Tom Daula points us to this article, “Mortgage-Backed Securities and the Financial Crisis of 2008: A Post Mortem,” by Juan Ospina and Harald Uhlig. Not our usual topic at this blog, but then there’s this bit on page 11:
We break down the analysis by market segment defined by loan type (Prime, Alt-A, and Subprime). Table 5 shows the results and documents the third fact: the subprime AAA-rated RMBS did particularly well. AAA-rated Subprime Mortgage Backed Securities were the safest securities among the non-agency RMBS market. As of December 2013 the principal-weighted loss rates AAA-rated subprime securities were on average 0.42% [2.2% for Prime AAA same page]. We do not deny that even the seemingly small loss of 0.42% should be considered large for any given AAA security.
Nonetheless, we consider this to be a surprising fact given the conventional narrative for the causes of the financial crisis and its assignment of the considerable blame to the subprime market and its mortgage-backed securities. An example of this narrative is provided by Gelman and Loken (2014):
We have in mind an analogy with the notorious AAA-class bonds created during the mid-2000s that led to the subprime mortgage crisis. Lower-quality mortgages [that is, mortgages with high probability of default and, thus, high uncertainty] were packaged and transformed into financial instruments that were (in retrospect, falsely) characterized as low risk.
OK, our paper wasn’t actually about mortgages; it was about statistics. We were just using mortgages as an example. But if Ospina and Uhlig are correct, we were mistaken in using AAA-rated subprime mortgages as an example of a bad bet. Analogies are tricky things!
P.S. Daula adds:
Overall, I think it fits your data collection/measurement theme, and how doing that well can provide novel insights. In that vein, they provide a lot of detail to replicate the results, in case folks disagree. There’s the technical appendix which (p.39) “serves as guide for replication and for understanding the contents of our database” as well as (p.7) a replication kit available from the authors. As to the latter, (p.15) footnote 15 guides the reader to where exactly where to look for the one bit of modeling in the paper (“For a detailed list of the covariates employed, refer to MBS Project/Replication/DefaultsAnalysis/Step7”).
The paper you cite is a good one, but it depends on the time at which you value the bet. From 2004 (or whenever issued) to 2013, the AAA subprimes kept paying their interest, Ospina-Uhlig find, but if these are based on 30 year mortgages, the show ain’t over yet. In 2007, prospects looked good. In 2009, prospects looked bad. In 2013, prospects look good again.
A big part of the financial crisis was that these securities are hard to value— the math is complicated and the data for the parameters for extrapolation was based on just a few years with no recession. Hard tho it is to believe, it seems most people on Wall Street didn’t realize the supposed expert quants doing the valuation didn’t really know what they were doing— and the realization came suddenly.
It shouldn’t be hard to estimate the losses (gains?) per class with that data over the time period. If the losses are highly concentrated in time in AAA, for example, that has a potential policy implication that says the opposite of the paper’s purpose, which is that ratings did not work during the loss periods. My first reaction – and I don’t know if they’re right or wrong having just glanced at the first few pages – was that you can send a regiment into combat and it may lose half its strength and then you refill the regiment with new troops so you have the same name and number but different people and thus different fitness, bad and good, for combat. This makes even better sense if you say this regiment is great but it’s actually half full of relatively untrained soldiers who’ve never seen combat or worked in units in combat: yes, the regiment continues and at some point it may well be battle hard again but not when they’re getting their butts kicked.
Good point.
The paper calculates ex-post losses, so these results only show that the turnaround came quickly enough to keep the losses from eating into the AAA layers. Table 4 shows that everything below AAA got hammered. (It helps to remember that the different securities are just different tranches of the same mortgage pools. So the lower rated securities take the hit first and the losses eat their way upwards in the structure over time. It just so happens that the turnaround came before the AAA tranches got hammered.)
But, in 2009, it wasn’t at all clear that the damage was going to stop at the AAA/AA border, so the AAA tranches were quite risky in 2009. This is why the markets suffered such severe convulsions. Investors couldn’t tell if many institutions were economically solvent because they couldn’t tell how much the market value of their RMBS had fallen. Also, many market mechanisms that relied on treating AAA RMBS as nearly risk-free were thrown into turmoil when AAA’s became risky.
There are market losses and then there are underlying security losses. They’re obviously not the same. A stock is a security that continues – until and unless the company is bought or goes under or goes private, etc. – so you can treat it as a limited life security but situationally and thus with caveats. (E.g., I could have bought Apple at $6 a share because people expected it would go out of business so you can treat that as a value crash, as a rating, etc. but it’s not truly a time limited security.) Any kind of mortgage related security is specifically time limited so you can value it at projected end and then at actual end when and if you actually get the money back (or the property back). The value in one without the other is what?
Good points, but more a question of market valuation and liquidity, not credit losses. The primary point of the paper is that AAA subprime outperformed relative to prime (0.42 vs 2.2 percent losses). These are unexpected losses. Both classes of AAA securities had subordination that was supposed to reduce their risk to essentially zero over a 5 year period. Obviously the subordination on the subprime was larger than the prime due to the higher risk.
Therefore, they argue the crisis is better characterized by the larger unexpected losses on a larger pool of borrowers in the prime market than the losses in the subprime market.