The Structurer
Why this chapter
To understand how property and debts disappear from the accounts while remaining somebody's very real problem.
An analogy
Imagine there is a shed full of junk in your yard and you would rather your guests did not see it. The first way is honest: sell the junk to a neighbor, take the money, forget about it. The second way is cunning: sign the shed over to your nephew, promising him that if anything goes wrong you will buy it all back. On the outside the yard is clean either way, but in the second case the junk is still yours — it is just written down in small print. The seventh chapter is about exactly that difference. It explains why the same tricks built the modern mortgage market and brought it down, and how an honest sale differs from one in disguise.
How it works
An SPV is an envelope company: assets are sold to it so that they “drive off” the balance sheet (a true sale). The SPV issues securities against them — securitization. Tranches: senior (paid first, rated AAA) → mezzanine → equity (burns first).
The key trick of 2008: a bank buying an ABS tranche creates money exactly the way it does when lending. Securitization does not “move risk” out of the system — it lets the system create even more credit on the same capital.
The dark side: Repo 105 (a loan disguised as a sale), Enron's SPEs (debts hidden in “independent” envelopes with a secret guarantee from the parent company). We study this in order to recognize it.
The words of this chapter
SPV
A vehicle company created for a single deal: property is sold to it so that it legally stops belonging to the seller, and it issues securities against that property for investors.
It was invented to separate specific assets from the fate of their former owner: if that owner goes bankrupt, the pool that was bought stays with the investors.
Securitization
Packaging many loans into a single pool and issuing securities against it: illiquid claims turn into tradable ones.
It was invented to free up a bank's capital: sell the pool and you can make new loans on the same equity.
A tranche
A floor of an issue. A senior tranche is paid first and gets a high rating; a junior one, on the contrary, takes the first losses and promises a high yield.
It was invented to make securities for very different buyers — cautious and bold — out of one and the same pool.
Covered bond
A bond secured by a pool of loans that stays on the bank's balance sheet: the investor has a claim both on the pool and on the bank itself.
It was invented as the opposite: here you cannot sell the risk and run, so the bank still has an interest in the quality of the loans.
The maturity gap
The situation where long investments are financed by short borrowings that constantly have to be rolled over.
It arises not out of stupidity but out of profit: short money is cheaper, and the difference looks like pure income.
Margin call
A demand to post more collateral urgently, when a position has fallen in price or a participant's rating has dropped. Margin is the collateral behind a deal, and its size is revised as you go.
It was invented so that a counterparty does not end up with an unsecured claim if the market moves against them.
A rating
An assessment of a security's reliability, issued by a specialized agency according to its own model.
It was invented so that an investor does not have to pick through thousands of loans inside a pool themselves.
What happens to the balance sheet
Follow one pool of loans across three balance sheets. At the bank it was property and it demanded an owners' cushion. After the sale to the vehicle the bank had money instead of loans, and the cushion was freed up. The vehicle issued securities divided into floors and sold them to investors. If the investor turned out to be another bank, it paid for the purchase with a new entry — and new money appeared in the system against that very same pool. Nothing disappeared: the risk simply spread along the chain, and in every link of it less capital was left than there had been at the start.
Where the money comes from here
The selling bank earns a fee and the capital it frees up, which can go into new lending. The vehicle earns the difference between the yield on the pool and the coupons on the securities it issued. The investors earn the yield of the floor they chose, and that yield honestly reflects their place in the queue for losses. The one who loses is whoever bought reliability and got somebody else's model: the senior floor burns last, but it burns. Here is why this matters to you: in any packaging, ask not “what is the risk here” but “on what day will real money be asked for on this obligation, and from whom.” Both of the chapter's biggest failures are not about losses but about maturity.
What people usually get wrong
The common belief. If a bank has sold its loans, it has got rid of the risk on them.
What is actually true. Not necessarily. The junior floor it kept for itself, the obligation to service the pool, and a promise to support the vehicle all bring the bank back into a deal it has formally left.
The common belief. Packaging loans into securities carries the risk out of the banking system.
What is actually true. If another bank buys the securities, it pays for the purchase with a new entry. That very same pool ends up financed twice on the same capital of the system.
The common belief. The top rating means a security almost cannot let you down.
What is actually true. It means that somebody else's model calculated it that way. The model rested on the assumption that trouble for borrowers in different towns is not connected. In 2008 it turned out to be connected.
After this chapter you will be able to
- tell a true sale apart from a pledge by who is left with the claim on the borrower
- see that a bank buying any asset from a non-bank creates new money
- find the maturity gap between an asset and its funding, and name the day it will blow up
- ask about any obligation not “how big is the risk” but “on what day will real money be asked for on it”
- check the assumption about correlation that a structured security's high rating rests on
Check yourself
The bank sold a pool of mortgages to a vehicle and kept the junior floor of the issue. Has the risk left its balance sheet?
Mostly no.
The junior floor takes the first losses. Formally the pool has been sold; in fact the bank kept for itself exactly the part that will suffer first. Selling an asset and selling a risk are different events.
Another bank bought the securities of this issue. What happened to the amount of money in the economy?
It grew.
The bank paid for the purchase by crediting the seller's account, that is, with a new entry. This is the same engine as when a loan is made. In the end one pool of loans turned out to be financed twice.
A company's obligations were off its balance sheet, and then its rating fell. What happened next?
Real money was demanded from it immediately.
A demand to post more collateral does not depend on where the obligation is recorded. It was exactly this construction that killed the largest insurer: the loss was in the future, and the money was needed today.
In short
A bank can gather many of the loans it has made into a heap, sell that heap to a separate company and issue securities against it. The securities are split into floors: some are paid first and promise little, others take the first losses and promise a lot. That is how a bank frees up room for new loans and how the buyers get their income. The trouble is that the risk does not disappear: it spreads along the chain, and in every link less of a safety margin is left. And when trouble arrives for everyone at once, even the upper floors burn.
Reading is half the job. In the academy this chapter is worked through by posting entries: you run the operations yourself and see whether the balance sheet still balances.