Arnold Kling  

Time to Re-think Securitization?

Why is College Expensive?... A Snark on Health Care Cost Re...

The New York Times reports,

On Wall Street and in bank boardrooms, the question of whether investors can force banks to buy back, or "put-back," the bad mortgages to the banks that sold them is dominating the debate and worrying analysts, money managers and banking executives.

The story says that the Federal Reserve, which now owns a lot of mortgage securities, may be joining the put-back crowd.

When mortgage lenders sell loans to Freddie Mac and Fannie Mae, they make "reps and warrants" (representations and warrants) that the loans comply with rules and guidelines. If a loan is later found not to be in compliance, the lender is required to buy it back.

This is just one of the many mechanisms within the securitization process that are designed to address the principal-agent problem. We are seeing evidence of this problem everywhere, including in the foreclosure mess.

Since the financial crisis began, I have been inclined to think that we need to go back to traditional mortgage lending and stop trying to use government policy to sustain the securitization model. Each month, the principal-agent costs seem to mount, adding to the disadvantage of securitization.

To me, it is obvious that the overall costs of the securitization model are high relative to the costs of traditional lending. But what is obvious to me is unthinkable in the two places that matter--Wall Street and Washington.

Comments and Sharing

COMMENTS (16 to date)
Bill N writes:

Are there some common issues to the mortgage secularization fiasco and medicare/medicaid?

One that comes to mind is the importance of distributed private information in dealing with inherently complex domains. Secularization distilled the decision process down to a few (not so simple?) rules. Many of the fixes seem to be little more than patching the rules without leveraging distributed/local knowledge. (e.g. as opposed to requiring the loan originators keeping and maintaining the mortgage)

In medical care, is this the trend? Will it get better or worse? Surely benefits of procedures can vary widely, a mean, median, or mode is not sufficient information for an individual decision. How are costs allowed to vary?

Peter Finch writes:

Bill N:
Did you really mean "secularization"? I assume not.

Bill N writes:


Thanks for that catch. Ouch, spell checkers.

Peter Finch writes:

No problem. If I was in charge of an economics blog spellchecker, I'd make sure "securitization" was in there. :)

To the original post: so why is roll-back the answer? I mean, caveat-emptor, right? The entities that bought MBSs were not widows and orphans. They were generally sophisticated.

Arnold Kling writes:

When, say, Freddie, buys a mortgage from Sleeze-E Mortgage Banking Corp., the contract says that the loan was underwritten to certain standards. Freddie does not at that time check to see if it was underwritten to those standards, because that is too expensive. Later, if it does a random audit on that loan, or if the loan goes bad quickly and as a result they decide to do an audit, that audit may turn up violations of the standards. At that point, whether the loan is performing or not, the contract says that Freddie can require Sleeeze-E to repurchase the loan.

My guess is that some provision like that exists in all securitization contracts. Otherwise, you would have to re-underwrite every loan at every step in the process.

Peter Finch writes:

Yeah, poor Freddie. They obviously didn't know what Sleeze-E Mortgage was doing. They had no incentive to ignore the flaws in the process.

I'm not saying they don't have a legal out here. I am not a lawyer.

Sleeze-E mortgage is clearly not blameless in your hypothetical, but neither is Freddie.

Peter Finch writes:

I'll put it another way and try to soften my tone:

In my experience, these things are not decided based on how a normal human would read a single line in a contract. There's often enough blame to go around, and a judge, jury or arbitrator is going to have to be convinced of the reasonableness of a solution, or a settlement that is not-too-disagreeable to both parties will have to preempt an imposed solution.

These big disputes are not normally resolved by "Perry Mason Moments" in which one key fact is uncovered in court.

Hyena writes:

I think you're right.

If securitization was really the way to go, you'd think that diligence would get progressively streamlined and loan documentation easier to check. That's not what happened at all, the principal-agent gap widened substantially instead.

Unless, of course, there's a legal barrier here as there is with titles. Certain aspects of portable diligence would likely violate privacy rights, especially once shaped into a database of complete credit information.

Steve Sailer writes:

Securitization tends to turn into Secretization.

Gary Rogers writes:

If there are no regulations preventing mortgages from being sold across state lines, is there any reason for securitized loans? If I remember correctly, Freddie Mac was created solely to create securitized loans to solve the problem of balancing states that were growing fast with other states that had investment money. At that time banking could not cross state lines.

Things like this almost always come down to government making some seemingly clever decision that has unintended consequences then patching the solution until it collapses. Hidden in the mess somewhere is also the abilitiy to hide "affordable loans" which have also proven to be a problem.

Raja writes:

Securitization is still a valid concept. The principal-agent problems are no worse than anywhere else in society, ie banks/FDIC/govt or public equities vs shareholders. The problem is that our whole society has been laid waste by the corrupt and the incompetent.

Joe In Morgantown writes:


Securitization is a neat concept. But the agent problems are worse.

Auditing all the loans is expensive; having ever party in the assignment chain perform them is prohibitive.

Worse, the principal-agent problems seem to be fundamentally harder here. See _Computational Complexity and Information Asymmetry in Financial Products_ by Sanjeev Arora et al.

Push-backs seem to be a reasonable partial work around. In practice the non-agency push-back procedures are more cumbersome and may not be practical in the current cases. Never mind cases where Countrywide the servicer would have to push-back onto Countrywide the packager.

I don't think securitizations are practical going forward without trustworthy packagers. I believe the existence of trustworthy packagers is, to date, an unsolved problem.

[Link replaced with original source--Econlib Ed.]

Joe Mamma writes:

This article, and some comments, are typical examople of a quasy-intellectual Ludite movement against securitization. Capitalism, as the most elaborate system of thievery in the history of mankind is fa ramework where all the frauds are happening. Yet, nobody talks about capitalism.

Carl The EconGuy writes:

I don't think the securitizers are at fault here. The problems were not necessarily securitization; rather, there were two antecedent faults. One, local agents were clearly overvaluing properties for loans -- fraudulently, in order to get commissions on loans. That problem lay with local banks not checking on these evaluations closely enough, and not with those who bundled mortgages into securities for later sale. Two, the rating agencies that have a legislated monopoly on evaluating the risk of the mortgages. Securitizers relied on the rating agencies for their bundling of mortgages into tranches. Therefore, it's not clear at all that the problem is the securitization. Rather, it's the lack of attention given to the fraudulent evaluations at the local level, both by loan originators and rating agencies. Yes, in a chain of decisions like this there will be principal-agent problems, but it does not strike me immediately as unreasonable that securitizers would rely on local real estate agents doing the evaluations, on the banks that accepted these, and on the rating agencies that validated them. The farther upstream you are from the source of the problem, the lower your responsibility ought to be. Even without securitization, tons of mortgages would have been sold to people who should not have bought them, even at steady-state prices, never mind the inflated prices fed by the speculation mania. The ultimate responsibility lies at the first and perhaps second links of the chain: local agents, originating banks, and rating agencies. I think securitizers will have a good legal defense if they are sued for take-backs -- they relied, as they traditionally have, on the honesty and competence of loan originators and rating agencies.
One of the features of the crisis was that many loan-originating banks kept investing in securitized assets -- so they bought into their own folly and never looked into the risks they took. If those with the strongest ability to look into the values of the properties and the abilities of borrowers to carry the debt burden failed, isn't it there we should begin to look for remedies? in fact, if I were building a legal defense for the securitizers against take-backs, I'd use this as an argument for why they relied on local banks -- the locals were themselves buying the overvalued securities.
The securitizations only led to more capital being available for lending, which federal agencies and Congress (Barney Frank and the Doddster from Connecticut) thought was just dandy, exactly what they wanted. But the real problem feeding the frenzy was fraud, ignorance, and excessive risk taking at the primary lending level. That's not a principal-agent problem, that's outright incompetence in mortgage origination. Bad local bankers will cause havoc -- and that's the main lesson here.

Carl The EconGuy writes:

Another thought struck me. While I stand by the conclusion I reached above that the securitizers are not really the primary source of the problem, they will nevertheless get hit, and will have to take back a lot of things they've sold. The reason is that they are the last guys standing, since all the primary culprits are pretty much out of assets. Deep pockets, and all that.

Eric Rasmusen writes:

The reasoning is shaky. Banks loans to businesses have even bigger principal-agent problems. By the same logic, shouldn't we ban them?

What government policies encourage securitization? Maybe the implicit promise to bail out Fannie Mae? I wouldn't count laissez faire inaction, such as allowing banks to securitize, as a policy.

Comments for this entry have been closed
Return to top