The Recourse Rule

I want to know if this is true:

Under the Recourse Rule, an AA- or AAA-rated asset-backed security, such as a mortgage-backed bond, received a 20-percent risk weight, compared to a zero risk weight for cash and a 50-percent risk weight for an individual (unsecuritized) mortgage. This meant that commercial banks could issue mortgages—regardless of how sound the borrowers were—sell them to investment banks to be securitized, and buy them back as part of a mortgage-backed security, in the process freeing up 60 percent of the capital they would have had to hold against individual mortgages. Capital held by a bank is capital not lent out at interest; by reducing their capital holdings, banks could increase their profitability.

Is it true that this is what the rule change required and is it true that this is how banks responded? Also, is there a way to quantify the impact of this rule change (assuming this paragraph is a correct summary of what happened)?

3 thoughts on “The Recourse Rule”

  1. Those are certainly the old Fed requirement that was incorporated into the first Basel Accord, so I have no reason to believe they were not continued in Basel II compliance. (Which does leave the question of why the pure speculative bubble only dates from ca. 2004 at the earliest–that hints at checking the 2003 Tax Act, not the 2001 ruling.)

    So I’m inclined to argue that isn’t “what the [2001] rule change required” so much as standardization of the previous process.

    On a side note, one of the better reasons for the difference between capital requirements for a single loan and a securitized package of loan–other than the obvious–was that securitization was only a significant business for Fannie, Freddie, and Ginnie. And any loan they would reject was, rather definitionally, riskier than one they would buy. (This was in the days before there was significant private securitization, and long before Daniel Mudd decided to bankrupt Fannie by chasing bad paper.)

    There’s probably a legitimate argument that a nonsecuritized loan isn’t likely to be 2.5x as risky as one that’s in a Fannie 8 or even a Ginnie 7.5, but it’s clearly riskier and some regulators remember all those TX loans that were securitized by the value of the oil in the land–which turned out to have no oil in it.

    The ability to overcollateralize leads to dubious “AA” paper, but a variation of that practice has been standard in the ABS market since at least the late 1980s (e.g., take $100MM worth of paper that yields 2% and market it as a $25MM 8% note).

    So the only “solution” would have been to tier securitized paper (e.g., GSE’s can be AAA, nothing else any better than AA except in extreme cases)–not really workable in the abstract. Or to have all the regulators and raters be people who bought a house on Lon Gisland in 1989 and hadn’t been able to move in the following 15-20 years because house prices never got back up to the level they paid.

Comments are closed.