We are still in a financial pickle triggered by the subprime mortgage meltdown.
The mess is becoming clearer. What isn’t clear is—what the hell happened, and how did we get here in the first place?
First, the present.
Two excellent articles describe the situation as of 2007.
The Wall Street Journal follows the sorry path of a single mortgage, from a truckdriver in Denver who loses his job and condo and turns suicidal, to his mortgage’s eventual home in the mutual fund portfolio of a Tennessee financial hotshot gone cold (though probably not suicidal).
Fortune dissects just one of those subprime funds—Goldman Sachs’ GSAMP Trust 2006-S3, a fund of second mortgages—tranche by tranche, detailing the infectious rot of loans gone bad (and defining “tranche” on the way).
At every step these articles read like the Wild West. No controls; no principles; no overview. You can’t help but ask: how did things get this way?
Set the wayback machine to 1995, Sherman.
To a Harvard Business School Press book, The Global Financial System: a Functional Perspective. It is a functional analysis of the financial system by a Who’s Who of financial academia, including Dwight Crane, Kenneth Froot and Robert Merton.
Crane’s Chapter (“The Transfer of Economic Resources”) describes the history of mortgage securitization. Mortgage loans in most countries were historically made by local institutions. This system encouraged risk management—the local S&L knew its borrowers, and owned its loans. It also meant high loan cost, and immobile funds. A low risk, but high cost, system. A classic cottage industry.
Securitization aimed to reduce cost by increasing efficiency—the classic Western formula for economic development. Interestingly, the government—via Fannie Mae and Ginnie Mae—led the way to securitizing mortgages; they set product standards and made guarantees, and became models for private-sector securitization.
Securitization made mortgage lending a national, even global, market. It made more money available, at lower rates, with greater accessibility. It did what it set out to do.
The shift was dramatic. Local savings institutions held 58% of outstanding US mortgages in 1950—that dropped to 15% in 1993. And by 1993, 63% of mortgages were securitized.
But what about risk?
The local model kept risk low through long-term relationships. Lenders knew borrowers personally, over a long time; lenders kept loans for the long-term; and borrowers owned houses for the long-term.
Here is Crane in 1995, explaining why risk management was under control in a securitized world:
In the modern market a) the criteria for loan approval are easily spelled out in terms of appropriate loan-to-value ratios and other variables; b) criteria used for mortgages to be put in the pool are also clearly specified, and c) there is an audit process that checks for compliance. In addition, the issuer of the securities has an incentive to manage the quality of mortgages put into the pool because d) a good reputation allows future deals to be done…[and] e) the buyers have an incentive to maintain the property since they retain ownership.
Fast forward to 2007.
a. Loan-to-value ratios may have been spelled out, but no one cared—Goldman’s GSAMP trust had an average loan-to-value ratios of 99%;
b. Criteria for loans to be put in the pool is anything but clear;
c. 58% of the mortgages in the GSAMP product were no-doc mortgages. The “audit” process had defaulted to S&P and Moody’s, both of whom were either wildly deluded or denied their responsibility for the role—or both;
d. If Goldman’s incentive to manage portfolio quality was its reputation, then its reputation was being priced awfully low. On the other hand, the short-sale bets Goldman (successfully) made against the very sort of subprime mortgage pools it was peddling looks like pretty powerful incentive. Why invest in managing a reputation when you can lay off the risk just by placing a bet against your own team?
e. Rather than “maintaining” an owned property, many buyers were in it to flip it.
How far apart had things fallen? The average equity held by the borrowers in Goldman’s GSAMP Trust 2006-S3 was 0.71%. That’s right—the ratio of the loan to the value of the underlying property was 99.29%. Picture you taking out a 99.29% mortgage.
As Fortune puts it:
A total of 93% [of GSAMP] was rated investment grade. That’s despite the fact that this issue is backed by second mortgages of dubious quality on homes in which the borrowers (most of whose income and financial assertions weren’t vetted by anyone) had less than 1% equity and on which GSAMP couldn’t effectively foreclose.
The price we paid for efficient markets was higher risk. Was it a good deal overall?
Very possibly so, even given the dislocations; remember the S&L crisis? Mortgage securitization helped replace that mess.
But good grief, couldn’t we have done better? Yes. In retrospect, the system assumed that better information flow would enable trust.
It didn’t happen.
Financial theorists tend to describe capital markets in terms of information. If you can package information, subject it to standards and audits and controls, then you can “trust” it.
That emperor is looking naked. Trust is not about information alone. It is about people, and in particular about people’s relationships to other people.
That’s what made the old system work. For those who think trust is scalable through information alone, the subprime crisis should be sobering food for thought.
Data won’t kill trust. But a steady diet of data alone will starve trust soon enough.