The Meta-Analysis as CDO

I don’t know who Dan Phelps is, and that annoys me, because I owe him a beer.

The clarity of this spun me round and kicked me in the unmentionables, because that’s about as good as an analogy gets.

For those of you lucky enough not to be involved in it, the global financial crisis (GFC) was a gigantic clusterfuck. I’ll try to explain it, badly. Don’t yell at me, I’m not a banker.

(But do, in fact, yell at bankers.)

1. In the early 2000s, interest rates were very low, construction and building materials had a pretty regular cost, and housing inventory was OK. More than anything else, house prices kept going up. People thought they would rise indefinitely, so did banks. So if anyone got under water on a mortgage, you could just sell or refinance. Hack financial analysts frequently encouraged everyone with at least two teeth to buy as many houses as possible.

2. As a consequence, anyone with two nickels to rub together could get a mortgage. (Literally – one of the types of mortgage was called a NINJA loan; no income, no job / assets). Now, say you’re a bank and you own that mortgage (remember, that’s what mortgages are, a commodity that’s bought and traded… someone sold mine last year, now I make repayments to a completely different set of people).

3. Then, there was a collective misunderstanding. Traditionally, mortgages are a very safe asset to own. People pay them back because they like living indoors. An equity made out of a massive bag of regular mortgages like mine is a very safe asset that just pays you money over time. But these weren’t safe old-school mortgages, they were risky as hell.

4. Instead of selling the individual mortgages (like they did with mine), you’d package thousands of mortgages up into mortgage-backed securities (MBS). These could be bad, a la ‘hey, this guy is a part-time bear trainer in a vegan circus, and he has a mortgage on a 5-bedroom house?’

5. Now, my mortgage is good – I’m going to pay it back. But many of them are less good – more money, at a higher interest rate, to people with less ability to pay. What do we do with those? We put them into a different but related type of security. These are great big piles of different debts that are all rolled together (similar to MBS) called collateralized debt obligations (CDOs).

6. CDOs are filled with vegan circus mortgage loans (which are likely to be defaulted on i.e. not paid back), but the ratings agencies (who ‘independently’ determine how risky these assets are) rated them as being good, reliable assets. Why? (1) the people who constructed CDOs paid them (2) the ratings agencies were in competition with each other, and whoever gave the best ratings got more business (3) they had a drastically outdated conception of the reliability of mortgage debt: historically solid (see 5.), but contemporaneously explosive (see 2.).

7. Let’s not bother with tranches, but synthetic CDOs are important: these are also tradable assets, but they aren’t bags of mortgages, they’re bags of credit default swaps betting on mortgages. Basically, they’re insurance contracts. And here’s the secret sauce: because you don’t need new mortgages to trade them, you can just design a new one from scratch and let people buy it again and again.

8. Then everyone started to default on the loans, and the housing market started to soften. This means the MBS market fail, the CDO market fails, and a massive ripple extends through the gigantic insurance market betting on the CDOs.

9. Splat.

Having established the above, we can now write an entire explanation of both the CDO’s role in the 2008 financial crisis starting in the housing market and the growing unreliability of meta-analysis using a common term: the mortgagepaper.

“Historically, mortgagepapers were solid, boring, and reliable. However, by both circumstance and design, at some point we made mortgagepapers far too easy to produce, and we started to make a lot of unreliable ones.

The explosion in the number of available mortgagepapers changed their nature as an asset. Instead of thinking about them singly, it started to become more common to contain them into collective vehicles.

So, we started to think about the market for mortgagepapers in terms of big collections of them, which were regarded as being more reliable because they were ‘diversified’ but had a common outlook.

This however does not ensure reliability if it just hides the underlying flaws in the mortgagepapers themselves. If you carefully examine the constituent mortgagepapers in any given individual collection, you realise there are some good ones and a far higher proportion of bad ones that we would initially expect.

This changes the nature of a collection of mortgagepapers from ‘generally more reliable’ to ‘a single thing which has a great inherent weakness’. Our opinion on the trustworthiness of the mortgagepaper collection needs to be updated before something terrible happens.”

Yes.

Sorry.

My previous on the same topic.

Sleep well.