He writes An Epistemology of the Financial Crisis, appearing in the current issue of Critical Review.
Subprime mortgages retained by U.S. banks for their own portfolios performed at least as badly as those they securitized for sale to others….A study by Wei Jiang, Ashlyn Nelson, and Edward Vytlacil [various versions available] examined one large lender and concluded that mortgage “loans remaining on the bank’s balance sheet are, ex post, of owrse quality than sold loans.” They concluded that RMBS [residential mortgage backed securities] investors had information advantages over banks. This is the very opposite of the view that banks sold the worst loans to unsuspecting third parties.
Robb’s theme is that the crisis was caused by imperfect knowledge rather than greed/adverse incentives. I have many comments
1. This same theme may be found in Jerry Muller’s similarly titled Our Epistemological Depression and my The Financial Crisis: Moral Failure or Cognitive Failure?, neither of which hare cited by Robb.
2. It is no surprise that Robb did not see the articles by Muller or me, nor is it likely that many people will see Robb’s. The mainstream narrative is available to everyone, while our narrative has been relegated to the most obscure publication outlets.
3. Robb writes,
One clue to what went wrong comes from a study that Fitch conducted [may be found here] on borrowers who defaulted within six months of taking out a mortgage. The study looked closely at 45 “early payment defaulters” from 2006. [It] found that 66 percent of them committed “occupancy fraud,” falsely claiming that they intended to occupy the home.
There are some forms of fraud that are often the fault of the lender, and the borrower is relatively blameless. Overstating borrower’s income is an example. But when it comes to occupancy fraud, you have to blame the borrower, and for lenders it is one of the most difficult forms of fraud to detect prior to making the loan. However, blaming the borrowers runs counter to the conventional narrative.
4. Contrary to what I have written, Robb argues that lenient risk-based capital rules for highly-rated mortgage securities were not implemented soon enough to be implicated in the financial crisis. On the Recourse Rule, a 2001 regulation that some of us believe encouraged subprime securitization, Robb writes,
While on the margin the Recourse Rule encouraged investment in highly rated ABS [asset backed securities], the incentives were not so great as to justify holding these securities unless banks thought they were safe.
5. Robb writes,
An unambiguous regulatory failure was the decision to allow Lehman Brothers to fail. The market expected the U.S. Treasury to cobble together a last-minute rescue over the weekend….largely because…Bear Stearns, had been rescued six months earlier by being absorbed into JPMorgan Chase. Lehman was 25 times large4r than Bear Stearns and far more interconnected…the Securities and Exchange Commission had no plan for an orderly transfer of clients’ assets…institutional clients with claims over $5000,000 had to wait until the summer of 2013.
6. According to Robb, the financial community never expected house prices to decline. But I would like to point out that you cannot just look at what investors were expecting on average. Instead, think of investors as assigning probabilities to various paths for house prices. I think it is fair to say that investors under-estimated the probability of a large decline in house prices. However, a sophisticated investor would not have assigned a zero probability to a path in which house prices declined.
7. Robb views post-crisis risk aversion as a major source of problems. He provides examples of assets that in hindsight were ridiculously undervalued by investors.
during a crisis, they learn to be skeptical of the probabilities, no matter how those probabilities are presented.
All in all, it is one of the most provocative essays I have read on the financial crisis.
I don’t have any reason to doubt cognitive failure among the regulators, but I’ve yet to see any evidence that moral failure wasn’t a major issue at banks and probably the ratings agencies as well (in addition to cognitive failure). You write in your Harvard Law paper that:
“[t]he Managing Director of the surveillance area for RMBS did not believe loan level data was necessary and that had the effect of quashing all requests for funds to build in‐house data bases.”20 This position is consistent with the cognitive narrative.
The way I look at it, the MD’s “belief” about loan level data doesn’t rule out either explanation. If he was acting in response to bad incentives, then he still would have said he didn’t believe it needed to be built.
There are a few things that go on within the industry that aren’t picked up in any of the work I’ve seen on moral hazard. First, there are people within bank risk management teams who could see what was coming but knew their superiors would never deliver their message. These folks tend to get weeded out, either voluntarily because they don’t like being part of something they don’t believe in (I’ve interviewed and hired such people), or involuntarily because they don’t give the “right” answers. Second, people responding to bad incentives will avoid any of the work or information that might give them the answer they don’t want to hear. This is the willful ignorance (and plausible deniability) piece that might make it sound like a cognitive error but, more fundamentally, it’s a moral error. And even without willful ignorance, you still won’t hear the moral failure story when thing go wrong because people won’t admit to that – they’ll claim they didn’t know.
I should add that this doesn’t make me inclined to support govt. intervention per your conclusion, although I understand your concern. It seems to me that Admati & Hellwig (Banker’s New Clothes) have the right idea – trying to regulate banks on the asset side of the balance sheet is almost impossible, but that would be okay if we could somehow get back to capitalization ratios of 30, 40, 50% on the liability side.
Maybe not precisely on point, but I may not get a better opportunity to ask this question.
As I understand your theory of the crisis, it depends on the idea that one must explain why sophisticated investors bought securitized mortgages rather than traditional ones, given the agency problems with the former, and you refer to the greater profitability of the former under the capital regs. In a recent Econtalk podcast, Calvo seems to suggest an alternative explanation: securitized mortgages, unlike traditional ones, were thought to be almost as liquid as cash, and unlike cash, they provided significant income.
The idea seems obvious enough that you have probably addressed it, but if so I don’t remember the response.
funding mortgages with deposits also makes them liquid
Arnold: If you read the study by Jiang, Nelson and Vytacil, you will see that they DO NOT conclude that investors knew more than lenders. They very rigorously show that the reason that bank loans are worse than loans sold to investors is that many of these “bank loans” are actually intended for sale. But, due to early defaults on several such loans, selling becomes impossible. Thus, many loans that are stuck on balance sheet of banks are indeed of very bad quality; that is why they could never be sold. Comparing them to loans that were sold is like comparing apples and oranges!!