Sentences to scare you the most important blog post in the world today

According to the report, between 1990 and 2006 — the year in which issuance of Asset-Backed Securities (ABS) peaked — assets with the highest credit rating rose from a little over 20 per cent of total rated fixed-income issues to almost 55 per cent. Think about it. More than half of the world’s debt securities were, for all intents and purposes, considered risk-free. In 2006, that was nearly $5,000bn of assets.

The financial crisis had a lot to do with triple-A ratings being slapped on to subprime securities which didn’t warrant them, we know that. The report says between 1990 and 2006 ABS accounted for 64 per cent of the total growth in the amount of AAA-rated fixed income, compared with 27 per cent attributable to the growth in public debt, 2 per cent to corporate and 8 per cent to other products.

But watch what starts happening from 2008 and 2009.

The AAA bubble re-inflates and suddenly sovereign debt becomes the major force driving the world’s triple-A supply. The turmoil of 2008 shunted some investors from ABS into safer sovereign debt, it’s true. But you also had a plethora of incoming bank regulation to purposefully herd investors towards holding more government bonds, plus a glut of central bank liquidity facilities accepting government IOUs as collateral.  Where ABS dissipated, sovereign debt stood in to fill the gap. And more.

It’s one reason why the sovereign crisis is well and truly painful.

It’s a global repricing of risk, again, but one that has the potential for a much larger pop, so to speak.

The link to the entire blog post is here, with a good graph, hat tip goes to Annie Lowrey.  We weren’t as safe as we thought we were.  And we still aren’t.


A random thought: should ratings agencies be required to rate on a curve, i.e. no more than x percent AAA, etc. Would handle the problem of their incentives to please those who pay for the rating.

Ratings are supposed to represent probability of default. For example: Given a basket of 10,000 AAA rated securities, 2 of them (0.02%) are supposed to default within the next 5 years. Given a basket of 10,000 BB rated securities, 73 of them (0.73%) are supposed to defualt within the next 5 years. "Grading on a curve" would screw up this correspondence, because the mix of default liklihoods in the market is not necessarily the same at all times.


The writer of the original post is ignorant about securitization. The idea is this: there are a lot of mediocre credits out there - home borrowers, companies, etc. But investors demand to buy strong credit, so that they can be reasonably sure they'll get the money back to pay pensions etc. (or, they demand it simply because the regulators told them they can only buy strong credit bonds). What to do? Well, pool these mediocre credits and 'tranche' them (divide in, say two sections): the first tranche will take all the losses, up until it is completely gone, in necessary. That protects the second, 'senior' tranche. Say, you expect 5% of borrowers to default (with total loss), over the lifetime of the loans. No problem, set the loss taking tranche at 10% of total, so that the senior tranche is well protected. Then, it deserves strong credit rating, say 'AAA'. There's nothing magical about it, nor sinister.

The root of the crisis was the unprecedented decline in home prices. In the US, nationwide average of house prices has not declined YoY, since the Great Depression, and back then only because of the Depression itself! So, the credits that were supposed to be mediocre turned out to be horrible - in the example above, instead of 5% expected losses, you got 50%! Now, I know that all of the esteemed readers of this blog feel outraged that these losses were not predicted by the rating agencies, but I would ask them to consider that even a Cassandra predicting housing meltdown would be legitimately challenged to produce historical data that show probability of such a scenario. I guess the moral is that we live in an imperfect universe (from an investor's POV).

Well, house-prices may not have fallen on a huge scale nation-wide since the GD, but then, house-prices have never risen at such a pace, after inflation, nationwide. You can't have it both ways. A simple look at who actually takes out these mortgages could have been helpful, too.

a) no capital requirements/regulations should be based on ratings
b) if a is not possible, then how about forcing the agencies to grade on a curve?

As a lawyer, I can't understand how ratings agencies get away with "No liability for our ratings - they are constitutionally protected opinions." There are lots of "opinions" that lawyers and doctors give, that if they are wrong, they are liable for malpractice - why not ratings agencies?

They wouldn't get away with it if plaintiffs with the proper standing, those who relied on that advice to their detriment, would sue them. A private-sector plaintiff's right to recover would not be affected by constitutional freedom of speech. Unfortunately, those most likely to be in a position to sue are discouraged from doing so either because their financial representatives (hedge and mutual fund operators, bankers) fear the consequences of admitting how much of the fund assets are at risk or because Supreme Court restrictions on class-action lawsuits make it uneconomical to do so. The government is estopped from suing as a criminal matter by constitutional restrictions, although a test case for NRSRO obligations under the Investment Advisors Act of 1940 would be interesting. That statute needs to be amended to define the public obligations of ratings agencies more clearly.

Should have said Securities Exchange Act of 1934, since NRSROs were exempted from the IAA by the Credit Ratings Agency Reform Act of 2006.

Michael Lewis' 'The Big Short' and other books I can't come to grips with the fact that professional money managers need ratings. From the perspective of an investor, if I pay you several millions USD, then you as a fund manager, better actually check the fundamentals yourself rather than relying on some dodgy underpaid bunch within a company that obviously has several conflicts of interest. It is beyond any comprehension to me that people watching over billions of investor money can't be bothered to do actual research themselves and rely on that AAA. One of my favourite lines from his book is (to paraphrase) 'Dusseldorf. Stupid Germans buy those CDO's. They still believe in rules and ratings.'

Having said that, I would like to thank anyone who's excessively relied on those agencies. Without people like you my performance would have been much worse. Thanks.

if I pay you several millions USD, then you as a fund manager, better actually check the fundamentals yourself rather than relying on some dodgy underpaid bunch within a company that obviously has several conflicts of interest. It is beyond any comprehension to me that people watching over billions of investor money can’t be bothered to do actual research themselves and rely on that AAA.

This is exactly right. You don't pay a manager to say, "Well, it's AAA so it must be fine. Buy it and let's go have lunch."

Unfortunately, this seems to be exactly what too many managers did.

The people selling securities might be willing to disclose more to a ratings agency than they would to an investment firm. The idea is they disclose once, to a secure entity that isn't an investor, and the analysis gets done and disseminated without having to tell everybody all your secrets. For example for a CDO, your secrets might include what your portfolio looks like, so giving that info to an investment manager would be giving away the farm.

Anyway, that was part of the motivation.

"For example for a CDO, your secrets might include what your portfolio looks like"

Isn't that the whole point? When you put money into a CDO you have to have knowledge of its actual contents and the loans etc behind it. Products need to be

What kind of secrets would states only trust rating's agencies to handle?

"The people selling securities might be willing to disclose more to a ratings agency than they would to an investment firm."

No, they would disclose the same data. It's all in the "prospectus". As I said above, it was very difficult to predict a sharp decline in house prices and to act on that prediction, because it never happened before. You don't need other explanation.

the key seems to me to get rid of the conflict of interest: the issuer pays the rater, who obviously wants to rate the next issuance, and so therefore, is unlikely to rate it in a way that the issuer doesn't want. If Moody's did that, Goldman would take the next issuance to S&P etc.

Why we can't have rating agencies that sell their reviews to bond funds is beyond me.

Sorry, I wasn't scared by that. It's missing the denominators.

I am curious how much and which Sovereign wealth funds hold sovereign debt (can't find the info at the moment). If the holdings are spread out among enough entities it could mean less of a shock if a bubble pops.

So there is far more demand for AAA debt than there is supply. This typically resolves itself with price adjustment - which could ultimately mean zero-coupon T-bills trading above face value.

I could be missing something but I see no reason in principle why that couldn't happen, though I do wonder about what arbitrage opportunities might open up, especially for those with access to the fed's lending and borrowing facilities for financial institutions.

That must be why the Fed had to buy all those US Treasuries, then! I guess if you include their artificial demand (designed to keep interest rates low) then yeah, sure.

The fundamental problem is that we live in a world of Knightian -- "true" uncertainty (relatively unique situations with big players, etc.) and so attempts to measure it will be fraught with many difficulties. Attempts to honestly come to grips with true uncertainty should produce, I imagine, a variety of opinions. One reason we don't see this variety is that the raters are trying to transform uncertainty into more easily measurable objective risk and they use similar techniques to do so.

Perhaps the rating on a "curve" idea is a good one. It might force raters to make intuitive distinctions where the measures come out the same. But I really don't know.

was waiting for this post... perfect companion to grades inflation post

Is this really a problem? I know that auto ABS during that time performed about as expected. Just because mortgage ABS had problems, does that mean all ABS is bad and overrated?

But suppose some of the sovereign debt was properly rated lower than AAA. The risk of default would be the same, no? And it certainly appears as if bond buyers are already grading sovereign debt on a curve, demanding higher rates from some 'AAA countries' than others. So it seems like the problem here is simply the elevated sovereign default risk -- regardless of whether the debt carries a AAA rating.

We're perfectly safe. The Fed and ECB will, sooner or later, print dollars and Euros for every one of those bank and sovereign debt bonds you are holding. Of course, you might not be able to buy more than a packet of cream for your morning coffee with their toilet paper currency by then. . .

As a person ignorant of finance, am I reading that chart correctly to say that in 1990, only a few billion dollars were considered "AAA" risk, that means, if what David writes above is accurate, that only .02% of them will default in the next 5 years?

Then, over the next 16 years that number went to something like 5 trillion dollars. How (and why?) did that happen? How did no one at the rating agencies look at this chart in say, 2005, and think to themselves "hm, is it really true that there is 5 trillion bucks of .02% defaulting debt?" And looking at it now, how can they justify saying that 60% of all debt in the world is that risk free?

I assume they are not downgrading them for some good reason (panic?) and I get why the comment above about "grading on the curve " might now work, but surely there's no way on earth that's all triple A investments?

Are these just stupid questions?

Your misreading it. In 1990, the entire US government debt was rated AAA, for example, and even then that was more than a few billion dollars.

Okay, good to know -- so what does the height of the bars represent, in terms that I'd understand? (not rhetorical, I really don't know anything about this)

The bars are claimed to represent new AAA debt issued in the corresponding year.

Which still doesn't make any sense, since the US government ran a $200+ billion deficit in 1990, so I would expect the sovereign debt bar for that year to be about $0.2 trillion... and it's just not.

Not at all a stupid question! Even if you want to grade for objective performance rather than grade on a curve, a radical change in the mix of grades in your class should still be a red flag that something you need to understand is afoot.

Okay, then I guess I have more questions, since you've been kind enough to reply!

First, why didn't anyone notice this? I guess hindsight is 20/20, but it seems like if a car insurance company looked at its books and realized that it was giving out an increasingly large (like 100 fold increase) of "no accidents, spotless driving" policies, someone might ask some questions. I guess the answer might be "because they're horrible people" but that's not really satisfying. :) Surely the ratings agencies have incentives to be accurate, why didn't they work?

Second, to my untrained eye, it seems like this really is the entire issue of the recession: that there were these things that everyone thought were worth 100 bucks because they were backed up by triple AAA stuff, but it turns out they were more risky, so they only have 60 bucks -- and then all the ripples from everyone else figuring that out, and then realizing they had planned to pay for lots of cool stuff with that other 40 bucks they don't have any more, and so on down the chain. If they had been assessed properly for their risk, there never would have been a problem. Is it really that simple: inaccurate risk ratings?

This graph makes me want to invest in hand crank water purifier, shotgun and can opener futures. At least I don't live in a country with a large population of people who're likely to riot. Seriously though, if I were on Wall Street I'd think seriously about getting out know. The scapegoating is going to be atrocious when this all goes down.

Are we overplaying the AAA element of this? Looks like the portion of AAA debt, while likely heavily inflated, has been pretty stable since, say, 1994 at slightly above 50%. The risk seems to me to be simply that we're issuing so much new debt of any stripe each year. In 1994, we issued $1Tr but in 2009 we issued $16.7Tr -- that's an increase of 18.7%/year continuously compounded!

If you are saying that there can't be more than a certain percentage of AAA securities at any one time, you are basically saying that there is just a fixed level of systematic risk we can never avoid. That is a perfectly fine statement when we are talking about private businesses engaged in risky economic activities. But Sovereign debt is not the same thing as private debt and systematic risk probably doesn't influence public debt the same way it influences private debt. In that case you can legitimately have the total percentage of safe debt increase.

Either I don't understand the graph, or it is wrong. I'm pretty sure the amount of sovereign debt increased rapidly in the early 1990s, but it doesn't show up on the graph. Why not?

The truth is these Asset backed securities are made AAA NOT by their assets, but by the structure of the securities. To talk about the distribution of ratings is pretty much ignoring that simple fact. The rating agencies require the structure to have enough over capitalization to support the rating. The simplest form just has a subordinate tranche set up to fund the AAA tranche in the event of shortfall. SO - the question is not whether asset backed loans warrant AAA ratings but why there isn't enough subordination or other credit support built into the structures. Of course, that requires research beyond summarizing some flashy stats. Or if you just want to point a finger, point it at the rating agencies.
Good luck with that.

Begun, the great internet eduatocin has.

«Are we overplaying the AAA element of this? Looks like the portion of AAA debt, while likely heavily inflated, has been pretty stable since, say, 1994 at slightly above 50%. The risk seems to me to be simply that we’re issuing so much new debt of any stripe each year.»

The credit explosion was deliberate policy to create asset bubbles, but it needed AAA rated securities to work, because there are regulatory and accounting issues with holding lots of unsafe debt. However AAA securities are essentially FREE MONEY:, because having capital as AAA securities is as good as having it as cash, and you can lever it up a lot.

The complement to all this discussion is that at the same time as the percentage and absolute value of issues AAA securities exploded, the leverage of most large financial companies also exploded, from something like 10 to 30-40 times capital.

The really big deal was not the "unexpected" rise in risk of ABSes, but that the holders of supposedly safe ABSes were so enormously leveraged that even the tiniest fluctuation in the risk and then value of those ABSes would wipe out their capital, triggering insolvency and the shutdown of their trading operations.

Sure, imaginarily rated AAA securities did enable colossal leveraging, but the leveraging was a political decision, and the AAA rating were sold to support that decision, not viceversa.

«assets with the highest credit rating rose from a little over 20 per cent of total rated fixed-income issues to almost 55 per cent»

To continue the point above: the increase in their absolute value was also colossal, because they became 55 per cent of a much much bigger debt market. And the derivatives built on top grew even faster and to much bigger values.

The past 20-30 years have been the result of several policies aimed at driving multipliers up, multipliers like P/Es for company shares and leverage ratios for large financial companies, presumably in a long term attempt to create low tax or tax-free capital gains for politically important constituencies, in one of the most gigantic pump-and-dump schemes (or "wealth creation" from the point of the insiders) in history.

Comments for this post are closed