-off asset, and they tend to rise in value whenever the world starts looking scary. They’re the place you put your money when you’re not thinking — they’re the default option, the safe end of the barbell. No one buys Treasury bonds because they want to go long US credit risk: they buy Treasury bonds because they don’t want any credit risk at all, and because they want the most liquid instrument in the world.

If the US were to lose its triple-A rating, all that would change. Treasuries would remain highly liquid, simply by dint of the sheer quantity of the things that there is outstanding. And they would continue to be used as collateral in repo operations and the like, just because there’s no alternative. But their fundamental nature would change forever. Look, for instance, at all the investors who are standing on the sidelines right now, ready to jump in and buy Treasuries if they sell off sharply. That’s a speculative, risk-on move: you try to catch the falling knife, and then ride the rebound.

A downgrade probably wouldn’t cause an immediate selloff in Treasury bonds, not least because there’s no other asset in the world which has nearly the same amount of size and liquidity. But the effect on non-Treasury triple-A debt, including municipal bonds, might well be larger. And the long-term effect is likely to be huge: when looking for a safe haven, investors will henceforth have to stop and think about which one of various options are the least risky. The knee-jerk flight-to-quality trade will be a thing of the past, because nobody will really know what “quality” is any more. Specifically, would triple-A debt be safer than Treasuries, if Treasuries aren’t triple-A?

Again, this is something it’s impossible to answer ex ante. My gut feeling is that triple-A issuers like Johnson & Johnson or the World Bank would still trade at higher yields than the US Treasury, even if the US was downgraded and they weren’t. But in the first instance, there would be bound to be a lot of volatility, with weird price action and very unexpected correlations.

There is one interesting alternative, however. What would happen if, instead of downgrading the US specifically, the ratings agencies simply decided to abolish the triple-A rating entirely? The rating is dangerous and impossible: a triple-A default is a much more significant event than a double-A default. Taking the entire fixed-income universe and capping it at double-A would send a message to the markets that, sorry, there’s no such thing as risk-free debt: you can worry about rates all you like, but everything you own has some element of credit risk built in.

Taking the triple-A out of the ratings universe would, at one level, be purely cosmetic. It would be like removing the pinstripes from the Yankees’ uniform, or changing the markings on Nigel Tufnel’s guitar amp so that they only go to 10 rather than 11.

But appearance matters, especially in something as fragile and important as the global confidence game that is the fixed-income market. And abolishing the triple-A rating would be an admission on the part of the ratings agencies that in creating it, they built a monster. It is a monster which has been very good to the US, over the years; its death throes would be extremely painful. But a world without triple-A debt is a world where investors are more alive to the risks that they’re running. And ultimately that’s probably a good thing.

17 comments so far | RSS Comments RSS

Felix, this proposal smells of a symbolic move more than anything else.

“If there’s credit risk in triple-A bonds, it’s too small to price or to worry about, and so, by convention, it’s ignored.”

All risk is relative. The credit risk for AAA bonds need not be zero, it need only be less than the credit risk for AA bonds. If you truly wanted “zero-risk dollars” (an oxymoron on the face of it), wouldn’t you hold physical greenbacks? Or the electronic equivalent on deposit with the Fed? If you are investing those dollars, and earning a fraction of a percentage of interest, then that interest is paying you for the risk of default above and beyond the risk of inflation that you would face either way.

The MBS disaster was another issue entirely. Those bonds were seriously *mis*-rated, and people unquestioningly accepted those ratings. Lazy people, charging millions of dollars for “managing” money that belongs to others.

Posted by TFF | Report as abusive

If AA+ were the highest rating, would people start thinking of that as risk-free? Would introducing a AAA+ rating and refusing to bestow it on anything stop people from thinking of AAA as risk-free? If there was yield enhancement over Treasuries available from AAA securities, was this entirely a liquidity premium, or was there some small acknowledgement that the credit risk was different from that of Treasuries?

Credit ratings seem to have some effect on spreads, but changes in spreads lead changes in credit ratings; I’m not as expert on this as you are, but my impression is that fixed-income types lean more on market indicators of risk than they do on ratings. (Insofar as this, taken to the extreme, is entirely circular, it may be that one guy making quotes based on credit ratings is enough that market prices are simply conveying the credit ratings; however, see the previous sentence.) Insofar as they measure different things — to the extent I can figure out what a credit rating is even supposed to estimate, it’s probability of a default event in the next year or five years or something like that — the market indicators are more useful, including size of loss; if the US did default next week, and paid all investors in full on August 3, I expect any significant effect on Treasury’s ability to borrow would be due to a reduction in the mystique surrounding Treasury bonds; the agencies might say “Russia defaulted in 1998; Argentina defaulted in 2001; the US defaulted in 2011; they’re all the same,” but investors would draw a more nuanced conclusion.

Posted by dWj | Report as abusive

My problem with all the bond discussions, is that I look back at all the stock discussions, and in retrospect, there was no logic, no underlying objective factual basis for how tech stocks (and I suspect all stocks) were/are priced. It was just a herd saying buy this…because the price is going up….because the herd it buying it…
Take Greece. Is anything happening there that wasn’t foreseeable a month ago…6 months ago…a year ago? Ditto Italy…
I imagine US bonds can be sold for quite some time at very low interest rates. I imagine at some point, much like the emperor’s clothes, someone will figure out that losing 1 or 2 or 4 percent is foolish, and that not losing money is preferable to losing money. The market will figure this out precisely two seconds before it occurs.

Posted by fresnodan | Report as abusive

I don’t disagree, but you’re obscuring the bigger point here, which is that a downgrade of the US is absurd. It would be a political move by the ratings agencies, not a meaningful reflection of true credit risk.

You’re quite right that Treasuries should continue to trade inside of even the highest quality corporates. Does it even make sense for, say, Wal-Mart to be of higher creditworthiness than the US government? In general, companies can’t have better credit than the government that issues the currency they use to repay their debts.

The US government can raise taxes or print money to pay its obligations. Present nonsense from Congress notwithstanding, there is no real alternative.

Posted by loudnotes | Report as abusive

“Does it even make sense for, say, Wal-Mart to be of higher creditworthiness than the US government?”

Walmart wouldn’t be my choice of an example. Don’t Exxon, J&J, and Microsoft all have AAA ratings? (Microsoft, of course, has almost no debt.)

That said, the risks are different. Sovereign default, especially for a country that controls the currency in which the debt is written, is wholly a matter of political will. If the country chooses to default, there isn’t much anybody can do about it. Corporate default depends on actual economic fundamentals.

Posted by TFF | Report as abusive

(Not, of course, that turning out like Japan is exactly desirable.)

Hmm. Sure, the Japanese are having a rough year, but with half the unemployment of the US, universal health care/insurance (at 1/3 the price with better results), and way lower crime than the US, turning out like Japan ain’t bad at all.

Posted by David239 | Report as abusive

Felix, your post reminds me of something you told the FCIC:

“Value-at-risk (VAR) was a way of allowing banks to ignore the tails. And the Gaussian copula function was a way of allowing banks to shove enormous amounts of risk into those tails. And when you put the two together, it’s absolutely disastrous.”

http://fcic.law.stanford.edu/resource/in terviews#S

So, if AAA-ratings are the current means by which tail-risks are priced too cheaply (consciously or not), to where is that tail-risk being shoved?

Posted by dedalus | Report as abusive

Honestly, the day Treasuries are a risky investment is the day I’m more worried about surviving my new Mad Max existence.

This whole ceiling debate is a complete fabrication.

This is like starving in a room full of food.

And the United States doesn’t “borrow” money from anybody. Who gives a rat’s ass what Moody’s thinks?

Posted by petertemplar | Report as abusive

“Because everybody believes Treasury bonds to be risk-free”. Not really: http://www.dagongcredit.com/dagongweb/en glish/pr/show.php?id=102&table=web_e_zxz x

Posted by Kamekon | Report as abusive

For some reason you ignored the real reason AAA is so important and that is due to regulation – you know the regulation that apparently doesn’t exist. A significant portion of investors have severe limitations on what they can buy. They are not too lazy or överworked to look at other bonds, they usually cannot buy them hence the tranching and structured products business. For instance UK Pension funds did not think that a notional yield of 3.48% was a great punt for a 50 year gilt but because of the way pension fund liabilities were regulated, magically anyone who bought a 50 year gilt at ANY price who consider more solvent and better funded which reduced the amount the sponsors had to put into the fund. That is a real cash incentive for companies to do something stupid whilst knowing it is stupid.

The ratings are written into virtually every single derivative contract, again due to regulation. Ratings are written in capital regulation. They are written into most loan covenants. All of which are positive feedback triggers waiting to send a company or country into a death spiral once they hit an arbitrary barrier. Due to their pervasiveness it is hard to see how they could be removed.

PS dedalus, VaR is a requirement for regulation. It is not the banks who were using it to ignore risks rather it was the way the regulators were measuring risk and so the banks paid alot of attention to it, because again there was a cash reward for managing this figure.

Posted by Danny_Black | Report as abusive

As has been mentioned, there is no way corporates are going to have better ratings than the government that issues their currency, and backs their banks. If the US government losses AAA, then so will all US based companies. All AAA companies will go down a notch, and so will all AAs and As, and so on down. If this results in higher borrowing rates for these companies, there’s not much they can do other than domiciling themselves in some other AAA rated country.

Posted by rogueecon | Report as abusive

“If the US government losses AAA, then so will all US based companies.”

The large “US based companies” do most of their business overseas. They may report in dollars (as do some companies that are NOT based in the US), but most of their revenues and much of their expenses are in other currencies.

Some might even benefit from a crashing dollar, since overseas revenues exceed overseas expenses? In a strong-dollar environment you see regular notes about adverse currency impact on their earnings.

Posted by TFF | Report as abusive

“The large “US based companies” do most of their business overseas. “

Quite right, and this might actually mean a seachange in raters’ perception that a company is subject to a specific country risk. So what does this mean if US government debt gets downgraded? Perhaps multinational US companies will just raise money abroad where country risk is better rated than in US.

But what does this mean for them if they want to still raise money in the US? US banks, US money managers, US intermediaries, all use US currency, expect to be paid in US currency, and by US banks backed by US government. Won’t the lower rated sovereign still affect the dynamics (and the perceived risk) of US financings?

Posted by rogueecon | Report as abusive

“Perhaps multinational US companies will just raise money abroad where country risk is better rated than in US.”

Quite possibly. If people lose faith in the dollar, then dollar-denominated debt becomes very costly. The multinationals could then choose to write bonds in Euros, yen, or renminbi. Some do that already.

“But what does this mean for them if they want to still raise money in the US?”

They can borrow money in another currency, then exchange it for dollars. But why the heck would they want to raise money in the US? They aren’t investing here…

For that matter, the top-rated multinationals don’t NEED to borrow money at all. They only choose to do so because it is cheaper than equity. J&J, for example, has $9B of long-term debt outstanding. They have some $11B of net cash and $14B of free cash flow. They could pay off every penny of their outstanding debt over a two-year period and still make their dividend. Should their access to cheap borrowing be choked off, they will do exactly that. (Maybe not quite so abruptly — their bonds don’t mature THAT rapidly.)

I suppose you can imagine a scenario in which J&J is forced into default, but we will see hundreds of other corporations (and regimes) topple before that comes to pass. Is this truly more likely than Tea Partiers winning control of Congress and deciding that they don’t care to honor the national debt?

Only reason that J&J debt trades at higher yields than US Treasury debt is that the latter is more liquid.

Posted by TFF | Report as abusive

“But why the heck would they want to raise money in the US? They aren’t investing here”

That about sums it up. And why indeed would they raise money in another currency, to exchange into USD, when foreigners will know to charge a higher rate if they know the proceeds will go into USD, and will be paid back by a company earning in USD.

Posted by rogueecon | Report as abusive

Felix, you beat me. As I read your article, I was thinking, this is the reverse of Spinal Tap “It goes to 11.” and then you write:

Taking the triple-A out of the ratings universe would, at one level, be purely cosmetic. It would be like removing the pinstripes from the Yankees’ uniform, or changing the markings on Nigel Tufnel’s guitar amp so that they only go to 10 rather than 11.

If we didn’t have AAA and lower ratings to help the regulators with credit risk monitoring, they’d have to invest something worse to replace it, like the NAIC SVO.

Posted by DavidMerkel | Report as abusive

Post Your Comment

We welcome comments that advance the story through relevant opinion, anecdotes, links and data. If you see a comment that you believe is irrelevant or inappropriate, you can flag it to our editors by using the report abuse links. Views expressed in the comments do not represent those of Reuters. For more information on our comment policy, see http://blogs.reuters.com/fulldisclosure/2010/09/27/toward-a-more-thoughtful-conversation-on-stories/