Like every other fool with a blog, I’m trying to understand what caused the financial crisis.
Let’s start by looking at the collapse of Northern Rock in some detail and try to work backwards from there. What happened to this bank?
Here’s the consolidated balance sheet from the Northern Rock 2007 annual report (with categories simplified and combined and figures rounded to the nearest billion pounds; see the report for the details):
|Assets||Loans and advances to banks||1||6|
|(a)||Loans and advances to customers||99||87|
|Liabilities||Bank of England loan||28||0|
So (a) in 2007 Northern Rock issued a net £12 billion in “loans and advances to customers” (almost all in the form of residential mortgage loans: about 90% to owner-occupiers and 10% buy-to-let), but were unable to increase their long-term sources of funding.
Northern Rock could raise money to make these loans in four ways: (b) ordinary bank deposits from customers; (c) mortgage-backed securities, issued by subsidiary company Granite Finance Trustees Limited; (d) bonds (that is, wholesale loans from other financial institutions like pension funds, hedge funds and other banks). Page 72 of the annual report suggests that the plan was to pay for the net £12 billion of loans primarily by issuing mortgage-backed securities totalling (for January–September, anyway) £15 billion.
But this didn’t work: the third issue of mortgage-backed securities, scheduled for , was never sold. It’s not clear to me exactly how this went down, but presumably Northern Rock were unable to find underwriters for this issue of securities, and couldn’t raise any other funds in the wholesale money markets, leaving them no choice but to go to the Bank of England for an emergency loan on Saturday (using the unsold securities as security) which was followed on by a run on the bank, and finally by nationalization on .
Why couldn’t Northern Rock underwrite the third issue of mortgage-backed securities? Here, I’m afraid I enter the realm of conjecture. The prospectus for the failed issue appears not to be available online, but here’s the prospectus (all 419 pages of it) for the preceding (May 2007) issue of £4.5 billion worth of securitized mortgage loans. I doubt the failed 2007-3 prospectus was very different. With the benefit of hindsight several warning signs are apparent, for example the issuing of self-certified (“liar”) loans:
In March 2004, [Northern Rock adopted] a new set of procedures under which verification of a borrower’s income was not required in certain circumstances. For mortgage loans with an LTV ratio no greater than 85% (or no greater than 80% in the case of mortgage loans in excess of £500,000) a borrower receiving a medium to high credit score did not need to provide proof of income. [See page 104.]
And although the mortgages making up the security are carefully tabulated by type of property, location of property, loan-to-value ratio, outstanding balance, mortgage product type, employment status, interest rate, type of repayment plan, there’s no tabulation by debt-to-income ratio (surely a crucial predictor of default) and no tabulation by degree of certification of income (for all we know, 55% of the loans in the prospectus might be self-certified).
Despite these warning signs, the 2007-2 issue was underwritten in May 2007 by Lehman Brothers (who filed for bankruptcy in September 2008), Merrill Lynch (sold to Bank of America in September 2008), and Barclays (still extant at time of writing).
So what changed between May and September? What’s clear to me from reading the prospectus is that there’s really no way to tell from a securitization document like that whether the underlying security is a good investment or not: the data necessary to work it out is missing. (Whether that’s by bad luck or by design I don’t know.) If you are going to invest in the security, you do so because you trust that the issuing bank has done its job and applied suitable underwriting standards, and is not trying to mislead you. By September 2007 it was clear that the banking system in general had completely failed to be worthy of this trust, and that meant that no-one was willing to purchase Northern Rock’s securities.
For all I know, maybe Northern Rock’s loans have been made prudently and the risks from their “fast track” procedure are small in practice. But since Northern Rock had 8% of the mortgage business in the UK in 2007 it seems more likely that they suffered from a good admixture of fraudulent and over-extended loans.
So, to summarize the story so far:
(I think this is the conventional wisdom, but I feel somewhat more confident having worked through the evidence.)
This story raises some questions. How did Northern Rock come to rely on securitization, and why did it relax its lending standards? Since every issue of securities from 2004 very likely had some problems with self-certified loans, why didn’t those issues fail too?
Undoubtedly the greed, stupidity, and short-sightedness of Northern Rock’s management are going to feature in the explanations, but since Northern Rock is far from the only bank where roughly the same sequence of events transpired, it seems to me that we need a general explanation.
Which I am confident will support my politics.