pageok
pageok
pageok
ARMs and the Interest Rate Spread:

This is pretty interesting:

The bars show the percentage of new mortgages at any given time that are adjustable rate mortgages. As you can see, this ratio has fluctuated over time, going over 60% in the 1980s and over 50% at times in the 1990s. Right now it is very low (under 10%).

The line shows the "spread" between the prevailing interest rate on 30 year fixed rate mortgages versus ARMs. ARMs are always lower because with a FRM the borrower is essentially paying the lender to bear the risk of interest-rate fluctuations as well as (in the U.S.) the unlimited right of the borrower to refinance when interest rates go down means that the lender bears the prepayment risk as well. Borrowers pay a lot for this insurance--as you can see, the spread is usually in the neeighborhood of 100 to 150 basis points, although it fluctuates higher and lower as well.

Note the general pattern here--the percentage of mortgages that are ARMs almost perfectly tracks the spread between the interest rates on ARMs and FRMs. As ARMs become less expensive relative to FRMs, the percentage of ARMs rises. The artificially low rates on ARMs as a result of easy money policies during the early 2000s created the gap between short and long term rates.

The problem, of course, is that this spread can disappear in one of two ways. Either the rates on FRMs can come down, or the rates on ARMs can go up. In the 1980s and mid 1990s, the FRM fell. This last go around the rate on ARMs rose. Which has helped to spur the foreclosure problems we see, especially in areas of the country with a lot of ARMs.

Note also that this is not an issue of subprime v. prime--the regularity of the interaction between ARMs and FRMs held prior to the existence of the subprime market, and in fact, the percentage of ARMs in the market was much higher at times in the past.

What got me thinking about this more specifically is Stan Leibowitz's article "Anatomy of a Train Wreck," where he emphasizes the role of ARMs in the foreclosure crisis. The article is excellent and I agree with almost the whole thing with one caveat. Stan argues that the rise in ARMs is a proxy for a rise of speculation in the real estate market. His argument is that speculators disproportionately selected ARMs with an intention of flipping the property before the interest rate reset and that the are thus also disproportionately represented in foreclosure. I agree that the role of speculators is important and that before we do anything drastic with respect to foreclosure relief we want to figure out the extent to which speculators are disproportionately represented in foreclosure.

What this chart seems to suggest, however, is in the larger picture the ARM issue is separate from the speculator issue. The popularity of ARMs is driven by the interest rate spread between the interest rates on ARMs and FRMs and in the past we have seen ARMs become popular even for prime borrowers in real estate markets that weren't as crazy as we've seen this past several years.

This is something that I suspect we knew intuitively, yet it is striking to see it on a graph like this. Teh only real anomaly seems to be in the early to mid-1990s, when the spread rose dramatically yet ARMs did not, before exploding in a wave of ARMs around 1995 or so.

As I discuss in my forthcoming article on "The Law and Economics of Subprime Lending," ARMs are standard in most of the world and it is the United States that is an outlier in terms of having 30 year fixed-rate mortgages with unlimited prepayment rights.

John T. (mail):
There are lots of quite different causes of the crisis. One has to separate out the fundamental causes of the housing price bubble from the various causes that made it worse and ensured that some financial institutions would take particular baths.

You note that the US is an outlier in terms of having unlimited prepayment rights. But the UK, Spain, Germany, and other countries also saw a housing bubble, so it's obviously not all the role of financing.

OTOH, some US states saw little to no housing bubble; prominent among them are the otherwise rapidly growing states of Texas and North Carolina. It would be worth studying why these states didn't see the bubble. If you listen to people like Ed Glaeser of Harvard, or Paul Krugman, you might be willing to blame zoning and land-use restrictions. There's an extremely strong correlation between growth-management planning (or inelasticity of housing in general) and the housing bubble. Florida and Arizona, despite their rapid growth of housing, were growing even more rapidly in population, and both have regional growth-management planning laws in effect.

Growth management encourages bubbles. In the "best" case, housing just becomes unaffordable and never comes down. In the worst case, there's enormous volatility as soon as demand slackens or when houses, delayed by red tape, finally come on the market.
10.25.2008 4:07pm
kietharch (mail):
Something that is not usually mentioned in the great housing bubble
postmortem:

When my wife and I bought our first home we took out a fixed rate mortgage and the paper was held by the seller. The mortgage liability that my wife and I assumed was not limited to the value of the house and property; if we defaulted the seller was entitled to (what I think was called) a "deficiency judgment", i.e., the difference between the remaining value of the mortgage less the salvage value of the home. As I understand it, the liability of current day mortgages is limited to the salvage value of the home (perhaps because of Federal regulations).

This is, by itself, a significant loosening of credit standards and a mild encouragement to speculation.
10.25.2008 4:09pm
kietharch (mail):
" the percentage of mortgages that are ARMs almost perfectly tracks the spread between the interest rates on ARMs and FRMs"

Maybe I am missing something but that does not look like a perfect track to me. I am surprised infact that the percentage of ARMs stays constant for a year or so while the spread is going up so dramatically. I would assume that a bad economy would do that. I would surely like to see actual interest rates in that graphic.
10.25.2008 4:21pm
Fub:
Note the general pattern here--the percentage of mortgages that are ARMs almost perfectly tracks the spread between the interest rates on ARMs and FRMs. As ARMs become less expensive relative to FRMs, the percentage of ARMs rises. The artificially low rates on ARMs as a result of easy money policies during the early 2000s created the gap between short and long term rates.

...

This is something that I suspect we knew intuitively, yet it is striking to see it on a graph like this. Teh only real anomaly seems to be in the early to mid-1990s, when the spread rose dramatically yet ARMs did not, before exploding in a wave of ARMs around 1995 or so.
Yet there is a similar, but not as dramatic, anomaly from about 2001 to 2005.

During the mid-1990s the %ARM mortgages appears to have actually declined slightly as the spread increased.

During the 2001-2005ish period, the %ARM mortgages just did not increase as rapidly as the spread, until a surge of ARMs beginning around 2004.

IANA economist or econometrician, but I have a dumb question.

Would looking at the ratio (spread / %ARMs, or vice versa) vs time better pinpoint in time the surges in %ARMs that could be argued dependent upon something besides just the spread?

This would be somewhat like looking at the residuals over time for a linear regression model of %ARMs vs spread.

The reasoning would be that during periods of great departures from the average ratio between spread and %ARMs (or greater magnitude residuals from the regression) something besides the spread was driving the %ARMs.

But I am way out of my depth here, even risibly so, which is why I warned it was a dumb question.
10.25.2008 6:31pm
Elizabeth Allemann (mail):
Yes, but were the loans that got so many people in trouble traditional ARM's? The ARM I have for the mortgage on my office has the interest rate adjusted yearly, based on the Prime rate. In addition, there are no balloon payments, and there was the 20% down payment. I am lucky to live in a small town and my small town bank keeps its own loans--doesn't bundle or sell them, so if I have trouble making a payment, I can call my loan officer who knows me by my first name--we've known each other for 20 years. I have a checking and savings accounts at the same bank. I had to submit proof of insurance and they want to see my tax returns every year. No matter what happens to the value of the property, I can still pay off my loan (and, since I use the office to make my living, so long as it continues to stand and the neighborhood is still decent, the building has the same value to me). The only use of the loan is to allow me to own my office rather than rent it.

The ARM's that were so problematic had no or low down payments, balloon payments, and a ballooning interest rate that is unrelated to the prime rate. I hear they were called "liar loans" in that loan officers encouraged people to exaggerate their income and assets in order to qualify for the loans. In short, these loans were designed to make sense ONLY in an environment of rapidly escalating real estate values. And they were foisted off on people who didn't realize what they were signing up for. In addition. people were encouraged to take out second mortgages on their homes to take trips, buy cars, buy furniture, etc. The people who sold these loans had no interest in owning the loans, but in bundling and selling them to other people who had no way of assessing the ability of the borrowers to actually pay back the loans or the value of the property.

So, unless I don't understand the situation, it isn't really the difference between FRM and ARM. Instead we need to compare the "Baily Savings and Loan" deal--lend money to someone you know for a property you can drive across town and see, so that you can earn the interest yourself and work with the person to pay off the loan. This is compared with encouraging people to take on debt unrelated to the value of the asset used as collateral in the interest of selling the loan. Speculation. Leverage. Lies. Loss of Relationship. Not the same things at all
10.25.2008 7:40pm
JDS:
Many loans in the current crisis are "option ARMs" in which the borrower can skip payments and just increase the loan balance. For a large number of loans in default, the borrower never made a single payment.
10.25.2008 7:43pm
Zywicki (mail):
Fub:
That's a good point--I see there is a lag in the 2001-2005 period too and then it jumps all of a sudden (as in the early 1990s).
10.26.2008 5:20am
paul lukasiak (mail):
the lag in 2001-2005 is easily explained -- fixed mortgage rates were extremely low in historic terms, and that meant that ARM rates that were 2 or 3 points below fixed rates had nowhere to go but up.

As with others, I don't see the pattern here -- and I'd suggest that the popularity of ARMs has far more to do with the prevailing rates on fixed mortgages than with the spread between fixed and ARMs (if fixed rates are at 10%, an ARM at 8% makes sense, because rates are as likely to go down as up -- but when fixed rates are at 6%, an ARM at 4% makes far less sense.)
10.26.2008 6:57am
Fub:
Zywicki wrote at 10.26.2008 4:20am:
That's a good point--I see there is a lag in the 2001-2005 period too and then it jumps all of a sudden (as in the early 1990s).
Thanks for confirming that my eyes weren't lyin' to me.

paul lukasiak wrote at 10.26.2008 5:57am:
As with others, I don't see the pattern here -- and I'd suggest that the popularity of ARMs has far more to do with the prevailing rates on fixed mortgages than with the spread between fixed and ARMs (if fixed rates are at 10%, an ARM at 8% makes sense, because rates are as likely to go down as up -- but when fixed rates are at 6%, an ARM at 4% makes far less sense.)
Good point. The same data from which the time series spread data derived should be useful to test a time independent %ARM = F(FixedRate, AdjRate) hypothesis.

Another dumb question -- Could some particular nonlinear %ARM = F(FixedRate, AdjRate) that determines well (ie: much better than a linear F) indicate something about the involvement or non-involvement of speculators?
10.26.2008 1:24pm
Harry Eagar (mail):
' For a large number of loans in default, the borrower never made a single payment.'

Reference? I'm not doubting you, but I haven't seen any statistics about this, just vague general claims.
10.26.2008 9:49pm
Vulpes Lagopus (mail):
The anomaly results from the end of the S &L crisis and the enactment of the Financial Institutions Reform, Recovery, and Enforcement Act of 1989. Bank management rewrote their bonus plan payout, which incentivized selling ARM's (much like WAMU's ballyhooed rewrite of payouts to remove mortgage losses at the end of last year) Afterward it became the tool of the house flipper.
10.26.2008 10:43pm
Dilan Esper (mail) (www):
As I discuss in my forthcoming article on "The Law and Economics of Subprime Lending," ARMs are standard in most of the world and it is the United States that is an outlier in terms of having 30 year fixed-rate mortgages with unlimited prepayment rights.

And truly, perhaps the fundamental divide between left and right in America about economic issues is about risk. Liberals believe risk is dangerous and needs to be mitigated; conservatives think risk fuels growth and should not be impeded.
10.27.2008 6:07pm