OK, Maybe Poor People Are the Problem

Commenter Voltaire may not agree with what I had to say on the topic of scapegoating poor home buyers over the weekend, but he raises a jumbo conforming question: Could the not-as-bad-as-advertised performance of subprime loans (relative to prime loans) be explained by fraud in Fannie Mae's accounting of prime loans?

My reason for not agreeing with Tim Cavanaugh's post is that apparently, since 1993, Fannie Mae and Freddie Mac have been increasingly systematically reporting as prime loans loans that were in fact subprime. It was the corruption of these GSE's that makes any statistical claims of the sort [Stan J.] Liebowitz makes suspect from the beginning.

The gist: Edward Pinto, a consultant who worked as chief credit officer at Fannie Mae during the Reagan Administration, is publicizing an enormous post-1992 increase in Community Reinvesmtent Act-guaranteed loan volume. Here's The Wall Street Journal explaining how Pinto… 

found that from the time Fannie and Freddie began buying risky loans as early as 1993, they routinely misrepresented the mortgages they were acquiring, reporting them as prime when they had characteristics that made them clearly subprime or Alt-A.

In general, a subprime mortgage refers to the credit of the borrower. A FICO score of less than 660 is the dividing line between prime and subprime, but Fannie and Freddie were reporting these mortgages as prime, according to Mr. Pinto. Fannie has admitted this in a third-quarter 10-Q report in 2008.

An Alt-A mortgage is one in which the quality of the mortgage or the underwriting was deficient; it might lack adequate documentation, have a low or no down payment, or in some other way be more likely than a prime mortgage to default. Fannie and Freddie were also reporting these mortgages as prime, according to Mr. Pinto.

It is easy to see how this misrepresentation was a principal cause of the financial crisis.

I'm not sure where that admission is in Fannie's third quarter 10-Q for 2008 [pdf], unless it's this reference:

Net investment losses were $1.6 billion in the quarter, compared with losses of $883 million in the second quarter. The third-quarter loss was driven by other-than-temporary impairments of $1.8 billion recorded primarily on private-label securities back by Alt-A and subprime mortgages, and reflected a reduction in expected cash flows for a portion of our private-label securities portfolio.

Anyway, Pinto claims about 19 million loans listed as prime in the GSEs' portfolio are actually high-risk, Option ARM, Alt-A and other junk. You can read him on the miscategorization of subprime loans at Slide 21 here, and you can watch him present his thesis at the Cato Institute's Apparatchik Fashion Show here.

I would not go to court on this evidence, but it's pretty intriguing. Pinto's broken-out numbers do not make a strong case for the CRA as the primary culprit in this Murder On the Orient Express plot wherein (spoiler) everybody is guilty. (If you want to read a very unpersuasive exoneration of the CRA by BusinessWeek's Aaron Pressman, here it is.) However, it would either shake or restore your confidence in Fannie's million-dollar men if there turned out to be this much bad medicine hidden in the mortgage giant's boot.

But what of poor Stan Liebowitz, whom we left arguing that mortgage resets, not borrowers' apparent quality, shaped the collapse? Liebowitz drew his data from the Mortgage Bankers Association, a private group of good reputation that, I hope, would not just gobble up whatever lies and propaganda the government gives out. Stay tuned.

And since the original had a cool clip, the rebuttal gets one too:

Comments

[Pendulum]

How dare you defame Countess Elena Andrenyi by implying her guilt in the death of Mr. Ratchett!

[Suki]

Good Morning reason!

[healthscarequotes]

i think willow is using too much magic in this video, who's with me?

[Brian Sorgatz]

By the way, that melody was sung with lyrics in English at the beginning of But I'm a Cheerleader.

[Hobo Chang Ba]

That's the remake "Chick Habit" by April March.

[Fluffy]

Two errors in this post:

First of all, in the context of agency loans it does not make any sense to set an arbitrary FICO cutoff of 660 and declare loans below that FICO "subprime".

The agency's definition of what made a loan investment grade did not depend on FICO scores, for the simple reason that the agency system predates the existence and use of FICO scores.

Agency loans were defined as investment grade based on the absence of mortgage lates, foreclosures and bankruptcies in the borrowers' credit histories within defined time frames, and based on the ability to document sufficient qualifying income to meet the ratio guidelines, sufficient assets to meet down payment and reserve guidelines, etc. "Back in the day" as long as you met those minimum guidelines, the kind of credit characteristics that would raise a credit score into the high 700's were just irrelevant. Credit reports were read negatively and not positively.

This means that during whatever period of time Voltaire thinks that Fannie was underwriting "properly" or "non-fraudulently" they were, in fact, purchasing very large numbers of loans that would have produced FICO scores below 660 had the scoring systems then existed. Hell, based on the credit use patterns of the 1970's I bet if you rescored the entire fossil agency portfolio from that time the average score would be below 660, since most borrowers did not have the opportunity to develop the extensive revolving credit histories they would need to achieve higher scores - the credit products just weren't that widely sold back then.

And with regard to the other half of the equation - borrowers with low documentation: when credit scores were introduced to agency underwriting, they were initially not used for risk-based pricing to the extent they are now. They were used to lower the documentation burden of certain high-score customers - a customer with a 740 score, for example, might be asked for a single paystub and a single bank statement, rather than 30 days of stubs, W-2's, tax returns, and a 90 day asset history. This was sold mainly as a convenience and not as an Alt-A guideline. Over time, as originators became used to the existence of the various low-doc categories, and as the agency automated underwriting systems became more and more forgiving and asked for less and less documentation, originators began steering Alt-A people into agency products, because if the automated underwriting engines weren't asking for much income documentation you could "game" that to have it be an Alt-A loan "in effect". So there's some validity to this half of Voltaire's point - but every last scrap of data Fannie had before 2006 said that high credit score customers always paid on time regardless of what the documentation burden placed on them was, so they had every reason to think that these loans were "prime". This was a data anomaly that was the fault of the Federal Reserve, though.

[Some Guy]

but every last scrap of data Fannie had before 2006 said that high credit score customers always paid on time regardless of what the documentation burden placed on them was, so they had every reason to think that these loans were "prime".

I'm sure compulsive gamblers always pay back their markers when the dealer busts every hand, too. That doesn't make it wise to lend to them in general.

And certainly doesn't excuse the fact that the government shouldn't be guaranteeing mortgages anyway.

[Fluffy]

That's very true. I'm just addressing the specific technical point that has come up between Voltaire and Tim.

So everything I'm saying here includes the standard libertarian disclaimers.

I participate in the argument over the proximate causes of the real estate disaster mainly because I believe the blame lies squarely on the Federal Reserve and its loose monetary policy after 2001, with some of the blame being shared with the Bush administration for their loose fiscal policy over the same period.

Every argument that seeks to direct blame away from these two entities amounts to blaming the victim. And the people who want to blame Fannie and Freddie are only doing so in order to be able to place ultimate blame on "speculators" [the people who bought mortgage securities underwritten to these standards] and "predatory lenders". Because the state always has to find private actors to scapegoat for its own errors. And I want to argue against that, in my own small way.

[robc]

Not entirely true. I agree with you that primary blame lays with the Fed. But I blame F&F also, not to place blame on "speculators" but on congress for creating them in the 1st place.

[Fluffy]

I agree that the agencies should not have been created, and that they had a distorting influence on real estate investment patterns and the economy in general.

They're also part of the great "Automobile society conspiracy" I am always getting on about, because they are one of the many ways in which the state chose to subsidize suburbia.

But if we're talking about the specific events of the real estate bubble of 2001 to 2006, then I think the blame lies elsewhere. The agencies are bad ideas, but the precise ways in which they are bad ideas are not necessarily bubble-inducing.

[robc]

They were bubble inducing in that they cause misallocation (I orginally typed that as mallocation and I cant remember the last time I programmed in C) of resources, hence bubblicious. Yeah, not as bubbly as cheap money, but ever little bit hurts.

[Russ 2000]

From what I know, the loose money policy of the Fed is the initial contributor, so I agree with Fluffy on that.

But after that initial contribution, circa 2004 it was killing the returns of Fannie and Freddie and the large institutional investors in F&F were begging the agencies for higher returns and, with the blessing of Congress, started buying more loans of poorer quality than they had in the past.

Simply put, so many loans were of the Alt-A/Option ARM type that the volume of typical mortgages was falling and F&F simply couldn't generate the profits of years past because the supply they typically bought were shrinking.

The blame for F&F going hog wild into buying lower quality loans rest squarely on them, their institutional investors (CalPERS, etc.) and the Fed and Congress. Just because the Fed started the game doesn't mean the GSE's had to blindly follow; F&F decided either on their own or with congressional cajoling that their mandate had changed from buying quality loans to providing large returns.

[Suki]

+1

[hmm]

The credit scoring algorithms have been known to be faulty for some time. They have been getting better, but many produce results that are not in line with the person being scored.

[Johnny Longtorso]

In case you missed it earlier:

The Financial Crisis and the CRA
...Bank of America reported that nonperforming CRA-eligible loans were a significant drag on its third-quarter 2008 income. Its earnings report states: "We continue to see deterioration in our community reinvestment act portfolio which totals some 7 percent of the residential book. . . . The annualized loss rate from the CRA book was 1.26 percent and represented 29 percent of the residential mortgage net losses."...

[shrike]

Your article is bogus.

Over 1/2 of foreclosures come from just four states (NV, AZ, CA, FL - source realtytrac.com) - hardly the "inner-city" red-lined areas affected by the 70's CRA legislation.

Your article claims $4.5 trillion of mortgages were "CRA eligible". What does that mean? Were the rows of Mac-houses in Stockton, Ca "CRA eligible"?

Cite some data on how the CRA disproportionately affected high-growth rural areas instead of red-lined inner city areas please.

Maybe that 32-story condo tower in Naples FL that lost its single tenant recently was CRA-eligible according to the author of your article.

[Johnny Longtorso]

You'll have to talk to Bank of America. You might want to read up on the "70s legislation" in the meantime. It was altered repeatedly.

[shrike]

Ahhh, since the CRA prohibited red-lining, the entire country is "CRA-eligible"!

That is some crafty propagandist!

[Johnny Longtorso]

The agency's definition of what made a loan investment grade did not depend on FICO scores, for the simple reason that the agency system predates the existence and use of FICO scores.

Given the number of new loans and the amount of refinancing that went on, how many existing loans are old loans?

[Fluffy]

Johnny, that's not the point.

I'm talking about how investment grade mortgage securities were and are defined.

The only reason a category distinction exists between prime and non-prime loans is because some loans were eligible for purchase by the agencies and some weren't. The subprime and alt-A product universes were invented to serve customers who could not get agency loans.

If the historical underwriting guidelines of the agencies, before the bubble and before any modern product innovation, accepted a certain type of loan, that type of loan was prime. If your loan met the agency guidelines for purchase in 1985, you were a prime borrower. It's tautological.

The credit score guidelines are a relatively modern invention. You can't point to the credit score spread and gasp theatrically and say, "The agencies started buying loans with credit scores below 660!" like this is some sort of bubble innovation when the vanilla loan guidelines from 30 and 40 years ago would have accepted many loans that would have scored below 660 had credit scoring existed at the time.

[Rich]

A little OT, perhaps, but --

I have been told that a borrower must be in arears 3 months in order to "prove" that s/he "qualifies" for mortgage "assistance".

I haven't (yet) been able to make sense of this assertion. Anyone have a quick pointer?

[Mango Punch]

You didn't link their 10-Q but rather their earnings press release.

10-Q

[Mango Punch]

http://www.sec.gov/Archives/ed.....2e10vq.htm

[Johnny Longtorso]

If your loan met the agency guidelines for purchase in 1985, you were a prime borrower. It's tautological.

The credit score guidelines are a relatively modern invention.

But what % of the loans outstanding were from 1985 and what % were made at a time when 'modern' prime scores were meaningful?

[Fluffy]

Johnny, I would imagine an extremely low % of loans from 1985 were still in existence during the bubble.

I wasn't trying to claim that their book of business was made up of these fossil loans. I'm saying that you have to account for those fossil loans when you're defining what a "prime" loan is and what a "subprime" loan is, regardless of the origination date of the loan in question.

I was just arguing that you can't use credit scores to show that the loans in question weren't prime when the very definition of prime loan, and the definition of investment-grade portfolio, that was employed for decades would have allowed the very same loans to be made.

You could get a Fannie Mae loan [in the days before credit scoring] with recent collections, recent revolving credit lates, tax liens, etc. on your credit report as long as you wrote "explanation letters" and cleared any arrears either prior to or at closing. A credit report with those kinds of dings on it in the credit scoring era will score below 660. But Fannie still wanted them, if the employment, income, asset, and appraisal guidelines were met. Those were "prime" loans. That means they can't be "subprime" loans.

[Johnny Longtorso]

I wasn't trying to claim that their book of business was made up of these fossil loans. I'm saying that you have to account for those fossil loans when you're defining what a "prime" loan is and what a "subprime" loan is, regardless of the origination date of the loan in question.

And I'm saying that if Freddie/Fannie were calling a loan prime, when it didn't meet the standards of prime AT THE TIME OF THE LOAN, then they are open to the charges above.

Your claim is basically that if an agency predates a rule, they get to ignore that rule forever.

[brotherben]

If I read fluffy right he is saying that the definition of prime has changed and now many of the loans once considered prime are now called subprime. The analogy that came to mind is that of autism rates. The guidelines for an autism diagnosis were made much broader and now we have more cases of autism.

[Johnny Longtorso]

Yes, but you don't get to use an older definition of autism when making the diagnosis within the past 10-15 yrs.

[P Brooks]

I participate in the argument over the proximate causes of the real estate disaster mainly because I believe the blame lies squarely on the Federal Reserve and its loose monetary policy after 2001

I think Ben Bernanke thinks you're wrong.

In a speech yesterday, Fed Chairman Ben Bernanke insisted that low interest rates were not the root cause of the most recent real estate bubble. The New York Times says he used his "strongest language yet" in defending the central bank's past decisions and emphasizing the importance of greater financial regulation moving forward.

Reuters

[P Brooks]

And Ben only says that because it's true; it's not like he's trying to cover his ass, or anything.

[Johnny Longtorso]

Lets try it this way. Freddie/Fannie call a loan prime today. Should I care what the definition of prime was in 1978?

[Fluffy]

Well, Voltaire is arguing that because Fannie made loans to borrowers with credit scores below 660, that means they were making "subprime" loans.

He's claiming that any loan with a credit score below 660 is "subprime" because he read that definition in a magazine someplace. Even though that was never the definition of a subprime loan, as an examination of the agency's historical underwriting guidelines would show.

Fannie recently began imposing dramatic pricing penalties on loans with credit scores below 680, which had the effect of driving most of those loans into the FHA program.

But this change is not them saying, "Oh shit you caught us we were making subprime loans!" It's more like them saying, "We're facing new limits on our loan portfolio size, so let's change the type of loan we're acquiring." They're using risk-based pricing to tweak the definition of prime and make it more of a continuum than a cutoff threshold, as it was for decades. But that doesn't mean that the loans they made in the past are now "exposed" as subprime.

[Fluffy]

Here's why this argument has to be shot down, straight out of the mouth of the Head Cocksucker in Charge, Ben Bernanke:

quoteWhat policy implications should we draw? I noted earlier that the most important source of lower initial monthly payments, which allowed more people to enter the housing market and bid for properties, was not the general level of short-term interest rates, but the increasing use of more exotic types of mortgages and the associated decline of underwriting standards. That conclusion suggests that the best response to the housing bubble would have been regulatory, not monetary. Stronger regulation and supervision aimed at problems with underwriting practices and lenders' risk management would have been a more effective and surgical approach to constraining the housing bubble than a general increase in interest rates. Moreover, regulators, supervisors, and the private sector could have more effectively addressed building risk concentrations and inadequate risk-management practices without necessarily having had to make a judgment about the sustainability of house price increases. [Hat Tip P Brooks]

He is either lying about, or is genuinely ignorant of, the fact that the bubble conditions created by Fed interest rate policy were the reason that underwriting standards declined.

People bought subprime securities because the data said those loans were being paid on time. The data said this because anyone who got in trouble in a subprime loan could just sell their property into the bubble, or refinance based on new higher bubble values.

We are being asked to reconfirm into the second most powerful job in the country a man who is either lying about, or ignorant of, the relationship between the macro interest rate picture and individual credit decisions.

[Tim Cavanaugh]

That's Robert Shiller's argument in his book The Sub-Prime Solution: The bubble created sub-prime, not the other way around.

[Johnny Longtorso]

They're using risk-based pricing to tweak the definition of prime and make it more of a continuum than a cutoff threshold

I.E. they changed the definition of 'prime' to make what they were doing look better when they were doing it.

[Fluffy]

No, they didn't. It was lower before.

If it was lower before there was a bubble, by motherfucking definition it can't possibly have been the source of the bubble. How fucking hard is that to understand?

[Fluffy]

Sorry, trying that again:

Here's why this argument has to be shot down, straight out of the mouth of the Head Cocksucker in Charge, Ben Bernanke:

[i]What policy implications should we draw? I noted earlier that the most important source of lower initial monthly payments, which allowed more people to enter the housing market and bid for properties, was not the general level of short-term interest rates, but the increasing use of more exotic types of mortgages and the associated decline of underwriting standards. That conclusion suggests that the best response to the housing bubble would have been regulatory, not monetary. Stronger regulation and supervision aimed at problems with underwriting practices and lenders' risk management would have been a more effective and surgical approach to constraining the housing bubble than a general increase in interest rates. Moreover, regulators, supervisors, and the private sector could have more effectively addressed building risk concentrations and inadequate risk-management practices without necessarily having had to make a judgment about the sustainability of house price increases. [/i][Hat Tip P Brooks]

He is either lying about, or is genuinely ignorant of, the fact that the bubble conditions created by Fed interest rate policy were the reason that underwriting standards declined.

People bought subprime securities because the data said those loans were being paid on time. The data said this because anyone who got in trouble in a subprime loan could just sell their property into the bubble, or refinance based on new higher bubble values.

We are being asked to reconfirm into the second most powerful job in the country a man who is either lying about, or ignorant of, the relationship between the macro interest rate picture and individual credit decisions.

[Fluffy]

Why the fuck aren't tags working anymore?

[ChrisH]

I blame threads. Or Canada.

[Johnny Longtorso]

No, they didn't. It was lower before.

The article says they were reporting subprime loans as prime. A lower definition of prime in 1978 or so has nothing to do with that.

"In general, a subprime mortgage refers to the credit of the borrower. A FICO score of less than 660 is the dividing line between prime and subprime, but Fannie and Freddie were reporting these mortgages as prime, according to Mr. Pinto. Fannie has admitted this in a third-quarter 10-Q report in 2008."

[hmm]

Since you're playing data, data, how valid is the data the next question woulb be where did the MBA get their numbers. Since systemic is such a catchy word was it F%F that were just lying or was everyone changing the definitions.

It still kils me people are looking for a magic bullet to hang this mess on. We already have the magic bullet that will save us, so we need something to blame.

[Johnny Longtorso]

The argument is that the redefinition of prime to meet affordable housing goals threw off the number of prime vs. subprime loans in everyone's risk calcs, throwing those calcs off, leading to problems (as opposed to "deregulation" leading to problems).

The various govt causes of the problem aren't mutually exclusive. Zero interest rates, "smart growth" causing housing shortages in the areas impacted most, and affordable housing initiatives could all have led to our govt caused crash working together.

[brotherben]

How much less would the economy have been damaged if the subprime and so-called ninja loans weren't bundled together in CDOs and MBSs that were given AAA ratings? Isn't that where the real catastrophic damage occurred?

[Johnny Longtorso]

Have the top tranches of those bonds failed yet? The idea is that if you bundle a bunch of lower quality loans together, the "90% have to fail before these bonds don't get paid" group is fairly safe.

[P Brooks]

I would like to know if the Mortgage Bankers Association has a nice graph of "mortgage balance in relation to household income" which could be overlaid on the default graph.

That might be interesting.

[brotherben]

Jonny Longtorso,
I am really rather ignorant of the whole process. The program I saw suggested that many of the bundles were 100% high risk loans and those were still given a AAA rating. (I'm trying to find the source giving the rating company) It suggested that this was where the real damage occurred. I am asking about it out of ignorance of investments and a healthy dose of skepticism. The folks that made the program may have been slanting it to blame the evil greedy rich.

[Johnny Longtorso]

"The last 10% of the loans in this bunch to fail" can be AAA if the loans are individually less than AAA. You're calculating the chance of more than 90% fail instead of the chance of a single one of them failing.

Last I heard those AAA tranches hadn't failed yet. That may be wrong, but it is a fair question to ask how they're doing before blaming them for anything.

[P Brooks]

Many years ago, during the S&L Crisis! I saw a fascinating report which said the single most important indicator of successful repayment of a mortgage loan was the size of the down payment.

Unfortunately, the people whose incentives were based on short term churn in the market (realtors, mortgage brokers, ________) were not particularly interested in the long term success of the loans.

[R C Dean]

The program I saw suggested that many of the bundles were 100% high risk loans and those were still given a AAA rating.

Could be that the rating was issued based on the provision of insurance against default on the bonds (issued by AIG, mostly).

[Hobo Chang Ba]

Thank you for the France Gall - my favorite song by one of my very favorite pop artists. Her entire mid-60s catalog is one of the most brilliant spans of pop music ever.

[brotherben]

A 5 page pdf analysis on the U.S. mortgage market here

If you read it, let me know if you think it is accurate. thanks

[coetsee]

Affiliate Marketing On The Internet
Affiliate Marketing is a performance based sales technique used by companies to expand their reach into the internet at low costs. This commission based program allows affiliate marketers to place ads on their websites or other advertising efforts such as email distribution in exchange for payment of a small commission when a sale results.

http://www.onlineuniversalwork.com

[Jlarsen]

The reporting of prime loans vs. Sub-prime loans is not as clear cut as expected. Contrary to popular (and wrong) thinking. Sub-prime is not defined as a FICO of under 650.

Sub-prime is the lower 40% (ish) of consumers on a bell curve. This means that not everyone who applies for a short loan are sub-prime.

So, it's no surprise that Fanny and Freddie are reporting sub-prime loans today that were prime a few years ago.

(or is that reverse?)

[Reverse Mortgage loan]

Reverse Mortgage - also known as HECM - is a government insured mortgage loan program designed for senior homeowners to release a portion of their home's equity without the burden of monthly mortgage payments while in retirement.
http://www.reversemortgagelend.....formation/
http://www.reversemortgagelend.....annuities/