I keep meaning to do a links round-up, so before I forget, reviews of Reckless Endangerment, the book on the GSEs (that cover similar ground as the Wallison/Pinto argument) by Gretchen Morgenson and Josh Rosner, are starting to happen. Jeff Madrick at NYRB’s Blog, Bob Kuttner at The American Prospect and Kevin Drum all have interesting, very critical, things to say about the book.
Meanwhile, Tyler Cowen asks How much did Fannie and Freddie cause the financial crisis?, trying to reground the debate over the GSEs in firmer soil with five points. Many others are answering this question by looking at the collapse moment; I want to focus on the origination of bad mortgage debt and the bubble itself. I pinged David Min to see if he wanted to respond, and I throw his arguments in with my own below.
Cowen: “1. It is not denied that the mortgage agencies were guaranteeing about half of all U.S. mortgages right before the crisis…”
David Min: “This conflates outstanding mortgages vs. annual mortgage originations (sort of like a snapshot vs. a video). The GSEs were guaranteeing nearly half of all mortgages in the late 1990s and early 2000s. But from 2002-2005, they saw a fairly precipitous drop in market share, going from about 50% to just under 30% of all mortgage originations. Conversely, private label securitization [PLS] shot up from about 10% to about 40% over the same period. This is, to state the obvious, a very radical shift in mortgage originations that overlapped neatly with the origination of the most toxic home loans.
Moreover, this point also papers over the fact that PLS loans have defaulted at over 6x the rate of GSE loans, as well as the fact that private label securitization is responsible for 42% of all delinquencies despite accounting for only 13% of all outstanding loans (as compared to the GSEs being responsible for 22% of all delinquencies despite accounting for 57% of all outstanding loans).”
I think that’s right. So far the debate has been focused on the idea of “other high-risk mortgages,” mortgages with an LTV > 90, for instance. The real split should be between PLS, the brand new private-market way of putting together the mortgage market, and non-PLS.
For #2 Cowen brings up leverage.
Min notes: “PLS was allowing far more leverage than the GSEs, which seems to contradict this claim. I do agree that absent the GSEs, we would have seen fewer 30 year FRMs…FRM have performed exponentially better than ARMs, and 30 year FRMs have performed far better than any other type of product.”
I’d say if we want to understand how leverage got into the mortgage market specifically, we should look at subprime as well as second-liens and other junior-lien mortgage claims, and what arguments were given for supporting them. My favorite document of what that argument looked liked at the time is something like Charles Calomiris and Joseph R. Mason 1999 AEI paper: High Loan-to-Value Mortgage Lending: Problem or Cure? That is all about how people should lever up as much as they can as a signaling device, as the junior claims will disrupt bad debt workouts and act as a straightjacket on credit writedowns. This is explicitly not an affordability mission, but instead a very conservative vision of bankruptcy law and debtor-creditor relationships. We are living through the consequences of that call. This extra option to re-lever to consistently re-bid against property appreciation goes hand-in-hand with dynamics of a bubble.
For #3, Arnold Kling has clarified his previous thoughts on the GSEs here in a way that I think works better with the data.
Tyler’s #4, “Following the crisis, banks recovered and paid back virtually all of their bridge/bailout.”
Min brings up: ”The GSEs are essentially single-purpose entities (particularly with the heightened restrictions on their direct investments following the crisis). They securitize conforming mortgages, and guarantee timely payment of interest and principal on those securities. They do not engage in derivatives dealing, market making, trading, or any of the other activities that have generated the bulk of bank profits since the 2008 crisis…GSE loss severity on its mortgage exposure (which is its entire business) is exponentially better than that of their private competitors.”
Also another point I don’t see brought up enough is that the GSEs had political pressures to purchase private-label mortgages in 2007 as the credit market was freezing up, in a desperate attempt to unlock it.
This argument, and the graph above, is developed in Roosevelt Institute senior fellows Rob Johnson and Tom Ferguson’s Too Big to Bail: The ‘Paulson Put,’ Presidential Politics, and the Global Financial Meltdown Part II. One could easily assume that they were buying the worst loans here, aggregating losses on the GSEs. I’d love to see much more investigations into this (more on that paper the the situation here).
Meanwhile, research moves on. Mian and Sufi’s The Consequences of Mortgage Credit Expansion: Evidence from the U.S. Mortgage Default Crisis is likely to be with us for a while as a research standard, and it is telling us to look at the securitization chain.