Arnold Kling  

Mortgage Loans with Low and High Risk

Robin Hanson on Market Failure... Demeanor and Brutality...

David Min criticizes the analysis of Peter Wallison and Ed Pinto of the role played by Freddie Mac and Fannie Mae in the housing bubble. (Pointer from Mark Thoma.)

The centerpiece of Min's critique is a chart that shows the serious delinquency rate for four categories of loans:

Subprime 28.3 %
Freddie 620-650 credit score: 10.04
Freddie over 90% loan-to-value ratio: 8.45%
Conforming: 6.8 %

Min's point is that the middle two categories, which Pinto classifies as high risk, seem to perform about as well as conforming loans. Therefore, it is wrong to classify them as high risk.

I am not sure what to make of this. The 6.8 percent serious delinquency rate on the conforming loans is horrible. The way Min breaks down the loan categories, every category is high risk.

If I were doing this work, I would try to find categories of loans for which the serious delinquency rate is under 1 percent. Now, they might be loans with loan-to-value ratios under 70 and FICO scores over 720, and only loans to purchase an owner-occupied home or refinance it to reduce the interest rate (in other words, no second mortgages, investment properties, or cash-out refis). I have no idea. But when I was with Freddie Mac in the late 1980's and early 1990's, we would not have said we were happy with a portfolio of loans with a serious delinquency rate anywhere near 6.8 percent, under any scenario.

If my view is correct, then Wallison and Pinto probably understate the shift toward high-risk lending that took place at Freddie and Fannie between, say, 1990 and 2007. By the same token, the extent to which this shift can be attributed to trying to meet affordable housing goals is overstated.

I feel pretty confident in arguing that Freddie and Fannie could have stopped the housing bubble by holding onto their credit standards of the early 1990's. However, I would not be comfortable attributing their relaxation of credit standards to the affordable housing goals. I think that the management attitude toward risk changed exogenously. In part, this was due to new CEO's with less experience in dealing with mortgage credit risk. As home prices rose, all sorts of high-risk loans performed well, and senior management misread this to indicate that it was safe to lower credit standards. Certainly, the political environment reinforced that decision. But the numerical affordable housing goals were not the dominant factor.

Comments and Sharing

COMMENTS (28 to date)
Lord writes:

The big problem was unqualified lending as it raises prices above value which make even qualified lending quite risky. Securitization generated false beliefs risk was being reduced rather than hidden.

Ilya writes:

Min claims that calling loans with a loan-to-value of over 90% high-risk is misleading. I give up.

Hugh writes:

I tried to find some international comparisons to get a handle on these percentages: unfortunately the graphs I googled up were old (2009/2010) but they still showed European rates about half the US rates - and that includes countries like Spain that are having a simply dreadful time.

Here in Europe almost all mortgages are with-recourse and that makes a huge difference. If we could understand the differences between US and European mortgage practices I think we would be 50% of the way to understanding what went wrong.

steve writes:

1) Let us clear up the semantics. Wallison is trying to equate 90%LTV with subprime loans. They are not even close.

2) All loans look bad, even the primes. Wasnt this inevitable with the drop in housing value? Given the magnitude of the drop in areas like Nevada, Arizona and California, I suspect you are looking at LTVs of 60 to avoid serious default risk. Is that kind of ratio doable in a modern economy?

3) You miss one of Min's primary points. Why do Wallison and Pinto ignore ARMs? Those were some of the worst performing loans. Did they conveniently ignore them because they were largely made by the private sector?

4) You are too nice to Fannie and Freddie execs. I think there is fairly good evidence that they were knowingly committing accounting fraud. Money was sufficient incentive for their actions.


David Min writes:

Dr. Kling,

In a vacuum, you'd be correct. But you're overlooking the fact that we've experienced 30+% peak-to-trough home price drops, stagnant wages, and a persistently high unemployment rate. By way of comparison, as I note in my piece, the calculated rate of default for urban mortgages during the Great Depression, the most comparable economic shock we experienced in this country (data for all home mortgages is unavailable) was about 50%.

I feel confident saying that a loan that only had a LTV>90 risk characteristic, but which was otherwise well underwritten, would have a serious delinquency rate of less than 3-4% during a "normal" economic period.

Moreover, I think you're ignoring the key issue here, which is what caused the financial crisis. The Pinto thesis is that Fannie/Freddie/CRA caused the crisis by causing a deterioration of underwriting standards. The problem with that thesis is that by and large, PLS (subprime, but for that matter, prime ARMs as well) are defaulting at exponentially higher rates than the allegedly "high risk" categories invented by Pinto.

David Min writes:

Also, to clarify, I agree that Fannie and Freddie took on more risk than they should have, and for what it's worth, I think I agree with you on the diagnosis for why they did so. But that is a separate question from the one being addressed, which is the Pinto argument that Fannie and Freddie (along with CRA and other government policies) were driving the "high risk" lending that caused the financial crisis. I think it's fairly clear that the "high risk" loans attributed to Fannie and Freddie are quite distinct from the high risk loans attributed to PLS (subprime etc.), something that the FCIC noted in its analysis of Pinto's work (also noted by Harvard's Joint Center for Housing Studies and others), and combining these loans into one large hybrid "subprime" category is disingenuous.

Thomas DeMeo writes:

I think that the real lesson is that markets simply cannot deal with this level of instability in real estate valuations, regardless of lending rules. A major drop in valuations would destroy markets even if the banks were insulated by conservative loan to value ratios. The effect on the banks matter, but the effect on consumption is more important. Perhaps most importantly, unusual increases in valuations need to be viewed with the same alarm as decreases.

Real estate valuations need to be watched and adjusted for as tightly as inflation.

J Mann writes:

Min's clarification is interesting, and important. So if I understand it:

1) The delinquency rate on Fannie and Freddie's conforming loans is just under 6.8%, which is generally lower than the national average of 9.1%

2) Fannie and Freddie also wrote a bunch of non-conforming loans. Less than 1% of their book was what Min calls arguably "high risk" (FICO score 90%) has a 8.5% delinquency rate.

3) Literally subprime loans have a much worse rate: 28.3%

That's very helpful, but if I want to determine how much of a role Fannie and Freddie played in the debacle, I have some more questions.

1) What percentage of all delinquent home mortgages in the country are Fannie or Freddie? What percentage are subprime?

2) If Fannie and Freddie's book is so much better than the national average, why are they giant money pits with Iraq-war-sized needs for taxpayer infusions? Am I wrong to think that if I need to give $350 billion to the companies, they must have made some imprudent loans?

Seth writes:

So, take away the political goals of home ownership and we still get the mortgage bubble?

steve writes:

"So, take away the political goals of home ownership and we still get the mortgage bubble?'

40% of our growth in the aughts came from the financial sector. If incentives really matter, wouldn't that much money be adequate incentive?


David Min writes:

J Mann, good questions. Answers are in my pieces, but here's a quick reply:

1) Fannie and Freddie are responsible for 57% of all loans, 22% of delinquent loans. Conversely, PLS is responsible for 13% of all loans but 42% of all delinquent loans.
2) The reason Fannie and Freddie are taking such heavy losses is primarily because they are such a large percentage of the mortgage market. This raises a serious question about reducing their footprint, of course.

Arnold Kling writes:

@David Min:

If a "normal economy" is one in which house prices do not fall, then in a normal economy there is no such thing as a high-risk loan. Worst case, the borrower sells the house rather than handing you the keys. Even a lot of fraud gets covered up by rising house prices.

So risk has to be defined relative to a scenario where prices fall. (They could fall locally, of course. They do not have to fall everywhere.) From that perspective, even a prime loan is risky if the loan-to-value ratio is above 90 percent. I would actually say that it's high when the LTV is above 80 percent. And in a cash-out refi, it's high when the LTV is above 70 percent, and even lower if the appraisal is hard to corroborate.

Costard writes:

Default rates may have popped the bubble but it was the assumption of low risk that created it in the first place. Private lenders merely leveraged themselves upon what Freddie, Fannie, the Fed, and Congress had implicitly guaranteed: that housing prices would not drop.

Furthermore since these loans were all against the same asset class - real estate - it absolutely does not matter that F&F's mortgages were relatively conservative. Because of their size and semi-public status, they WERE the market; they provided the current; and if everyone else swam with it, what does this prove vis a vis responsibility?

Michael P. Manti writes:

As far as I can tell, Min focuses on mortgages originated for private label securitization versus mortgages originated for GSE securitization. Is GSE purchase of private label securities or participation in structured transactions accounted for?

When I was at Freddie, they did loosen standards for originations. But the big deal was participation in the supposedly AAA tranches of structured transactions that the ratings agencies were giving their seal of approval. The motivation for such deals was, of course, profit, but they were also cited as being "goal-rich"--i.e., they brought Freddie closer to compliance with the HUD affordable housing goals.

Patrick R. Sullivan writes:

I stopped taking Min seriously here:

To support his claim that the Community Reinvestment Act, which requires regulated banks and thrifts to provide credit nondiscriminatorily to low- and moderate-income borrowers, caused the origination of 2.24 million outstanding “high-risk” mortgages, Pinto includes many loans originated by lenders who were not even subject to CRA. In fact, most of the “high-risk” loans Pinto attributes to CRA were not eligible for CRA credit.

Because, while not technically covered by the CRA they were covered by the HUD 'Best Practices Initiative', its evil twin. This kind of quibble is a dead giveaway.

Michael P. Manti writes:

I checked one of the links in Min's article to answer my own question. I do not see how the argument that PLS purchase by the GSEs isn't a major factor is credible. A third of Freddie's retained portfolio! And almost certainly overrepresented in Freddie's losses.

As an aside I find Min's citation of the GAO, who unquestioningly quote FHFA's Lockhart on the PLS issue, as nonpartisan, to strain credulity. Cripes--this all happened on Lockhart's watch. What the heck do you think he'd say?

Yeah, I think Wallison and Pinto's work is lousy. But the arguments in favor of the GSE housing goals' being benign just don't pass muster with me. Just about every questionable transaction I knew about had a housing goal justification.

Pete writes:

I'm interested to hear why you think Wallison's work is "lousy"

Pete writes:

I know that Pinto believes there is a large swath of loans with a credit score below 660 that the MBA considers "Prime" and are included in their 6.8% above.

If so, that casts a dark shadow over the legitimacy of this data. Obviously you can not compare the delinquency rates of loans with FICO scores between 620-659 with those of "Prime" loans that also include scores between 620-659.

FYI, I don't know if its accurate or not, but Pinto claims the real number is 3.4% for conforming loans if you back out those in his 620-659 category.

Michael P. Manti writes:


I was unfair to Wallison and Pinto. I think Wallison goes too far when he says that only government housing policy was at fault in his blog post. I shouldn't have said their work as a whole was lousy.

Patrick R. Sullivan writes:


Wallison isn't saying that only govt. housing policy is at fault. He's saying that it is the sine qua non; it's the wanting nail that caused the shoe to be lost....

Min's quibbling over semantics is irrelevant. Almost no one was making these non traditional loans prior to the early 1990s. Without govt. interference in the industry incentivizing them they probably never would have become more than a handful of each one or two hundred mortgage loans.

And, Min's attempt to separate Wallison from his three Republican colleagues is equally disingenuous. That's merely a division of labor, not any disagreement between the two dissents.

Shayne Cook writes:

Following and expanding on Costard comment above, the GSE's had always had an implied warrant of Federal Government intervention in case of problems. I suspect the CRA and other subsequent related Government actions merely reinforced that implication for investors, i.e., that the Federal Government would warrant/assume all risk associated with any form of credit paper, as long as it was in some way or another related to housing goals. That, more or less irrespective of who/what participated in the origination/participation of housing related paper, inasmuch as housing related credit paper could be "legitimized" through the GSE's at any time. The implication was that, credit paper would be backed by full faith and credit of the U.S. government, as long as it was housing related.

But, as Arnold and others have noted, the GSE's themselves were not the only participants. A substantial if not majority of mortgage originations/securitizations of the period were outside the GSEs - specifically, via the Investment Banking community. Enter the 1999 repeal of the Glass-Steagall provisions for separation of Commercial versus Investment banking (established in 1933). Additionally, the subsequent Federal Reserve actions of 2006-2008 (as well as tacit support of the LTCM bailout) to prop up Investment banks, there was implicit and explicit warrant that the Federal Reserve would intervene in the case of problems.

So the entire investment community had (and still has) an eminently lucrative investment target - housing related paper - that had (and still has) implicit and explicit warrant by both the U.S. Federal Government and the U.S. Central Bank against any and all risks!

Shayne Cook writes:

Follow-up ...

For those still wondering why the U.S. economy is not enjoying a robust recovery, given all the monetary and fiscal stimulus applied, I would steer you to my comment above, in conjunction with the Capital Asset Pricing Model (CAPM).

At its foundation, CAPM has the notion of the expected returns of the "risk-free" asset. Unfortunately, too many folks apply the paltry returns of U.S. Treasury debt as the standard of the "risk-free" returns. With that paltry return standard, confusion arises as to why the investment and business communities haven't ratcheted up growth and spending.

Well, the CAPM isn't broken. What is broken is the assumption that U.S. Treasury debt yields are the current "risk-free" rate of return. In fact, given my comment above, current housing related paper returns are "risk-free", by virtue of both Federal Government and Federal Reserve warrant. And housing related paper returns range from a low of about 14% to something over 20%.

Plug those expected returns into CAPM and see what you get for nearly any non-housing related investment.

Cahal writes:

I know this point isn't a proper response to your post, but it's relevant:

In general, I'm reluctant to buy the CRA/Fannie/Freddie argument, simply because other countries didn't have them and experienced the same thing. Spain, UK , Iceland, Ireland, U.S. - the common theme in the worst affected countries was financial deregulation, not particular policies or even loose monetary policy.


'So, take away the political goals of home ownership and we still get the mortgage bubble?'

Firstly, you have to question if those goals were advanced by banks via the government as a way of making money. The financial deregulation/tax cuts of the 80s do, in fact, coincide with the rise of home ownership.

Secondly, perhaps not - but does it matter? S & L crisis, dot com bubble, Mexican currency crisis, East Asian crisis - the banks will always find something to exploit and loot. The fact that it was housing in this instance may not be that relevant.

Steve Sailer writes:

The problem with most debate on this topic of who is to blame for the mortgage meltdown is that there are two general camps:

Republican / libertarian / pro-private enterprise / not wanting to be called anti-"diversity" (for example, Wallison's American Spectator article only refers to lower income goals, and not to racial quotas).


Democrat / regulationist / anti-private enterprise / pro-"diversity" (champing at the bit to tar opponents as "racist")

Wallison and Pinto fall into the first camp, so they have to strain to attribute everything to Fannie and a few other culprits.

Personally, although I've voted more Republican than not over the years, this screw-up has got more Republican than Democrat fingerprints on it. Democrats like James Johnson, Henry Cisneros, and Barney Frank loaded the gun, but George W. Bush pulled the trigger at his October 15, 2002 White House Conference on Increasing Minority Homeownerships, where he denounced down-payment requirements as the primary impediment to racial equality in attaining the American Dream. When a Republican President plays the Race Card, his employees (federal regulators) get the message in a hurry that the boss wants them to go easy on dubious mortgage. And that's what happened, with the bubble inflating in the wake of Bush's conference and going into hyperdrive after Congress endorsed it in 2004 by passing the American Dream Act providing downpayment assistance.

Similarly, I'm not a libertarian. I'd like to be, but I spent 18 years in the private sector, and so I'm not as naive about just how far people would go to make a quick killing. Therefore, I'm in favor of prudent regulation by the government of the financial sector.

So, that leaves me without a team in this debate. Yeah, it was, more than anything else, Bush's fault for undermining regulation, but he did it in the sacred name of diversity, and until we admit all the risks associated with our national worship of "diversity" as a holy thing, stuff like this is going to keep on happening. Smart guys like Angelo Mozilo are going to keep on figuring out how to use diversity rhetoric to justify heads-I-win / tails-the-taxpayer-loses risktaking.

Ted Craig writes:

What's the definition of delinquency? That can vary by lender.

keatssycamore writes:

Arnold Kling writes,

"So risk has to be defined relative to a scenario where prices fall."

This seems wrong. Wouldn't risk have to be defined relative to a scenario (reality) where home prices rise and fall?

James A Donald writes:

Beverly hills Bank was told that it had to provide "innovative" loans - that is to say, no money down loans to people with no income, no job and no assets - and its failure to do so put it in violation of the CRA.

That is not "a numerical affordable housing goal" - but it is an affordable housing goal.

Todd Schaal writes:

"Beverly hills Bank was told that it had to provide "innovative" loans - that is to say, no money down loans to people with no income, no job and no assets - and its failure to do so put it in violation of the CRA."

Um...Not quite

The lending test performance criteria from the CRA that you are refering to reads:

(5) Innovative or flexible lending practices. The bank's use of innovative or flexible lending practices in a safe and sound manner to address the credit needs of low- or moderate-income individuals or geographies.

See that bit about "safe and sound"? Also, the report you linked doen't cover residential 1-4 family loans, because "The institution does not originate residential mortgages." But hey, it was worth a shot.

Comments for this entry have been closed
Return to top