Why are CDOs hard to value?

I know that there have been threads done on this, but I couldn’t find them. Anyway, why are CDOs so hard to value or rate? AIUI, they are basically bonds with a face value and a rating. What made them toxic? Again, I understand that that means that they are worthless on the secondary market, but their hold-to-maturity value is some unknown proportion of their face value.

Thanks,
Rob

CDO’s are essentially a bundle of other assets. The problem is they were bundling good and bad assets together and there is not a clear paper trail to determine what every CDO has as a part of it. Essentially they used smoke and mirrors to hide the crappy debts in these things and the ratings agencies played along and would rate these things as good risks.

Unbundling them is a huge job and not easy.

The “scare quotes” around the title are FULLY deserved.

Why is that? Is it difficult to compose a prospectus stating that 1% of the value of this instrument comes from a loan to Good Reputation Industries and 1% comes from a loan to Fly-By-Night-Co, etc.? I know that some of the value came from the idea that if the mortgagee defaults, the bank got the asset back which they could resell. Is it just the case that these instruments were unregulated in terms of required disclosure?

Thanks,
Rob

Most of them are amalgamations of hundreds or even thousands of mortgages or commercial loans. You have to investigate each one individually, for its current status, the credit-worthiness of the borrower, and the updated value of the collateral. Big job.

The original thinking was that amalgamation meant safety, since out of a large group, a predictable number would default. (Just like insurers know that out of a big pool, a predictable number of drivers will have auto accidents.) Of course this ignored the high degree of correlation and contagion–sometimes, like now, huge numbers of people will all default at once. Now all amalgamation means is that these monsters are impossible to re-value.

Also, many of these instruments were originally purchased in a highly leveraged manner, with the purchaser betting heavily that the values would only go up.

If a bank purchased a bunch of these assets at a 20-1 leverage, a 5% drop in their value wipes out the original investment. A 10% drop in value means that not only is the investment wiped out but they are in debt for an amount equalling that of the original investment. If the value drops 20, 25 ,30% then the “assets” turn into a giant money sucking sinkhole.

As part of this discussion, can anyone explain asset-based securities and CDOs in plain English that a 12 year old would understand?
How are these pools of assets divided into bits that fit different risk preferences? Are they intrinsically confusing, which is why few realized that there were problems, or was there some deliberate obfuscation going on? What are tranches?

Links are fine. I just am so confused about how these work.

I, Freddy the Pig, have a pot of money and I’d like to get into the mortgage lending business. Mortgage loans offer a nice return, they’re long term, and they’re historically fairly safe. But it’s not as if I can go out and originate loans myself. I’m not a banker. I don’t know beans about qualifying customers or appraising property. What to do?

The solution–let someone else originate the loans–someone who hopefully knows what they’re doing–and bundle them into a “mortgage-backed security” in which I can invest. I still don’t know anything about mortgages, but I don’t have to. The loan originators will handle it.

What if they exercise crappy judgment, and loan to a bunch of deadbeats, and all the mortgages of which I own a small piece default? Well, that didn’t seem very likely. Mortgage loans were an old business with well understood lending standards (X% of income and all that.) Even deadbeats didn’t default all that often, because (a) people don’t like to lose their homes; and (b) when in danger, they could flip the property for a higher price and pay off the loan. And in extremis, when people did default, the loan originator could flip the property and reclaim the balance. That’s why mortgages are asset-backed or collateralized debt obligations.

In fact, believe it or not, for a long time the greatest risk associated with mortgages was the risk of early payment. When interest rates fall, borrowers pay off and refinance, cheating the lender out of long-term gain.

As a further refinement, therefore, the repayments from a pot of mortgages could be separated into tranches with different risk characteristics. One tranche might have greater risk of early payment, another greater risk of default (not because of the mortgages themselves, but because of how the payments were assigned to tranches.) The safest tranches were considered safe indeed.

By now, you can probably see the problems with this model. First, for a variety of reasons (moral hazard, pressure and encouragement by government) lending standards deteriorated. Second, the property bubble burst, so houses at risk of default could no longer be flipped. And third, the economy teetered, so that more people lost their jobs and couldn’t make payments. Even the safest tranches not only looked like shit, but shit that was impossible to value.

Mortgages are just one example. Other “asset-backed securities” involve commercial real estate and commercial loans. All involve the same problems–deteriorating loan standards, more defaults, and suddenly worthless collateral.

Let me see if I can explain, although I am by no means an expert. Tranch is French for slice. I believe what banks do is to order loans by risk and gather together, say, $10 million worth of the least risky. They then divide that into X number of what are essentially bonds. Then they do the same to the next least risky instruments. The first group are, I believe, called A tranches (least risky) and the last are called ZZ tranches (most risky). Then, they repeat the procedure, taking a part of each tranch and combining it together into a CDO. As the previous poster said, the idea here is to spread risk around such that each CDO was based on a balance of risky and safe loans. Now, how do rate them in order to determine their liquid value? That is my question which was answered above.

There was some shenanigans going on in terms of mortgage brokers and predatory lenders overrating borrowers which throws off the risk calculation, but mostly, I think it was a problem with analysts being unable to come up with a value because it is hard (after all, they can factor in fraud risk).

Do I have that about right?

FWIW,
Rob

My GF works for a rating agency that rates Residential Mortgage Backed Securities (RMBS). These are bonds of bundled mortgages that are often bundled with other debt instruments into CDOs.

I’ve met a lot of her coworkers and I can only come to the conclusion that they are hard to value because they are being valued by morons.

But one of the main problems is that a significant number of the underlying mortgages were issued under false pretenses and that the models where they calculated the ability of the homeowners to pay them back were incorrect and didn’t take a number of factors into account (ie that homes can LOSE value). One thing that I have both heard from people in the industry and seen in my own career investigating corporate fraud, is that quite often some of these models people create are bullshit. They go through multiple revisions with no clear QA process and a lot of them really are innaccurate or incomplete and no one can really explain how they work.

That throws off the true risk associated with the RMBS and exagerates their ratings, which then propagates through each layer of the system (remember that the lenders, the rating agencies, the banks and the various securities firms are all separate and distinct companies.

There is also a tendency for these companies to gloss over irregularities when they are bundling these securities together so they can sell them.

And of course, there are people like this one idiot I know who don’t even bother to do proper due dilligance and just assume because a security has a particular rating that makes it good.

At the risk of sounding stupid, why is that a big job? Does it all have to be done by hand?

Also, can you explain correlation and contagion a little further? After all, not everyone is defaulting. In the example above, the instruments are still worth $95 billion, are they not?

Thanks for your help,
Rob

There is nothing wrong with making an awful (high risk) loan on the face of it. If you want to accept the risk then that is your business.

The problem is moving the risk away from the people making the loan. The mortgage brokers didn’t give a shit if they made an awful loan as they would sell the risk to someone else. Ordinarily that would not be profitable as others would want terms to accept that risk that you could not meet or just that no one would buy it and you’d be stuck with it.

But in this case it was not a problem. Downstream these things would get bundled and details on the real risk washed out of the system. They knew the risk was greater and that defaults would rise on these things but they felt it would go from (making numbers up) an average 8% default to a 12% default. They STILL made money on it so keep going with the bad loans!

The real kicker however is the ratings agencies (such as Moody’s or Standard & Poors) were giving these things AAA ratings. They knew these were shitty but guess who pays S&P? The people who want their CDOs rated. If S&P wanted to give a lower rating they’d be told to stuff it and they’d go to Moody’s.

In short you had the foxes guarding the hen house. This managed to keep working as long as the housing bubble kept growing. Once the bubble popped the defaults skyrocketed and the backing assets were shown to be crap. Now you have a problem.

Then you come along with a pile of money to invest. You have no freaking clue what is in that CDO but it has a AAA rating so you think great! Nice and safe and buy it. Then the housing market implodes and you wonder WTF just happened?

Amazing no one isn’t going to jail yet over any of this but I guess everyone kept marching along with a wink and a nod and everyone shares some guilt.

Don’t investors know that ratings agencies are paid by bond issuers? Why did investors pay attention to these agencies if they felt that the ratings were inflated (and wouldn’t that be fraud)? Sure, little old me might be suckered, but a big, sophisticated investor would know better. Had they developed a good reputation for accurately assessing risk and then decided to squander that trust?

Thanks,
Rob

It is hard because no one really knows what the value is. Technically the value of a thing is worth whatever someone is willing to pay for it. Thing is few people are buying. Is the value the $500,000 you took a loan for to buy the property? Is the value the remaining $320,000 you owe the bank on the home? Is the value the $120,000 a buyer is willing to pay for your $500,000 house? Depending who you ask (bank, investors, home owners, prospective buyers) you will get different answers as each one wants what is best for them.

And for contagion the issue is having negative equity in your house. Let’s say you bought a house for $500,000 and put $100,000 as a down payment. You have $100,000 in equity in that house.

Now the bubble bursts. Your house is valued at $300,000 today. You now have a negative $100,000 in equity in the home. It will take you years to just climb back to zero.

As such it can actually make fiscal sense to walk away from the debt. You shed the debt.

So, as housing prices spiral down in an area it depresses property prices and more people walk away and depresses prices further. And so on and so on. Contagion.

I dunno. That is how business was done. Presumably the rating agencies operated on their reliability and reputation so, in theory, would not make bogus ratings. The ratings are heavily relied upon by all investors including (even especially) commercial banking.

Thing is people would shop them and the ratings companies want to make money too. So if S&P wanted to give my CDO a low rating I could go to Moody’s and ask if they could do a little better and they might be willing to nudge it a bit for the business. Round and round you go and eventually it just became this great big circle jerk and no one could really stop.

If S&P gives a low rating, doesn’t that get published? And don’t they get paid up front? Also, are we talking about jacking a rating up from junk status to AAA? It seems like you could make a case for fraud if that occurred.

Thanks,
Rob

No cite but I heard talk of criminal investigations over this whole debacle. Whether the rating agencies will be looked at like this I do not know.

As for the rest I do not know either. In general there are ways to game the system in business that while dodgy are not outright illegal. How the whole process works though is a mystery to me.

As for how far they bumped these things up the ratings ladder I do not know either but remember you can probably make a case that these things were ok back in the day. Like the insurance comparison above they had reason to believe a given CDO would only have an X% default rate and those were pretty accurate as long as the whole thing kept marching along. So is that a good or bad risk? These things are all bundled together and some defaults are expected but overall the bundle was a safe bet.

The model was not complete though and apparently did not account for what happened and the house of cards came tumbling down. Are they criminal for that? Should they have known? Did they know and ignore it?

I don’t know.

It gets more complicated than said so far. Say we start with 1000 mortgages for a total cost of $100,000,000 and an average interest rate of 6%. You pool them all together and then split them up into 26 slices. Each slice does not contain 10 mortgages. They just have different interest rates and default conditions. The bonds sold as the A slice have the lowest interest rate, say 3%, but will pay out as long as less than 90% of the mortgages default. The Z slice, on the other hand pays out a great rate (say 9%), but if anyone at all defaults, you lose out first. So far so good. This is pretty straight forward, and barring illegal shenanigans, anyone can look at the type of underlying loan and figure out their default rate and put a value on this like any other bond. It’s what happens next that screws things up. Its easy to get a good price on the A and B slices, but when you get down the Z slices no one wants to pay very much for them. So what do you do?

We take the Z level slice from this CDO, and a bunch of others as well. We pool another $100,000,000 of them together and slice them up again. All of these started out as Z level slices, but now we are calling some of them A’ and B’ slices. We get the rating agencies to rate these as AAA and sell them. Of course the Z slices of these (let’s cal them Z’) are still hard to sell. Well we know what to do with with those! Soon we have A’’’ and B’’’ and even Z’’’. If housing prices keep going up and none of the original mortgages has an issue, then the mortgage payments go to the first instrument to be paid out as coupon (the interest payments on bonds are usually paid out on a fixed schedule instead of compounding like a bank account). These payments then go to pay out the the next level and so on.
The bundlers make money on every step in fees. Any attempt to revalue bonds from the a Z slice affect all the A’,A’’,…slices, and it becomes difficult to calculate the default rate on a bond based on slices of other bonds, based on slice of other bonds, …, based on slices of multiple pools of mortgages. One house going into foreclosure could effect the repayment of thousands of bonds as it works it way through the system.

If everyone held on to the assets they would get some value back from them, depending on which sliced of what they were based on. Some will get full value, some zero value, but right now, no one can tell who will get what in the future. And all bond prices are based on what you think you will get over the life of the bond.

How much would you pay for an investment that could pay back anything from $0-1000 dollars? Last time I looked, lottery tickets are selling for $1, and we know what the probability is of winning.

Jonathan

Let’s make it really simple.

  1. CDO’s or asset backed are called that because they contain some sort of ownership of the underlying asset. For example, 1,000 residential home loans bundled together (derivatives are named because the derivative value is based on the underlying asset but it doesn’t actually own the underlying asset. Like betting on a horse race)
  2. These are sliced and diced to be sold to different investors. For example, investor A needs 20 year maturity and investor B wants 30 years, so the get those respective slices
  3. Ratings are important because many funds have to invest in only investment grade A ratings. AAA or triple A rating is the best private with US Treasuries as the most risk free investment with the highest rating.
  4. Rating agencies get paid to assign a rating, so there is some bias.
  5. Turning shit into shinola: you can also take a CDO made out of the worst assets possible and give them an investment grade rating ***IF ***there is insurance. In other words, if someone with a good rating guarantees the CDO will pay the interest and principle at maturity, then the rating agencies can assign an investment grade rating.
  6. Now that *no one *has a clue where the property/stock market/economy will bottom, there is no way to value the underlying CDO assets like those 1,000 residental loans. And because there have been so many defaults, those companies providing the CDO insurance can not possibly pay all the claims.
  7. Therefore a huge chunk of those “safe” “investment grade” holdings that school districts, pensions, unit trusts have are now worth 10 cents on the dollar, and all those investors thinking they had something safe are SOL.

AIG Financial Products made a very large percentage of the guarantees in #5, and that’s why Uncle Sugar is bailing them out to keep the entire global Ponzi scheme from collapsing. AIG itself is a decent business, but the 200 people in AIG FP are the ultimate enablers of this global mess.

Caveat Emptor or buyer beware was coined a few thousand years ago that you should understand what you are buying or suffer the consequences. Read the damn prospectus and use your brain. Just because your broker says it’s safe doesn’t make it so.

Finally, if ***anything ***has a better return (or higher interest) than US treasuries, then you ***are ***taking on risk. And in times of stress like we are in right now, the normal rules don’t apply and risk gets amplified. Getting that extra 0.5% interest may cost a 50% loss of the principle as many school districts are learning.

Well, keep in mind that the procedure that Freddy gave is for valuing one tranche. What happens when you own tens of thousands of tranches? Or millions? How many person-hours does it take to value millions of loans? And how much will the value of those loans change in the time it takes you to value them all?