Here's a simple idea: When you invest you should probably actually look at what you're investing in.
Strangely, some of the most sophisticated investors in the world have been violating this simple idea. Last month at the 2008 ASF conference in Las Vegas, I asked one of the Collateralized Debt Obligations ("CDO") investors who was on a discussion panel if he had any plans to change how he would go about analyzing the collateral under CDO's in the Future. In front of maybe 500 people, his answer was a surprising and somewhat defensive, "We have our procedures in place for selecting managers and those will not change." Umm. So hang on a second here. My jaw almost hit the floor. Let me get this straight. The CDO market itself is melting down right in front of our eyes and may even (especially in the area of "ABS CDO's) vanish and go the way of the dinosaur. The contagion effect from sub-prime has infected all other areas of the CDO Market including those CDOs that have experienced no defaults whatever in their underlying collateral such as "Collateralized Loan Obligations" (also known as "CLO's") and this guy is telling me he is not going to change the way he operates. Bet that this guy is gonna vanish along with all of the CDO's out there that are imploding right now.
Well I'll be !! I did not know what I expected to hear but "do nothing" was the last thing I expected to hear.
So let's see what I mean by "looking" at what you're investing in:
For ABS bond investors, this would mean being able to look through to the underlying loans and see what this stuff is that is behind the ABS bond payments. See all the characteristics of all of those loans such as mortgage type, FICO scores, original appraisal values, interest rates on the mortgages, geographical distribution, who originated the loans, who is serving the loans – in short, all of the standard attributes. A lot of this information can be found in the original prospectus for the bonds. It also means realizing that very quickly the "origination" data becomes stale quite fast. In particular, the FICO scores and the original property appraisal values get stale. Because the data vendors that are out there such as Loan performance and Intex do not give you updated property values what should you do when a deal has 2005 loan mandates and here it is 2008? Here are a few things:
1. Get "Home Price Index" data from various places such as OFHEO's web site (free); Case Shiller (S & P bought them and has freebie data for some areas of the states); Purchase a subscription to the non-freebie Case Shiller listings or go out to one of the other vendors such as Loan Performance's new Home Price Indices which are provided to be quite good or Radar Logic's indexes etc. In short, you're trying to find something that will tell you how homes, in general, have been performing in the area that the loans backing your particular ABS are in? That way you can possibly work out:
A. What is the current estimate of home value right now.
B. Use some basic projection formulas to work out what values could be into the future.
These things are being done by quite a few people, but not enough people do it because they do not have access to a full database of expensive Loan Performance or Intex data and so they are, to that degree, flying blind and depend on others For their research and analysis and "transparency" into the deals themselves.
2. Another thing to do is to retrieve ALL loans from that Zip code across ALL securitized deals – not just the single ABS bond you're analyzing – and look at the # of loans and the payment status of each of those loans (Current, 30 days delinq, 60 days delinq, 90+ days delinq, Foreclosure or REO) to try to gauge how affected is that particular zip code by the sub-prime mortage market crash. Also, you could use some mapping services to find out exactly where that Zip is. Is it in a highly populated area etc? Keep in mind that the services provided by Loan Performance give you only the securitized loan database. Many investment firms also have a whole other "inventory" of loans that were never securitized and these loans will not be received reported on by Loan Performance or Intex. But it's definitely better than nothing and should serve to give you a better idea of what's out there in your loan's particular zip code and the surrounding areas.
3. Items # 1 and # 2 above can only really give you an estimate of the value of the homes because none of those data sellers give you data that has a "finer granularity" than the zip code of the individual loans. This means that you do NOT have a property address. Sounds somewhat reasonable because giving the property address probably violates some sort of privacy information laws. Ok, fair enough. So what if you really want to nail it down closer then the guestimate in steps 1 and 2. Is there any way around this without violating any laws? Well, there are now data vendors appearing on the scene who will sell you a "county records" database. It's expensive as hell. A cool annual subscription from one vendor of 1 million USD and from another 1.2 to 1.5 million. This database gives you detailed information about loans in a particular area and is publicly available information about who owns a particular property. Right down to the street address, when the loan was originated and the size of the loan etc. Wow! Not bad.
If you've got a million bucks to spend (and some hedge funds are doing just that from what I've heard) you can really get down and look at these loans. One problem is how to "marry up" the property data with the loan-level data provided by Loan Performance and Intex. One way that has been proposed is to simply run a process across all the loans in your database looking for Zip, Origination Date and Origination Dollar Amount and look for exact matches within the same zip code of the County Records database. If you find only one record that matches your loan from that Zip code, chances are really quite high that you've found the exact right property. Now you can drill into the street address and actually see your property using Google Earth (watch out for stale data) or other services and try to get a real sense of what the value of the property is. You also now know who owns the property and maybe you've crossed the line so watch out here. This would be a violation of privacy laws. Maybe some lawyer can tell me. Someone would have paid a lot of money for Loan Performance and a lot of money for the county records database but if these expenses translate into valuable investing insight, it's probably worth it.
Okay, let's go further into looking at our loans.
4. We should map all of the above loans onto mapping software so that you can see where your geographical concentrations are. The sets of loans would include:
A. The loans in the particular ABS deal you're looking at investing in
B. The loans across all securitized deals for the particular areas that are in the specific SPV you're looking at.
C. If you now have the property addresses you can look much more closely at the property
(I did not say it would not take some time to actually look at the data. But that is not that what good investors do? They perform "due diligence" on their investments before taking the plunge.)
5. Another approach to take is to always run your loans through standard "econometric models". There are various vendors out there who have
these models. You feed it the data about the loan and it comes back and gives you an "estimated loss" figure among other figures. Loan Performance has their own "analytic" model. Sum up all the estimated losses on your loans and you have some idea of a projected loss amount for your loans.
6. So let's step up into the next layer of atmosphere in the Structured Products arena. Let's say you're a CDO bond trader and you're likely to be looking at investing in a CDO bond. You should be able to do one or more of all the above items for every single ABS bond that backs the particular CDO Bond you're looking at investing in. When I spoke to one CDO trader he said: "We did not do any of that stuff. These things were just more bonds to be traded." Yikes! I think you could say this violates the idea of "looking" at ones investments.
7. Now let's get really extreme. For CDO's, that invested heavily in ABS bonds (which in turn had sub-prime mortgages in them), you should be able to run cash flows on every single ABS bond under that particular CDO given appropriate prepayment speeds, interest rate curves, default rates And loss severities. You should be able to stress each of those factors to get some kind of idea of how sensitive the prices of EACH of the ABS bonds are to various scenarios. Furthermore, if the CDO has some investments in OTHER CDO bonds as part of its collateral, then you should probably look hrough to each of those CDO's and see what sort of collateral underlies those bonds because if the losses on those underlying assets reach a certain level , It will most certainly result in writedowns (losses) on the CDO bonds themselves.
8. Now, obviously CDO's became a highly complicated and "layered" bit of stuff. Layers on top of layers on top of layers – but what was forgotten if the under founding foundation were to erode, ie. The individual underwriting mortgage loans themselves, the entire edifice would collapse. Hmmmm. Does not that basically just about sum up where we are at the current moment?
The bottom line really brings down to these simple ideas:
A. THE INDIVIDUAL COUNTS. As goes the individual, so goes the ENTIRE economy.
B. You have to actually LOOK at what you're investing in.