Juan Ospina and Harald Uhlig write,
AAA securities did ok: on average, their total cumulated losses up to 2013 are 2.3 percent. . .Losses for other rating segments were substantially higher, e.g. reaching above 50 percent for non-investment grade bonds. . .
Cumulative losses of 2.2% of principal on AAA-rated securities surely is a large amount, given that rating. Such losses after six years may be expected for, say, BBB securities, and not for AAA securities. AAA securities are meant to be safe securities, and losses should be extremely unlikely. From that vantage point, an average 2.2% loss rate is certainly anything but “ok”. We have chosen this label not so much in comparison to what one ought to expect from a AAA-rated security, but rather in comparison to the conventional narrative regarding the financial crisis, which would lead one to believe that these losses had been far larger. Ultimately, of course, different judgements can be rendered from different vantage points: our main goal here is to simply summarize the facts.
Authors’ emphasis. They also say,
these facts provide challenge the conventional narrative, that improper ratings of RMBS were a major factor in the financial crisis of 2008.
I object to this conclusion. The capital requirements for the securities depended on the ratings. Because of the AAA ratings, the capital requirement was less than what the loss percentage turned out to be.
As I see it, the facts in the paper support the conventional narrative. If the securities had been correctly rated, then there would have been no financial crisis. If the securities had been properly assigned BBB ratings, or any ratings below AA, banks could not have bought the securities without having at least five times the amount of capital that was required for AAA.
The charitable interpretation is that the authors do not appreciate the significance of capital regulations. The uncharitable interpretation is that they are trolls.
I thought the same damn thing!
I tired to read their paper, but I kept coming up with questions I could not seem to answer. One was the time period starting in 08. Why then? The delinquency rate started accelerating in 2006.
“Mortgage analyst Laurie Goodman estimated that private label securitizations issued during 2005-2007 incurred a loss rate of 24%, whereas the GSE loss rate for 2005-2007 vintage loans was closer to 4%.”
https://www.americanbanker.com/opinion/gse-critics-ignore-loan-performance
I also agree with Arnold’s leverage comments, so I cannot figure out this paper.
As in foreclosure starts hit 200,000 in 06 and 400,000 in 07.
https://www.marketwatch.com/story/where-have-all-the-foreclosures-gone-2016-07-26
” The capital requirements for the securities depended on the ratings. Because of the AAA ratings, the capital requirement was less than what the loss percentage turned out to be.”
May be a more powerful way to explain to the authors what they get wrong is to remind them that these differences make a difference when you lever your position 30:1 to 50:1. And this is exactly what financial players did and would do if they expected near zero default rates.
I think the focus on banks and leverage is a problem. Leverage is just a measure of how you are paying investors. Investors who expect a fixed income are creditors. To buy $1 billion of these securities, you need $1 billion in capital. Leverage isn’t a magical multiplication machine. I’m not accusing Arnold of making this mistake, but it seems like a common rhetorical implication in these conversations.
Commercial banks are considered leveraged because deposits are insured, so they are considered debt to the bank. So, maybe banks should have held more equity capital, but it seems wrong to me to look at institutions that take in cash, pay a low floating rate on it and invest the proceeds in generally safe assets that also pay a low rate, and to describe that in terms of risky leverage. As the authors say in the conclusion, these securities ended up with a total rate of return of about 2.4% to 3.3%, which wasn’t that much less than treasuries. The only reason a bank would have taken a loss is because they were selling at panicked, fire sale prices during the crisis. Furthermore, banks weren’t even that leveraged. Capital ratios had been rising.
https://fred.stlouisfed.org/graph/?g=ju78
The highly leveraged investment banks were mostly involved in underwriting and warehousing these securities. The leverage is part of a business model where they are profiting from the construction of the securities. It wouldn’t be possible to simply buy up these securities and borrow 97% of the capital from bondholders or lenders and somehow profit from the spread. And, for many investors, like pensions, money market funds, etc., leverage just isn’t really even part of the equation. They have a bunch of capital, and they are trying to match their assets to their respective payout and risk profiles.
Every conversation I see about these securities revolves around this “leveraged 30 to 1” idea, which adds more heat than light.
“As the authors say in the conclusion, these securities ended up with a total rate of return of about 2.4% to 3.3%, which wasn’t that much less than treasuries.”
Yeah, by the time the crisis was over and things settled down that may have been the case, but I still don’t know where they got that number and whether it was total portfolioed loans.
On the other hand, leverage played a huge role when delinquencies started from 2006 onward. My cash flow stops coming close to planned levels and my leverage rates balloon. Then when the bank run started most banks were technically insolvent, which is why the FED propped them up with around $15 trillion in loans and guarantees.
“Yeah, by the time the crisis was over and things settled down that may have been the case”
The collateral underlying these securities “settled down” at valuations several standard deviations outside any normal range, yet they still provided a net gain to investors that held them. This is a significant point.
The take-away from recent research like this is that there were securities that did perform very poorly – anything associated with CDOs, synthetic CDOs, etc. and the lower rated tranches of the MBS pools. Of all of those securities, the AAA rated MBS performed exceedingly well, all things considered.
I am talking about the mortgages held in portfolio. The banks owned these loans.
The delinquencies caused huge liquidity problems for the banks, which (in addition to the CDOs) demanded a bailout. At the same time, I see absolutely nothing in this paper that means anything at all. Yeah, I got it, the best paper sold to investors returned well less than expected.
Meanwhile, I have no knowledge of what mortgages the banks held themselves, but I can guarantee is was far, far from the best paper. Very typical of banks(even unintentionally, and they did a lot of intentionally) to place mortgages into a security with a rating that does not belong there.
When the investor finds out(not common, but really common by 2008), the investor kicks it back to the bank. If the bank cannot find a buyer(and by 2008 that was almost impossible), they are forced to portfolio the loan.
Think you should take a look at what constitutes Total Equity.
Further, I cannot seem to find any info on when these securities were issued, which is important in terms of comparing treasury return.
I know what they troll about.
If we had good general price stability, then the rating system works as advertised. I can see that the drop in dollar value leading up would have put the ratings system in error if the system made the macro assumption the dollar share of transaction are constant. A long declining dollar means the total exchanges are not observed and the capital in reserve suddenly more volatile. That was the thing about capital reserves, they suffer equally when transactions leave the dollar sphere.
I read your blog a lot and you often refer to capital requirements.
If I wanted to learn more about these without having a technical regulatory background, do you have any recommendations on where i might read up on them? X loan requires X capital for dummies?
Tim, try to find something that describes the requirements that were in effect in the early 2000s, called Basel 2. these answers might be helpful: https://www.quora.com/What-are-BASEL-1-2-and-3-norms-What-are-the-basic-differences-between-these-norms
Remember Long-Term Capital Management?
I hate to say add more regulations. But when big money dudes can leverage hundred to one… And make piles of money if they bet right but run away if they bet wrong
Remember the Commodity Futures Modernization Act?
That monstrosity of deregulation was a large factor in the financial crisis. Shadow banking stayed in the shadows(and still is to a large extent) until it crept out and helped to almost kill the entire banking system. Oh, and it also managed to exempt energy derivative trading from regulation, thus Enron.
I am sure why the authors did not look at 2005 – 2007 mortgages as these were the ones that busted in 2008. Anyway, it seems wise to say this mortgage securities rating had an impact to the Financial Crisis (10% – 20%) and if these institutions had lower leverages that would have controlled the damage. (So Bear Steans and Countrywide fail versus 20+ other institutions.) In viewing the Financial Crisis the hardest part is drawing a line that give investors more confidence. (1933 Bank Holiday and TARP, etc.)
I still think the main reason for the Housing Bust/Financial Crisis is our nation had 60 year Bull Run on housing and real estate (with some drops in real values in late 1970s and 1980s.) It was not just an economic system but the idea that if you did the right things you can have home ownership of a home for your family. (Remember the American dream during the 19th century was an average person can own their own farm for their family with hard work and this started to end during the Great Depression.)
“AAA securities are meant to be safe securities, and losses should be extremely unlikely. From that vantage point, an average 2.2% loss rate is certainly anything but “ok”. ”
The Law of Large Numbers and the Law of Truly Large Numbers paint an entirely different picture. In fact TLOTLN tells us that this was inevitable.
Some default rates:
https://www.thebalance.com/what-is-the-default-rate-416917
Default rates have been quite low in the corporate bond market over time, averaging 1.47% of all outstanding issues in the 32-year period measured. Investment grade bonds defaulted at a rate of just 0.10% per year, while the default rate for below-investment-grade (high yield) bonds was 4.22%.
Default at 2.2% instead of 0.10% is a … 22 times worse event; 2,200% more than expected. (100% more would be 0.20%).
Because banks ARE regulated by “capital requirements” (like Basel 2; now it’s Basel 3), the mis-rating of AAA instead of BBB meant that banks could give out more loans — which kept increasing the bubble of prices & even house building, right up thru 2005.
Building decreased in 2006. House prices stopped increasing then, and the smart speculators knew it was time to bail out.
The Big Short (Michael Lewis) remains must reading, and fun reading (haven’t seen the movie tho, hmm).
Michael Burry read the loan documents and knew it was a bubble. Probably a few more read the loan documents than read the computer acceptance agreements, but not a lot more (10% instead of 1%?).
The rating agencies needed to have more “skin in the game”, and should have all gone belly up / been sued successfully into bankruptcy.
Japan, 1989, shows what real estate can do – crash. Not to mention (again) LTCM — with a Nobel economist and one of the smartest bond traders from the earlier bond book by Lewis “Liar’s Poker” (an even better, more fun book).
All the big banks should have crashed, bankrupted (equity wiped), been taken over by interim gov’t agencies; stabilized (at much lower executive salaries) for bank business and had short term creditors become equity owners in recapitalized small banks. (Rich) equity investors losing their shirts would be demanding more capital & safety & insurance on AAA securities … and probably much more regulation, too.
For Main Street, the Fed could be making loans to all the smaller and mid-sized banks so that they have cash to loan out to companies with good projects.
Keep in mind that this article is talking about basic RMBS while the Big Short guys were dealing in synthetic CDOs made up of lower rated RMBSs. This article does not dispute that the lower rated tranches had very high default rates.
This is somewhat selective quoting. The next section shows that SUBprime AAA only had loss rates of about 0.40%, well below the 2.2% for Prime AAA. Part of the story is that subprime lending was not the problem the conventional narrative says it is.
Hindsite gives us the values (facts) we are now debating. There are three values we need to keep in mind – expected rate, rate trajectory, and leverage. The expected rate is not known but estimated. Its true nature, the true underlying distribution, is not known to any and is but an estimate (and best understood as fraught with endogeneity issues). Likewise, the rate trajectory is estimated and not known. Then only factor we choose is leverage. I would submit that the amount of leverage introduced into the systems impacts rates. To after the fact argue that there is material deviation in returns is burdened that their is a belief that the expected values are known with greater certainty than what is warranted. Nassim Taleb takes up this issue with great passion so I won’t belabor the point here.